Dental Age Estimation in Adults. An in-depth analysis of ...

Dental Age Estimation in Adults. An in-depth analysis of

secondary dentine deposition in different populations

and clinical conditions.

Talia Yolanda Marroquin Peñaloza DDS, MClinDent(Endodontist)

This thesis is presented for the degree of Doctor of Philosophy of The University of Western

Australia

School of Human Sciences

2017

1

THESIS DECLARATION

I, Talia Yolanda Marroquin Peñaloza, certify that:

This thesis has been substantially accomplished during enrolment in the degree.

This thesis does not contain material which has been accepted for the award of any other

degree or diploma in my name, in any university or other tertiary institution.

No part of this work will, in the future, be used in a submission in my name, for any other

degree or diploma in any university or other tertiary institution without the prior approval

of The University of Western Australia and where applicable, any partner institution

responsible for the joint-award of this degree.

This thesis does not contain any material previously published or written by another

person, except where due reference has been made in the text.

The work(s) are not in any way a violation or infringement of any copyright, trademark,

patent, or other rights whatsoever of any person.

The research involving human data reported in this thesis was assessed and approved by

The University of Western Australia Human Research Ethics Committee. Approval: (Ref:

RA/4/1/6797).

Written patient consent has been received and archived for the research involving

patient data reported in this thesis.

2

The work described in this thesis did not need funded of tertiary parties.

This thesis contains published work and/or work prepared for publication, some of which

has been co-authored.

Signature:

Date: 30th November, 2017.

3

SUMMARY

Dental age estimation has been an important issue for humanity since ancient times. In the

same way, different methods have been proposed through history for young individuals

and adults. Once forensic dentistry is an important branch of the forensic sciences, and

with the introduction of radiology in dentistry, it has been possible to propose non-invasive

methods for age estimation, based on the radiological analysis of the inverse relation of the

root canal size with age. Age estimation of adult individuals has always been a challenge,

as those methods based on the narrowing of the root canal, as a result of secondary

dentine production have shown wide variability in their results. This thesis analyses some

of the observed obstacles for dental age estimation in adults, and proposes new

approaches, aimed to overcome them.

The published papers presented in this thesis follow some of the previously proposed

methods for dental age estimation in adults, such as Kvaal et al’s method, and those

methods based on volume measurements of pulp and tooth, adapting, and testing them in

diverse groups of populations. Measurements have also been taken in individuals with

different dental characteristics, thereby challenging some of the established parameters for

dental age estimation as it is the exclusive analysis of totally sound teeth, in samples with a

unique ethnic background. Additionally, this thesis also includes what could be considered

to be the first literature review in forensic dentistry for dental age estimation in adults; a

study, with enormous contributions to this research field. This thesis also includes the

analysis of the possible factors that affect the accuracy of the tested methods.

4

CONTENTS

THESIS DECLARATION .................................................................................................................. 1

SUMMARY .................................................................................................................................... 3

CONTENTS .................................................................................................................................... 4

TABLE LIST .................................................................................................................................... 7

GRAPH LIST ................................................................................................................................... 9

ACKNOWLEDGEMENTS ............................................................................................................. 10

AUTHORSHIP DECLARATION. .................................................................................................... 12

GENERAL INTRODUCTION ......................................................................................................... 16

REFERENCES ............................................................................................................................... 25

SECTION ONE. Non-invasive methods for dental age estimation in adults: Systematic

Review. ....................................................................................................................................... 27

Preface ....................................................................................................................................... 27

1. Chapter 1. Age estimation in adults by dental imaging assessment systematic review28

SECTION TWO. Assessment of possible effect of orthodontic treatment on one of the

available methods for age estimation in adults: The Kvaal et al. method. ............................. 70

Preface ....................................................................................................................................... 70

2. Chapter 2. Changes in age estimation standards secondary to full fixed edgewise

orthodontic treatment .............................................................................................................. 76

3. Chapter 3. Orthodontic treatment: Real risk for dental age estimation in adults? ........ 95

SECTION THREE: New uses of the Kvaal et al method for age estimation in young

individuals and adults from different groups of population. ................................................ 111

5

Preface ..................................................................................................................................... 111

4. Chapter 4. Determining the effectiveness of adult measures of standardized age

estimation on juveniles in a Western Australian population. ............................................... 112

5. Chapter 5. Application of the Kvaal method for adult dental age estimation using Cone

Beam Computed Tomography (CBCT). ................................................................................... 132

SECTION FOUR: New technologies in dental age estimation. .............................................. 152

PREFACE ................................................................................................................................... 152

6. Chapter 6. Reliability and repeatability of pulp volumetric reconstruction through

three different volume calculations ....................................................................................... 153

7. Chapter 7. A new proposal to overtake population group differences for dental age

estimation in adults through pulp/tooth volume calculation. Pilot Study. .......................... 178

GENERAL DISCUSSION. ............................................................................................................ 196

GENERAL CONCLUSIONS ......................................................................................................... 204

APPENDIX 1. Age estimation in adults by dental imaging assessment systematic review. 209

APPENDIX 2. Changes in age estimation standards secondary to full fixed edgewise

orthodontic treatment ………………………………………………………………………………………………….. 219

APPENDIX 3. Orthodontic treatment: Real risk for dental age estimation in adults?.........226

APPENDIX 4. Determining the effectiveness of adult measures of standardized age

estimation on juveniles in a Western Australian population ……………………………..…………...231

APPENDIX 5. Application of the Kvaal method for adult dental age estimation using Cone

Beam Computed Tomography (CBCT)……………..…………………………………………………..………....242

APPENDIX 6. Reliability and repeatability of pulp volumetric reconstruction through three

different volume calculations………………………………………………………………………………………..…248

6

APPENDIX 7. A new proposal to overtake population group differences for dental age

estimation in adults through pulp/tooth volume calculation. Pilot Study. In press………….261

7

TABLE LIST

Table 1.1. Total studies reporting the use of the Cameriere et al. method. .......................... 38

Table 1.2. Total studies reporting the use of the Cameriere et al. method. ......................... 39

Table 1.3. Total studies reporting the use of volume calculation. ......................................... 40

Table 1.4. The Kvaal et al. method: Studies included in the quantitative analysis. .............. 41

Table 1.5. The Cameriere et al. method: Studies included in the quantitative analysis. ...... 42

Table 1.6. Volume calculation based methods. ...................................................................... 43

Table 2.1. Correlation coefficients before (1) and after orthodontic treatment (2). ................ 83

Table 2.2. Multiple regressions for estimation of chronological age (in years) for individual

maxillary and mandibular teeth in subjects before (1) and after (2) orthodontic treatment

.................................................................................................................................................... 84

Table 2.3. Multiple regressions for estimation of chronological age (in years) for the

combination of different teeth before orthodontic treatment .............................................. 85

Table 2.4. Multiple regressions for estimation of chronological age (in years) for the

combination of different teeth after orthodontic treatment ................................................. 86

Table 3.1. Contingency table to estimate if orthodontic treatment was a causal of

overestimation, when the Kvaal et al. method is used to age estimation. .......................... 102

Table 3.2. Comparison of over- and under-estimates obtained per tooth .......................... 103

Table 4.1. Correlation coefficients between chronological age and Kvaal dental

measurements and ratios, for individuals under 30 years of age. ........................................ 121

Table 4.2. Multiple regression for estimation of chronological age (in years) from individual

maxillary and mandibular teeth from individuals under 30 years of age. ............................ 122

8

Table 4.3. Multiple regressions for estimation of chronological age (in years) from the

combined maxillary and mandibular teeth. ........................................................................... 123

Table 5.1. Age and sex distribution for CBCT and i-CAT tomographs. .................................. 140

Table 5.2. Correlation coefficient measurements CBCT ........................................................ 140

Table 5.3. Linear regression analysis, linear regression formulae, correlation coefficient (r)

between age and the ratios of bucco-lingual and mesio-distal width measurements from

CBCT & i-CAT ............................................................................................................................ 141

Table 6.1. Volume results of using automatic segmentation (Vol 1), manual segmentation

using just the first and last slice (Vol 2), and manual segmentation each forth slice (Vol3), of

cone beam computer tomography from upper canines. (FDI 13, 23) .................................. 163

Table 6.2. Cone shape approach results using the tool polygon (Pol) and oval (Oval). Pulp

area (in mm2) at the CEJ level and volume (in mm3) calculation with the measured root

length (in mm). ........................................................................................................................ 164

Table 6.3. Setting parameter combinations used in part C. ................................................. 165

Table 7.1. Distribution of the individuals for age in years and country; Malaysia (Mal) and

Colombia (Col) used for the image segmentation and volume calculation in this study. ... 186

9

GRAPH LIST

Graph 1.1. Flow chart of the study selection for the Kvaal et al. method. ............................ 35

Graph 1.2. Flow chart of the study selection for the Cameriere et al. method .................... 36

Graph 1.3. Flow chart of the study selection for methods based on volume calculation. ... 37

Graph 4.1. Scatter plot showing a positive association between the estimated age in years

and the chronological age in years for the teeth 33/43 ........................................................ 122

Graph 6.1. Volumetric reconstructions, upper right canine using different methods. ..... 162

Graph 7.1. Scatter plot showing the correlation between age and the pulp/pulp+tooth

volume ratio. (P/PT ratio) R2=0.42 .......................................................................................... 186

10

ACKNOWLEDGEMENTS

This thesis is the result of an international collaboration between different eminences from

4 different countries: Australia, Colombia, Malaysia and Norway, from some of the most

prestigious institutions around the world. The National University of Malaysia, The National

University of Colombia, The University of Oslo, and The University of Western Australia. It

has been an honor to count upon their support through this journey, which started only as

an academic project, but that now is an incredible personal and professional achievement.

I want to thank my supervisors, Dr. Shalmira Karkhanis, Dr. Sigrid I. Kvaal, Dr. Estie Kruger,

Dr. Sivabalan Vasudavan, and Dr. Marc Tennant, who enlightened me with their diverse

knowledge, supported me in those difficult moments, inspired me with their different

goals, and leaded me in my professional growth. I thank Professor John McGeachie for his

collaboration with the final editing of this document. I also want to thank Dr. Nurul Firdaus,

Dr. Mohaed Shahida Mohd Said, Dr. Shalini Kanagasingam, Dr. Manuel Gonzalez, and Dr.

Claudia Garcia, who trusted on me and gave me access to their data bases. I want to thank

to Dr. Daniel Franklin and Dr. Ambika Flavel for their support through the Forensic Science

Center at The University of Western Australia. Also, I want to thank to all those

anonymized patients, who accepted that their records were used for the good of science.

I want to bring special recognition to Edwin Castelblanco, Marty Firth, and Alethea Rea,

who supported me in the statistical analysis of the data.

11

I also want to thank Colfuturo and Colciencias – Colombia, for having awarded me with the

international scholarship program, which helps Colombian students to continue with their

academic formation undertaking research degrees in the best universities around the

world. I hope, this thesis contributes to the academic and scientific development of our

country. Finally, I want to thank my friends and family who supported me during all these

studying years, this achievement would not have been possible without their support.

Talia Yolanda Marroquin Peñaloza

12

AUTHORSHIP DECLARATION.

This thesis contains work that has been published, accepted for publication, and submitted

for publication, and follows the format of my thesis by a series of papers, as stipulated by

The University of Western Australia for the Degree of Doctor of Philosophy,

In this thesis, all seven papers (7 primary) listed have undergone peer review, 6 have been

already published, and in all papers, the candidate was responsible for the work. The

candidate collected and analyzed the data.

The data were processed, and this information enabled the candidate to scrutinize existing

theories. With this knowledge, the candidate created new knowledge generated new

concepts, methodologies and understandings. With permission, the candidate was given

authority to submit the work for publication.

This thesis is comprised of a series of published papers in peer reviewed journals of which

all of them have been co-authored.

PhD Candidate, Talia Yolanda Marroquin Peñaloza. 30th , November, 2017.

Coordinating Supervisor, Winthrop Professor Marc Tennant. 30th November, 2017.

13

PUBLICATIONS ARISING FROM THIS THESIS

For all the papers of this thesis the contribution of the first author was estimated to be at

least 85% as co-author:

1. Chapter 1

Marroquin, T. Y., S. Karkhanis, S., Kvaal, S. I., Vasudavan, S., Kruger, E., & Tennant, M.

Age estimation in adults by dental imaging assessment systematic review. Forensic

Science International Journal. Accepted for publication. March 2017.

2. Chapter 2.

Marroquin, T. Y., Karkhanis, S., Kvaal, S. I., Vasudavan, S., Castelblanco, E., Kruger, E., &

Tennant, M. (2016) Changes in age estimation standards secondary to full fixed

edgewise orthodontic treatment. European Journal of Forensic Science, Accepted for

publication. June 2016.

3. Chapter 3.


Tennant, M. (2016). Orthodontic Treatment: Real Risk for Dental Age Estimation in

Adults? Journal of Forensic Science. Accepted for publication. October 2016.

4. Chapter 4.


Tennant, M. (2016). Determining the effectiveness of adult measures of standardised

14

age estimation on juveniles in a Western Australian population. Australian Journal of

Forensic Sciences. Accepted for publication. March 2016.

5. Chapter 5.

Marroquin, T.Y., Karkhanis, S., Kvaal, S. I., Vasudavan, S., Nurul F., Kanagasingam S., D.

Franklin., E., Kruger, E., & Tennant, M. (2016) Application of the Kvaal method for adult

dental age estimation using Cone Beam Computed Tomography (CBCT). Journal of

Forensic and Legal Medicine, Accepted for publication. October 2016.

6. Chapter 6.


Tennant, M. Reliability and repeatability of pulp volumetric reconstruction through

three different volume calculations. Journal of Forensic Odontostomatology. Accepted

for publication. March 2017.

7. Chapter 7.


Franklin., E., Kruger, E., & Tennant, M. A new proposal to overtake population group

differences for dental age estimation in adults through pulp/tooth volume calculation.

Pilot Study. In press.

Student signature: Talia Yolanda Marroquin Peñaloza

15

The external supervisors, Shalmira Karkhanis, Sigrid Kvaal, and Sivabalan Vasudavan,

agreed with the contents and papers published in this thesis.

I, Marc Tennant certify that the student statements regarding their contribution to

each of the works listed above are correct

Coordinating supervisor signature: Estie Kruger

Date: 30th November, 2017.

16

GENERAL INTRODUCTION

Forensic science has been defined as the application of science to the law or legal

scenarios. In the same way, forensic dentistry refers to the application of dentistry to the

law to interpret dental and related evidence in criminalistics fields, having important value

for identification of individuals [1]. More, specifically, forensic dentistry is a specialized

branch of dentistry that uses specialised knowledge to collect, manage, interpret, evaluate

and present dental evidence for legal proceedings. This is applicable to dental

identification, sex determination, cheiloscopy and palatoscopy, molecular biomarkers, bite

mark analysis, human abuse and neglect, dental malpractice and negligence, dental

anthropology and archaeology, and the main scope of this thesis is concerned with age

estimation [2].

Age is one of the most important features to establish the identity of any individual, next to

sex, height, and ethnicity [3]. The need to confirm the age of different individuals has been

a crucial topic in anthropological, legal, and later in forensic scenarios. One forensic case

that has been stipulated as the first using dental age estimation with forensic purposes [1],

was narrated in the book “L’Art Dentaire en medicine legale” by Dr. Oscar Amoedo in 1898

[4], the father of forensic dentistry. This is related to the story of Louis-Charles Capet, or

Louis XVII of France, son of Louis XVI and Marie-Antoniette.

He passed away at the age of ten years two months, in prison, in 1795. However, the coffin

with his remains could not been found. It was believed that he had been substituted with

17

other boy his age, and that the prince was still alive; with this, a number of individuals

unsuccessfully claimed to be Louis XVII. In 1849, in the old cemetery of St Margarite a lead

coffin was found, with a skeleton, believed to belong to the young prince.

The first examinations, by Dr. Milcent, determined that these were the remains of Louis

XVII, but Dr. Recamier rejected this hypothesis, based on a dental exam, and claimed that

the individual was 15 to 16 years old. This little story, that could have been a fragment of a

fairy tale, is only one simple example that highlights the relevance of forensic dentistry and

the role of the forensic dentist to confirm the age and identity of different individuals.

Up to date, many methods have been proposed for age estimation, using different

indicators which vary somewhat, since the study of bone changes with maturation up to

more invasive techniques, based on chemical analysis of different tissue samples. All these

methods, have displayed a wide variability in their accuracy and reliability, which is even

more questionable when it comes to the analysis of individuals that have reached their

physical maturity.

Dentistry has contributed substantially to forensic sciences, in regards of age estimation,

having a relevant and positive impact, helping in the analysis and resolution of many cases

around the world. Among the different available methods for age estimation, those that

are minimally invasive are of interest owing to the major advantage of not requiring tooth

extraction, making them appropriate for the identification of unknown deceased

18

individuals, and also for the identification of living individuals with doubtful documents of

identification.

Teeth analysis for age estimation and identification of individuals has different advantages

over the analysis of osseous tissue, such the high resistance of teeth to decomposition and

taphonomic processes, which make them suitable for study long after the individual has

deceased [3]. Furthermore, age-related changes in teeth can be observed and studied on

dental radiographs, offering minimally invasive methods, which is not always possible with

bone analysis.

The genetically controlled process of tooth formation provides reliable data for the age

estimation of individuals, that have not completed tooth maturation. In this way, there are

different methods to estimate the age of individuals, from the 14th week of gestation, up to

the age of 21 years, where due to the analysis of third molars formation it is possible to

establish if they are over or under the age of 18 years [5].

Nevertheless, age estimation in individuals with their dentition, totally formed, is still a

challenge. Numerous methods have been proposed for dental age estimation in adults.

Although, in certain cases, the accuracy and reliability could be questionable, each one has

been an important step for the proposal of more modern, easier, and accurate methods.

In 1925, Bodecker [6] identified a dental change that is related with age and can be

observed once the root development has finished. This is the apposition of secondary

19

dentine, on the walls of the root canal. The formation of secondary dentine commences

once the tooth formation has finished, making the root canal or pulp chamber smaller [7].

In this way, a negative correlation between the age of the individual and the dimensions of

the root canal may be established [8].

In 1950, Gustafson [9] proposed a method that included the analysis of dental attrition,

periodontosis, cementum apposition, root resorption, root transparency, and the

formation of secondary dentine, which for his study, had to be observed on microscopic

sections. The methodology proposed a system of points depending on the status of the

different age indicators.

The sum of points and the relation with age was recorded; this revealed an increase of

points with age. With this, the author affirmed that it was possible to draw a regression

line for the relationship between age and points; however, this system was not applicable

for teeth in broken-down condition.

The author also recommended that those who chose to use this method had to create their

own standard curve. Since then, the statistical rigor of methods has noticeably increased,

as well as the number of participants in different studies. However, the use of linear

regression models has remained constant and is still an important method for the

development for age estimation in adults.

20

As the dimensions of the root canal can be measured in dental radiographs or dental

tomography, different methods have been proposed. In 1995 Kvaal et al. [8], published a

method based on linear measurements of tooth, root and pulp chamber length and the

measurement of the width of the tooth, and root canal at three different levels on the

roots. All these measures had to be taken from six different single rooted teeth in

periapical radiographs.

The original method demanded the use of Vernier calipers to measure the length, and a

stereomicroscope with a measuring eyepiece to the nearest 0.1mm with which to measure

the width of pulp and root. It was proposed that the calculation of pulp/tooth ratio

dimensions and the application of linear regression models were adequate to produce a

formula for age estimation in individuals older than 20 years.

Up until now, this method has been tested in at least 29 studies in 12 different countries;

some studies have modified or adapted the original method to the new emerging imaging

technology in dentistry. In this thesis, this method was tested in new scenarios and

adapted to new technologies.

Following the same principle of secondary dentine production with age and the negative

correlation between root canal size and age, Cameriere et al. [9], proposed a method based

on area measurements of root canal and tooth. As the available technology, at the time of

the elaboration of this method, was different, this method requires the use of digital

radiograph and the access to specific software.

21

This method has also been tested in at least 29 studies in 11 countries. It is also based on

the negative correlation of the pulp/tooth dimension ratio with age, but requires area

measurements. One of the last proposals, for age estimation in adults, is the calculation of

pulp and root canal volume from computed tomography (CT) or cone beam computed

tomography (CBCT).

This last method, is still under development. However, at least 16 studies from 7 countries

have been published using different techniques and software. In this thesis, some of the

published methodologies were analyzed, and a new method is proposed to estimate dental

age in adults. The above-mentioned methods, besides being based on the production of

secondary dentine with age, have another similar feature: the proposal of linear regression

models and equations to estimate the age of the individuals.

The use of the pulp/tooth ratio calculation, not only helps the researchers to observe how

the size of the root canal or pulp chamber changes with age, but also, as established by

Kvaal et al., it is useful to diminish the effects of image magnification [8], and probably sex.

On the other hand, the brief description of some of the different methods for dental age

estimation testifies to the strong relation between forensic dentistry and radiology. The

use of radiology in dentistry, and forensic dentistry dates back to 1895, when Wilhelm

Conrad Röntgen discovered the X-rays. One year later, in 1896, there were significant

advances for the application of radiology in dentistry.

22

Some of the most relevant contributors were Otto Walhoff, who obtained a radiograph of

his own maxillary sinuses, and a dental radiograph using dried human skulls. Furthermore,

Edmund Kells, who is recognized as a dental pioneer and scientific genius [10], introduced

radiology into dentistry, obtaining images from living patients.

Also in 1896, it was reported that one of the first uses of radiology for forensic applications

to confirm the presence of bullets inside the head of a woman, who had been attacked by

her husband [11]. For those days, to obtain a radiographic record, it was necessary to

expose the individual to X-radiation for up to 70 minutes [12]. During the following years,

the technology and new techniques in radiology have not only reduced the exposure times,

but also have provided forensic science with new and better techniques, to support the

resolution of different legal cases.

Now, in a digital era, plus the easy access to the Internet, to obtain, analyze, duplicate and

share images, it is faster than ever, promoting the proposal of new emergent methods for

identification and age estimation. This has allowed forensic teams, and academic

researchers to collaborate, using two and three-dimensional images. As it has been

described above, in forensic dentistry and age estimation, there have been major

developments since the time of the young prince, Louis XVII of France.

At this stage, it is important to clarify that those studies on dental age estimation for young

and older individuals have been based essentially, on the analysis of totally sound teeth,

and the vast majority of the published studies have obtained data from individuals who

23

belong to the same population. However, there are different challenges for dental age

estimation in adults that should be considered.

Firstly, there is no available systematic review that compares the accuracy of the above

mentioned three different methods for age estimation. This thesis presents the first

systematic review in forensic dentistry for age estimation in adults that follows the

parameters established by the Cochrane handbook for systematic reviews-methodology, in

the absence of guidelines for forensic sciences.

Secondly, it has been overlooked that the production of secondary dentine can be affected

by external factors, generating the production of tertiary dentine, the two of which are

undistinguishable in dental radiographs. Furthermore, it has been suggested that for

dental age estimation it is necessary to use only totally sound teeth, excluding those

individuals who have received orthodontic treatment, a dental procedure that is becoming

popular and that also alters pulp chamber size and tooth length, parameters used for the

application of the Kvaal et al. method.

Thirdly, tooth formation finishes, in general, at the age of 14 years. For age estimation of

older individuals (adolescents and sub-adults), the available methods are based on the

radiological analysis of third molar root formation. However, there is a problem with some

young individuals with agenesis of third molars.

Fourthly, despite the introduction of tomography in dentistry, there are no studies testing

the Kvaal et al. method on CBCT, an analysis that is useful to perform, as it could provide

24

more tools for dental age estimation in adults, performing linear measurements not only in

a coronal, but also a sagittal plane, inaccessible in routine dental radiographs.

Fifthly, although the introduction of tomography to dentistry and forensic dentistry has

promoted the proposal of different methods based on volume reconstruction, no study has

considered the different technological aspects, unique to each software and techniques

used to undertake the respective volume measurement.

Finally, there are distinct advantages of available imaging techniques, which allow

researchers to obtain volume measurements of any anatomic region, as well as the

advantages that telecommunication brings to share images and data.

The above factors, could be considered as challenges when different non-invasive methods

are used for dental age estimation in adults. The first aim of this thesis is to look for an

answer to these challenges. To this end, a series of 7 papers are presented and their

results are discussed in 4 different sections that this thesis encompasses.

25

REFERENCES

1. Senn DR. Forensic Dentistry. 2nd Ed. Boca Raton, FL: CRC; Press, 2009.

2. Rai B., & Kaur, J. Dental Age Estimation. In Evidence-Based Forensic Dentistry.

Springer Berlin Heidelberg; Press, 2012.

3. Franklin D. Forensic age estimation in human skeletal remains: Current concepts

and future directions. Leg Med. 2010;12:1-7.

4. Silva RF, Franco A. L'art dentaire en médecine legale. RBOL. 2016;3,1.

5. Senn DR, Weems RA. Manual of Forensic Odontology. 5th ed. Boca Raton, FL:

CRC; Press, 2013.

6. Bodecher CF. A consideration of some of the changes in the teeth from young to

old age. Dental Cosmos. 1925;67:543-9.

7. Panchbhai AS. Dental radiographic indicators, a key to age estimation.

Dentomaxillofacial Radiol. 2011;40:199-212.

8. Kvaal SI, Kolltveit KM, Thomsen IO, Solheim T. Age estimation of adults from

dental radiographs. Forensic Sci Int. 1995;74:175-85.

9. Cameriere R, Ferrante L, Cingolani M. Variations in pulp/tooth area ratio as an

indicator of age: a preliminary study. J Forensic Sci. 2004;49:317-19.

26

10. Jacobsohn P, Kantor M, Pihlstrom B. The X-ray in dentistry, and the legacy of C.

Edmund Kells A commentary on Kells CE. The X-ray in dental practice. J Natl

Dent Assoc 1920; 7:241-272. J Am Dent Assoc. 2013;144:15S-9S.

11. Thali MJ, Brogdon B, Viner MD. Forensic radiology, Second edition. CRC Press;

2002.

12. Chandrasekhar T, Vennila P. Role of Radiology in Forensic Dentistry. J Indian

Acad Oral Med Radiol. 2011;23:229-31.

27

SECTION ONE. Non-invasive methods for dental age estimation in adults: Systematic

Review.

Preface

Being one of the main issues in forensic and anthropological sciences, age estimation in

adults has been the target of researchers in the field. In the same way, the analysis of

dental age-related changes in adults has been widely explored, based on the clinical

observation of certain dental changes such as tooth attrition and periodontal recession,

which are not reliable age indicators.

The aim of this thesis is the analysis of non-invasive dental age estimation in adults, based

on the radiological measurement of the age-related dimensional change of the root canal,

due to secondary dentine production through life. To this end, a systematic review was

carried out, collecting all the available publications on linear measurements of tooth and

pulp width and length, area measurements of tooth and pulp, and also pulp/tooth volume

quantification from dental computed tomography.

The aim, methodological design and findings of this systematic review are presented in

Chapter 1. This systematic review is one notable achievement from this thesis as, to date,

there is no similar publication in forensic dentistry or in the anthropological science

literature.

28

1. Chapter 1. Age estimation in adults by dental imaging assessment systematic review

This chapter was published as:

Marroquin, T. Y., Karkhanis, S., Kvaal, S. I., Vasudavan, S., Kruger, E., & Tennant, M. Age

estimation in adults by dental imaging assessment. Systematic review. Forensic Science

International Journal. Accepted for publication, March 2017.

29

ABSTRACT

Importance: The need to rely on proper, simple, and accurate methods for age estimation

in adults is still a world-wide issue. It has been well documented that teeth are more

resistant than bones to the taphonomic processes, and that the use of methods for age

estimation based on dental imaging assessment are not only less invasive than those based

on osseous analysis, but also have shown similar or superior accuracy in adults.

Objectives: To summarise the results of some of the most recently cited methods for

dental age estimation in adults, based on odontometric dental imaging analysis, to

establish which methods are more accurate, accessible, and simple.

Evidence review: A literature search from several databases was conducted from January

1995 to July 2016 with previously defined inclusion criteria.

Conclusion: Based on the findings of this review, it could be possible to suggest pulp/tooth

area ratio calculation from first, upper canines and other single rooted teeth (lower

premolars, upper central incisors), and a specific statistical analysis that considers the non-

linear production of secondary dentine with age, as a reliable, easy, fast, and predictable

method for dental age estimation in adults. The second recommended method is the

pulp/tooth width-length ratio calculation. The use of specific population formulae would

be recommended, but to include data of individuals from different groups of populations in

the same analysis is not discouraged. A minimum sample size of at least 120 participants is

recommended to obtain more reliable results. Methods based on volume calculation are

time consuming and still need improvement.

Key words: Forensic dentistry, adults, age estimation, dental imaging, systematic review.

30

INTRODUCTION

Age is one of the most important characteristics used to establish the identity of any

individual in different legal, forensic, or anthropological research context [1]. To this end,

forensic teams depend on osseous analysis based methods, which have acceptable results

for young individuals or in their early adulthood [2], and dental development based

methods, which are highly reliable in individuals under 21 years of age [3].

However, these methods have some disadvantages: the poor resistance of bones to the

taphonomic process [4], and once the individual reaches the threshold of 21 years of age,

and third molars development concludes [3], the currently available dental development

based methods are not applicable. In individuals with the congenital absence of third

molar teeth, this threshold is not applicable for individuals older than 14 – 15 years of age.

To respond to the need of an aging population, and with the evident resistance of teeth to

the taphonomic process, alternative methods for dental age estimation in adults have been

proposed. Primarily, these are based on the formation of secondary dentine, studied since

1950 [5], and the subsequent narrowing of the pulp cavity, which can be observed in dental

radiographs, leading to the proposal of minimally invasive methods.

This systematic review focuses on three methods based on odontometric analysis of the

pulp cavity, performing length and width measurements [6], area measurements [7], and

lastly, volume calculation [8]. The objective of this review is to summarise the results of

31

these recently most cited methods for dental age estimation in adults, to establish which

method is more accurate, accessible, and simple.

Description of the problem

Different methods have been published for dental age estimation in adults, based on the

pulp/tooth dimensions’ ratios. Nevertheless, the results of the application of some of these

methods for dental age estimation, in adults from different population groups, exceeds the

accepted threshold in forensic sciences, which indicates that the standard deviation of a

method for adult’s age estimation should preferable be below a standard deviation (SD) of

years ±10 years [9].

Description of the methods being investigated

The methods for dental age estimation in adults, analysed in this paper, were selected

based on their minimally invasive nature, not a requirement for the extraction of teeth to

be performed, and pulp/tooth ratio calculation, which have been applied in individuals

from different populations. The Kvaal et al. [6] method is based on the analysis of linear

measurements of the pulp, tooth, and root length as well as root and pulp width

measurements at three different root levels, initially applied on periapical radiographs, and

later on panoramic radiographs and tomographs. The Cameriere et al. [7, 10] method is

based on the analysis of pulp and tooth area measurements on periapical, and panoramic

radiographs. Finally, different methods for dental age estimation in adults, based on

pulp/tooth volume ratio from cone beam computer tomography [8, 11-14], were included

in this systematic review.

32

How these methods might work

The methods in this study are based on a negative correlation between age and the pulp

chamber size, as well as on the tooth/pulp ratio calculation, regardless of the nature of the

measurements: length/width, area, or volume. In other words, all measurements evaluate

the association of one age related phenomenon, as it is the formation of secondary dentine

with the decrease of pulp chamber size with age. This has been accepted as an age

indicator that is observable and measurable with different dental imaging techniques.

Ideally, the accuracy of the methods should not exceed the threshold of a SD ±10 years [1,

4].

Why it is important to do this review

The relevance of this review is grounded on the need to recommend a method for dental

age estimation with the following characteristics: simple, fast, non-invasive, non-expensive,

and reproducible, and over all, accurate; that can be systematically used in different

academic and forensic scenarios. The objective is to aid the reconstruction of identity

profiles of unidentified deceased individuals or living individuals with doubtful identity

documents.

METHODS

Criteria for considering studies for this review

Qualitative analysis of the original human studies reporting the use of any of the listed

methods for dental age estimation, based on pulp/tooth ratio calculation (length/width,

33

area, volume) were included; in particular, those that preferably reported intra-inter

observer calibration, generating population specific formulae.

Studies with samples including individuals younger than 14 years of age were excluded, as

well as studies with small sample sizes (n<50), and those using extracted teeth. Also,

studies that did not report the use of specific population formulae were excluded for the

quantitative analysis.

Search methods for identification of studies

The information was searched though data-bases available at The University of Western

Australia, which included the collections of:

Directory of open access journals (DOAJ), Medline/PubMed (NLM), OneFile (GALE),

ProQuest, collection, (Web of Science), Science Direct Journals (Elsevier), Social Sciences

Citation Index (Web of Science), Scopus (Elsevier), Social Sciences Citation Index (Web of

Science), Wiley (Cross Referenced), Wiley Online Library. Also, Google Scholar, by papers

that referenced the original study performed by Kvaal et al., Cameriere et al., and those

methods for age estimation that referred to the use of CBCT and volume reconstruction of

pulp chambers and teeth.

The search key words were as follows: Kvaal and dental age estimation, Cameriere and

dental age estimation, age, and tooth volume. The literature search included publications

by these key authors after the publication of their original papers up until July 2016. The

34

search was conducted during the years 2014 to 2016. This systematic review follows the

Cochrane handbook for systematic reviews-methodology and, where possible, the RevMan

software recommended by the Cochrane handbook. (Graphs 1.1-1.3)

DATA COLLECTION AND ANALYSIS

Selection of studies

The initial selection was based on the title and abstract. Papers with titles that referred the

inclusion of minors only (children) were excluded. Studies reporting the use of third molars

or developing teeth were excluded. Studies that included the use of other methods for

dental age estimation in children were excluded (Demirjian), or any other invasive method

for dental age estimation were excluded. (Graphs 1.1-1.3, Tables 1.1-1.3)

Data extraction and management

The data were organized in an Excel spreadsheet as follows: Author, year, country, number

of participants (male and female), age, intra and inter-observer agreement assessment.

Also recorded were: the imaging technique, measuring instrument, best correlation

coefficient between age and the different age predictors, best result in terms of accuracy

per individual tooth or per set of teeth, when possible, as well as the highest error recorded

by a set of teeth or individual tooth (Table 1.4-1.6).

Assessment of risk of bias in included studies

To avoid bias in this systematic review, and to avoid false positive or false negative

35

Graph 1.1. Flow chart of the study selection for the Kvaal et al. method.

36

Graph 1.2. Flow chart of the study selection for the Cameriere et al. method

37

Graph 1.3. Flow chart of the study selection for methods based on volume calculation.

38

Table 1.1. Total studies reporting the use of the Cameriere et al. method.

Author Year Country Total* Age

Kvaal et al. [6] 1995 Norway 100 20-87

Kvaal et al. [16] 1999 Sweden 21 20-60

Willems et al. [9] 2002 Belgium 29 26-85

Soomer et al. [1] 2003 Caucasian 20 14-95

Bosmans et al. [20] 2005 Belgium 197 19-75

Paewinsky et al. [19] 2005 Germany 168 18-81

Meinl A. [31] 2007 Austria 44 13-24

Avendaño et al. [18] 2009 Colombia 107 21-50

Landa M. [32] 2009 Portugal 100 14-60

Shetty et al. [15] 2010 India 100 20-70

Sharma et al. [71] 2010 India 50 15-60

Saxena. [30] 2011 India 120 21-60

Saxena, Tiwari [40] 2011 India 160 21-60

Chandramala et al. [21] 2012 India 100 20-70

Kanchan et al. [35] 2012 India 100 25-77

N. Agarwal. [17] 2012 India 50 20-70

Erbudak et al. [29] 2012 Turkey 123 15-57

Thevissen et al. [33] 2012 Belgium 450 15-23

Limdiwala et al. [22] 2013 India 150 20-55

Parkin et al. [28] 2013 India 30 15-60

Kostenko. [72] 2013 Ukraine 64 ---

Karkhanis et al. [24] 2014 Australia 279 20-62

Misrlioglu et al. [25] 2014 Turkey 114 17-72

Patil et al. [69] 2014 India 200 20-50

Edward et al. [23] 2014 Sudan 99 15-30

Muszynska et al. [34] 2015 Poland 3 ---

Marroquin et al. [26] 2016 Australia 74 12-28

Mittal et al. [27] 2016 India 152 14-56

Rajpal et al. [70] 2016 India 50 15-57

29 papers 12 Countries 3254 12-95

*Number of individuals reported in each study. Sample size.

39

Table 1.2. Total studies reporting the use of the Cameriere et al. method.

Author Year Country Total Age

Cameriere et al. [7] 2004 Italy 100 18-72

Cameriere et al. [52] 2006 Italy 33 ---



Cattaneo et al. [64] 2008 Ethiopia 1 52

Cameriere et al. [37] 2009 Portugal and Italy 229 20-84

Singaraju et al. [38] 2009 India 200 18-72

Babshed. [4] 2010 India 178 20-70

De luca et al. [2] 2010 Spain – Italy 73 ---

Babshed et al [50] 2011 India 61 21-71

Cameriere et al [73] 2011 Italy 90 50-79

Jeevan et al. [39] 2011 India 228 16-72

Saxena [30] 2011 India 120 21-60

Vodanovic et al. [65] 2011 Croatia 192 ---

Zaher et al. [41] 2011 Egypt 144 12-60

De luca et al. [51] 2011 Mexico 85 18-60

Cameriere et al. [42] 2012 Spain 606 18-75

Cameriere et al. [46] 2013 Portugal 116 18-74

Charis et al. [43] 2013 India 120 20-70

Cameriere et al. [46] 2013 Portugal 116 18-74

AC Azevedo. [44] 2014 Italy 81 19-74

Misirlioglu et al. [25] 2014 Turkey 114 17-72

Azevedo et al. [44] 2015 Brazil 443 20-78

Ravindra et al. [47] 2015 India 308 9-68


De Angelis et al. [67] 2015 --- 1 ---

Fabbri et al. [66] 2015 Italy 18 ---

Sakhdari et al. [48] 2015 Iran 120 >12

Torkian. [75] 2015 Iran 120 >12


*Number of individuals reported in each study: Sample size.

40

Table 1.3. Total studies reporting the use of volume calculation.

*Number of individuals reported in each study: Sample size.

Author Year Country Total* Age

Vandervoort et al. [8] 2004 Belgium 52 24-66

Yang F. et al. [11] 2006 Belgium 19 23-70

Someda et al. [12] 2009 Japan 155 12-79

Agematsu et al. [61] 2010 Japan 258 teeth 20-79

Aboshi et al. [13] 2010 Japan 50 20-78

Star et al. [55] 2011 Belgium 111 10-65

Tardivo et al. [56] 2011 France 58 14-74

Jagannathan et al. [14] 2011 India 188 10-70

Sakuma et al. [53] 2013 Japan 136 14-79

Tardivo et al. [57] 2014 France 210 15-85

Sasaki et al. [54] 2014 Japan 363 15-70

Mendonca et al. [60] 2015 Brazil 72 22-70

Ge et al. [62] 2015 China 403 12-69

De Angelis et al. [58] 2015 Italy 91 17-80

Pinchi et al. [59] 2015 Italy 148 10-80

Ge et al. [59] 2016 China 240 16-63


41

Table 1.4. The Kvaal et al. method: Studies included in the quantitative analysis.

Study Sample Measuring instrument SEE ± years per tooth or group of teeth (FDI)

6 teeth Upper Lower 11/21 12/22 15/25 32/42 33/43 34/44 11/21 32/42

13/23

Kvaal [6] 1995, Norway.

100 PER 20-87 years

Stereomicroscope and Vernier calipers.

8.6 8.9 9.4 9.5 10 11 10.5 11.5 11.5

Bosman [20] 2005, Belgium.

197 OPG 19-75 years

Adobe Photoshop software 9.5 9.2 9.9 9.7 9.8 9.3 11.6 8.2 8.1

Paewinsky [19] * 2005, Germany.

168 OPG 14-81 Years

Hipax program 5.6

6.4

Avendano [18] 2009, Colombia

107 PER 21-50 years

Scion Image software

7.1

Shetty [15] 2010, India.

100 PER 20-70 years

SV- 4 mini slide viewer, digital Vernier calipers, stereomicroscope

10.5 10.2 11.8 11.5 12.4 11.8 12.7 13.3 11.8

Erbudak [29] 2012, Turkey.

123 OPG 14-57years

Image J software 10 10.1 8.7 8.82

Saxena**[40] 2011, India.

120 OPG 21-60 years

Autocad2005 software 3.63

Kanchan [35] 2012, India.

Group A, 47 PER 25-75 years

Adobe Photoshop software

12 12 12.4 12.7 13.2 11.8 13.3

13.2

Group B, 43 PER 25-77 years

11.9 11.2 12.4 11.1 13.4 12.6 13.4 13.8 12.7

Limdawala [22] 2013, India.

Group A, 100 PER 25-50 years

Kodak Dental Imaging Software 8.3 8.2 9

Group B, 50 PER 25-50 years

9.4 9.5 9.8

Misirlioglu [25] 2014, Turkey.

144 OPG 17-72 years

Easy-Dent PC software

5.8

7.3 7.5 6.9

Patil [69] 2014, India.

200 PER 20-50 years

Image-Pro Plus II software

6.5

Karkhanis [24] 2015, Australia.

200 OPG 20-62 years

Image J software 8.9 9.6 8.3 9.3 9.6 9.5 10.2 10.9 10.5

Mittal [27] 2016, India. 152 OPG 14-60 years Vistascan DBSWIN software 7.9 8.5 7.5 8.1 8.5 7.8 8.8 7.9 7.5

Rajpal [70] 2016, India 50 PER 15-57 years Kodak Dental Imaging Software 6.4 7.3 7.8

*Error reported as standard deviation. **Error reported as the difference between the chronological age and the estimated. PER=Periapical radiographs. OPG=Panoramic radiographs. Tooth numeration FDI

42

Table 1.5. The Cameriere et al. method: Studies included in the quantitative analysis.

*Error reported as mean absolute error (MAE), **Error reported as the module of the differences between the chronological and estimated age, ***Error reported as mean error, PER=Periapical radiographs. OPG=Panoramic radiographs. Tooth numeration FDI (Federation Dentaire International)

Study Sample Measuring instrument SEE ± years per tooth (FDI)

32/42 11/21 12/22 31/41 32/42 33/43 34/44 35/45 13/23

Cameriere et al. [7]* 2004, Italy.

100 OPG 18-72 years

AutoCAD2000

3.27

Cameriere et al. [37] 2009, Portugal and Italy.

229 PER 20-84 years

Adobe Photoshop

4.33

4.24

Babshed. [4] 2010, India.

178 PER 20-70 years

Adobe Photoshop AutoCAD 2004

10

10

Babshed et al. [50] 2011, India.

61 PER 21-71 years

Adobe Photoshop AutoCAD 2004

12.22

12.28 13.08 12.45

Jeevan. [39]* 2011, India.

228 PER 16-72 years

Adobe Photoshop

6.39

4.28

Zaher et al. [41] 2011, Egypt.

144 PER 12-60 years

AutoCAD2008

2.63 1.94

Cameriere et al. [42] 2012, Spain.

606 OPG 18-75 years

Adobe Photoshop

6.38 5.75

Cameriere et al. [46] Portugal. 2013

116 PER 18-74 years

Adobe Photoshop

7.03 6.64 10.8 10.9

Charis et al [43]* 2013, India.

120 PER 20-70 years

Jenoptic ProgRess Version ss2.7.

5.4

AC Azevedo [44]** 2014, Italy.

81 PER 19-74 years

Jenoptic ProgRess Version ss2.7.

3.05

Misirlioglu et al. [25] 2014, Turkey.

114 PER 17-72 years

Adobe Photoshop

6.75

Azevedo et al. [44] 2015, Brazil.

443 PER 20-78 years

Adobe Photoshop

6.41

5.79

43

Table 1.6. Volume calculation based methods.

Author Sample Image segmentation method Measuring instrument

Accuracy pert tooth (FDI)

Star et al. [55] 2011, Belgium.

111 CBCT 10-65 years

automatic segmentation (threshold), and manual correction

Simplant® 11/21 SD=12.8 13/23 SD=13.1 15/25 SD=8.4

Tardivo et al. [57] 2014, France.

210 CT 15-85 years

Semiautomatic Mimics® 13/23 MAE=3.4 33/43 MAE=4.6

De Angelis et al. [58] 2015, Italy.

91 CBCT 17-80 years

Manual Osirix software prediction interval = ±28 years

Pinchi et al. [59] 2015, Italy.

148 CBCT 10-80 years

cone shape approximation Osirix® software SEE=±11.45

CBCT: Cone beam computed tomography. SD: Standard deviation. MAE: Mean absolute error. SEE: Standard error of estimation. Tooth numeration FDI (Federation Dentaire International)

44

conclusions, it was necessary to analyse the possibility of author bias. This was determined

by the participation of the same authors in repeated publications. To this end, the results

were analysed comparing individual papers, and then grouping them per author.

Additionally, in certain cases where there were doubts about the origin of the sample,

which means the likelihood of finding different studies by the same authors that had used

the same sample, the authors were contacted to confirm the origin of the sample. In case

that two studies had the same sample, the authors were asked to suggest which study

should be included in the meta-analysis. Likewise, to deal with missing data, the authors

were contacted for more information. There was another possible source of bias observed

in this study: the use of non-specific population regression models and equations, which

could generate the wrong assumption on the accuracy of the analysed methods. To

overcome this problem, only studies using specific population formulae were used in the

quantitative analysis. However, there is not enough evidence to state that the production

of secondary dentine varies among individuals of different population.

RESULTS.

Kvaal et al. [6] (29 papers, 3254 participants)

Qualitative analysis

The original paper on the Kvaal et al. [6] method was published in 1995. Since then, this

method has been applied to a global sample of 3254 individuals, from 12 different

45

countries (Graph 1.1, Table 1.1). The initial methodology suggested the use of periapical

radiographs from six different teeth, using Vernier calipers to measure the maximum tooth

length, pulp length and root length, and a stereomicroscope with a measuring eyepiece to

the nearest 0.1mm, to measure pulp and root width at three different levels previously

described by Kvaal et al. [6].

Only a few studies have followed the original methodology, with slight adaptations [1, 6, 9,

15-18]. In 2005, this method was applied to panoramic radiographs [20], which became

more popular [20-34]. The use of digital imaging in dentistry also introduced the use of

different software to perform the odontometric measurements.

The most commonly used software programs are: Adobe Photoshop (different versions

n=5), Image J (n=4) and Kodak dental imaging software (n=2). From those studies that

tested the effect of sex on the accuracy of the age estimated: three found that sex does not

affect the models [15, 19, 30]; and two that it does [18, 23]. In the original study, sex was

included as a factor for the mandibular lateral incisor, indicating that pulp cavity size

changes occurred faster in males, as females need another 6 years to get the same age as

males for this tooth [6].

Quantitative analysis

From the total sample, only 16 papers met the inclusion criteria for the quantitative

analysis. (Table 1.4). The most accurate result (SD=5.6, r=-0.95) was obtained from

panoramic radiographs using the Hipax program (version 3.01) software, 168 participants

46

(77 female and 91 male) in a Caucasian group [19]. The largest error reported error was

SEE >13 years r2=0.1 [15, 35] obtained from the lower canine, following original Kvaal et al.

methodology on periapical radiographs [35] and the Adobe Photoshop 6.0 software in

panoramic radiographs [15].

Kvaal et al. measurements have been reported to have a high degree of intra and inter-

observer agreement indicating the reproducibility of the measurements, even in those

studies that did not follow the original methodology. Only one study reported significant

differences in the root length and pulp width at the level B and root width at the level A

measurement, although this study reported a SEE=±13.8 (r2=0.38), the average error of age

estimation was ±18-21 years using periapical radiographs. [35]

Note: The studies that did not meet the inclusion criteria for the quantitative analysis also

reported high intra and inter-observer agreement [31-33]. One study reported that the

lower lateral incisor produced lower intra-observer correlation [32], and one affirmed that

the use of a stereomicroscope improved the accuracy of the age estimates [9].

Cameriere et al. [7] (29 papers, 4167 participants)


The original paper documenting the Cameriere et al. method [7] was published in 2004. In

the present systematic review, 29 papers using this method were studied, having a global

sample of 4167 individuals from 11 countries (Graph 1.2, Table 1.2). This method was

47

initially applied to digital panoramic radiographs, using (AutoCAD2000, Install Shield3.0,

1997) and different versions of this software have been used in 11 studies. The most used

software, to perform the pulp/tooth area measurements, was Adobe Photoshop (13

studies).

This method was designed to be applied to single rooted teeth, especially canines, but it

has also been tested in premolars, central, and lateral incisors. From those studies that

assessed the effect of sex on the regression models, the vast majority reported that sex

does not affect the regression model [4, 7, 25, 30, 36-45]. Only 4 studies reported that sex

affects the regression model [46-49].


Twelve studies met the inclusion criteria for the quantitative analysis (Table 1.5). The use

of this method has a standard error of estimation (SEE) from±1.2 [41] to ±12-13years

(r=0.2-0.4) [50]. From these studies, two reported a median error <±5 years [7, 37], six

reported an error from ±5 to ±8.5 years [25, 39, 41-43, 45], and three studies reported an

error from ±10 to ±13 years [4, 46, 50]. From the studies included in the qualitative

analysis, only one reported significant intra-observer differences in lateral incisors and first

premolars (p<0.05) [50].

Note: All the studies that did not meet the inclusion criteria for the qualitative analysis and

that included intra and inter-observer calibration, reported high intra/inter-observer

agreement [36, 48, 51, 52].

48

Volume (16 papers, 2444 participants)


Owing to the relatively few studies doing pulp/tooth volume reconstruction, in the

qualitative analysis those studies using extracted teeth were included. The first study doing

pulp/tooth volume reconstruction reported the use of micro-focus X-ray in 2004 [8]. In this

current systematic review, 16 studies used the pulp/tooth volume reconstruction, by

means of micro-focus computer tomography, CT scan or CBCT, and the use of different

software to do manual or semiautomatic volume reconstruction, such as Tri/3D Bon,

Mimics®, ITK-SNAP 2.4, Osirix, Amira among others, in a global sample of 2444 individuals

from 7 countries (Graph 1.3, Table 1.3).

However, the methodology used by some studies, required the use of extracted teeth [8,

12, 13, 53, 54]. With regard to the effect of sex on the regression models, nine studies

reported that sex has no effect [8, 14, 25, 53, 55-59]. Three reported that the models are

more accurate in females than in males, having a higher determination coefficient for

women than for men [12, 60, 61]. The comparative data are as follows: r2=0.6 for males

and r2=0.7 for females [12], or r2=0.29 for males and r2=0.15 for males [60], and r2=0.67 for

males and r2=0.75 for females when the mandibular central incisor is used, or r2=0.56 for

males and r2=0.58 for females when the second premolar is used.

Two studies, which reconstructed only the pulp cavity volume, also reported significant

differences between tooth type and sex [62, 63]. One of them found a significant

49

difference in the pulpal volume between both genders (p=0.028 <0.05), and a stronger

relation between pulp chamber and age for females (r2=0.6) than for males (r2=0.5) [62].

The other one reported a signifficant difference in volume between the two genders for 12

types of teeth (p=0.000) [63].


Four studies meet the inclusion criteria for the quantitative analysis. The error of these

studies varies between ±3.47 years [57] to ±28 years [58]. A high intra-observer agreement

was reported [59]. Among those studies that did not meet the inclusion criteria for the

quantitative analysis, seven also reported high intra- and inter-observer agreement [8, 11-

14, 56, 62, 63].

DISCUSSION

Age estimation in adults is a challenge in all forensic investigations, especially in cases that

require the use of non-invasive methods. The formation of secondary dentine and the

resultant non-linear narrowing of the root canal with age [11], is one age predictor

measurable in dental radiographs and tomographs, leading to the proposal of different

methods for age estimation in adults as an alternative to more invasive methods, and as a

complement to osseous analysis [2, 16, 34, 64-67].

In the same way, many methods have been suggested for dental age estimation in adults.

Nevertheless, there are few available studies that compare their accuracy [68].

50

In the lack of a consensus on a uniformly applicable method for age estimation in adults,

the need to perform a systematic review was timely. This systematic review summarizes

and compares the results of some of the most used methods for dental age estimation in

adults.

In terms of the qualitative analysis, it has been reported that there are no significant

differences between right or left teeth [14, 56]. Another aspect to test among the different

papers was the effect of sex of the different methods. Although not all the included studies

assessed the effect of sex on age estimation, those that did (n=38), 71% (n=27) affirm that

sex has no statistically significant effect on age estimation, when pulp/tooth ratio is

calculated. However, if only pulp volume is measured and used as age indicator, there is a

significant difference between the accuracy for males and females [62, 63].

In the light of the above evidence one could suggest that ratio calculation not only

diminishes the effects of tooth magnification in dental radiographs, as reported by Kvaal [6]

et al., but also reduces the effect of sexual dimorphism related to tooth size. The age of

the participants is also another relevant aspect to include.

In this systematic review, those studies including participants under 14 years of age were

excluded for two main reasons. Firstly, there are other methods more reliable for aging

teenagers and infants, and secondly, before 14 years of age, not all the teeth have

completed the apex closure, which by definition, is a requirement for the formation of the

secondary dentine [76].

51

For the qualitative analysis, characteristics related to sample size and inclusion criteria

established in different studies were compared. As compared to previous efforts, no

significant correlation between sample size and the accuracy of the method was found,

when all the data were analysed together (r2=0.1, p-value <0.05, sample size of the

analysed studies n=50 to 604) or when the data were analysed individually for each

method.

It is necessary to highlight that from the studies included in this current analysis (n=30) only

4 (13.3%) had a sample smaller than 100 individuals (n=50 to 91 individuals), which may

suggest that in any study for age estimation the minimum sample should include data from

at least 100 participants, or over 100 teeth from different individuals in those studies using

only one tooth type. Some of the excluded studies, for the quantitative analysis, fail in this

aspect [55, 56, 60, 61], including data from several teeth of the same individual and

counting them as different samples, which may affect the reported results.

A previous study recommended a minimum of 120 participants in order to achieve 80% of

power and 5% of significance [30]. One almost universal inclusion criterion was the use of

only totally sound teeth. Only two studies did not follow this parameter, using teeth with

lesions to do volume calculations [54], or using panoramic radiographs that did not meet

the inclusion criteria suggested by Kvaal et al. [22]. Both studies found no significant effect

of these situations on the accuracy of the results.

52

Another aspect to consider is the description of the different methods, in terms of the

procedure to measure the different pulp/tooth dimensions, as well as the statistical

processing of the data. In certain cases, it was necessary to re-read the papers several

times to understand the proposed methodology and how the accuracy was reported.

In the quantitative analysis two main characteristics were compared among the studies:

repeatability of the measurements, in terms of intra and inter-observer agreement; and

their accuracy, reported as standard error of estimation (SEE), mean absolute error (MAE)

or standard deviation (SD). Due to this variation in the approach to evaluate and to report

the obtained accuracy, it was not possible to perform a deeper statistical analysis, as a

meta-analysis to confirm that one method was superior to another. However, in this

review the studies using pulp/tooth area ratio reported a lower error (table 1.4, 1.5 and

1.6).

All the studies reported significant intra- and inter-observer agreement, regardless of the

method and the measuring instrument, which suggests that after adequate training any of

these methods is reproducible. However, with regard to the simplicity of the technique,

and the access to the required tool to measure and calculate the respective pulp/tooth

ratio dimensions, Kvaal et al., and Cameriere et al. are more convenient.

By using periapical radiographs or panoramic radiographs, as the radiological technique

does not affect the accuracy of the measurements, as long as the quality of the image

53

allows the observer to clearly observe the boundaries of the root canal and tooth surface

[22].

The volumetric reconstruction of anatomic structures involves more technical and time-

consuming training as well as the use of more complex software. These factors could be

considered as obstacles for the larger scale application of some of these methods.

Likewise, the user spends more time doing the volumetric reconstruction per tooth, which

varies from 4 hours [8] to 15 minutes [60], when reported.

Another important aspect, with regard to the accuracy of the method, is the use of specific

population formulae, which has been believed to improve the results of the age estimates,

independently of the used method [14, 29, 41, 45, 63, 69]. As in those studies, using non-

specific population formulae the reported error was notably larger (error > ±20 years) [31,

32]. But the most recent findings about age mimicry, explained in Chapter 7, could reveal

that this assumtion is called into question [77].

The use of data from different populations in the same statistical analysis, to generate a

unique age estimation equation, is achievable [2, 20, 46]. Furthermore, it has been

reported that the pulp chamber size variation can be detected only each 10 years, which

opens a question about the reliability of those methods reporting a lower error.

In the specific analysis by the Kvaal et al. method, although requiring the inclusion of six

different single rooted teeth per idividual, the use of only one tooth [18], only upper

54

central incisor [69], canines [25], mandibular teeth [32], or different combinations of teeth

to apply the odontometric analysis described by Kvaal et al [24, 26], has also been

reported, with acceptable results (SEE < ±10 years).

It has also been reported that its accuracy depends on the quality of the image and on the

precision of the measurements [22]. Several studies reported that tooth length is not

strongly correlated with age [15, 17, 19, 22], as tooth length would depend on tooth wear,

bruxism, and food habits, rather than a physiological aging related process [15].

In regard to the Cameriere et al method, it only requires the use of one tooth, and the

majority of the studies report the use of canines. However, one study using three different

lower teeth (lateral incisor, canines, first premolars) found that lower canines had the

poorest correlation coefficient with age (r2=-0.2) [50]. Another study found more accurate

estimations from upper lateral incisors, claiming that the narrowing velocity in the pulp

chamber of this tooth is twice as fast as in lower incisors [46].

The Cameriere et al. method reported the lowest error (total average ±5.6 years), and a

sample of at least 120 individuals was reported as being adequate to obtain more accurate

age estimates [30].

The combined use of the Kvaal et al. width ratios and Cameriere area ratios has been

reported on canines measured from digital panoramic radiographs. More accurate age

estimates were achieved when the pulp/tooth width ratio at the mid-root level (i.e. half-

55

way between the enamel-cementum junction, and the apex of the root level) and

pulp/tooth area ratio were used together [30], or when the area ratio calculation from

canines is used individually [25].

In this systematic review the following advantages about the use of pulp/tooth area

calculation over other methods were noted: area measurements on digital radiographs

with different software are faster; and the most recent study using this method proposes

software for the automatic selection of pulp and tooth borders. These steps minimize the

time taken to obtain the area of tooth and pulp chamber, and reduce the error associated

with the observer, when performing the area selection [74].

Linear measurements can also be performed on digital radiographs, but, the observer

needs to take nine measurements per tooth, in six teeth, and calculate several ratios. In

the method using pulp/tooth area measurements, only one ratio needs to be calculated

and allows the researcher to develop the respective statistical regression model. In the

same way, the use of canines, especially upper canines, to calculate pulp/tooth area ratio

has certain advantages, as facilitated by their longer survival, less wear, and the larger size

of the pulp chamber in comparison with other teeth [58]. Also, as observed in table 1.5,

there are not only more studies reporting the use of upper canines for pulp/tooth area

calculations but also, these studies present acceptable accuracy.

This is in contrast to what was reported by Kvaal et al when doing pulp/tooth length/with

ratio calculations; they observed that upper canines had the lowest correlation with age,

56

when using dental radiographs of extracted teeth [76], which was an exclusion criterion, in

this current systematic review.

Finally, although the use of pulp/tooth volume ratio calculations from non-extracted teeth

still requires more research. The initial results of the analysis of pulp volume from

individuals of different age groups provide important and detailed information to

understand the formation of secondary dentine.

It has been reported that, the pulp/tooth volume ratios in the cervical area, were more

correlated with age, and that this correlation decreased towards the tooth apex [13]. Also,

that the most marked reduction in volume ratio was observed between the second and the

fifth decades of life in lower first and second premolars, and between the second and the

third decades of life in lower first premolars [13]. These clear findings may enlighten the

prospect of future and more accurate methods for age estimation in adults.

CONCLUSION

Narrowing of the root canal caused by the formation of secondary dentine is a well-

accepted age indicator in adults. This systematic review observed that age estimation

methods based on pulp/tooth area ratio calculation provided more accurate results, even

when one tooth is analysed per individual. However certain conditions need to be fulfilled:

a sample of minimum 120 individuals; older than 14 years of age; assessment of the

57

accuracy of the observer. The inclusion of data of individuals from different populations in

the same analysis is not discouraged.

Future studies for dental age estimation in adults, must consider the non-linear deposition

of secondary dentine through life in their statistical analysis, and also that the use of linear

regression models would generate the over-estimation of age in young individuals and the

under-estimation of age in older individuals [78]. This systematic review also recommends

the use of dental age estimation methods: firstly pulp/tooth area ratio calculation of single

rooted teeth: upper canines and other teeth (lower premolars, upper central incisors); and

secondly pulp/tooth length/with ratio calculation, following the methodology proposed by

Kvaal et al.

These factors need to be considered in combination with other methods that include

diverse age indicators to produce more reliable age estimates. The author of this

systematic review also recommends future studies to report their results in terms of

standard deviation, mean absolute error and standard error of estimation simultaneously.

Such a display of data will facilitate more meaningful meta-analyses in future reviews.

58

REFERENCES

1. Soomer H, Ranta H, Lincoln MJ, Penttila A, Leibur E. Reliability and validity of

eight dental age estimation methods for adults. J Forensic Sci. 2003;48(1):149-

52.

2. De Luca S, Alemán I, Bertoldi F, Ferrante L, Mastrangelo P, Cingolani M, et al.

Age estimation by tooth/pulp ratio in canines by peri-apical X-rays: reliability in

age determination of Spanish and Italian medieval skeletal remains. J Archaeol

Sci. 2010;37(12):3048-58.

3. Shahin KA, Chatra L, Shenai P. Dental and craniofacial imaging in forensics.

JOFRI. 2013;1(2):56-62.

4. Babshet M, Acharya AB, Naikmasur VG. Age estimation in Indians from

pulp/tooth area ratio of mandibular canines. Forensic Sci Int. 2010;197(1–

3):125.e1-.e4.

5. Gustafson G. Age determination on teeth. J Am Dent Assoc. 1950;41(1):45-54.


dental radiographs. Forensic Sci Int. 1995;74(3):175-185.


indicator of age: a preliminary study. J Forensic Sci. 2004;49(2):317-19.

59

8. Vandevoort FM, Bergmans L, Van Cleynenbreugel J, Bielen DJ, Lambrechts P,

Wevers M, et al. Age calculation using X- ray microfocus computed

tomographical scanning of teeth: a pilot study. J Forensic Sci. 2004;49(4):787.

9. Willems G, Moulin-Romsee C, Solheim T. Non-destructive dental-age calculation

methods in adults: intra- and inter-observer effects. Forensic Sci Int.

2002;126(3):221-6.

10. Cameriere R, Ferrante SS, Belcastro MG, Bonfiglioli B, Rastelli E, Cingolani M.

Age Estimation by Pulp/Tooth Ratio in Canines by Mesial and Vestibular Peri-

Apical X-Rays. J Forensic Sci. 2007;52(5):1151-5.

11. Yang F, Jacobs R, Willems G. Dental age estimation through volume matching of

teeth imaged by cone-beam CT. Forensic Sci Int. 2006;159:S78-S83.

12. Someda H, Saka H, Matsunaga S, Ide Y, Nakahara K, Hirata S, et al. Age

estimation based on three-dimensional measurement of mandibular central

incisors in Japanese. Forensic Sci Int. 2009;185(1–3):110-4.

13. Aboshi H, Takahashi T, Komuro T. Age estimation using microfocus X-ray

computed tomography of lower premolars. Forensic Sci Int. 2010;200(1–3):35-

40.

14. Jagannathan N, Neelakantan P, Thiruvengadam C, Ramani P, Premkumar P,

Natesan A, et al. Age estimation in an Indian population using pulp/tooth

volume ratio of mandibular canines obtained from cone beam computed

tomography. J Forensic Odontostomatol. 2011;29(1):1-6.

60

15. Ranjani S, Ashok L, Sujatha GP. Age estimation in adults using intra oral

periapical radiographs in Indian population using Kvaal's method. Med Legal

Update. 2010;10(2):73-7.

16. Kvaal SI, During EM. A dental study comparing age estimations of the human

remains from the Swedish warship Vasa. Int J Osteoarchaeol. 1999;9(3):170-81.

17. Agarwal N, Ahuja P, Sinha A, Singh A. Age estimation using maxillary central

incisors: A radiographic study. J Forensic Dent Sci. 2012;4(2):97.

18. García GA, García YM, Velásquez LDE. Estimación de la edad por aposición de

dentina secundaria en una muestra de la población de Bogotá entre 21 y 50

años de edad (Spanish). Ages estimation base on seconday dentine aposition in

a sample from Bogota between 21 and 50 years old. Univ Odontol.

2009;28(60):29-38.

19. Paewinsky E, Pfeiffer H, Brinkmann B. Quantification of secondary dentine

formation from orthopantomograms—a contribution to forensic age estimation

methods in adults. Int J Legal Med. 2005;119(1):27-30.

20. Bosmans N, Ann P, Aly M, Willems G. The application of Kvaal's dental age

calculation technique on panoramic dental radiographs. Forensic Sci Int.

2005;153(2–3):208-12.

21. Chandramala R, Sharma R, Khan M, Srivastava A. Application of Kvaal's

Technique of Age Estimation on Digital Panoramic Radiographs. Dentistry.

2012;2012.

61

22. Limdiwala PG, Shah JS. Age estimation by using dental radiographs. J Forensic

Dent Sci. 2013;5(2):118-22.

23. Ayad E, Hamid HM, Abdalla EA, Kajoak SA. Estimation of age for Sudanese adults

using Orthopantomographs. GJMR. 2014;14(1).

24. Karkhanis S, Mack P, Franklin D. Age estimation standards for a Western

Australian population using the dental age estimation technique developed by

Kvaal et al. Forensic Sci Int. 2014;235(0):104.e1-.e6.

25. Misirlioglu M, Nalcaci R, Adisen MZ, Yilmaz S, Yorubulut S. Age estimation using

maxillary canine pulp/tooth area ratio, with an application of Kvaal’s methods

on digital orthopantomographs in a Turkish sample. Aust J Forensic Sci.

2014;46(1):27-38.

26. Marroquin T, Karkhanis S, Kvaal S, Vasudavan S, Castelblanco E, Kruger E, et al.

Determining the effectiveness of adult measures of standardised age estimation

on juveniles in a Western Australian population. Aust J Forensic Sci. 2016:1-9.

27. Mittal S, Nagendrareddy SG, Sharma ML, Agnihotri P, Chaudhary S, Dhillon M.

Age estimation based on Kvaal's technique using digital panoramic radiographs.

J Forensic Dent Sci. 2016;8(2):115.

28. Parikh N, Dave G. Application of kvaal's dental age estimation technique on

Orthopantomographs on A Population of Gujarat–A Short Study. Bujod.in 2013.

62

29. Erbudak HO, Ozbek M, Uysal S, Karabulut E. Application of Kvaal et al.'s age

estimation method to panoramic radiographs from Turkish individuals. Forensic

Sci Int. 2012;219(1-3):141-6.

30. Saxena S. Age estimation of indian adults from orthopantomographs. Braz Oral

Res. 2011;25:225-9.

31. Meinl A, Tangl S, Pernicka E, Fenes C, Watzek G. On the applicability of

secondary dentin formation to radiological age estimation in young adults. J

Forensic Sci. 2007;52(2):438-41.

32. Landa MI, Garamendi PM, Botella MC, Aleman I. Application of the method of

Kvaal et al. to digital orthopantomograms. Int J Legal Med. 2009;123(2):123-8.

33. Thevissen P, Galiti D, Willems G. Human dental age estimation combining third

molar(s) development and tooth morphological age predictors. Int J Legal Med.

2012.

34. Lorkiewicz-Muszyńska D, Przystańska A, Kulczyk T, Hyrchała A, Bartecki B,

Kociemba W, et al. Application of X-rays to dental age estimation in medico-

legal practice. Arch Med Sadowej Kryminol. 2015;65(1):1.

35. Kanchan-Talreja P, Acharya AB, Naikmasur VG. An assessment of the versatility

of Kvaal's method of adult dental age estimation in Indians. Arch Oral Biol.

2012;57(3):277-84.

63

36. Cameriere R, Ferrante L, Belcastro MG, Bonfiglioli B, Rastelli E, Cingolani M. Age

Estimation by Pulp/Tooth Ratio in Canines by Peri‐Apical X‐Rays. J Forensic Sci.

2007;52(1):166-70.

37. Cameriere R, Cunha E, Sassaroli E, Nuzzolese E, Ferrante L. Age estimation by

pulp/tooth area ratio in canines: Study of a Portuguese sample to test

Cameriere's method. Forensic Sci Int. 2009;193(1–3):128.e1-.e6.

38. Singaraju S, Sharada P. Age estimation using pulp/tooth area ratio: A digital

image analysis. J Forensic Dent Sci. 2009;1(1):37-41.

39. Jeevan MB, Kale AD, Angadi PV, Hallikerimath S. Age estimation by pulp/tooth

area ratio in canines: Cameriere's method assessed in an Indian sample using

radiovisiography. Forensic Sci Int. 2011;204(1–3):209.e1-.e5.

40. Saxena S, Tiwari S, Bhambal A. Variations in Morphological Variables of Canine

as Indicators of Age Among Indian Adults. Indian J Stomatol. 2011;2(1):18-20.

41. Zaher JF, Fawzy IA, Habib SR, Ali MM. Age estimation from pulp/tooth area ratio

in maxillary incisors among Egyptians using dental radiographic images. J

Forensic Leg Med. 2011;18(2):62-5.

42. Cameriere R, De Luca S, Alemán I, Ferrante L, Cingolani M. Age estimation by

pulp/tooth ratio in lower premolars by orthopantomography. Forensic Sci Int.

2012;214(1–3):105-12.

64

43. Joseph CC, Reddy BS, Cherian NM, Kannan SK, George G, Jose S. Intraoral Digital

Radiography for Adult Age Estimation: A Reliable Technique. JIAOMR.

2013;25(4):287-90.

44. Azevedo AC, Michel-Crosato E, Biazevic MGH, Galić I, Merelli V, De Luca S, et al.

Accuracy and reliability of pulp/tooth area ratio in upper canines by peri-apical

X-rays. Leg Med. 2014;16(6):337-43.

45. Azevedo AdCS, Alves NZ, Michel-Crosato E, Rocha M, Cameriere R, Biazevic

MGH. Dental age estimation in a Brazilian adult population using Cameriere's

method. Braz Oral Res. 2015;29.

46. Cameriere R, Cunha E, Wasterlain SN, De Luca S, Sassaroli E, Pagliara F, et al.

Age estimation by pulp/tooth ratio in lateral and central incisors by peri-apical

X-ray. J Forensic Leg Med. 2013;20(5):530-6.

47. Ravindra SV, Mamatha GP, Sunita JD, Balappanavar AY, Sardana V.

Morphometric analysis of pulp size in maxillary permanent central incisors

correlated with age: An indirect digital study. J Forensic Dent Sci. 2015;7(3):208-

14.

48. Sakhdari S, Mehralizadeh S, Zolfaghari M, Madadi M. Age Estimation from

Pulp/Tooth Area Ratio Using Digital Panoramic Radiography. JIDAI. 2015;27(1):1.

49. Kumar NN, Panchaksharappa MG, Annigeri RG. Digitized morphometric analysis

of dental pulp of permanent mandibular second molar for age estimation of

Davangere population. J Forensic Leg Med. 2016;39:85-90.

65

50. Babshet M, Acharya AB, Naikmasur VG. Age estimation from pulp/tooth area

ratio (PTR) in an Indian sample: A preliminary comparison of three mandibular

teeth used alone and in combination. J Forensic Leg Med. 2011;18(8):350-4.

51. De Luca S, Bautista J, Alemán I, Cameriere R. Age-at-Death Estimation by

Pulp/Tooth Area Ratio in Canines: Study of a 20th-Century Mexican Sample of

Prisoners to Test Cameriere’s Method. J Forensic Sci. 2011;56(5):1302-9.

52. Cameriere R, Brogi G, Ferrante L, Mirtella D, Vultaggio C, Cingolani M, et al.

Reliability in age determination by pulp/tooth ratio in upper canines in skeletal

remains. J Forensic Sci. 2006;51(4):861-4.

53. Sakuma A, Saitoh H, Suzuki Y, Makino Y, Inokuchi G, Hayakawa M, et al. Age

Estimation Based on Pulp Cavity to Tooth Volume Ratio Using Postmortem

Computed Tomography Images. J Forensic Sci. 2013;58(6):1531-5.

54. Sasaki T, Kondo O. Human age estimation from lower-canine pulp volume ratio

based on Bayes' theorem with modern Japanese population as prior

distribution. Anthropol Sci. 2014;122(1):23-35.

55. Star H, Thevissen P, Jacobs R, Fieuws S, Solheim T, Willems G. Human Dental

Age Estimation by Calculation of Pulp–Tooth Volume Ratios Yielded on Clinically

Acquired Cone Beam Computed Tomography Images of Monoradicular Teeth. J

Forensic Sci. 2011;56(s1):S77-S82.

66

56. Tardivo D, Sastre J, Ruquet M, Thollon L, Adalian P, Leonetti G, et al. Three-

dimensional Modeling of the Various Volumes of Canines to Determine Age and

Sex: A Preliminary Study. J Forensic Sci. 2011;56(3):766-70.

57. Tardivo D, Sastre J, Catherine J-H, Leonetti G, Adalian P, Foti B. Age

determination of adult individuals by three-dimensional modelling of canines.

Int J Legal Med. 2014;128(1):161-9.

58. De Angelis D, Gaudio D, Guercini N, Cipriani F, Gibelli D, Caputi S, et al. Age

estimation from canine volumes. Radiol Med. 2015;120(8):731-736.

59. Pinchi V, Pradella F, Buti J, Baldinotti C, Focardi M, Norelli G-A. A new age

estimation procedure based on the 3D CBCT study of the pulp cavity and hard

tissues of the teeth for forensic purposes: A pilot study. J Forensic Leg Med.

2015;36:150-7.

60. Porto LV, Celestino da Silva Neto J, Anjos Pontual Ad, Catunda RQ. Evaluation of

volumetric changes of teeth in a Brazilian population by using cone beam

computed tomography. J Forensic Leg Med. 2015;36:4-9.

61. Agematsu H, Someda H, Hashimoto M, Matsunaga S, Abe S, Kim H-J, et al.

Three-dimensional Observation of Decrease in Pulp Cavity Volume Using Micro-

CT: Age-related Change. Bull Tokyo Dent Coll. 2010;51(1):1-6.

62. Ge ZP, Ma RH, Li G, Zhang JZ, Ma XC. Age estimation based on pulp chamber

volume of first molars from cone-beam computed tomography images. Forensic

Sci Int. 2015;253:133.e1-.e7.

67

63. Ge ZP, Yang P, Li G, Zhang JZ, Ma XC. Age estimation based on pulp

cavity/chamber volume of 13 types of tooth from cone beam computed

tomography images. Int J Legal Med. 2016;130(4):1159-67.

64. Cattaneo C, De Angelis D, Ruspa M, Gibelli D, Cameriere R, Grandi M. How old

am I? Age estimation in living adults: a case report. J Forensic Odontostomatol.

2008;26(2):39-43.

65. Vodanović M, Dumančić J, Galić I, Savić Pavičin I, Petrovečki M, Cameriere R, et

al. Age estimation in archaeological skeletal remains: evaluation of four non-

destructive age calculation methods. J Forensic Odontostomatol. 2011;29(2):14.

66. Fabbri PF, Viva S, Ferrante L, Lonoce N, Tiberi I, Cameriere R. Radiological

tooth/pulp ratio in canines and individual age estimation in a sample of adult

neolithic skeletons from Italy. Am J Phys Anthropol. 2015;158(3):423-30.

67. De Angelis D, Gibelli D, Fabbri P, Cattaneo C. Dental Age Estimation Helps Create

a New Identity. Am J Forensic Med Pathol. 2015;36(3):219-20.

68. Jeon HM, Jang SM, Kim KH, Heo JY, Ok SM, Jeong SH, et al. Dental Age

Estimation in Adults. JOMP. 2014;39(4):119-26.

69. Patil S, Mohankumar K, Donoghue M. Estimation of age by Kvaal's technique in

sample Indian population to establish the need for local Indian-based

formulae.(Original Article)(Report). J Forensic Dent Sci. 2014;6(3):171.

68

70. Rajpal PS, Krishnamurthy V, Pagare SS, Sachdev GD. Age estimation using

intraoral periapical radiographs. J Forensic Dent Sci. 2016;8(1):56.

71. Sharma R, Srivastava A. Radiographic evaluation of dental age of adults using

Kvaal′s method. J Forensic Dent Sci. 2010;2(1):22-6.

72. Kostenko YY, Goncharuk-Khomyn MY. Clinical and experimrntal study for

improving methods of determining the age of adults by dental status.

Morphologia. 2013;7(1):85-8.

73. Cameriere R, Ferrante L. Canine pulp ratios in estimating pensionable age in

subjects with questionable documents of identification. Forensic Sci Int.

2011;206(1–3):132-5.

74. Cameriere R, De Luca S, Egidi N, Bacaloni M, Maponi P, Ferrante L, et al.

Automatic age estimation in adults by analysis of canine pulp/tooth ratio:

Preliminary results. JOFRI. 2015;3(1):61-6.

75. Torkian A. Age estimation using digital panoramic radiography. IJBPAS.

2015,4(9) Special Issue:124-129

76. Kvaal S. and Solheim T., A non-destructive method for age estimation. J Forensic

Odontostomatol. 1994;12: 6-11.

77. Rolseth V MA, Dahlberg PS, Ding KY, Bleka Ø, Skjerven‐Martinsen M, Straumann,

GH DG, Vist GE. Demirjian’s development stages on wisdom teeth for estimation

of chronological age: a systematic review. Oslo: Folkehelseinstituttet, 2017.

69

78. Solheim, T., & Vonen, A. (2006). Dental age estimation, quality assurance and

age estimation of asylum seekers in Norway. Forensic Sci Int. 2006, vol. 159,

S56-S60.

70

SECTION TWO. Assessment of possible effect of orthodontic treatment on one of the

available methods for age estimation in adults: The Kvaal et al. method.

Preface

Section two answers one of the main aims of this thesis: the effect of external factors on

the accuracy of the Kvaal et al. method. This section presents two studies that tested this

method in individuals who had received orthodontic treatment. For chapter 2, the Kvaal et

al. method was tested on two different data sets: the first, collected from individuals that

had not received orthodontic treatment; and the second from individuals in whom

orthodontic treatment was completed.

Two sets of formulae to estimate the age of individuals were proposed: one for individuals

who did not receive orthodontic treatment; and the second for those who had received

orthodontic treatment; comparisons of the results were performed. It is necessary to

highlight that this is the first study that applies one age estimation method for adults, in

individuals whose dental conditions have been altered. This is in contrast to previous

studies that used data from individuals with sound teeth or who had not been exposed to

other dental procedures.

In Chapter 3, the statistical approach for the analysis of the data is innovative, as it follows

the design of epidemiological studies to assess risk factors. As in the study presented in

Chapter 2, it was observed that there was no statistically significant difference for the age

71

estimation from individuals with and without orthodontic treatment. Chapter 3 presents a

detailed analysis, at the tooth level, of the effect of orthodontic treatment on the accuracy

of the Kvaal et al. method testing formulae that were previously published for dental age

estimation in adults.

The chapters presented in this section share the methodological design with regards to the

inclusion and exclusion criteria of the participants: analyzed teeth, use of panoramic

radiographs, and measurements collection as follows:

Inclusion criteria

All subjects who had a recorded panoramic radiograph of high quality with respect to

factors of image brightness, contrast, and sharpness. In addition, all teeth included in the

analysis were clinically sound, with completed root formation, and at the occlusal plane

level.

Exclusion criteria

Any radiographs that did not meet the inclusion criteria, or had observable failings (Eg.

image distortion, poor contrast, overlap or superposition of tooth structure, or improper

positioning) were excluded. Teeth with rotations, incomplete root formation, dilacerations,

pulpal pathologies and endodontic treatment, and/or restorations were excluded. Teeth

with developmental abnormalities in size, shape and tooth structure, as well as previous

pulpal and periodontal pathologies, and endodontic treatment were excluded. Teeth with

72

large areas of enamel overlap between neighboring teeth were excluded in the studies that

are included in this section.

Teeth Selection

In accordance to Kvaal et al. [1], three single rooted upper teeth (with Federation Dentaire

International (FDI) notation 11/21, 12/22. 15/35) and three single rooted lower teeth (FDI

32/42, 33/43, 34/44) were measured. Kvaal et al. [1] reported that there no significant

differences between teeth from the left or the right side of the arch.

Consequently, in the absence of one of the teeth, the contralateral tooth was used, as long

as it met the inclusion criteria. In accordance with Karkhanis et al [2], panoramic

radiographs that presented any combination of the required teeth were included. Both

studies also shared the procedures to perform the measurements on digital panoramic

radiographs, the calibration of the observer, and the methods to test the measurement

precision as described below:

Measurements

All the panoramic radiographs were obtained in digital format, the measurements were

performed using Image J software (version 1.48 19 April 2014 - National Institute of Health,

USA). Following the methodology used by Kvaal et al. [1], the following measurements

were recorded: tooth, pulp and root length, as well as the pulp and tooth width at three

different levels: A (cemento-enamel junction in the mesial surface of the roots); B

(midpoint between the points A and C); and C (midpoint between the cemento-enamel

73

junction and root apex). After the collection of the figures in an Excel (2013) spreadsheet, a

series of ratios was calculated: (T) tooth/root lengths; (R) pulp/tooth length, (P) pulp/root

length, and root width/ pulp width ratio at levels A, B and C.

The ratio calculation and use, was proposed by Kvaal et al. [1] with the aim to reduce the

effect of magnification and angulation inherent in most radiographs [1, 3]. The different

mean values calculated with these ratios provided the age estimation predictors. M: mean

value all ratios, W: mean value width ratios B and C, L: mean value ratios P and R, W-L:

difference between W-L).

In Chapter 2, the predictors M and W-L were later used to formulate the age estimation

models by means of multiple regression analysis. In Chapter 3, the predictors M and W-L

were later used to estimate the age of all the participants, as established by Kvaal et al. [1].

The measurements were completed by a single observer (TYM). All data were collected

using Excel (version 2013 Microsoft Redmont, USA) and statistical analyses were completed

using the program R Core Team version 3.1.3 (2015). (R: A language and environment for

statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL

http://www.R-project.org/).

Calibration

Intra and inter-observer calibration was estimated using five randomly selected panoramic

radiographs from the final study sample. These panoramic radiographs were measured on

five different days with at least one-day interval. With the aim to avoid the recall of the

measurements and landmarks, five new panoramic radiographs were also measured on

http://www.r-project.org/

74

each occasion. Data were recorded using the software Image J by TYM and SK. To

determine intra-observer precision, three estimates of precision were calculated: the

technical error of measurement (TEM <1.0), relative technical error of measurement (rTEM

<5%) and coefficient of reliability was calculated to quantify the inter- and intra-observer

measurement error [4].

Measurement precision

After the estimation of intra- and inter-observer error, the obtained values were as follows:

intra-observer error: TEM <1.0 (0.92), rTEM<5% (2.99%) and R>0.75 (0.95). Inter-observer

error (between SK and TYM) was: TEM<1.0 (0.80), rTEM<5 (2.37) and R>0.75 (0.98)

showing comparable intra and inter-observer measurement precision; within acceptable

standards for all the measurements.

75

REFERENCES


dental radiographs. Forensic Sci Int. 1995; 74:175-85.



Kvaal et al. Forensic Sci Int. 2014;235:104.e1-.e6.



2012;57:277-84.

4. Weinberg SM, Scott NM, Neiswanger K, Marazita ML. Intraobserver error

associated with measurements of the hand. Am J Hum Biol. 2005;17:368-71.

76

2. Chapter 2. Changes in age estimation standards secondary to full fixed edgewise

orthodontic treatment



Tennant, M. (2016). Changes in age estimation standards secondary to full fixed edgewise

orthodontic treatment. European Journal of Forensic Science, Accepted for publication.

June 2016.

77

ABSTRACT

Objective: The aim of this cohort study was to evaluate if the side effects of orthodontic

treatment have any significant impact when the Kvaal et al. method is used for dental age

estimation. The objective was to observe the potential effects of orthodontic treatment on

the accuracy of the age estimation, as performed by using the Kvaal et al. standards when it

was applied on individuals before and after orthodontic treatment.

Methods: Following the methodological approach of Kvaal et al., odontometric

measurements were acquired and the data were statistically analyzed to develop age

estimation regression models. The total number of radiographs analyzed was 182 (64%,

n=58) female and 36% (n=33) male. The ages ranged from 12 to 50 years for females

(mean age 22 years), and 12 to 52 for males (mean age 22 years), before starting the

treatment. The average length of the treatment was 2.2 years for both females and males.

Results: The standard error of estimate (SEE) for the regression models did not change

dramatically for the pre- and post-treatment data.

Conclusions: It is recommended that similar analyses be performed for other methods of

dental age estimation, for example methods based on cone beam computer-tomography

or micro-focus computer tomography. These novel approaches, although more accurate,

are also potentially more sensitive to dental changes caused by orthodontic treatment.

Key words: Forensic sciences, dental age estimation, secondary dentine formation, adults,

orthodontic treatment, root resorption.

78

INTRODUCTION

Kvaal et al. [1] developed a non-invasive method for the estimation of age in adults with

measurement of dimensions of the pulp chamber and tooth length to formulate regression

models for the estimation of age. Originally, this method was proposed to be used on

periapical radiographs. However, more recent work has focused on the use of the Kvaal et

al. [1] method in panoramic radiographs [2-4]. During tooth development, secondary

dentine deposition commences once the root formation is complete, and the tooth erupts

to the level of the occlusal plane [5].

Annually, the rate of secondary dentine formation has been documented to be 6.5 μm/year

for the crown, and 10 μm/year for the root [6]. The rate of secondary dentine formation, is

influenced by external stimuli such as occlusal forces, trauma and caries [7]. These external

factors also generate the production of tertiary dentine [5], which is indistinguishable from

secondary dentine on dental radiographs [8].

The estimation of dental age has been a valid tool in forensic and anthropological research

[1, 9, 10]. For younger persons, including newborns, infants and adolescents, there are

dental age estimation methods, generally based on tooth maturation, which are considered

the most accurate to estimate of chronological age in sub-adults [3, 11-13].

In adults, and once the root formation is deemed complete, methods for dental age

estimation are broadly based on the analysis of secondary dentine deposition and the

79

subsequent narrowing of the pulp chamber [1, 6, 14]. A potential confounding factor in the

estimation of dental age is the high incidence of orthodontic treatment in developed

countries [15]. The judicious application of forces is central to successful orthodontic

treatment and is well known to induce biological changes in dental hard tissues, including

the formation of secondary dentine [7] and root shortening [16]. The corollary being a

change in the dental morphological features analysed in the estimation of dental age in

adults [7, 17].

Given the potential of the development of tooth structure in adolescents and adult patients

to be influenced by the application of orthodontic forces, and these structures being the

foundation of age estimation methods, it is important to recognise and quantify such

influences. The first aim of this study was to evaluate if the side effects of orthodontic

treatment had any significant impact on dental age estimation when using the Kvaal et al.

[1] method. The objective of this study was also to observe the potential effects of

orthodontic treatment on the accuracy of age estimation, as performed by using the Kvaal

et al. [1] standards when applied to individuals before and after orthodontic treatment.

MATERIALS AND METHODS

Ethics approval was obtained from the Human Research Ethics Committee of The

University of Western Australia (Ref: RA/4/1/6797) prior to commencement. The design of

the study, data collection and analysis has been performed in accordance with the ethical

standards laid down in the 1964 Declaration of Helsinki.

80

Sample

This cohort study consisted of a series of subjects who consecutively presented for

orthodontic treatment at a private specialist orthodontic clinic (SV). All panoramic

radiographs were acquired from the same machine and in digital format, with the informed

consent of the participants. All of the panoramic radiographs were unidentified, and the

only information known was the sex and age of the participants at the starting and finishing

time of the orthodontic treatment.

Sample population

This study analyzed a total of 182 pre-, and post-orthodontic treatment panoramic

radiographs from 91 participants, 64% (n=58) female and 36% (n=33) male, from a Western

Australian population. All participants presented for orthodontic treatment that, after

diagnosis and treatment planning, was completed with fixed appliances. The ages ranged

from 12 to 50 years for females (mean age 22) and 12 to 52 for males (mean age 22) before

starting the treatment. The average length of treatment was 2.2 years for both females

and males. Digital panoramic radiographs were taken as part of routine clinical orthodontic

treatment to evaluate the initial condition of the participants and the final treatment

results.

Statistical analysis

The strength of linear association between chronological age and the Kvaal et al. [1] dental

ratios was determined by Pearson correlation coefficient (r2). Following this stage,

regression analysis was applied to evaluate the relationship between the ratios and the

81

chronological age. The final results were later used to generate the statistical models for

age estimation. To quantify the predictive accuracy of the models, the standard error of

estimation (SEE) was used in the cross-validation sample. Multivariable linear regression

analyses were used to examine the association between the independent and outcome

variables. The outcome variables included chronological age (dependent variable), and the

predictors M and W-L (independent variables). Different equations were generated to

estimate the chronological age, by individual teeth and by groups of teeth, as follows:

Three maxillary teeth with M and W-L values (6 predictors) individually calculated for each

tooth, and the average of M and W-L for the same teeth (2 predictors).

Three mandibular teeth with M and W-L values (6 predictors) individually calculated for

each tooth, and the average of M and W-L for the same teeth (2 predictors). All six teeth,

with M and W-L values (12 predictors) individually calculated for each tooth, and the

average of M and W-L for the same teeth (2 predictors). All regression models were built

using the ordinary least squares approach. R-square values were computed to examine the

amount of variance explained by the predictor variables. The formula proposed by

Preoteasa et al. [18] was used to calculate the percentages of teeth with root shortening.

RESULTS

Age distribution

The age distribution before orthodontic treatment for females was: 12-19 years (62%,

n=36), 20-39 years (28%, n=16), >40 years (10%, n=6), and after treatment it was: 15-19

82

years (57%, n=33), 20-39 years (33%, n=19), >40 years (10%, n=6). For males, before

orthodontic treatment was: 12-19 years (52%, n=17), 20-39 years (42%, n=14), and >40

years (6%, n=2), and after orthodontic treatment it was: 13-19 years (48%, n=16), 20 to 39

years (45%, n=15) and >40 years (6%, n=2).

The correlation coefficient between chronological age and the dental ratios was calculated

for the data before and after orthodontic intervention (Table 2.1). In both cases, the

correlation coefficient related with the width ratios (A, B and C) were more significant than

those related with length (P, T, R and L). As established by Kvaal et al. [1], the predictors M

and W-L are required to be included in the final equation for dental age estimation. Based

on the Pearson correlation coefficients, it could be observed that after orthodontic

treatment, the significance of the values for the predictor W-L decreased.

Individual tooth regression models, and multiple teeth regression models, were proposed

for individuals that had not received orthodontic treatment (Tables 2.2 and 2.3

respectively), and for individuals after finishing the treatment (Table 2.3 and 2.4). In regard

to the models for age estimation using individual teeth, the SEE tends to increase after

orthodontic treatment, especially for the mandibular canine, where the increase was ±2

years. Furthermore, while the results for the SEE provided by the mandibular canine

showed a lower SEE (± 8.26 years) in the data analysis before treatment, the same tooth

showed the greatest value of SEE (±10.23 years) for individual tooth analysis, after


83

Table 2.1. Correlation coefficients before (1) and after orthodontic treatment (2).

Correlation coefficients between chronological age before(1) and after(2) starting orthodontic treatment and the Kvaal et al. [1] dental ratios (P: pulp length/root length, T: tooth length/root length, R: Pulp length/tooth length, A: pulp width/root width at level A, B: pulp width/root width at level B, C: pulp width/root width at level C) and predictors (M: mean value all ratios, W: mean value width ratios B and C, L: mean value ratios P and R, W-L: difference between W-L). p< 0.05* p<0.01** NS non-significant

ratio

Tooth number (FDI tooth numeration)

11/21 12/22 15/25 32/42 33/43 34/44

P1 0.15 0.07 -0.08 0.08 -0.08 0.2

P2 -0.26 -0.13 -0.06 -0.02 -0.22 -0.12

T1 0.03 0.03 -0.13 -0.18 -0.23 -0.11

T2 -0.16 -0.12 -0.14 -0.09 -0.08 -0.15

R1 0.16 0.04 0.07 0.32* 0.17 0.38**

R2 -0.09 -0.01 0.13 0.07 -0.14 0.03

A1 -0.44** -0.41** -0.42** -0.41** -0.32** -0.16

A2 -0.31* -0.38** -0.47** -0.17 -0.21 -0.19

B1 -0.43** -0.41** 0.02 -0.21* -0.62** -0.59**

B2 -0.44** -0.49** -0.44** -0.22* -0.48** -0.52**

C1 -0.34** -0.34* -0.32* -0.18 -0.51** -0.45**

C2 -0.34** -0.43** -0.36* -0.32* -0.33* -0.43**

M1 -0.13 -0.33* -0.28* -0.19 -0.56** -0.37**

M2 -0.31* -0.40** -0.41** -0.22* -0.38 -0.44**

W1 -0.44** -0.41** -0.14 -0.23* -0.52** -0.57**

W2 -0.45** -0.51** -0.46** -0.29* -0.43** -0.52**

L1 0.18 0.07 -0.13 0.19 0.01 0.33

L2 -0.22* -0.11 -0.14 0.01 -0.21 -0.07

W-L1 -0.36 -0.32 0.02 -0.32* -0.48** -0.57**

W-L2 -0.01 -0.23* -0.05 -0.22* -0.26* -0.41**

84

Table 2.2. Multiple regressions for estimation of chronological age (in years) for

individual maxillary and mandibular teeth in subjects before (1) and after (2)


Teeth FDI n R R2 Equation SEE ± years

11/21(1) 83 0.22 0.21 Age= 39.87-76.79(M) -51.75(W-L) 8.57

11/21(2) 83 0.14 0.12 Age= 54.50-73.67(M)-30.64(W-L) 9.06

12/22(1) 81 0.18 0.16 Age= 68.21-99.54(M) -33.52 (W-L) 8.92

12/22(2) 81 0.25 0.23 Age= 83.07-128.42(M)-42.68(W-L) 8.48

15/25(1) 58 0.08 0.04 Age= 79.8105-81.5881(M)-0.5444(W-L) 9.99

15/25(2) 58 0.21 0.18 Age= 90.22-129.66(M)-18.62(W-L) 9.29

32/42(1) 87 0.12 0.10 Age= 15.40-37.51(M)-43.05(W-L) 9.73

32/42(2) 87 0.09 0.74 Age= 49.10-77.55(M)-37.65(W-L) 9.89

33/43(1) 63 0.47 0.45 Age= 111.80-238.35(M)-132.7(W-L) 8.26

33/43(2) 63 0.19 0.16 Age= 76.60-159.75(M)-105.79(W-L) 10.26

34/44(1) 68 0.35 0.33 Age= 21.67-69.84(M)-69.59(W-L) 8.40

34/44(2) 68 0.31 0.29 73.52-137.96(M)-64.43(W-L) 8.71

R2, coefficient of determination. Standard error of estimation (SEE) in years. M: mean value all ratios, W: mean value width ratios B and C, L: mean value ratios P and R, W-L: difference between W-L.

85

Table 2.3. Multiple regressions for estimation of chronological age (in years) for the combination of different teeth before orthodon tic

treatment

Teeth N R R2 Equation (FDI tooth numeration) SEE ± years

3 max 2 pds 50 0.27 0.24 Age= 80.40- 145.51(M) -47.69 (W-L) 7.78

3 max 6 pds 50 0.31 0.21 Age= 63.704 - 6.412 (11/21M) - 17.216 (11/21W-L) - 32.812 (12/22M)-19.332 (12/22W-L) - 77.969 (15/25M) + 10.594 (15/25W-L)

7.93

3 mdb 2 pds 46 0.4 0.37 Age= 13.11-36.64 (M) - 19.83 (W-L) 9.12

3 mdb 6 pds 46 0.6 0.54 Age= 124.528 + 81.230 (32/42M) - 7.784 (32/42W-L) - 328.184 (33/43M) - 121.021 (33/43W-L) - 37.503 (34/44M) - 9.855(34/44W-L)

7.8

6 teeth 2 pds 36 0.45 0.41 Age = 133.86 - 238.98 (M) - 67.88 (W-L) 7.6

6 teeth 12 pds 36 0.72 0.58 Age = 130.727 + 15.058 (11/21M) + 23.906 (11/21W-L) + 61.249 (12/22M) - 29.091 (12/22W-L) - 100.481 (15/25M) - 2.441 (15/25W-L) + 54.517 (32/42M) + 3.089 (32/42W-L) - 268.438 (33/43M) - 89.805 (33/43W-L) - 41.009 (34/44M) - 25.32 (34/44W-L)

6.3

R2, coefficient of determination. Standard error of estimation (SEE) in years. M: mean value all ratios, W: mean value width ratios B and C, L: mean value ratios P and R, W-L: difference between W-L. Age predictors (pds). Maxillary teeth (max). Mandibular teeth (mdb).

86

Table 2.4. Multiple regressions for estimation of chronological age (in years) for the combination of different teeth after orthodontic

treatment

Teeth n R R2 Equation (FDI tooth numeration) SEE ± years

3 max 2 pds 50 0.38 0.36 Age= 102.46 -102.16(M) -63.13(W-L) 7.21

3 max 6 pds 50 0.42 0.34 Age= 114.631 - 0.876 (11/21M) + 2.218 (11/21 W-L) + -115.748 (12/22M) - 38.699 (12/22W-L) - 91.406 (15/25M) -20.963(15/25W-L)

7.31

3 mdb 2 pds 46 0.33 0.30 Age=98.68-213.10 (M) - 105.05 (W-L) 9.63

3 mdb 6 pds 46 0.44 0.36 Age= 42.761 + 81.109 (32/42M) + 1.905 (32/42W-L) - 111.346 (33/43M) - 173.252(33/43W-L) - 151.769 (34/44M) - 43.162 (34/44W-L)

9.23

6 teeth 2 pds 36 0.49 0.46 Age = 174.12 -288.42 (M) -59.22 (W-L) 7.37

6 teeth 12 pds 36 0.6 0.39 Age= 154.569+9.383(11/21M) + 19.718 (11/21W-L)-70.675 (12/22M)-39.284(12/22W-L)-39.932(15/25M) - 4.881 (15/25W-L)+47.114 (32/42M)-17.071 (32/42W-L)-58.568 (33/43M) -29.183 (33/43W-L) -149.635(34/44M)-0.842 (34/44W-L)

7.8

R2, coefficient of determination. Standard error of estimation (SEE) in years. M: mean value all ratios, W: mean value width ratios B and C, L: mean value ratios P and R, W-L: difference between W-L. Age predictors (pds). Maxillary teeth (max). Mandibular teeth (mdb).

87

When multiple regression models were formulated using different combinations of teeth, it

is also noticeable that in a few cases the SEE decreased after orthodontic treatment:

maxillary lateral incisor (±0.437), maxillary first premolar (±0.698), the multiple tooth

regression models using three maxillary teeth with two (±0.574) and six (±0.613)

predictors, and in the multiple regression models using six teeth and two predictors

(±0.237): In both circumstances, the multiple regression models using the different teeth

combinations improved the accuracy of estimation of chronological age, consistent with

previous studies [2].

Root shortening was detected in in 62% of the upper central incisors, 53% of upper lateral

incisors, 48% of upper second premolars, 48% of lower lateral incisors, 50% of lower canine

and 51% of lower second premolars, with a formula previously proposed [18].

DISCUSSION

This is the first study specifically designed to apply the Kvaal et al. [1] method to a group of

orthodontically treated subjects. Also, the first study evaluating the impact of the side

effects of orthodontic treatment on one of the proposed methods for dental age

estimation in adults based on the formation of secondary dentine and the pulp/tooth

dimension ratios.

Previous investigations failed to disclose whether subjects had previously received

orthodontic treatment. There are well known biological changes that occur secondarily to

88

the application of orthodontic forces, including apical root resorption and secondary

dentine formation.

External root resorption is one unavoidable, unpredictable, and undesirable sequela [19-

21] of orthodontic treatment. The reported frequency of external root resorption is about

100% when it is examined under microscopy. When it is quantified on periapical or

panoramic radiographs, the frequency falls to 70%, with a mean reported value of root

shortening of 1.42 ±0.44mm, and 1% to 5% of the cases with a loss of 4mm or one third of

root length [18].

In addition, maxillary teeth are more sensitive than mandibular teeth to root resorption.

The most frequently affected teeth, according to severity, are the maxillary laterals,

maxillary centrals, mandibular incisors, the distal root of mandibular first molars, second

premolars, and maxillary second premolars [20, 22], and all these teeth have been used in

different methods for dental age estimation in adults [1, 2, 23]. In this study, a higher

percentage of root resorption for upper teeth was also observed.

Confirming the findings of Karkhanis et al. [2], the length ratios (P, T, R, and L ratios) in this

study have a non-significant correlation coefficient with age, (p>0.05). Furthermore, the

percentage of negative correlation coefficients (Table 2.1), closer to -1, is higher in the

observed values after orthodontic treatment (23% before vs 66% after).

89

As the Kvaal et al. [1] method is based on the negative correlation between age and the

above-mentioned ratios, it is possible to observe that after orthodontic treatment this

negative correlation is more evident, but not statistically significant. In this study, there

were no significant variations in the correlation coefficients of the calculated length ratios

with age among the different teeth. Furthermore, as the Kvaal et al. [1] method is based

on length ratio calculations, not the linear measurements per se, it is possible that the

effect of root shortening on the estimated ages had been diminished when the ratios are

calculated.

In terms of secondary dentine formation due to orthodontic treatment (a phenomenon

that has not been as widely investigated as root resorption), there are reports of complete

obliteration of the canals in maxillary incisors [24]. In one study analysing the tooth

changes in terms of root length and pulp chamber within maxillary central incisors (tooth

11 and 21), statistically significant changes in the width of the pulp chamber at the

midpoint of the dental root [25] were found, equivalent to the point C in the Kvaal et al. [1]

method.

Another study, using cone beam computer tomography in maxillary teeth, found a

statistically significant difference between the volumes before and after orthodontic

treatment, with the highest mean volume loss observed in the upper left lateral incisor

(3.86 mm3), and the least for the upper right central incisor (3.04 mm3) [7].

90

In this study, the correlation coefficients of the pulp/root width ratios (A, B, C, and W)

maintained their negative correlation with age before and after the treatment, without

showing the same variation observed for the length ratios. It would still be necessary to

assess if the root shortening after orthodontic treatment, displaced the location of the

reference points B and C towards the crown (where the pulp chamber is wider), and its

relationship with the observed variation of the SEE before and after orthodontic treatment,

which in both cases is acceptable in forensic terms (SEE=± 10 years).

CONCLUSION

This current investigation demonstrated an alteration in tooth and pulp chamber

morphology and dimensions following orthodontic treatment. However, these changes did

not serve to significantly influence the validity and accuracy (SEE) of the Kvaal et al. [1]

method when applied in the assessment of panoramic radiographs to estimate

chronological age. Furthermore, it is not possible to affirm that the instrument used, the

dental panoramic is sensitive enough to detect such small changes in the pulp chamber

dimensions. It is recommended that further analysis of other methods for dental age

estimation in adults based on the formation of secondary dentine, such as developed by

Cameriere et al. [3], or the more recently proposed methods based on volumetric

reconstructions of the tooth and pulp chamber in CBCT [26, 27], which could be more

susceptible to be affected by orthodontic treatment sequels.

91

REFERENCES

1. Kvaal S.I., Kolltveit K.M., Thomsen I.O., Solheim T.. Age estimation of adults from


2. Karkhanis S., Mack P., Franklin D.. Age estimation standards for a Western


Kvaal et al. Forensic Sci Int. 2014;235:104.e1-6.

3. Panchbhai A.S., Dental radiographic indicators, a key to age estimation.

Dentomaxillofac Radiol. 2011;40(4):199-212.

4. Kanchan-Talreja P., Acharya A.B., Naikmasur V.G., An assessment of the

versatility of Kvaal's method of adult dental age estimation in Indians. Arch Oral

Biol. 2012;57(3):277-84.

5. Hargreaves KM. Cohen's Pathways of the Pulp Expert Consult. St. Louis: Mosby;

2010.



2012;214(1–3):105-12.

7. Venkatesh S, Ajmera S, Ganeshkar SV. Volumetric Pulp Changes after

Orthodontic Treatment Determined by Cone-beam Computed Tomography. J

Endod. 2014;40(11):1758-63.

92

8. Nanci A. Ten Cate's Oral Histology-Pageburst on VitalSource: Development,

Structure, and Function. St. Louis, MO: Elsevier Health Sciences; 2007.

9. Sarkar S, Kailasam S, Mahesh Kumar P. Accuracy of estimation of dental age in

comparison with chronological age in Indian population – A comparative

analysis of two formulas. J Forensic Leg Med. 2013;20:230-3.

10. Senn DR, Weems RA. Manual of Forensic Odontology. 5th ed. Boca Raton, FL:

CRC Press; 2013.

11. Yusof MY, Thevissen PW, Fieuws S, Willems G. Dental age estimation in Malay

children based on all permanent teeth types. Int J Legal Med. 2014;128(2):329-

33.

12. Ebrahim E, Rao RK, Chatra L, Shenai P, Veena KM, Prabhu RV, et al. Dental Age

Estimation Using Schour and Massler Method in South Indian Children. Sch J

Appl Med Sci. 2014;2(5C):1669-74.

13. Demirjian A, Goldstein H, Tanner JM. A new system of dental age assessment.

Hum Biol. 1973;45(2):211.


15. Pretty I, Sweet D. Forensic dentistry: A look at forensic dentistry–Part 1: The role

of teeth in the determination of human identity. Br Dent J. 2001;190(7):359-66.

16. Baumrind S, Korn EL, Boyd RL. Apical root resorption in orthodontically treated

adults. Am J Orthod Dentofacial Orthop. 1996;110(3):311-20.

93

17. Ramazanzadeh BA, Sahhafian AA, Mohtasham N, Hassanzadeh N, Jahanbin A,

Shakeri MT. Histological changes in human dental pulp following application of

intrusive and extrusive orthodontic forces. J Oral Sci. 2009;51(1):109-15.

18. Preoteasa CT, Ionescu E, Preoteasa E, Tenreiro Machado JA, Baleanu MC,

Baleanu D. Multidimensional Scaling for Orthodontic Root Resorption. Math

Probl Eng. 2013;2013;1-6.

19. Ponder SN, Benavides E, Kapila S, Hatch NE. Quantification of external root

resorption by low- vs high-resolution cone-beam computed tomography and

periapical radiography: A volumetric and linear analysis. Am J Orthod

Dentofacial Orthop. 2013;143(1):77-91.

20. Campos MJ, Silva KS, Gravina MA, Fraga MR, Vitral RWF. Apical root resorption:

The dark side of the root. Am J Orthod Dentofacial Orthop. 2013;143(4):492-8.

21. Harris DA, Jones AS, Darendeliler MA. Physical properties of root cementum:

Part 8. Volumetric analysis of root resorption craters after application of

controlled intrusive light and heavy orthodontic forces: A microcomputed

tomography scan study. Am J Orthod Dentofacial Orthop. 2006;130(5):639-47.

22. Brezniak N, Wasserstein A. Root resorption after orthodontic treatment: Part 2.

Literature review. Am J Orthod Dentofacial Orthop. 1993;103(2):138-46.

23. Mathew DG, Rajesh S, Koshi E, Priya LE, Nair AS, Mohan A. Adult forensic age

estimation using mandibular first molar radiographs: A novel technique. J

Forensic Dent Sci. 2013;5(1):56-9.

94

24. Delivanis HP, Sauer GJR. Incidence of canal calcification in the orthodontic

patient. Am J Orthod. 1982;82(1):58-61.

25. Tobón D, Aristizabal D, Álvarez C, Urrea J. Cambios radiculares en pacientes

tratados ortodoncicamente. (Spanish). Root changes in patients treated

orthodontically (English). CES Odontol. 2014;27(2):37-46.

26. Ge ZP, Ma Rh, Li G, Zhang JZ, Ma XC. Age estimation based on pulp chamber


Sci Int. 2015;253:133.e1-7.




95

3. Chapter 3. Orthodontic treatment: Real risk for dental age estimation in adults?



Tennant, M. (2016). Orthodontic Treatment: Real Risk for Dental Age Estimation in Adults?

Journal of Forensic Science. Accepted for publication. October 2016.

96

ABSTRACT

Dental age estimation becomes a challenge once the root formation has finished. In living

adults, the most commonly used methods are based on the formation of secondary

dentine. One of the side effects of orthodontic treatment is the formation of secondary

dentine and root shortening.

The aim of this study was to establish if the secondary effects of orthodontic treatment

generate a statistically significant difference in dental age estimations. The study sample

included 34 pairs of pre- and post-orthodontic treatment panoramic radiographs, from

different individuals with exactly the same age and sex distribution. Females (n=22, 65%)

age range 15-50 years old, median 17.5, and males (n=12, 35%) age range 16-37 years old,

median 22.5 were included.

After data collection, dental age was estimated per tooth using formulae previously

published. The risk of obtaining over-estimation of age was calculated (RR=1.007). The

changes caused by orthodontic treatment do not have any significant effect on age

estimation when the Kvaal et al method is applied on panoramic radiographs.

Key words: Forensic science, dental age estimation, secondary dentine formation, adults,

orthodontic treatment, root resorption.

97

INTRODUCTION

The analysis of teeth for age estimation has been scientifically reported since the early

1800s [1]. Methods based on tooth formation in juveniles have shown high reliability [2-4].

Once the root formation has finished, (generally at the age of 14 years) [5], dental age

estimation becomes a challenge, partly as a result of third molar formation variability [6].

The most reported non-invasive methods for dental age estimation are based on the

formation of secondary dentine and the decrease of pulp chamber dimensions.

These features are measurable in periapical radiographs [7, 8], panoramic radiographs [9,

10], micro-focus computed tomography [11], computed tomography [12], and cone beam

computed tomography [13]. These methods proposed different formulae to be used in

specific populations. An important characteristic of these studies is that in their analysis,

they included only totally sound teeth.

It is well known that orthodontic forces generate irreversible changes on tooth structure,

such as root shortening [14], and secondary dentine formation [15]. These two biological

changes in tooth structure may directly affect the features used for dental age estimation

in adults, especially the method proposed by Kvaal et al. [7].

This method is based on the measurement of tooth/pulp length, root canal and root width

at different levels, followed by ratio calculations and linear regression analysis for dental

age estimation. It would be expected that with the mentioned changes, secondary to

98

orthodontic treatment, age estimation in participants post orthodontic treatment, would

show a higher estimation error and higher overestimation, compared to that of non-

treated participants. If that were to be the case, it would be necessary to develop specific

standards for orthodontically treated participants, not only for the Kvaal et al. [7] method,

but for any method based on secondary dentine formation and the variation of pulp/tooth

dimensions with age.

In the event of the results being different to the expected, it would mean that the Kvaal et

al. method could be used despite the evidence of anatomic changes related to orthodontic

treatment. The aim of this study was to establish if orthodontic treatment could generate

changes in dental age estimations at the tooth level when the Kvaal et al. [7] method is

applied per individual tooth.


Ethics approval was obtained from the Human Research Ethics Committee of The

University of Western Australia (Ref: RA/4/1/6797) prior to commencement.

Panoramic radiographs, orthodontics, and dental age estimation

Initially, the Kvaal et al. [7] method was proposed to be used on periapical radiographs.

However, recent studies have also applied this method on panoramic radiographs [9, 10].

Panoramic radiographs have more image distortion than periapical radiographs, but it has

been reported that when root resorption associated with orthodontic treatment is

99

measured on panoramic radiographs, it is significantly higher than when measured on

periapical radiographs [16].

This is the first study to see the effect of orthodontic related changes on age estimation

therefore, no power calculation was achievable. However, based on the size of previous

research assessing pulpal changes associated to orthodontic treatment [17], a sample size

in a similar range was deemed appropriate.

Sample Selection

The study cohort consisted of a series of participants who consecutively presented for

orthodontic treatment at a private specialist orthodontic clinic (SV). The initial sample for

this study was 91 participants, who had a pre, and post-treatment panoramic radiographs

from a Western Australia population. From the initial sample, a total of 34 pre-, and 34

post-orthodontic treatment panoramic radiographs, were selected, resulting in a final

sample of 34 pairs of panoramic radiographs (n=68).

Sample size reduction corresponded to the need of age and sex matching between the

different individuals in each pair, one of them without orthodontic treatment and the other

one with a concluded treatment. Females 65% (n=22, for both groups), age range 15-50

years old, median 17.5. Males 35% (n=12, for both groups) age range 16-37 years old,

median 22.5. The minimum length treatment was 1.2 years, the maximum 3.6 years,

median 2.1 years.

100

Special characteristics of this study

The purpose of this study was to assess the biologic variation effects of orthodontic

treatment at the tooth level. Therefore, the tooth-by-tooth formulae for the Kvaal method

were applied. [10, 18]. Measurements were taken before categorizing the sample into two

groups: pre for panoramic radiographs obtained from non-treated participants, and post

for treated participants, in order to avoid observation bias.


After the completion of the measurements, dental age estimation was calculated using the

age estimation equations from individual teeth. In participants 30 years or older the

equations used were those reported by Karkhanis et al. [10]. And for participants 29 years

or younger, the equations used were obtained from a different group of non-treated

participants (n=74 aged 12-28 years). Both set of equations were obtained from a Western

Australian population. Then, the respective analysis for risk assessment was performed.

RESULTS.

After age estimation per individual tooth, there was a total of 189 age estimates for group

pre, and 196 estimates for group post. The age estimates obtained were then compared

between participants with exactly the same age and sex in group pre, and group post. In

this way 183 pairs of age estimates were obtained and compared 1 to 1. For 79% (n=145)

of the total observations, the fluctuation of the data was the same, regardless if there was

101

over or under estimation of the age. In terms of over-estimation, in 47% of these

observations, the over estimation was higher after the orthodontic treatment.

Although the fluctuation of the data was the same in groups, pre-and post, there was a

clear difference between the ages. Before the age of 25, the large majority of results

showed a slight over estimation of age which did not exceed the reported SEE. In contrast,

the majority of age estimation data obtained from older participants showed under

estimation of the age that notably exceeded the reported SEE.

A contingency table (Table 3.1) was developed to test if the secondary effects of

orthodontic treatment were a real risk to produce over-estimation of the age when the

Kvaal et al. method is applied. The relative risk calculated was RR=1.0071 (Low risk).

The fluctuation of the data in regard to the chronological current age of the individual was

analyzed per individual tooth (Table 3.2), showing that there was not a significant

difference in the percentages of over estimation and under estimation between both

groups (Pearson correlation coefficient=0.78).

Although the fluctuation of the data was the same in groups, pre-and post, there was a

clear difference between the ages. Before the age of 25, the large majority of results

showed a slight over estimation of age which did not exceed the reported SEE. In contrast,

the majority of age estimation data obtained from older participants showed under

estimation of the age that notably exceeded the reported SEE.

102

Table 3.1. Contingency table to estimate if orthodontic treatment was a causal of

overestimation, when the Kvaal et al. method is used to age estimation.

Orthodontic treatment Over-estimation Under-estimation TOTAL

YES

Post-treatment

48%

n=94

52%

n=102 n=196

No

Pre-treatment

48%

n=184

52%

n=99 n=189

TOTAL n=184 n=201 n=385

Incidence over-estimation group A=47.9%. Incidence under-estimation group B=47.6%. Relative risk of presenting overestimation owed to orthodontic treatment secondary effects when Kvaal et al. method is used for dental age estimation: RR=1.0071 (No or low risk).

103

Table 3.2. Comparison of over- and under-estimates obtained per tooth

TOOTH (FDI) 11/21 12/22 15/25 32/42 33/43 34/44 O

VER

ES

TIM

ATI

ON

Pre-treatment 39%

(n=13)

50%

(n=)16

52%

(n=17)

47%

(n=16)

54%

(n=13)

47%

(n=15)

Post-treatment 38%

(n=13)

45%

(n=15)

48%

(n=16)

44%

(n=15)

61%

(n=20)

52%

(n=15)

UN

DER

ES

TIM

ATI

ON

Pre-treatment 61%

(n=20)

50%

(n=16)

48%

(n=16)

53%

(n=18)

46%

(n=11)

53%

(n=17)

Post-treatment 62%

(n=21)

55%

(n=18)

52%

(n=17)

56%

(n=19)

39%

(n=13)

48%

(n=14)

SEE ± years (<30 years) 4.3 4.2 4.5 4.4 3.7 3.7

SEE ± years (> 30 years) 9.3 9.6 9.5 10.2 10.9 10.5

Reported standard estimation error (SEE) per individual tooth, before and after orthodontic treatment, for participants younger than 30 years (<30 years) and over 30 years of age (> 30 years).

104

DISCUSSION

The reliability of dental age estimation in adults, based on the formation of secondary

dentine, have shown superior accuracy to other methods, based on the analysis of other

age related dental or osseous changes [8, 10]. However, teeth are subjected to different

changes through life, related with pathology, biological, chemical, or mechanical trauma or

dental procedures.

It has been reported that orthodontic treatment causes mechanical trauma to the

periodontal ligament and induces pulpal reactions [20]. The most frequently reported side

effect is external root resorption, a process characterized by the destruction of root

structure [20], with the subsequent diminution of root length (mean reported value of 1.42

±0.44mm) [21].

Another side effect is the reduction of pulp chamber dimensions, owed to secondary

dentine formation [15]. With these changes, it was expected to obtain significant

differences between group pre, and post, with higher percentages of over estimates of age

in treated participants.

Analysis of the obtained data facilitated the calculation of the potential risk of having over

estimation in participants after finishing the orthodontic treatment (Table 3.1). After

calculating the incidence of overestimation in groups pre, and post, there was no evidence

of association between over-estimation of age and orthodontic treatment (RR=1.0071).

105

However, it is necessary to mention that in this study none of the tested teeth showed

signs of severe apical root resorption.

Previous studies have found that, maxillary teeth are more affected by root shortening due

to orthodontic treatment, especially, lateral and central incisors [22, 23]. In this study, they

also presented the higher percentage of underestimates of age for both groups, followed

by the mandibular lateral incisor. In the case of under estimation, it was observed in a

higher percentage in lower canines for group pre, and group post, 54% and 61%

respectively.

Previous studies using the Kvaal et al. method and their proposed formulae, reported a

constant under-estimation of age, from 18 to 20 years [19] up to 47.10 years [24]. These

studies did not use population specific formulae. In this study, the formulae used were

obtained from the same Western Australian population. There was a clear cutline (24 years

for both groups) where the estimates would notably exceed the reported SEE (1 to 5

years), having statistically significant under estimates.

It is necessary to determine in future studies if this finding could be related to the fact that

the apposition of secondary dentine does not occur in a linear manner through life [24],

causing a larger decrease of pulp dimensions between 20 to 40 years of age, than between

40 to 60 years of age [25]. A limiting factor to clearly examine this relation in the current

study was the lack of individual over 40 years of age.

106

As the main objective of this study was to establish if orthodontic treatment would

generate changes in dental age estimation, and as different teeth are affected with

different severity degrees, in this study previously published formulae were used for dental

age estimation for individual teeth rather than per set of teeth or per individual.

The use of the Kvaal et al [7] individual teeth formulae to estimate dental age has been

reported on extracted teeth [26]. In the same way, there are other methods based on the

assessment of a single tooth per individual to generate dental age estimation models with

acceptable results in a forensic framework [5, 7, 13].

CONCLUSION.

In this current study, orthodontic treatment did not affect the final results when the Kvaal

et al. method was used for dental age estimation. Although it has been previously

established that, to use the pulp complex as a biomarker for general ageing, the analysed

teeth have to fulfil the requirements of being in normal and functional occlusion, totally

sound and free from dental procedures [7], the real effect of different dental conditions on

methods based on the formation of secondary dentine, for dental age estimation, has not

been tested. The results of this current study allow forensic dentists to use the Kvaal et al

method in participants who have had previous orthodontic treatment, when there are no

signs of severe apical root resorption.

107

REFERENCES

1. Saunders E. The teeth a test of age. Lancet. 1838;30(774):492-6.

2. Ebrahim E, Prasanna KR, Laxmikanth C, Prashanth S, Veena K, Rachanna V, et al.

Dental Age Estimation Using Schour and Massler Method in South Indian

Children. Sch. J. App. Med. Sci. 2014;2(5C):1669-74.


Hum Biol. 1973;45(2):211.


Dentomaxillofacial Radiol. 2011;40(4):199-212.



Sex: A Preliminary Study. J Forensic Sci. 2011;56(3):766-70

6. Celikoglu M, Kamak H. Patterns of third-molar agenesis in an orthodontic

patient population with different skeletal malocclusions. Angle Orthod.

2012;82(1):165-9.





108






Kvaal et al. Forensic Sci Int. 2014;235:104. e1-e6.







13. Ge Z-p, Ma R-h, Li G, Zhang J-z, Ma X-c. Age estimation based on pulp chamber


Sci Int. 2015; 253:133.e1-.e7.

14. Baumrind S, Korn EL, Boyd RL. Apical root resorption in orthodontically treated

adults. Am J Orthod Dentofacial Orthop. 1996;110(3):311-20.



Endod. 2014;40(11):1758-63.

109

16. Sameshima G. Asgarifar K. Assessment of root resorption and root shape:

periapical vs panoramic films. Angle Orthod. 2001;71(3):185-189

17. Lazzaretti, Bortoluzzi, Torres Fernandes, Rodriguez, Grehs, and Martins

Hartmann. Histologic Evaluation of Human Pulp Tissue after Orthodontic

Intrusion. J Endod.2014;40(10):1537-540.

18. Marroquin TY, Karkhanis S, Kvaal SI, Vasudavan S, Castelblanco E, Kruger E,

Tennant M. Determining the effectiveness of adult measures of standardised

age estimation on juveniles in a Western Australian population. Aust J Forensic

S. 2016 May 22:1-9.



2012;57(3):277-84.

20. Krishnan V, Davidovitch Ze. Cellular, molecular, and tissue-level reactions to

orthodontic force. Am J Orthod Dentofacial Orthop. 2006;129(4):469.e1-.e32.

21. Preoteasa CT, Ionescu E, Preoteasa E, Tenreiro Machado JA, Baleanu MC,

Baleanu D. Multidimensional Scaling for Orthodontic Root Resorption. Math

Prob Eng. 2013;2013.

22. Brezniak N, Wasserstein A. Root resorption after orthodontic treatment: Part 2.

Literature review. Am J Orthod Dentofacial Orthop. 1993;103(2):138-46.

110

23. Silva Campos MJ, Silva KS, Gravina MA, Fraga MR, Vitral RWF. Apical root

resorption: The dark side of the root. Am J Orthod Dentofacial Orthop.

2013;143(4):492-8.

24. Meinl A, Tangl S, Pernicka E, Fenes C, Watzek G. On the applicability of

secondary dentin formation to radiological age estimation in young adults. J

Forensic Sci. 2007;52(2):438-41.

25. Oi T, Saka H, Ide Y. Three-dimensional observation of pulp cavities in the

maxillary first premolar tooth using micro-CT. Int Endod J. 2004;37(1):46-51.


methods in adults: intra-and inter-observer effects. Forensic Sci Int. 2002;

126:221-226

111

SECTION THREE: New uses of the Kvaal et al method for age estimation in young individuals

and adults from different groups of population.

Preface

The studies presented in this section explore the use of the Kvaal et al method in two new

scenarios. In Chapter 4, this method is applied on a sample that contains data from

panoramic radiographs of young individuals (14 to 30 years of age) from a Western

Australian Population. The objective was to address one of the issues faced by forensic

dentists: individuals that present agenesis of third molars, which limits the use of methods

based on tooth formation.

Chapter 5 presents a new proposal to implement the Kvaal et al method on computed

tomography (CT) and cone beam computer tomography (CBCT). It also explores the

pulp/tooth width ratio calculation not only in the coronal view of the tooth, observable in

dental radiographs, but also in the sagittal view, which is only observable with this last

imaging techniques. Simultaneously, in this chapter, one of the controversies, or previously

stated principles of dental age estimation (that only sound teeth should be used), was

tested by the use of teeth with a variety of different conditions.

112

4. Chapter 4. Determining the effectiveness of adult measures of standardized age

estimation on juveniles in a Western Australian population.



Tennant, M. (2016). Determining the effectiveness of adult measures of standardised age

estimation on juveniles in a Western Australian population. Australian Journal of Forensic

Sciences. Accepted for publication. March 2016.

113

ABSTRACT

The estimation of chronological age through assessment of dental radiographs is well-

established and a useful method to assist in the identification of persons in forensic and

anthropological scenarios.

The objective of this investigation was twofold: i) to validate the Kvaal et al. age-estimation

method on a sample of Western Australian subjects, and ii) to increase the range of

chronological ages to which the Kvaal et al. method can be applied. The sample size

included panoramic radiographs from 74 subjects (aged 12-28 years).

A set of ratios was calculated and then used to apply different statistical models of linear

regression, to generate a final formula to estimate age. The most accurate estimations

were obtained from the models generated by the mandibular canine measurement (SEE

±3.708 years), and for the 3 mandibular teeth (SEE ±3.388 years). The results indicate that

inclusion of juveniles did not affect the final results, and the method still produced

estimates acceptable in a forensic framework.

Keywords: Forensic dentistry, dental age estimation, secondary dentine formation, young

adults.

114

INTRODUCTION

The accurate approximation of chronological age, secondary to sex determination, is an

integral factor in constructing a biological profile for forensic [1] and anthropological

purposes, including the confirmation of identification at times of mass disasters, crimes,

accidents and of unknown remains [2-4]. Further, the estimation of the chronological age

in the living is also needed in situations such as immigration and refugee [4]. Saunders

originally proposed the estimation of chronological age based on different stages of tooth

eruption [5], since then several more methods have been suggested to estimate dental age

in children.

Amongst these various methods, those that are based on the radiological examination of

permanent teeth development, are the most accurate [6, 7]. The Demirjian et al. [8, 9]

classification is reported as the best method for dental age estimation in children and

adolescents, thanks to its high observer agreement and correlation between the defined

stages and age [10]. The Demirjian et al. method has been adapted to different

populations, showing a mean difference between the chronological age and the dental age

of around one year in different studies [6, 9, 11].

The completion of tooth development is marked with the closure of the root apex, with the

third molar being the last tooth to complete its eruption and root development at

approximately 17-21 years of age. Following this stage, the accurate estimation of

chronological age based on dental formation becomes challenging [12], leading to the

115

development of alternative methods in adults. Bodecker [13] identified the relation of the

continued apposition of secondary dentine and the subsequent change in the morphology

of the pulp chamber, with chronological age. Accordingly, several dental age estimation

methods have been developed, based on the relation between secondary dentine

formation, and the subsequent narrowing and change in pulp chamber dimensions and

shape with age [1, 2, 14, 15].

Dental analysis using radiographs for the purpose of age estimation presents numerous

advantages over other histological and biochemical methods; it is relatively simple, non-

invasive, and economically viable [16]. The method developed by Kvaal et al. [1] is a

relatively non-invasive method and has been validated in diverse populations [17, 20].

Karkhanis et al. [17] applied the Kvaal et al. [1] method in a Western Australian population

to develop age estimation standards with acceptable results in a forensic framework (±10

years) [2].

Originally, the Kvaal et al. [1] method was proposed to be applied in an adult population.

More recently, the method has been applied in younger populations with varying degrees

of success. The level of error in estimates of ages remains high in younger populations with

Landa et al. reporting a standard deviation approaching 15 years [20].

The aim of the present study was to validate the applicability of the Kvaal et al. [1] method

in a younger population (Western Australia sample) and to provide further refinement of

the level of variation in younger people. This also assessed the potential of this method as

116

an additional tool for dental age estimation in juveniles, where methods based on the

analysis of tooth development cannot be used.


This study received ethics approval from the Human Research Ethics Committee of The

University of Western Australia (Ref: RA/4/1/6797).

Sample Selection

The study cohort consisted of a series of subjects who consecutively presented for

orthodontic treatment at a private specialist orthodontic office. From the initial sample, 74

Western Australians were aged less than 30 years, of those 63.5% (n=47) were female and

36.5% (n=27) were male, with a median age of 16 years for both genders. All panoramic

radiographs were acquired from the same machine in digital format.

Analyses were completed with Image J software (version 1.48 19 April 2014 - National

Institute of Health, USA) and the measurements were completed by a single observer (TM).

All data were collated using Excel (version 2013 Microsoft Redmont, USA), and statistical

analysis was completed using R Core Team version 3.1.3 (2015). (R: A language and

environment for statistical computing. R Foundation for Statistical Computing, Vienna,

Austria. URL http://www.R-project.org/). All subjects with a panoramic radiograph of high

quality with respect to factors of image brightness, contrast and sharpness were included.


117

Any radiographs that did not meet this requirement or had observable failings (eg. image

distortion, poor contrast, superposition of tooth structure, or improper positioning) were

excluded. All teeth included in the analysis were clinically sound with completed root

formation and at occlusal plane level. Teeth with rotations, incomplete root formation,

dilacerations, pulpal pathologies, and endodontic treatment, and/or restorations were

excluded. Finally, teeth with large areas of enamel overlap between neighboring teeth

were excluded in this study.

Teeth analysed

Following the parameters as provided by Kvaal et al. [1], the teeth analysed were the

maxillary central incisors (with Federation Dentaire International (FDI) notation 11 and 21)

lateral incisors (FDI notation 12 and 22), second premolars (FDI notation 15 and 25),

mandibular lateral incisors (FDI notation 32 and 42), mandibular canines (FDI notation 33,

and 43) and first mandibular premolars (FDI notation 34 and 44). In the original study,

Kvaal et al. [1] analysed only those periapical radiographs from individuals where all the six

teeth were present and suitable for examination.

In the present study, panoramic radiographs with any combination of the required teeth

available for measurement were included. Previous research [1, 17, 19] has demonstrated

the absence of bilateral asymmetry in the deposition of secondary dentine. Consequently,

teeth from either side that fulfilled the inclusion criteria were analysed for the purpose of

developing age estimation standards.

118

Measurements

Odontometric data were acquired following the methodological approach of Kvaal et al. [1].

In this way, the data were obtained from measuring: maximum tooth length (from the

incisal border or cusp tip to the root apex), maximum pulp length (from the pulp chamber

roof to the root apex), and maximum root length (from the cemento-enamel junction (CEJ)

on the mesial surface of each tooth to the root apex). The pulp chamber and root width

measurements were collected at the points A (CEJ), B (mid-point between the points A and

C) and C (mid-point between the CEJ and root apex).


A pilot set of 30 panoramic radiographs, which were not included in the final analysis, was

used to perform an initial training and benchmarking to an expert in the field (SK). Six of

the 30 radiographs were measured on 6 different days, in addition to 4 randomly selected

panoramic radiographs, thus resulting in the analysis of 10 panoramic radiographs per day.

Based on this, intra- and inter-observer error calculations were performed [17].

The measurements were acquired from all the panoramic radiographs with a minimum of

one day between each session to minimize the possibility of memorizing reference points

and/or measurements in each observer. A second intra- and inter-observer calibration was

performed using 5 randomly selected panoramic radiographs from the final study sample,

and recording the measurements on 5 different days, also using 4 additional panoramic

radiographs per day, to avoid measurements memorizing, with at least one day between

each evaluation.

119

Statistical Analysis

Simple descriptive statistics (mean, standard deviation, and frequency distributions) were

used to summarize the data. Initial tests for normality (assessment for skewness, kurtosis

and Shapiro-Wilk) were performed to determine, where appropriate, parametric and non-

parametric univariate analysis testing for the continuous variables (Table 4.1). Pearson’s

correlation coefficient (r) was calculated to assess the strength of correlation between age

and dental ratios based on the Kvaal et al. [1] method (Table 4.1).

The correlation coefficients calculated (using the width ratios) presented significant

correlation values. The most representative were observed for the ratio calculated

between pulp and root width at the B point of the root, and for the predictor W-L. The

mandibular canines showed the highest correlation coefficient for the predictors B (-0.570)

and W-L (-0.625).

To examine the association of several factors with chronological age, multivariable linear

regression analyses were used to look at the association between the independent and

outcome variables. The outcome variables included chronological age. The predictor

variable M and W-L were used as the independent variables. Age estimation models were

developed individually for each tooth (Table 4.2), and the following tooth combinations

(Table 4.3): Three maxillary teeth with the predictors M and W-L individually introduced for

each tooth, (6 predictors), and the average of M and W-L (2 predictors).

120

Three mandibular teeth with the predictors M and W-L individually introduced for each

tooth, having 6 predictors in the equation, and the average of M and W-L resulting in 2

predictors. And mandibular and maxillary teeth M and W-L predictors, introduced

individually in the equation (12 predictors) and the average of M and W-L (2 predictors). All

regression models were built using the ordinary least squares approach. R-square values

were computed to examine the amount of variance explained by the predictor variables.

All statistical tests were two-sided and a p- value of less than 0.01 was considered to be

statistically significant. Corrections were made for multiplicity using a modified Bonferonni

method to reduce the likelihood of Type I errors; an alpha threshold for statistical

significance for all comparisons was set at 0.01. Univariate analyses and multivariable

linear regression analyses were performed using R Core Team version 3.1.3 (2015). (R: A

language and environment for statistical computing. R Foundation for Statistical

Computing, Vienna, Austria. URL http://www.R-project.org/).

RESULTS


The technical error of measurement (TEM), relative technical error of measurement (rTEM)

and coefficient of reliability were within acceptable standards for the intra- and inter-

observer measurement precision (TEM<1.0, rTEM<5%, R>0.75). These values were 0.92,

2.99% and 0.95 respectively for intra-observer precision analysis. Similarly, the inter-


121

Table 4.1. Correlation coefficients between chronological age and Kvaal dental

measurements and ratios, for individuals under 30 years of age.

TOOTH NUMBER (FDI numbering system)

RATIO 11/21 12/22 15/25 32/42 33/43 34/44

P 0.13NS 0.23 NS 0.13 NS 0.16 NS 0.45* 0.15 NS

T 0.04 NS -0.07 NS -0.002 NS -0.13 NS 0.17 NS -0.12 NS

R 0.29* 0.29* 0.15 NS 0.33* 0.28* 0.39**

A -0.13 NS -0.28* -0.004 NS -0.26* -0.25 NS -0.14 NS

B -0.38* -0.30* -0.28* -0.35* -0.57** -0.38*

C -0.34* -0.38* 0.01 NS -0.09 NS -0.61** -0.5**

M -0.05 NS -0.2 NS 0.0001 NS -0.12 NS -0.11 NS -0.26*

W -0.41** -0.39** -0.15 NS -0.23* -0.12 NS -0.49**

L 0.15 NS 0.29* -0.002 NS 0.25* 0.47** 0.29*

W-L -0.31* -0.42** -0.05 NS -0.36* -0.62** -0.54**

*= p<0.05 **=p<0.01 NS= Not significant

Abbreviations: P, ratio between length of pulp and root; T, ratio between length of tooth and root; R, ratio between length of pulp and root; A , ratio between width of pulp and root at CEJ (Level A); B, ratio between width of the pulp and root at mid-point between level C and A (level B); C, ratio between width of pulp and root at mid-root level (level C); M, mean value of all ratios (first predictor); W, mean value of width ratios from levels B and C; L mean value of the length ratios P and R; W-L, difference between W and L (second predictor)(1).

122

Graph 4.1. Scatter plot showing a positive association between the estimated age in

years and the chronological age in years for the teeth 33/43

Table 4.2. Multiple regression for estimation of chronological age (in years) from

individual maxillary and mandibular teeth from individuals under 30 years of age .

Tooth (FDI) n R R2 Equation SEE ± years

11/21 69 0.15 0.12 Age= 22.915-28.78 (M) -22.376 (W-L) 4.3

12/22 67 0.19 0.17 Age= 19.724-25.935 (M) – 23.631(W-L) 4.2

15/25 68 0.003 -0.27 Age= 17.78-3.962(M)-2.204(W-L) 4.5

32/42 71 0.13 0.10 Age= 4.644-5.363(M)-22.558(W-L) 4.4

33/43 48 0.41 0.39 Age= 11.26-41.33(M)-86.06(W-L) 3.7

34/44 71 0.32 0.31 Age= 10.513-27.075(M)-38.176(W-L) 3.7

Note. *R2, coefficient of determination. SEE, standard error of estimation in years. See table 4.1. for abbreviations.

12

3

Table 4.3. Multiple regressions for estimation of chronological age (in years) from the combined maxillary and mandibular teeth.

Teeth (FDI) n R R2 Equation SEE ± years

3mx (2pds) 60 0.15 0.12 Age= 28.702-46.338(M) - 24.233(W-L) 4.1

3mx(6pds) 60 0.23 0.15 Age=24.394-30.751 (11/21M) -12.090 (11/21W-L) - 29.2257 (12/22M)- 12.894 (12/22W-L) + 2.611 (15/25M) - 0.631 (15/25 W-L)

4.1

3mdb(2pds) 45 0.41 0.38 Age= -27.44 + 10.48(M) – 63.82 (W-L) 3.6

3 mdb (6pds) 45 0.53 0.46 Age= 1.894 + 13.428 (32/42M) -5.931 (32/42W-L) - 51.067 (33/43M) -66.495 (33/43W-L) - 5.114 (34/44M) -16.806 (34/44W-L)

3.3

6 teeth (2 pds) 40 0.25 0.21 Age= 31.91 - 63.34 (M) - 63.34 (W-L) 4.1

6 teeth (12 pds) 30 0.54 0.33 Age= 2.627 -58.42 (11/21 M) + 4.526 (11/21 W-L) + 32.169 (12/22 M) - 12.913 (12/22 W-L) + 25.460 (15/25M) + 1.098 (15/25 W-L) + 5.270 (32/42 M) + 0.096 (32/42 W-L) - 28.692 (33/43 M) + 1.686 (33/43 W-L) - 13.857 (34/44 M) - 49.336 (34/44 W-L)

3.7

*R2, coefficient of determination. SEE, standard error of estimation in years. See table 4.1 for abbreviations.

124

observer calibration (between TM and SK) showed comparable precision [21] (TEM 0.80,

rTEM 2.37 and R 0.98). In this way, individual (Table 4.2) and multiple (Table 4.3) tooth

regression models were obtained. The accuracy to predict the age was quantified by the

standard error of estimation (SEE ±years). The most accurate age estimation model based

on the analysis of individual tooth was for mandibular canines (SEE ±3.708 years) (Graph

4.1). Prediction accuracy improved with the combined analysis of teeth; it was highest for

the model developed with the three mandibular teeth (6 predictors, SEE ± 3.388 years).

DISCUSSION

The estimation of chronological age has been a very relevant topic of discussion through

human history. For this reason, diverse skeletal and dental methods have been developed,

showing variable levels of reliability in different populations. Those methods, based on

dental changes, are well accepted and have shown their potential for forensic and

anthropological applications [22].

Once root formation has been completed, progressive dental changes such as secondary

dentine formation can be radiographically assessed based on the narrowing on the pulp

chamber. Odontometric and morphometric measurements can thus be acquired to

quantify secondary dentine deposition, and are the foundation for non-invasive adult age

estimation methods such as Kvaal et al. [1], and Cameriere et al. [23].

125

The Kvaal et al. [1] method has been widely validated in different populations, because of

numerous benefits such as its relatively non-invasive approach, applicability in live

individuals, and low cost, among others. Previous research has applied this method to a

Western Australian population [17], and showed significant results that are valid under

forensic standards. Although the Kvaal et al. method [1] was initially developed to estimate

age in individuals over 20 years old, previous studies have tested this method in younger

populations [19].

The present study applied the Kvaal et al. method [1] in Western Australian participants

under 30 years of age. The primary aim was to assess the applicability of this method in a

younger population to broaden the age range that the method can be applied to, without

compromising the age estimation accuracy.

To this end, regression models were developed based on the data acquired from individual

teeth (Table 4.2) and a combination of teeth (Table 4.3). The accuracy to predict age was

quantified by the standard error of estimation (SEE ±years). The most accurate model for

individual teeth was for the mandibular canines (±3.708 years), in contrast to the study of

Karkhanis et al. [17], where the same tooth presented the highest SEE (±10.903 years).

Prediction accuracy improved when multiple teeth were included in the regression models

(Table 4.3). In this study, the highest level of accuracy was obtained when the equation

included 3 mandibular teeth and 6 predictors, (SEE ±3.388 years), in comparison with the

126

study of Karkhanis et al. [17], where the most accurate model was for the combined

analysis of the 6 teeth and 12 predictors (SEE ±7.963).

Age estimation in individuals older than 14 years is difficult, as all permanent teeth, except

the third molars (when present), have completed their development [19]. Some examples

of previous studies amongst people under 30 years of age, using the Kvaal et al. method [1]

on panoramic radiographs, are: Erbudak et al. [24] in a Turkish population, including in their

sample 75 participants between 14 to 35 years of age, obtaining a SEE= ± 8.73 years at

best, using the regression formulas of Paewinsky et al. [25] and Kvaal et al. [1].

Further such studies were published by: Landa et al. [20] in a Spanish population with an

age range of 14 to 60 (n=100,) from which 40 idividuals were aged younger than 31 years

of age, and with an underestimation of age when using the regression formulae of Kvaal et

al. [1], and Paewinsky et al. [25], and a standard deviation of 12.53 at best, when the Kvaal

et al [1]. regression equation was used.

The last example is the study of Meinl et al. [19], which was conducted in an Austrian

population (n=44) with an age range of 13 to 24 years, and resulted in a mean

underestimation of 31.44 years when the Kvaal et al. [1] regression models were applied,

or a mean overestimation of 20.88 years when Paewinsky et al. [25] formula was applied.

In contrast, the present study (n=74) shows that the Kvaal et al. [1] method provides

acceptable results [2] in a sample composed of sub-adults and young adults.

127

Further research is warranted, in other populations with similar age ranges and using larger

samples. It is also insightful to compare the results of this study with other methods based

on third molar development as an estimator of chronological age. One of the most

remarkable studies is the research performed by Lewis and Senn [10]. They reported a

standard deviation of no more than 3 years, when different methods are applied in a

North-American population. Another method based on third molars is the examination of

the periodontal membrane in lower third molars in a German population [26], which found

a standard deviation of between 1.9 to 4.8 years. However, third molar agenesis has been

reported to be between 14% up to 51% in different studies [27]. The Kvaal et al. [1]

method can be considered as an alternative in these cases.

CONCLUSION

Panoramic radiographs are a unique diagnostic tool, but also provide useful information,

valid for forensic purposes. It has been well established that dental records are a highly

precise instrument for determining an individual’s identity. The use of the Kvaal et al. [1]

method was initially purposed on periapical radiographs, obtained from adults, but lately

this method was applied using panoramic radiographs, and included juvenile participants.

The validation of the Kvaal et al. [1] method for age estimation, using panoramic

radiographs obtained from a population including juvenile individuals, enlarges the range of

ages where the ratio between secondary dentine production and decrease of pulp

chamber can be applied. This also presents the opportunity to examine the application of

this method in other juvenile population groups.

128

REFERENCES

1. Kvaal SI, Kolltveit KM, Thomsen IO, Solheim T. Age estimation of adults from dental

radiographs. Forensic Sci Int. 1995;74(3):175-185.

2. Mathew DG, Rajesh S, Koshi E, Priya LE, Nair AS, Mohan A. Adult forensic age

estimation using mandibular first molar radiographs: A novel technique. J Forensic

Dent Sci. 2013;5(1):56-59.

3. Sheelavant SS, Chandra GYP, Harish S. Estimation of age from the physiological

changes of teeth in adults between 25 - 60 years. An autopsy study. J Forensic Med

Toxicol. 2014;8(1):202-207.



2005;153(2):208-212.

5. Saunders E. The teeth a test of age. The Lancet. 1838;30(774):492-496.

6. Yusof MY, Thevissen PW, Fieuws S, Willems G. Dental age estimation in Malay

children based on all permanent teeth types. Int J Leg Med. 2014;128(2):329-333.


and future directions. Leg Med. 2010;12(1):1-7.


Human Biol. 1973;45(2):211-227.

129

9. Feijóo G, Barbería E, De Nova J, Prieto JL. Dental age estimation in Spanish children.

Forensic Sci Int. 2012;223(1–3):371.e1-.e5.

10. Lewis JM, Senn DR. Dental age estimation utilizing third molar development: A

review of principles, methods, and population studies used in the United States.

Forensic Sci Int. 2010;201(1–3):79-83.

11. Kırzıoğlu Z, Ceyhan D. Accuracy of different dental age estimation methods on

Turkish children. Forensic Sci Int. 2012;216(1–3):61-67.


Dentomaxillofacial Radiol. 2011;40(4):199-212.

13. Bodecker CF. A consideration of some of the changes in the teeth from young to old

age. Dental Cosmos. 1925;67:543-549.



15. Drusini AG, Toso O, Ranzato C. The coronal pulp cavity index: a biomarker for age

determination in human adults. Am J Phys Anthropology. 1997;103(3):353-363.

16. Sarkar S, Kailasam S, Mahesh Kumar P. Accuracy of estimation of dental age in

comparison with chronological age in Indian population – A comparative analysis of

two formulas. J Forensic Leg Med. 2013;20(4):230-233.

130

17. Karkhanis S, Mack P, Franklin D. Age estimation standards for a Western Australian

population using the dental age estimation technique developed by Kvaal et al.

Forensic Sci Int. 2014;235(0):104.e1-.104.e6.

18. Kanchan-Talreja P, Acharya AB, Naikmasur VG. An assessment of the versatility of

Kvaal's method of adult dental age estimation in Indians. Arch Oral Biol.

2012;57(3):277-284.

19. Meinl A, Tangl S, Pernicka E, Fenes C, Watzek G. On the applicability of secondary

dentin formation to radiological age estimation in young adults. J Forensic Sci.

2007;52(2):438-441.

20. Landa MI, Garamendi PM, Botella MC, Aleman I. Application of the method of Kvaal

et al. to digital orthopantomograms. Int J Legal Med. 2009;123(2):123-128.

21. Lewis SJ. Quantifying measurement error: Oxbow Books. 1999. Current and recent

research in osteoarchaeology 2: Proceedings of the 4th, 5th and 6th meetings of

the Osteoarchaeological Research Group, Anderson IS, editor; 54-55.

22. Manual of Forensic Odontology. 5th ed. Edited by, David R. Senn, Richard A.

Weems. Boca Raton, FL: Boca Raton, FL : CRC Press; 2013.



2012;214(1–3):105-112.

131

24. Erbudak HO, Ozbek M, Uysal S, Karabulut E. Application of Kvaal’s et al. age

estimation method to panoramic radiographs from Turkish individuals. Forensic Sci

Int. 2012;219(1-3):141-146.




26. Olze A, Solheim T, Schulz R, Kupfer M, Pfeiffer H, Schmeling A. Assessment of the

radiographic visibility of the periodontal ligament in the lower third molars for the

purpose of forensic age estimation in living individuals. Int J Legal Med.

2010;124(5):445-448.

27. Celikoglu M, Kamak H. Patterns of third-molar agenesis in an orthodontic patient

population with different skeletal malocclusions. The Angle Orthodontist.

2012;82(1):165-169.

132

5. Chapter 5. Application of the Kvaal method for adult dental age estimation using

Cone Beam Computed Tomography (CBCT).




dental age estimation using Cone Beam Computed Tomography (CBCT). Journal of Forensic

and Legal Medicine, Accepted for publication. October 2016.

133

ABSTRACT

Different non-invasive methods have been proposed for dental age estimation in adults,

with the Kvaal et al. method as one of the more frequently tested in different populations.

The purpose of this study was to apply the Kvaal et al. method for dental age estimation on

modern volumetric data from 3D digital systems. To this end, 101 CBCT images from a

Malaysian population were used. Fifty-five per cent were female (n=55), and forty-five

percent were male (n=46), with a median age of 31 years for both sexes. As tomography

allows the observer to obtain a sagittal and coronal view of the teeth, the Kvaal pulp/root

width measurements and ratios were calculated in the bucco-lingual and mesio-distal

aspects of the tooth.

From these data, different linear regression models and formulae were built. The most

accurate models for estimating age were obtained from a diverse combination of

measurements (SEE ±10.58 years), and for the mesio-distal measurements of the central

incisor at level A (SEE ±12.84 years). This accuracy, however is outside an acceptable range

for forensic application (SEE ±10 years), and is also more time consuming than the original

approach based on dental radiographs.

Key words: Adults, age estimation, tomography, Kvaal method.

134

INTRODUCTION

Currently, the need for developing more accurate and non-invasive methods for age

estimation, as part of the identification of adult individuals in situations of forensic

scenarios is increasing globally [1]. Gustafson first proposed a method for dental age

estimation in adults, [2] since then, the analysis of dental changes in adults relative to

chronological age has continued. The adult dentition is relatively resistant to

environmental and chemical influences, and methods based on the assessment of teeth are

considered advantageous for this reason [3, 4].

Furthermore, methods based on odontometric measurement of pulp and tooth structures

from dental radiographs or tomographs are non-invasive/non-destructive and can be

applied to both living or deceased individuals [5]. Kvaal et al, proposed a method for age

estimation in adults which has been tested in different populations around the world since

1995, but all reported a high standard error (±8.5 to ±13 years) [1, 5, 6]. The basis of this

method is the analysis of the narrowing of the pulp chamber with age, which is observable

and measurable in dental radiographs. However, there is scope for improvement in the

accuracy of the age estimations.

With the emergence of computer tomography (CT) and cone beam computer tomography

(CBCT), new methods based on volumetric reconstruction of tooth and pulp volume and

ratio calculations, have been proposed for dental age estimation in adults [7, 8]. However,

those techniques have not provided better accuracy relative to the previously described

135

Kvaal et al. method, using dental radiographs. Moreover, pulp/tooth volume calculation is

labor intensive, [9] and can require the use of complex, and often costly computer software

[10]. Clearly, these more technical multi-dimensional systems provide considerably more

data, at a substantially more detailed level, and with a reduction in many of the

complications of simple plane film images.

The method described by Kvaal et al. [5] was initially proposed, with measuring dimensions

on periapical radiographs with a microscope and a pair of calligraphers. The applications of

this method have been expanded to include panoramic radiographs using different

analytical software [6, 11]. The purpose of the present study is to apply the Kvaal et al. [5]

method for dental age estimation using modern computer tomographic data from 3D

digital systems. The null hypothesis is that the substantially increased quantity (and

quality) of the data from these systems will provide greater accuracy of dental age

estimation in adults.


Sample

The study sample included a total of 101 tomographs obtained as part of normal treatment

from the radiology department of the University of Keibangsaan Malaysia. Almost half (n=

47) were obtained with a Kodak device, reference K9000-3D (exposure parameters:

scanning time: 10.8 s, 3-15 mA, 60-85kVp, field of view FOV: 5.2 cm x 5.2cm to 5.5cm to

5.5cm, 180° rotation, slice thickness 0.076-0.2 mm), and 54 were obtained with an i-CAT

136

device (Dental Imaging Technologies Corporation. i-CAT® FLXTM) (exposure parameters:

scanning time 4s, 5 mA, 120 kVp, FOV: 16cm x 16cm, Slice thickness 0.3 mm to 0.4 mm). Of

these tomographs, 55% were from female (n=55), and 45% were from male (n=46)

patients, with a median age of 31 years for both sexes (Table 5.1).

All images were received in DICOM format, and analysed using the Osirix® software

package (OnDemand 3D software CyberMed Inc, Seoul, South Korea). The study sample

included teeth with very small coronal restorations, as long as they did not compromise the

pulp chamber and secondary caries was not present. Teeth with shape abnormalities,

active caries, root canal treatment, prosthetic restorations, signs of pulp or periapical

pathologies, or open root apex, were duly excluded from the sample. Multi-rooted upper

second premolars were also excluded.

Teeth analysed

Initially, Kvaal et al. [5] proposed a set of linear measurements, including the length of the

tooth, pulp, and root, in addition to the width of the root and pulp chamber at different

levels, in six different teeth: upper central and lateral incisors, and the second premolar;

and the lower lateral incisor, canine and first premolar. In the present study, due to the

unclear definition of the pulp chamber and tooth boundaries for the lower teeth (most

likely related to their small dimensions) only upper teeth were included: central incisor,

lateral incisor, canine, and second upper premolar (when possible). As the dental images

are tomographs, it is possible to assess the teeth in sagittal and coronal views; accordingly,

the root length, and pulp/tooth width measurements proposed by Kvaal et al. [5] were

137

acquired in the mesio-distal and bucco-lingual aspects of the teeth (see below). Significant

differences between teeth of the left and right side of the maxilla have not been reported

previously, [5] and therefore in this study, teeth from either the left or right side were

measured.

Measurements

As proposed by Kvaal et al., [5] the mesio-distal measurements of the teeth (coronal view)

were performed as follows: first, the level A, was located at the CEJ (cemento-enamel

junction) on the mesial side of the tooth, then level C was located halfway between point A

and the root apex, and finally point B was located halfway between point A and C.

As this is the first study applying the Kvaal et al. method exclusively using tomographs,

bucco-lingual measurements were also acquired (sagittal view). Point A was located at the

vestibular CEJ, point C halfway between the vestibular CEJ and the root apex, and point B,

halfway between point A and point C. In this way, there were six pulp/tooth width

measurements per tooth. All measurements were recorded by one observer (TYM) using

the Osirix® software.


Intra-observer error was tested, four tomographs were randomly selected (two Kodak and

two i-CAT) and all the measurements were recorded by one observer (TYM). Upper

central, lateral and canine were measured on four separate occasions, with a minimum of

one day between repetitions ensuring that the observer did not remember the previously

138

recorded measurements. Measurement error was assessed by calculating the technical

error of measurement (TEM), coefficient of reliability (R) and relative technical error of

measurement (rTEM) [12].

All data were collected in an Excel spreadsheet, Excel (version 2013 Microsoft, Redmont,

USA), and statistical analysis was completed using R Core Team version 3.1.3 (2015). (R: A

language and environment for statistical computing. R Foundation for Statistical

Computing, Vienna, Austria. URL http://www.R-project.org/). Descriptive statistics were

used to summarise the data (mean, standard deviation, and frequency distributions).

Normality was also tested (skewness, kurtosis and Shapiro-Wilk) to establish whether

parametric or non-parametric analysis was required.

Pearson correlation coefficient (r) was calculated to assess the relationship between the

dental ratios and age, for data obtained from the Kodak CBCT and for data obtained from i-

CAT CBCT individually, and also when both data sets were combined. This was done with

the aim of testing if the difference between the resolutions of both devices could cause any

statistically significant differences (Table 5.2).

Linear regression models were built with age designated as the dependent variable.

Different age estimation equations were then formulated using different groupings of age

predictors: 1) the tooth/pulp ratios at the levels A, B or C individually; 2) calculating the

average of the pulp/tooth ratios of all the teeth at the different levels (A, B or C) obtaining a

formula for each level; 3) calculating the average of the pulp/tooth ratios of the lateral (Lat)


139

and central (Cent) upper incisor at the different levels (A, B and C); 4). Averages of the

pulp/tooth ratios from all the teeth at the levels A, B and C in the same formula (A+B+C);

and 5) same as 4, but excluding the ratios obtained from the canines (Cent-Lat). A total of

17 linear regression models were thus generated per data-set: sagittal (s) view (bucco-

lingual measurements); coronal (c) view; and the average of both (sc). The accuracy of the

models is reported as the standard error of estimation (SEE) [11, 13].

RESULTS

Intra-observer error results were considered to be within acceptable range. (TEM <1.0 and

the coefficient of reliability R>0.80%), regardless the source of the tomograph (Kodak or i-

CAT) or whether the measurements were recorded in the bucco-lingual or mesio-distal

view of the teeth. In terms of the relative technical error of measurement (rTEM), 72% of

the observations scored <5% and 28% had an rTEM < 10%. It was also observed that the

accuracy of the observer was better for the bucco-lingual measurements (sagittal view of

the teeth) regardless the source of the tomography.

Pearson correlation coefficient test (r) demonstrated a stronger correlation between the

pulp/tooth ratios and age for the bucco-lingual measurements (sagittal (s)) than for the

mesio-distal measurements (coronal view (c)) (Table 5.2). Although there were differences

between the correlation coefficients from the Kodak and i-CAT CBCT’s, scatter-plots

including the results of both sources do not show outliers associated with the specific use.

140

There was a clear and significant negative correlation between age and pulp/tooth ratios in

all the data sets.

Table 5.1. Age and sex distribution for CBCT and i-CAT tomographs.

Table 5.2. Correlation coefficient measurements CBCT

CBCT i-CAT CBCT & i-CAT

Tooth number Coronal Sagittal Coronal Sagittal Coronal Sagittal

11/21 A -0.05 NS -0.36* -0.634** -0.65** -0.45** -0.61**

11/21 B -0.26 NS -0.37* -0.421** -0.59** -0.32* -0.5**

11/21 C -0.29 NS -0.45** -0.288* -0.36* -0.23* -0.41**

12/22 A -0.53** -0.08 NS -0.326* -0.62** -0.31* -0.41**

12/22 B -0.35* -0.5** -0.3* -0.60** -0.22* -0.46**

12/22 C -0.24 NS -0.57** -0.013 NS -0.41** -0.05 NS -0.41**

13/23 A -0.25 NS -0.13 NS -0.153 NS -0.50** -0.04 NS -0.17 NS

13/23 B -0.59** -0.56** -0.329* -0.41* -0.28** -0.39**

13/23 C -0.52** -0.34** -0.348* -0.24 NS -0.28** -0.21*

*p< 0.05, ** p< 0.01, NS= non-significant

SEX FEMALE MALE TOTAL

AGE RANGE (YEARS) CBCT (n=23) i-CAT (n=32) CBCT (n=24) i-CAT (n=22)

15-20 2 1 7 2 12

20-29 12 9 5 8 34

30-39 3 5 2 2 12

40-49 4 6 8 5 23

50-59 2 6 2 0 10

60-75 0 5 0 5 10

TOTAL 55 46 101

141

Table 5.3. Linear regression analysis, linear regression formulae, correlation coefficient

(r) between age and the ratios of bucco-lingual and mesio-distal width measurements

from CBCT & i-CAT

Abbreviations: Tooth number (11/21, 12/22 and 13/23) with Asc, Bsc or Csc= pulp/tooth ratio at the levels A, B or C in the sagittal and coronal view (sc) calculated per tooth. Asc, Bsc or Csc: average of the pulp/tooth ratio at the respective level, from all the teeth. Asc, Bsc or Csc with Lat/Cent: average of the pulp/tooth ratio only from central and lateral incisors. Asc+Bsc+Csc=average from the pulp/tooth ratio from all the teeth at the respective level. Asc+Bsc+Csc=average from the pulp/tooth ratio from all the teeth at the respective level including only lateral and central teeth.

Sagittal and coronal

Age predictors n Linear regression formulae r r2 s

11/21 A sc 86 Age=99.74-72.07(11Ac)-161.63(11As) 0.4 0.4 12.2

11/21 B sc 86 Age=81.82-35.031(11Bc)-147.14(11As) 0.26 0.24 14

11/21 C sc 86 Age=71.92-29.16(11Cc)-134.69(11Cs) 0.17 0.15 14.9

12/22 A sc 89 Age=87.43-68.05(12Ac)-129.37(12As) 0.21 0.19 14.5

12/22 B sc 89 Age=73.98-24.34(12Cb)-127.46(12Bs) 0.21 0.19 14.5

12/22 C sc 89 Age=57.56+37.84(12Cc)-148.22(12cs) 0.18 0.17 14.7

13/23 A sc 95 Age=51.14-5.4(13Ac)-47.192(13As) 0.03 0.01 15.4

13/23 B sc 95 Age=69.51-47.15(13Bc)-70.44(13Bs) 0.17 0.15 14.2

13/23 C sc 95 Age=60.81-78.34(13Cc)-34.37(13Sc) 0.09 0.08 14.9

A sc 78 Age=100.41-33.11(Ac)-206.40(As) 0.34 0.33 13.5

B sc 78 Age=89.22-20.54(Bc)-178.29(Bs) 0.31 0.29 13.9

C sc 78 Age=69.46+16.4(Cc)-174.42(Cs) 0.17 0.15 15.2

A sc Cent/Lat 83 Age=105.42-72.96(Ac1/2)-187.22(As1/2) 0.4 0.38 12.8

B sc Cent/Lat 83 Age=86.31-19.86(Bc1/2)-180.19(Bs1/2) 0.29 0.27 13.9

C sc Cent/Lat 83 Age=70.41+42.09(Cc1/2)-206.7(Cs1/2) 0.23 0.21 14.4

Asc+Bsc+Csc 74 Age= 104.09 -17.65 (Ac)-135.48 (AS)-38.18(Bc) -105.54(Bs)+40.87(Cc)+10.93 (Cs)

0.4 0.35 13.3

Asc+Bsc+Csc Lat/Cent

82 Age=106.25-180.14(Asc1/2)-135.83(Bsc1/2)+51.73(Ccs1/2)

0.4 0.39 12.7

142

Correlation coefficients were calculated as follows: i) for the measurements obtained from

the Kodak tomographs; ii) for the measurements recorded from the i-CAT tomographs; and

iii) for the combination of both. The i-CAT tomographs showed the strongest correlation to

age, especially for the upper central incisor (r= -0.65), followed by the combination of the

Kodak & i-CAT data sets (r=-0.61). The correlation coefficients obtained from the Kodak

tomographs, although statistically significant, were lower than those obtained from the i-

CAT and combined Kodak & i-CAT images (Table 5.2).

In relation to the accuracy of the different linear regression models, all estimates were over

the threshold generally accepted for forensic application (SEE ±10 years) [14]. However,

the SEE is reduced when both data-sets are combined into a linear regression model (Table

5.3). There was only one equation with an acceptable level of predictive accuracy (SEE

10.58 years (n=72, r=0.59, r2=0.56)):

Age= 101.46 + 734.74 (Ac) -374. 81 (Ac(Cent-Lat)) – 1075.18 (C sc) + 327.07 (Bc) + 310 (Cc(Cent-

Lat)) – 159.25 (11/21 As)

Abbreviations: Ac: average of the pulp/tooth ratio in the coronal view at level A from the

three teeth. Ac(Cent-Lat): average of the pulp/tooth ratios from the lateral and central

incisors. Csc: average of the pulp/tooth ratio at level C from the sagittal and coronal views

(sc) from all teeth. Bc: average of the pulp tooth ratio at the level B from all the teeth in the

coronal view. (Cc(Cent-Lat)): average of the pulp/tooth ratios at the level C from the lateral

143

and central incisors in the coronal view. 11/21 As: pulp/tooth ratio at the level A from the

central incisor in the sagittal view.

The most accurate linear regression models obtained from the individual analysis of the

bucco-lingual measurements were as follows: the central incisor at level A (Age= 84.059 –

191.003 (11/21 As), r2=0.37, SEE=±12.84), the average of the bucco-lingual measurements

from the central incisor and lateral incisor at the level A, B and C (Age= 102.08-157.94 (As) -

73.0 (Bs) -36.81 (Cs), r2=0.39, SEE=±12.73 years).

When both data-sets were combined (Table 5.3) the highest accuracy was achieved from:

the average of the bucco-lingual (s) plus the average of the mesio-distal (c) measurements

at level A, (SEE ±12.17 years); the average of the bucco-lingual and mesio-distal

measurements at level A from the central and lateral incisors, (SEE ±12.76); and the

average of the bucco-lingual and mesio-distal measurements from the central and lateral

incisors at level A, plus level B, plus level C, (SEE ±12.7). The linear regression models

generated by the inclusion from only the mesio-distal (c) measurements did not result in an

SEE lower than ±14 years, and therefore these results are not included.

DISCUSSION

This is the first study testing the Kvaal et al. method exclusively in tomographs, and also the

first one applying Kvaal et al’s. pulp/tooth width ratio calculations in the bucco-lingual

aspect of the teeth. Computer tomography is increasingly being used in dental practice

144

and provides three-dimensional information about any area of interest, in a relatively quick

and cost-effective manner [15].

In this study, it was expected that application of the Kvaal et al. method using CBCT

diagnostic images would increase the array of methods available for the forensic profiling

of living and deceased individuals.

Until now, there has been only one other study proposing a similar method for adult dental

age estimation, based on the assessment of the pulp/tooth bucco-lingual dimensions (area)

in canines [16]. Although the results of the latter study were satisfactory (Mean prediction

error, ME ±2.8 years at best), this method required extracted teeth, and thus has restricted

application in a forensic context. The use of tomographs overcomes issues related to the

requirement for tooth extraction, and also permits the analysis of multiple teeth with less

radiation exposure.

Although, in this study, the number of individuals was small in certain age intervals, there

are well documented biological age-related changes in the human dentition, which have

allowed forensic dentists to propose and use different methods for dental age estimation

[17].

In previous publications applying the Kvaal et al. [5] method for dental age estimation in

adults, one of the principal limitations is the absence of a clear limit between the dentine

and the pulp camber, [11] which is observed as a grey zone instead of a distinct line [6].

145

This affects the accuracy of the measurements and intra/inter-observer agreement.

In the present study, neither Kodak nor i-CAT tomographs were exempt from this issue. In

certain cases, the border between the pulp chamber and dentine, or tooth and bone are

also more blurred than in periapical radiographs.

It is noteworthy that dental Kodak tomographs may have higher resolution than i-CAT

tomographs. However, in the present study the data obtained from the i-CAT had a higher

correlation to age compared to the Kodak scans (Table 5.2). Similarly, the bucco-lingual

width ratios had a higher correlation to age than the mesio-distal width ratios (with the

latter measurable in periapical radiographs).

It has previously been reported that Kvaal et al. length measurements are both: difficult to

register, and do not provide significant information to the final equation [5, 18]. It is

important to note that in the tomographs analysed in the present study, it was even more

difficult to observe the root apex compared to periapical radiographs.

The only length measurement registered was the root length from the CEJ towards the

apex to locate points A, B, and C. In this way, only pulp/tooth width, and ratios were used

as the only age predictors. The exclusive use of pulp/tooth ratios proposed by Kvaal et al.

[5] has previously been documented [18] on panoramic radiographs, showing more

accurate results (SEE=±10.02 years) than those obtained in this study. (SEE=±10.58 years at

best).

146

In a previous study that analysed the odontometric age-related changes in different teeth,

the rate of secondary dentine formation varied depending on tooth type: in canines, the

increase of coronal dentinal thickness was not as high as for the incisor, and premolars [19].

In the present study, it was similarly observed that although the correlation coefficient

between pulp/tooth ratios and age were significant, the exclusion of canine ratios when the

linear regression models were generated, improved their accuracy.

On the other hand, this is also the first study documenting the inclusion of teeth with small

crown fillings in the sample to generate linear regression formulae, in contrast to previous

publications using the Kvaal et al. method, [5] where one of the inclusion criteria was to use

only totally clinically sound teeth. There were no statistically significant differences caused

by inclusion of these teeth. This may relate to the nature of the formation of tertiary

dentine or reactionary dentine to external threats. Tertiary dentine formation is highly

dependent on the stimulus, and that caries mediators induce a focal dentine production,

[19-21] with a rate formation of about 3µm per day [22].

Regression analysis, has been chosen for age estimation due to its simplicity [23].

Nevertheless, it is necessary to note the limitations of the use of linear regression models

for dental age estimation; for example, the questionable believe that higher standard

deviations when the original formulae of Kvaal et al. are applied in populations of different

ethnic backgrounds [24], are related to the population or origin of the individual rather

than to the intrinsic characteristics of the sample. Also, it assumes that formation of

secondary dentine is a linear process, when it is more similar to a curve [25]. Additionally,

147

although using tomography resulted in obtaining more odontometric data from the same

teeth (mesio-distal and bucco-lingual measurements), this apparent advantage over the

use of dental radiographs, did not improve the final outcome. Also, the need to properly

align the teeth in the sagittal and coronal plane demands more time than the traditional

Kvaal et al. approach in dental radiographs.

Finally, it is necessary to mention that CBCT is not exempt from artefacts intrinsic to the

process of image acquisition [26]. This could explain, not only the high SEE obtained in this

study, but also the poor accuracy of those methods based on pulp/tooth volume

calculation, [9,10] when compared with more simple approaches such as the measurement

of pulp/tooth area on dental radiographs (Cameriere et al method, with a SEE=±3.27 years)

[16].

CONCLUSION

Radiological assessment of dental age-related changes is a valid tool for age estimation. It

is the case that new diagnostic technologies provide practitioners and forensic researchers

new opportunities to use and propose new methods for dental age estimation. However,

the use of tomographs in this study, applying the Kvaal et al method, did not improve the

results. It is recommended to also include teeth with small fillings in studies based on

pulp/tooth dimension ratios, as long as the pulp chamber is not compromised. The aim of

this is to further investigate the application of these different methods on those individuals

who do not have totally sound dentitions.

148

REFERENCES



2005;153(2–3):208-12.


3. Soomer H, Ranta H, Lincoln MJ, Penttila A, Leibur E. Reliability and validity of

eight dental age estimation methods for adults. J Forensic Sci. 2003;48(1):149-

52.

4. Limdiwala PG, Shah JS. Age estimation by using dental radiographs. J Forensic

Dent Sci. 2013;5(2):118-22.


dental radiographs. Forensic Sci Int.1995;74(3):175-85.



2012;57(3):277-84.



tissues of the teeth for forensic purposes: A pilot study. J Forensic Leg Med.

2015;36:150-7.

149



Sci Int. 2015;253:133.e1-.e7.


estimation from canine volumes. Radiol Med. 2015:1-6.



Endod. 2014;40(11):1758-63.




12. Lewis SJ. Quantifying measurement error. Anderson IS, editor. Current and

recent research in osteoarchaeology 2: Proceedings of the 4th, 5th and 6th

meetings of the Osteoarchaeological Research Group. Oxbow Books. 1999. p.

54-5.

13. Marroquin T, Karkhanis S, Kvaal S, Vasudavan S, Castelblanco E, Kruger E, et al.

Determining the effectiveness of adult measures of standardised age estimation

on juveniles in a Western Australian population. Aust J Forensic Sci. 2016:1-9.

14. Solheim T, Sundnes PK. Dental age estimation of Norwegian adults - a

comparison of different methods. Forensic Sci Int. 1980;16(1):7-17.

150

15. Oi T, Saka H, Ide Y. Three-dimensional observation of pulp cavities in the

maxillary first premolar tooth using micro-CT. Int Endodontic J. 2004;37(1):46-

51.


Estimation by Pulp/Tooth Ratio in Canines by Mesial and Vestibular Peri-Apical

X-Rays. J Forensic Sci. 2007;52(5):1151-5.


Dentomaxillofac Radiol. 2011;40(4):199-212.



methods in adults. Int J Leg Med. 2005;119(1):27-30.

19. Murray PE, Stanley HR, Matthews JB, Sloan AJ, Smith AJ. Age-related

odontometric changes of human teeth. Oral Surg Oral Med Oral Pathol Oral

Radiol Endod. 2002;93(4):474-82.

20. Hargreaves KM, Berman LH. Cohen's Pathways of the Pulp expert consult.

Elsevier Health Sciences 2015.

21. Bergenholtz G Hörsted-Bindslev P, Reit C, editors. Textbook of Endodontology.

2nd ed. Hoboken. Wiley; 2009.

22. The principles of endodontics. 2nd ed. ed. Patel S, Barnes JJ, Ovid Technologies

I, editors. Oxford University Press; 2013.

151

23. Darmawan MF, Yusuf SM, Abdul Kadir MR, Haron H. Age estimation based on

bone length using 12 regression models of left hand X-ray images for Asian

children below 19 years old. Leg Med. 2015;17(2):71-8.

24. Parikh N, Dave G. Application of kvaal's dental age estimation technique on

Orthopantomographs on A Population of Gujarat–A Short Study. BUJOD. 2013;

3(3);18-24.

25. Woods MA, Robinson QC, Harris EF. Age-progressive changes in pulp widths and

root lengths during adulthood: a study of American blacks and whites.

Gerodontology. 1990;9(2):41-50.

26. Schulze R, Heil U, Groβ D, Bruellmann DD, Dranischnikow E, Schwanecke U, et

al. Artefacts in CBCT: a review. Dentomaxillofac Radiol. 2011;40(5):265-73.

152

SECTION FOUR: New technologies in dental age estimation.

PREFACE

The main contribution of this section to the thesis is the study of those methods for age

estimation based on volume reconstruction from dental CBCT. These methods do not

surpass the results obtained from the analysis of dental radiographs, which means that

although there are more advanced imaging techniques, they deserve to be deeply explored

and studied for their application in forensic dentistry for dental age estimation in adults.

Chapter 6 presents a study on three previously proposed methods for volume

reconstruction in dental age estimation and the possible technological and methodological

causes of variation on their results.

Chapter 7 presents an ingenious proposal for age estimation in adults using CBCT from

individuals native to different countries. Such individuals had very different ancestry, and

ethnic backgrounds, thereby challenging one of the established requirements for the

creation of methods for dental age estimation that demand the use and elaboration of

specific population studies and formulae.

153

6. Chapter 6. Reliability and repeatability of pulp volumetric reconstruction through

three different volume calculations



Tennant, M. Reliability and repeatability of pulp volumetric reconstruction through three

different volume calculations. Journal of Forensic Odontostomatology. Accepted for

publication. March 2017.

154

ABSTRACT

Dental age estimation by means of the analysis of dental regressive changes is still a valid

tool in the forensic field. The latest methods are based on the analysis of CBCT and

volumetric reduction of dental pulp with advancing age in adults. These methods do not

show superior accuracy to those based on the analysis of conventional dental radiographs.

Objective: To test the variability of the volume measurements when different

segmentation methods are applied in pulp volume reconstruction.

Materials and methods: Osirix® and ITK-SNAP software were used. Different segmentation

methods (Part A) and volume approaches (Part B) were tested in a sample of 21 dental

CBCTs from upper canines. Different combinations of the data set were also tested on one

lower molar and one upper canine (Part C) to determine the variability of the results when

automatic segmentation is performed.

Results: Although the results show correlation among them (r>0.75), there is no evidence

that these methods are sensitive enough to detect small volumetric changes in structures

such as the dental pulp canal (Part A and Part B). Automatic segmentation is highly

susceptible to small variations in the setting parameters (Part C).

Conclusion: Although the volumetric reconstruction and pulp/tooth volume ratio has not

shown better results than methods based on dental radiographs, it is advisable to

persevere with the research in this area with new developments in imaging techniques.

Key words: Age estimation, pulp volume calculation, image segmentation.

155

INTRODUCTION

Since 1950 different methods have been proposed to estimate dental age in adults [1].

Most of them are based on the formation of secondary dentine and the decrease of the

pulp chamber size with age [2, 3]. Methods that have an invasive approach, (for example

aspartic acid racemization and cementum annulation), are not of preference, as they often

require the physical damage of the sample [4]. With the evolution of different diagnostic

imaging techniques, non-invasive methods for dental age estimation have become the

preferred methods [5], starting with the use of periapical dental radiographs, [2, 3] then

panoramic radiographs [6], and more recently cone beam computed tomography (CBCT)

[7].

Micro-Computer Tomography (µCT) was introduced in the early 20th century [8] to

determine age related three-dimensional changes in pulp cavities, from maxillary first

premolar teeth. Consistent with early histological research, µCT findings indicate that a

decrease in pulp volume is not linear. It demonstrates a quicker reduction between 20 and

40 years, and then it slows down thereafter. Further research correlating the ratio

between pulp and tooth volume with the chronological age, using µCT and linear regression

models to estimate dental age in adults, revealed promising outcomes [8].

However, in light of the high radiation doses associated with µCT, only extracted teeth can

be measured in this way [9]. More recently, CBCT has been used for dental age estimation

calculating the volumes from tooth and pulp chamber. Since then, volumetric

156

reconstruction from CBCT with different software has been reported in various studies for

dental age estimation in adults [10-13]. The most recent studies document the use of CBCT

from single- [5, 13] and multi-rooted teeth. [14]

These new methods have not shown superior accuracy to the methods based on dental

radiographs [2, 15]. It is evident that within each type of software there are sensitivity

settings that influence the mathematical approach to measure the baseline volumetric

data. These settings clearly have consequences for the outcome and may well be in part

the cause of the substantial variation. The primary aim of this study was to test the

variability of the volume measurements when different segmentation methods were

applied to pulp volume reconstruction. This study is designed to ensure that the

underpinning assumptions of a commonly used age estimation approach, based on pulp

volume calculation, are as accurate as possible.


Ethics approval for this study was obtained from the Human Research Ethics Committee of

the University of Western Australia (Ref: RA/4/1/6797). This study has been performed in

accordance with the ethical standards laid down in the 1964 Declaration of Helsinki.

Study sample

A total of 22 anonymized dental CBCTs were used in this study. All were recorded with

therapeutic purposes, with the consent of the participants, and there was no unnecessary

157

or repeated radiation exposure. All images were previously unidentified, and age and sex

of each individual was the only known information. From these, 57% (n=12) were female,

age range 17-63 years, mean age 33.7, and 43% (n=9) were male, age range=15-52 years,

mean age 36.7.

The CBCTs were obtained from the Radiology Department of the University of Kebangsaan,

Malaysia, and the research group INVIENDO from the National University of Colombia. In

both cases, the images were obtained using the 9000 3D Extra-oral Imaging System

(Carestream Dental, Atlanta, GA, USA) which had a 180˚ rotation and a field of view (FOV)

of 50 mm by 37 mm. The radiation exposure parameters were 8-10 mA, 70 Kv, with a slice

thickness of 0.076mm. Images were saved in DICOM format.

The selected tooth to apply the different segmentation approaches was the upper right

canine, or the left when the right was absent (with Federation Dentaire International (FDI)

notation 13 and 23) reconstructing one tooth per subject. Only sound teeth, free of any

restorations with completed apex formation were included. The measurement software

packages used were: ITK-SNAP version 3.4 (open source software, www.itksnap.org) and

Osirix® (OnDemand 3D software CyberMed Inc, Seoul, South Korea).

All segmentations were completed by a single observer (TYM). All data were collated using

Excel (version 2013 Microsoft Redmont, USA) and statistical analysis was completed using R

Core Team version 3.1.3 (2015). (R: A language and environment for statistical computing.

R Foundation for Statistical Computing, Vienna, Austria. URL http://www.R-project.org/.)


158

Description of the method

This study tested the comparative outcome of three different ways of estimating the pulp

volume, dental age indicator used in adults [5, 13, 14]. The tests were carried out on the

root pulp chamber of upper canines (n=21), to calculate the correlation between automatic

[14], and manual segmentation [5] (Part A) (Table 6.1) and cone shape approach volume

calculation [13] (Part B) (Table 6.2). The final part of the study (Part C) (Table 6.3) tested

the variability of the results obtained with different setting parameter values combinations

for automatic segmentation of the pulp canal of one multi-rooted tooth, as previously

reported in the literature [14] and one single rooted tooth. (Graph 6.1)

Part A. Automatic and manual segmentation.

Automatic segmentation was performed using the software ITK-SNAP. (Graph 6.1. A).

Manual segmentation used software Osirix® and its 2D viewer, (Graph 6.1. B), which allows

the observer to go slice by slice and manually select the boundaries of the space to

reconstruct, using the tool polygon. The automatic segmentation process in ITK-SNAP is

called “seed region-growing”. To apply it, the observer sets up a “seed” inside the

structure to reconstruct, in this case, the pulp chamber. This seed grows, and a volume is

finally obtained.

Automatic segmentation of the pulp was generated using the same combination of setting

parameters through the entire sample: Scale of Gaussian blurring: 3.0, edge transformation

contrast: 0.2, edge transformation exponent: 4.0, expansion (balloon) force: 1.0, smoothing

(curvature) force: 0.2, edge attraction (advection) force 2.0. In three cases, the edge

159

transformation exponent needed to be adjusted to 0.05. Regression models were run to

establish which of the setting parameters would have the most significant influence on the

final results. Little changes in the Scale of Gaussian generated the bigger changes in the

final values (p< 0.01).

With manual segmentation two approaches were used. Firstly, every fourth slice was

analysed as for the published method. Secondly, the first and last slices were used as

samples. To analyse the same length of root in the automatic and manual approaches, a

length defined in the automatic approach was set for the manual approach.

Part B. Cone shape geometric approximation

The geometric approximation of the root canal to cone is a simple conservative method,

and has been used for dental age estimation in adults. Using Osirix®, ovals of best fit were

formed in the root canal at the cemento-enamel junction (CEJ) level, following the reported

method [13]. Secondly, instead of best fit ovals, the boundaries of the pulp canal were

drawn using the tool polygon (Graph 6.1, C). The root length was measured from the CEJ

to the apex, (Graph 6.1, D), and finally two different volumes of cones were calculated per

tooth with a mathematical formula.

Part C. Automatic segmentation and different setting parameters combinations

As observed in part A, it is not possible to apply the same combination of setting

parameters values to all teeth with automatic segmentation. Considering the advantage of

this automatic process, [16] twenty combinations of different values for all the setting

160

parameters (Table 6.3) were used to reconstruct the pulp chamber of one lower first molar

(FDI 36, Male 23-year-old Malaysian origin, slice thickness 0.2mm) and an upper canine

(FDI 23, Male 31-year-old Colombian origin). The volumetric reconstruction with each

combination was completed twice per tooth. The variation among the volume from the

same tooth was calculated.


Shapiro Wilk normality test, mean, standard deviation (SD), paired t-Test and Pearson

correlation coefficient, were calculated to compare the results obtained using the different

segmentation methods (Part A), cone shape volume approach (Part B), and the obtained

results when different parameters are used to do the automatic segmentation of the same

tooth (Part C).

RESULTS

When automatic and manual segmentation using each fourth slice were compared, the

Pearson’s correlation coefficient (r2=0.83), shows a greater correlation between them, than

between automatic and manual segmentation using only the first and the last slices

(r2=0.75) or between the two manual segmentation methods (r2=0.79).

However, this does not mean that manual segmentation using every fourth slice and

automatic segmentation produces comparable results. The paired t test shows that there

is a statistically significant difference between the means of the results of both manual

161

segmentations and automatic segmentation (p<0.01).

When comparing the volumes, the manual segmentation produced volumes are larger than

those obtained with automatic segmentation. To facilitate the understanding of this

difference, a simple mathematical analysis was used, taking the obtained volume with

automatic segmentation as 100% of the volume. Then the difference between the volumes

obtained with automatic segmentation and both manual segmentation, were used in a rule

of three to calculate the percentage this difference represents.

Part A. Automatic and manual segmentation

Three data-sets were obtained from the volumetric reconstruction of the root canal of the

canines using the different methods (Table 6.1). The results of this calculation showed that

the volumes obtained using every fourth slice are on average 25% (1% to 53%) larger than

those obtained with automatic segmentation.

Also, using the first and the last slices are on average 57% (12% to 210%) larger than those

obtained with automatic segmentation. The average of the volumes from both manual

segmentation methods is in general 41% larger than automatic segmentation. The volumes

obtained doing manual segmentation of each fourth slice were on average 28% larger than

those obtained using only the first and the last slice. However, some of them were on

average 16% smaller (p>0.05). The mathematical analysis was like the previously

described, but in this case, the volume obtained with manual segmentation using only the

first and last slice was taken as 100% in the rule of three.

162

Graph 6.1. Volumetric reconstructions, upper right canine using different methods.

Volumetric reconstruction of tooth 13 from a female individual, 36 years old, using the different methodologies mentioned in this paper. A. Semiautomatic recontruction using ITK-SNAP software. B. Manual segmentation using Osirix ® software. C and D. Cone-shape geometric approximation, following the methodology reported by the author doing area measurement of the root canal at the level of the CEJ (D) and length measurement of the root from the CEJ up to the root apex (C), using Osirix ® software.

163

Table 6.1. Volume results of using automatic segmentation (Vol 1), manual

segmentation using just the first and last slice (Vol 2), and manual segmentation each

forth slice (Vol3), of cone beam computer tomography from upper canines. (FDI 13, 23)

Individual Vol 1 (mm3) Vol 2 (mm3) Vol 3 (mm3)

1 10.87 10.6 13.2

2 16.03 17 19.7

3 16.4 20.9 19.8

4 9.309 12.9 14.8

5 5.478 7.1385 8.04

6 11.25 12.5 12.6

7 13.26 25.9 16.3

8 16.72 17 23.2

9 37.01 32.6 41.9

10 16.98 23.4 28.9

11 11.77 19.2 17.7

12 5.857 8.7131 11.7

13 9.107 13.3 28.3

14 17.24 33 29.8

15 12.76 27.5 21.1

16 15.87 16.1 27

17 13.2 19.7 22.6

18 11.61 19.7 16.5

19 24.52 35.3 40

20 15.51 22.1 22.1

21 17.6 29.3 35.5

Mean 14.68 20.18 22.41

164

Table 6.2. Cone shape approach results using the tool polygon (Pol) and oval (Oval).

Pulp area (in mm2) at the CEJ level and volume (in mm3) calculation with the measured

root length (in mm).

area Area Pol Area Oval Root length Vol Pol Vol Oval

1 22 20 1.55 11.2 10.2

2 29 30.5 1.46 14.1 14.8

3 28 24.5 1.46 13.6 11.9

4 20 9.85 1.51 10.0 4.9

5 21 16 1.04 7.3 5.5

6 16 13.5 1.56 8.3 7.0

7 46 29 1.86 28.6 18.0

8 37 20.5 1.64 20.3 11.2

9 52 41 1.38 24.0 18.9

10 32 26.5 1.63 17.4 14.4

11 30 17 1.28 12.8 7.2

12 34 16 1.38 15.8 7.4

13 32 22 2.10 22.4 15.4

14 28 14 1.71 16.1 8.0

15 42 28.5 1.79 24.9 16.9

16 27 16 1.71 15.4 9.1

17 25 17.5 1.96 16.2 11.3

18 35 25.5 1.47 17.2 12.5

19 58 44.5 1.73 33.9 26.0

20 39 26.5 1.52 19.8 13.4

21 46 24.5 2.17 33.3 17.7

SD 8.7 10.86 7.45 5.2

165

Table 6.3. Setting parameter combinations used in part C.

Repetition

Scale of Gaussian

Edge contrast

Edge transformation exponent.

Smoothing force

Average molar volume mm3

Average canine volume mm3

1 2.14 0.09 4 0.2 38.415 36.58

2 3 0.05 3 0.2 45.775 49.7

3 2.05 0.05 3 0.2 34.525 49.7

4 2.5 0.05 3 0.2 46.075 41.81

5 2.5 0.1 2.5 0.2 49.38 48.87

6 2.5 0.2 2.05 0.2 53.74 48.94

7 2 0.1 2.5 0.2 43.55 39.27

8 2.15 0.09 4 0.2 36.53 40.26

9 2.15 0.09 4 0.3 38.94 39.76

10 2.15 0.09 4 0.35 42 39.38

11 3 0.05 3 0.3 47.29 50.89

12 3 0.05 3 0.35 43.55 48.92

13 3 0.05 3 0.4 44.69 48.9

14 2.05 0.05 3 0.3 38.69 35.98

15 2.05 0.05 3 0.35 38.79 35.89

16 2.05 0.05 3 0.4 38.46 36.04

17 1.5 0.05 2.51 0.2 34.3 31.92

18 1.5 0.05 2.5 0.2 34.95 30.08

19 1 0.1 3 0.2 31.73 29.83

20 2 0.05 4 0.2 36.40 35.10

SD 0.53 0.03 0.59 0.07 5.74 7.12 Variance 0.28 0.01 0.35 0.005 32.94 50.75

Note: There are 3 parameters available in ITK-SNAP software that are not listed, as they were constant for all the reconstructions: Edge attraction force (=2) and step size (=1) and smoothing force which was changed from 1 to 0.5 only for the 18th combination.

166

Part B. Cone shape geometric approximation

It was expected to obtain similar area and volume values when using ovals of best fit and

the tool polygon following the outline of the pulp chamber. Unexpectedly, with the tool

polygon the areas and volumes were noticeably larger (50%) than with the ovals (table 6.2).

However, there is a strong correlation coefficient between both measured areas using the

tools oval and polygon (r=0.84) at the CEJ and the volumes (r=0.88).

Repeatability of the automatic segmentation for the first and the second repetition for

both teeth was evaluated with the paired t test (r2=0.7, p=0.99), showing that there is no

statistically difference between both repetitions with a 95% level of confidence. The

average between the obtained volumes in both repetitions was calculated.

The volumes for the molar pulp chamber (FDI 36), ranged from 31.7mm3 to 53.74mm3,

with a SD of 5.7and a variance of 33. For the volumetric reconstruction of the canine (FDI

23), the volume value ranged between 29.8 mm3 to 50.8 mm3, SD=7.12 and variance of

50.75. The SD for the different parameters was not larger than 0.6 and the sample

variance was not greater than 0.3.

Part C. Automatic segmentation and different setting parameters combinations

When the areas using the two different tools are compared the data show a larger

variability and a high SD; this effect is diminished once the volume is calculated. In this

way, the length of the root is what really determines the final volume.

167

DISCUSSION

With the introduction of computer vision techniques for medical imaging and their

application in dentistry and forensic sciences, the possibilities to propose methods for age

and sex estimation are almost unlimited. Owing to the apposition of secondary dentine,

the anthropometric assessment of the variation in the ratio between size of the pulp

chamber and size of the tooth is still an acceptable dental age predictor [14].

The aim of this study was to test the variability of the volume measurements when

different segmentation methods are applied in pulp volume reconstruction. This is because

although it would be expected that with volume measurements the results for age

estimation were more reliable, the literature shows the opposite. To find out an

explanation to the lower accuracy of the methods based on volume reconstruction, in this

study, we tested three different methods doing only pulp volume reconstruction.

The methods employed in this thesis, using CBCT, do not seem to be more accurate than

those based on radiographs for dental age estimation in adults, based on the analysis of the

available literature. This is regardless of the sample size of the study, which ranges from 19

individuals, analysing only 12 canines [7] to 403 individuals analysing only the pulp chamber

of first molars [14]. This is clear when the reported accuracy of the different methods is

compared. The mean absolute error using automatic segmentation of the first molar’s pulp

chamber is 6.26 years at best [14]. The cone shape geometric approach of pulp and tooth

168

has a standard error of estimation of ± 11.45 years, [13] and finally, manual segmentation

of pulp and tooth, has a prediction interval of ± 12 years [5].

Other methods also using automatic segmentation of canines reported a prediction interval

of 3.47 years at best [17]. When these results are compared with some of the most

commonly used adult age estimation methods using dental radiographs, such as Cameriere

et al. [3] (with reported mean predictor errors of 2.37 years [18] to 11.01 years [19]

(tooth/pulp area ratio)), or the method developed by Kvaal et al. [2] (pulp/tooth-linear

measurements ratios with reported standard estimation errors of ±9.367 years [20]), it is

possible to observe that the use of CBCT and dental structures volume reconstruction, do

not improve the final results for adult age estimation. The relation between the lack of the

accuracy of the above mentioned methods for age estimation [5, 13, 14], and the different

methods for volume reconstruction are presented as follows:

Part A. Manual and automatic segmentation may produce similar results, as observed in

this study (r2=0.83). Manual segmentation is time consuming, difficult to perform and

prone to be influenced by the subjectivity of the observer, and personnel trained in this

aspect is necessary [21]. Additionally, it is not accurate enough to estimate the variation of

the volume in small anatomic structures, such as the root canal, as observed in this study.

Although there are different methods for automatic image segmentation, the first problem

faced in the bio-medical area, is the exclusive dependence on gradient or intensity analysis

without using anatomical information [21]. In the case of automatic segmentation, with

169

ITK-SNAP software, although the software chooses which pixels will be included in the

segmentation, the observer finally decides how to adjust the parameters to control the

algorithms in the segmentation. This makes this method prone to a certain degree of

subjectivity from the observers. These parameters may also change from one CBCT to

another.

Unfortunately, in the previous study using this software, the values of the setting

parameters were not mentioned [14]. In this study, it was also observed that although all

the CBCTs were obtained from machines with the same reference, and under similar

exposure conditions (slice thickness, Kv, mA, and time of exposure) and the same

parameters for automatic segmentation, the interaction of the seed region-growing with

the boundaries of the structure to reconstruct may also differ from one CBCT to another.

For this reason, it is not possible to follow the recommendations of the ITK-SNAP

manufacturer, who recommends keeping the same parameters among different CBCT, and

it is neither possible to warrantee the uniformity of the segmentations obtained [14].

In this study, it was also observed that small changes in the values of the setting

parameters can generate significant variation in the final volumes with automatic

segmentation, even though the segmentation was made on the same CBCT, same tooth,

and same anatomic region. The significance of this is that the final volume will be always

the same under the same parameters, in the same CBCT, and anatomic region, but these

parameters generally need to be modified from one CBCT to another, and in the same way,

small changes in the parameters alter the final volume.

170

Part B. With the aim of simplifying the measurement of the tooth structure volume, a cone

shape geometric approach was proposed (Part B) [13]. The developers of this method

were aware of the main disadvantages of this proposal: measurements inaccuracy and

subjectivity. They also reported that this method tends to underestimate the real volume

of the tooth structures (56% to 67%), which in theory would not affect the final results

after pulp/tooth ratio that was calculated to estimate the age.

To test this method, the current study used two differ approaches for doing the volume

calculation. The first, following the published methodology [13] displaying ovals at the CEJ,

and a second, selecting manually the outline of the root canal at the CEJ, using the tool

polygon. As observed in the results section, despite the difference between both area

measurements, at the CEJ, once the volume is calculated the difference decreases.

None of these approaches is recommended to do pulp volume measurements, firstly,

because of the subjectivity of the observer, who only counts with the naked eye to display

one or another, and secondly because both bring a distorted representation of the root

canal, ignoring the different variations in its internal shape.

Part C. In this study, there are different parameters that affect the final volumetric

reconstruction of any anatomical region. The scale of Gaussian, or blurrier factor [22] was

the most relevant when doing automatic segmentation. To increase the value of the

Gaussian scale, allows the observer to eliminate certain noises in the image, which affect

the “growing of the seed”. However, the bigger the value the Gaussian scale is, the more

171

principal structures, and details will be lost [23]. This would mean that the initial volume of

the pulp chamber will be magnified [16].

Although automatic segmentation is less prone to the subjectivity of the observer than

manual segmentation, in this study it was observed that it is not possible to keep the same

volume reconstruction parameters for all the CBCT, and that minimum changes may

produce significant variations, even though they are done on the same tooth. The volume

obtained with this tool must be understood as an approximation to the real volume of the

structure to measure.

Taking into consideration the reported rate of secondary dentine formation is 6.5 μm/year

for the crown, and 10 μm/year for the root [24], the methods in this study are not sensitive

enough to detect such small dimension changes, generating a big variation of the

measurements, which could explain the greater error of pulp/tooth volume based methods

for age estimation when compared with simpler methods based on linear [2] or area

measurements [3] on dental radiographs.

Previous research in the biomedical area has highlighted the problems related to

semiautomatic segmentation, and specifically, to the process that the software uses to

select which pixels or voxels to be included in the final volume reconstruction. To

overcome this problem, the probabilistic anatomic atlas has been elaborated, which

facilitates the segmentation of an organ of interest [21]. When doing the semiautomatic

segmentation of teeth, and pulp chambers, different obstacles were observed.

172

Firstly, owing to the similar density of dentine and bone, and the irregularity of the

periodontal ligament, it was not possible to obtain a volumetric reconstruction of the tooth

as it is for the pulp chamber. This is the reason why the automatic tooth volume

reconstruction was not analysed. Secondly, owing to the large variety in the pulp chamber

shapes in the apical third of the tooth, there was always a leak of the seed region -growing

in to the dentine.

To create a probabilistic atlas for dentistry could help to overcome these issues, and help

the dentist, and forensic dentist to implement more accurate, and easier methods for

automatic segmentation, and volume reconstruction. Thirdly, although the automatic

reconstruction of the pulp chamber of molars may be understood as a simple process, it

may be more reliable to use other methods for age estimation when using molars.

CONCLUSIONS

Based on the results of this study, it is possible to affirm that the volume measurements

with any segmentation method must be understood as an approximation of the measured

structure, and not as the real volume. In the same way, manual segmentation could

produce volume values that are even more inaccurate, and should be avoided in future

studies.

Although the volumetric reconstruction and P/T volume ratio has not shown better results

than methods based on dental radiographs, it is worth to persevere with the research in

173

this area with an interdisciplinary team to build software that can reliably reconstruct pulp,

and tooth volumes for forensic purposes.

174

REFERENCES













methods in adults. Int J Legal Med 2005;119(1):27-30.

7. Yang F, Jacobs R, Willems G. Dental age estimation through volume matching of

teeth imaged by cone-beam CT. Forensic Sci Int. 2006;159, Supplement(0):S78-

S83.

8. Saka H, Ide Y. Three-dimensional observation of pulp cavities in the maxillary

first premolar tooth using micro-CT. Int Endod J. 2004;37(1):46-51.

175

9. Agematsu H, Someda H, Hashimoto M, Matsunaga S, Abe S, Kim H-J, et al.

Three-dimensional Observation of Decrease in Pulp Cavity Volume Using Micro-

CT: Age-related Change. Bull Tokyo Dent Coll. 2010;51(1):1-6.

10. Someda H, Saka H, Matsunaga S, Ide Y, Nakahara K, Hirata S, et al. Age

estimation based on three-dimensional measurement of mandibular central

incisors in Japanese. Forensic Sci Int. 2009;185(1–3):110-4.

11. Jagannathan N, Neelakantan P, Thiruvengadam C, Ramani P, Premkumar P,

Natesan A, et al. Age estimation in an Indian population using pulp/tooth

volume ratio of mandibular canines obtained from cone beam computed

tomography. J Forensic Odontostomatol. 2011;29(1):1-6.

12. Star H, Thevissen P, Jacobs R, Fieuws S, Solheim T, Willems G. Human Dental

Age Estimation by Calculation of Pulp–Tooth Volume Ratios Yielded on Clinically

Acquired Cone Beam Computed Tomography Images of Monoradicular Teeth. J




tissues of the teeth for forensic purposes: A pilot study. J Forensic Legal Med.

2015;36:150-7.



Sci Int. 2015;253:133.e1-.e7.

176


Estimation by Pulp/Tooth Ratio in Canines by Mesial and Vestibular Peri-Apical

X-Rays. J Forensic Sci. 2007;52(5):1151-5.

16. Yushkevich PA, Piven J, Hazlett HC, Smith RG, Ho S, Gee JC, et al. User-guided 3D

active contour segmentation of anatomical structures: Significantly improved

efficiency and reliability. NeuroImage. 2006;31(3):1116-28.

17. Tardivo D, Sastre J, Catherine J-H, Leonetti G, Adalian P, Foti B. Age

determination of adult individuals by three-dimensional modelling of canines.

Int J Legal Med. 2014;128(1):161-9.

18. Cameriere R, Cunha E, Sassaroli E, Nuzzolese E, Ferrante L. Age estimation by

pulp/tooth area ratio in canines: Study of a Portuguese sample to test

Cameriere's method. Forensic Sci Int. 2009;193(1–3):128.e1-.e6.

19. Babshet M, Acharya AB, Naikmasur VG. Age estimation in Indians from

pulp/tooth area ratio of mandibular canines. Forensic Sci Int. 2010;197(1–

3):125.e1-.e4.




21. Dong C, Chen YW, Foruzan AH, Lin L, Han XH, Tateyama T, et al. Segmentation of

liver and spleen based on computational anatomy models. Comput Biol Med.

2015;67:146-60.

177

22. Tschirsich M, Kuijper A. Notes on Discrete Gaussian Scale Space. J Math Imaging

Vis. 2015;51(1):106-23.

23. Guerrero González N. Segmentación de imágenes empleando el espacio de

escala Gaussiano. Diss. Universidad Nacional de Colombia-Sede Manizales.

2006.

24. Murray PE, Stanley HR, Matthews JB, Sloan AJ, Smith AJ. Age-related

odontometric changes of human teeth. Oral Surg Oral Med Oral Pathol Oral

Radiol Endod. 2002;93(4):474-82.

178

7. Chapter 7. A new proposal to overtake population group differences for dental age

estimation in adults through pulp/tooth volume calculation. Pilot Study.

In press version, under revision.



differences for dental age estimation in adults through pulp/tooth volume calculation. Pilot

study. Journal of Forensic Sciences. With editor.

179

ABSTRACT

Objectives: The formation of secondary dentine on the walls of the root canal has been

used as an age indicator in adults. The aims of this study are: firstly, to propose a method

for dental age estimation in adults, regardless of their nationality or ethnicity; and secondly

to test if by volume reconstruction of only a fraction of the tooth it is possible to obtain

linear regression models and a unique equation for dental age estimation.

Materials and Method: Eighty-one anonymized cone beam computed tomography images

obtained from two different population groups were used. Only sex and age information

were kept. One group had a Latin-American background (Colombian individuals aged 23 to

71 years), and the other had an Asian background (Malaysian individuals aged 15 to 58

years). The tooth analyzed was the upper canine. The software used was ITK-SNAP version

3.4 (open source software, www.itksnap.org), to do automatic volume reconstruction of

the cervical third of root canal, and root. The images were shared via online under the

strictest security parameters to guarantee the integrity and safety of the data.

Results: The correlation coefficient between pulp/pulp+tooth volume ratio increased when

data from individuals of both populations were included in the same statistical analysis

(r2=0.42). It is possible to propose a formula for age estimation regardless the origin of the

individual: Age=67.104+(-434.741 x p/pt) (Standard error of estimation=±11.4 years).

Conclusion: It has been established that methods for age estimation must be population

specific. In this study is has been demonstrated that meaningful data may be obtained

from individuals that are ethnically different and widely separated geographically.

Key words: Computed tomography, secondary dentine, age estimation, ethnic variation.

180

INTRODUCTION

Much has been said in biological sciences about race and ethnicity since 1775, when

Johann Friedric Blumenbach defined five different races: Caucasians (whites), Mongolians

(East Asians) Malayans (South Asians), Negroids (Black Africans), and Americans (First

Nations) [1]. Since then, different studies followed this structure to analyze, interpret data,

and present results; a phenomenon that has been named “scientific racism” [2].

However, owing to the different nomadic and migration processes that humanity has

experienced across history, this racial classification is less accurate and practical [2].

Nevertheless, in forensic sciences it is still stipulated that the ethnic background of an

individual is of paramount relevance, to formulate a biological profile and to clarify the

identity of living or deceased individuals, next to sex, stature, and age [3].

In terms of age estimation, many methods have been proposed for individuals of different

ages, based on bone and tooth formation or deterioration with age [4]. Methods based on

the radiological analysis of dental development, have high accuracy, due to the genetically

controlled process of tooth formation, which seem to be similar for individuals from the

same group of population [5].

However, the most recent findings suggest that their results may be incurring on a

selection bias called “age mimicry”, [6, 7], which has caused the erroneous need to

generate specific population formulae for age estimation.

181

Simultaneously, age estimation in adults, by means of non-invasive analysis, is still a

challenge for scientists of different areas, especially, those who are involved in forensic

investigations. Although different methods have been proposed, there is no consensus

about a unique method to be applied to individuals from different countries, and it has

been claimed that methods for age estimation must be population specific [8] whouthout

testing the use of mixed samples. Additionally, the few changes that bones and teeth have

after-body maturation, are a setback for the accuracy of the existing methods.

One of the first methods for dental age estimation in adults established that the formation

of secondary dentine on the wall of the root canal could be considered an age predictor in

adults [9]. Based on the negative correlation of pulp/tooth ratio size with age, different

non-invasive methods were proposed: linear measurements [10]; area measurements; in

dental radiographs [11]; and lastly volume measurements from dental tomography [12].

The aims of this study are: firstly, to propose a method for dental age estimation in adults,

regardless of their nationality or ethnic background; and secondly to test if by

reconstructing only a fraction of the tooth it is possible to obtain linear regression models

and an equation for age estimation.


The sample for this cohort study encompasses 81 cone-beam computed tomography

(CBCT) obtained from two different population groups. Latin-American (Colombian

182

individuals aged 23 to 71 years, 60% female (n=23) and 40% male (n=15)) and Asian

(Malaysian sample individuals aged 15 to 58 years 46% female (n=20) and 54% male

(n=23)) previously collected for different dental diagnosis and treatment procedures, with

informed consent. All the images were anonymized keeping information about sex and

age.

Table 7.1 shows the age and country of origin distribution. Although the CBCT were

obtained in different institutions, in both cases the reference of the device was the same:

Kodak device, reference K9000-3D (exposure parameters: scanning time: 10.8 s, 3-15 mA,

60-85kVp, field of view FOV: 5.2 cm x 5.2cm to 5.5cm to 5.5cm, 180° rotation, slice

thickness 0.076-0.2 mm).

The tooth analyzed was the upper canine, left or right, one per individual, as previous

studies have shown no right/left difference in tooth development [13]. The reconstructed

root fraction was in all cases 5 mm from the highest point of the vestibular cemento-

enamel junction, towards the root apex, with the aim to include only the cervical third of

the root [12].

Two different proceedings were tested to obtain the pulp/tooth ratio. Firstly, by dividing

the volume of the pulp by the volume of the tooth, and secondly, dividing the volume of

the pulp by the sum of the volume of the tooth and pulp.

183

Inclusion criteria

The included roots portions were from sound teeth, teeth with no evidence of physical

trauma, free of active caries, free of evidence of periapical disease, without signs of pulp

pathology, and free of cervical lesion of any nature. The analysis also included teeth with

small coronal fillings, with at least 2 mm of dentine between the tooth restoration and the

pulp chamber, which is assumed to have negative effect on the root portion.

The software used was ITK-SNAP version 3.4 (open source, www.itksnap.org), consistent

with previous studies, doing automatic reconstruction of the root canal and root. In certain

cases, it was necessary to manually correct the semiautomatic root segmentation, as the

software included portions of bone, that were required to be manually deleted. The

measurements were collected by only one observer, who had been previously trained in

the use of this software.

All the data were collected in an Excel spreadsheet, using Excel (version 2013 Microsoft

Redmont, USA), before and after manual correction of the image segmentation. Data

analysis was performed with R Core Team version 3.1.3 (2015). (R: A language and

environment for statistical computing. R Foundation for Statistical Computing, Vienna,

Austria. URL http://www.R-project.org/).


184

RESULTS

Pearson correlation coefficient tests were calculated to observe if there was any difference

between the volume values obtained, for pulps and teeth doing automatic segmentation,

and after manual correction slide by slide. There was no significant difference between

both measurements (r2=0.98). The two established proceedings to obtain the pulp tooth

ratio were compared. The second approach, dividing the volume of the pulp by the sum of

the volume of the tooth and pulp, generated a higher correlation coefficient (r2=0.34 vs

r2=0.42).

In the same order, the determination coefficient between the pulp/pulp+tooth ratio (p/pt)

was calculated for the individuals of each group separately (r2=0.26 for the Colombian

sample and r2=0.38 for the Malaysian sample) and for the total sample using the data of

both groups in the same analysis, obtaining a higher correlation coefficient when all the

data are included in the analysis (r2=0.42). (Graph 7.1)

In regard to tooth condition, teeth with small crown fillings, that did not involve the root

portion, as were the used teeth, did not generate any outlier that could affect the linear

regression model (Graph 7.1). Linear regression models were generated to produce a

formula for dental age estimation. The final formula is:

Age=67.104+(-434.741 x p/pt)

185

where p/pt is the ratio between the volume of the pulp divided by the sum of the volume

of pulp and tooth. The standard error of estimation (SEE) is ±11.4 years (r2=0.42).

It is necessary to mention, that owed to the age distribution of the sample used for this

study, this formula could not estimate the age if individuals over 67 years of age.

DISCUSSION

It is undeniable that dentistry has made a valuable contribution to anthropological

sciences, bringing insights about human evolution. In the same way, dental analysis has

become an important tool for human identification in forensic sciences. In both fields,

anthropology and forensics, the concepts about racial and ethnic differences have deep

roots and a significant influence in research [14].

However, in a changing world, where day by day the biological boundaries between

individuals of different backgrounds are constantly vanishing, the proposal of methods for

age estimation that are racially or ethnically focused must be re-evaluated.

The last systematic review about Demirjian’s developmental stages of third molars [6] and

the Greulich and Pyle’s bone atlas [7] for age estimation in children and adolescents,

exposes a selection bias called age mimicry, present in numerous publications. This bias

explains the variation, in the accuracy of different studies, which has been wrongly

associated to the population origin and ethnicity of the participants, rather than to the

intrinsic characteristic of the used sample.

186

Table 7.1. Distribution of the individuals for age in years and country; Malaysia (Mal)

and Colombia (Col) used for the image segmentation and volume calculation in this

study.

Age range Mal Col Total %

15-20 12 0 12 15

21-30 13 2 15 18

31-40 6 11 17 20

41-50 9 12 21 26

51-60 3 7 10 12.3

61-71 0 6 6 0.7

Total 43 38 81 100%

Graph 7.1. Scatter plot showing the correlation between age and the pulp/pulp+tooth

volume ratio. (P/PT ratio) R2=0.42

187

To avoid age mimicry, studies must have a study population with an even age distribution,

large even samples, same number of individuals in each age group, appropriate upper and

lower age limits facilitating the observation of the different age-related changes.

Furthermore, and ideally samples that include data of individuals from different regions, to

conclude whether regional differences are important for age estimation in children [6] and

probably in adults. Subsequently, it is necessary to perform multi population studies in

adults.

The research in this thesis makes a contribution to disclose the population related

differences in age estimation in adults when secondary dentine production is analyzed as

an age predictor.

In terms of age mimicry in this study, it is possible to note that the age distribution is not

even when both samples are analysed separately: 58% (n=25) of the Malaysian sample

belongs to the age range of 15 to 30 years, while for the Colombian sample, 60% (n=23) of

the sample belongs to the age range 31 to 50 years. Once both samples are summed, the

age distribution is more even, as well as the general increase of the total data used to

perform the linear regression analysis, which could explain why the determination

coefficient is higher.

In this study, the production of secondary dentine and the narrowing of the root canal with

age, which is one of the used parameters for dental age estimation in adults, has a similar

behavior in individuals that are not only ethnically different but also geographically

188

separated. Also, that the separate generation of regression models, for each one of the

population groups, would have generated age mimicry, owed to the unequal age

distribution of both samples, as in the Malaysian group the younger group dominates,

while for the Colombian group the middle age dominates (Table 7.1).

By adding both samples it is possible to obtain a more even age distribution and also a

higher correlation coefficient between age and p/pt ratio (R2=0.42). However, for future

studies it is recommended to have a sample with the same number of individuals for the

different populations to analyse per age group, to obtain more reliable results. To follow

up on the current work, further studies will be necessary, to test the proposed formula, and

also, to perform a similar analysis in a much larger sample, that follows the established

requirements to avoid age mimicry (7).

Up to now different methods and formulae have been published to estimate dental age in

adults [15]. Nevertheless, few studies have tested the proposed formulae in groups of

populations different to those included in the original studies. In the few studies that have

done this analysis, it has been suggested that local formulae produce more accurate results

[15].

With the results of the present study it could be possible to affirm that it is necessary to

include the data of individuals of different populations in the same statistical analysis. This

would enable the observation that there is no significant difference between populations

189

or ethnical groups with regards to the production of secondary dentine, and to generate an

equation that can be used regardless the origin of the individual.

Furthermore, there are other factors that can affect the length, area, and the total volume

of the tooth, such as bruxism and attrition [17], among others, that also affect the

production of dentine in the crown and apical third of the root. These factors are not

ethnically related but individual dependent. By analyzing sound cervical root third of

canines, or any other teeth, the problems related to tooth attrition are diminished.

In this study, the maxillary canine was selected as the tooth to analyze, owing to the

reported longer survival, compared to other teeth [18]. It was also preferable to measure

the volume of only the cervical third, as it has been reported that pulp/tooth ratios, had a

higher correlation with age at the cervical root level and that the correlation with age

decreased towards the apex, for different types of teeth, when doing linear [19] or volume

[20] measurements.

In the first studies, that analyzed the production of secondary dentine with age [9], the

methods required not only the use of extracted teeth but also the destruction of the

sample, which is unethical in a modern context. The use of different radiological

techniques has allowed researchers to analyze age indicators in a non-invasive and more

conservative way. The use of CBCT to analyze and obtain the volume of any anatomical

structure is much easier, cheaper and faster, and has great potential for applications in

anthropological and forensic studies.

190

In this study, this radiological technique allowed the observer to separate only a fraction of

the tooth, calculate its volume, and guarantee not only the integrity of the individual but

also the uniformity of the samples size. In terms of accuracy, although the error exceeds

the threshold that must accomplish a method for age estimation in adults (±10) [21] the

results of this study (SEE=±11.4 years) are not so different to previous studies that used

data of individuals from the same country or belonging to the same ethnic group (±12.5

years at best) [22].

In the same way, other studies also based on volume reconstruction has been published.

Never the less, some of them are based only on pulp volume reconstruction [23, 24], or on

a geometrical opproach of rooth canal and root [25]. As their methods are not based on a

clear mesurement of the real volume of the root canal and root, and a pulp/tooth ratio

calculation, their results can not be compared with thos obtained from the present study.

Finally, confirming the tooth selection in our study, the latest study on pulp/tooth volume

ratio calculation for dental age estimation in adults, confirms that the upper canine is one

of the teeth, (next to upper central incsor), that presents a better correlation between age

and pulp/tooth ratio, which encourages future studies based on utilisation of this specific

tooth [26].

CONCLUSION

It has been claimed that in the modern world would testify the end of ethnicity, and that it

changes according to circumstances [13]. This study proposes the use of data from

individuals not only from different countries but also have been historically catalogued as

191

members of different racial and ethnical groups. By calculating the pulp/tooth+pulp

volume of only one part of the tooth there is no significant difference between these two

populations and that is possible to generate a formula to estimate the age for adult

individuals regardless their origin.

192

REFERENCES

1. Meer N. Key Concepts in Race and Ethnicity. London, UNKNOWN: SAGE

Publications; 2014.

2. Wade P. Race and Ethnicity in Latin America. London: Pluto Press; 2010.


and future directions. Leg Med. 2010;12(1):1-7.

4. Schmitt A, Cunha E, Pinheiro JES. Forensic Anthropology and Medicine. Totowa, NJ,

UNITED STATES: Humana Press; 2007.

5. Liversidge HM, Chaillet N, Mörnstad H, Nyström M, Rowlings K, Taylor J, et al.

Timing of Demirjian's tooth formation stages. Annals of human biology.

2006;33(4):454

6. Rolseth VM, Dahlberg PS, Ding KY, Bleka Ø, Skjerven‐Martinsen M, Straumann, GH

DG, Vist GE. Demirjians utviklingsstadier på visdomstenner for estimering av

kronologisk alder: en systematisk oversikt. [Demirjian’s development stages on

wisdom teeth for estimation of chronological age: a systematic review.]. Oslo:

Folkehelseinstituttet, 2017.

7. Dahlberg PS MA, Ding KY, Bleka Ø, Rolseth V, Straumann GH, Skjerven‐Martinsen M,

Delaveris GJM, Vist GE. Samsvar mellom kronologisk alder og skjelettalder basert på

Greulich og Pyle‐atlaset for aldersestimering: en systematisk oversikt. [Agreement

between chronological age and bone age based on the Greulich and Pyle‐atlas for

age estimation: a systematic review]. Oslo: Folkehelseinstituttet, 2017.

193

8. Senn DR. Senn, Weems R. Manual of forensic odontology. 5th ed. Boca Raton, FL:

Boca Raton, Press; 2013.



radiographs. Forensic Sci Int. 1995;74(3):175-85.



12. Star H, Thevissen P, Jacobs R, Fieuws S, Solheim T, Willems G. Human Dental Age

Estimation by Calculation of Pulp–Tooth Volume Ratios Yielded on Clinically

Acquired Cone Beam Computed Tomography Images of Monoradicular Teeth*. J




Sex: A Preliminary Study. J Forensic Sci. 2011;56(3):766-70.

14. Ansell A. Race and Ethnicity: The Key Concepts. Abingdon, Oxon, UNKNOWN: Taylor

and Francis; 2013.

15. Jeon HM, Jang SM, Kim KH, Heo JY, Ok SM, Jeong SH, et al. Dental Age Estimation in

Adults. J Oral Med Pain. 2014;39(4):119-126. 2014;39(4):119-26.

194

16. Patil S, Mohankumar K, Donoghue M. Estimation of age by Kvaal's technique in

sample Indian population to establish the need for local Indian-based

formulae.(Original Article)(Report). J Forensic Dent Sci. 2014;6(3):171.


Wevers M, et al. Age calculation using X- ray microfocus computed tomographical

scanning of teeth: a pilot study. J Forensic Sci. 2004;49(4):787.



19. Solheim T. Amount of secondary dentin as an indicator of age. Eur J Oral Sci.

1992;100(4):193-9.

20. Aboshi H, Takahashi T, Komuro T. Age estimation using microfocus X-ray computed

tomography of lower premolars. Forensic Sci Int. 2010;200(1–3):35-40.


methods in adults: intra- and inter-observer effects. Forensic Sci Int.

2002;126(3):221-6.

22. Landa MI, Garamendi PM, Botella MC, Aleman I. Application of the method of Kvaal

et al. to digital orthopantomograms. Int J Legal Med. 2009;123(2):123-8.

23. Ge ZP, Yang P, Li G, Zhang JZ, Ma XC. Age estimation based on pulp cavity/chamber

volume of 13 types of tooth from cone beam computed tomography images. Int J

Legal Med. 2016 Jul;130(4):1159-67.

195

24. Ge ZP, Ma RH, Li G, Zhang JZ, Ma XC. Age estimation based on pulp chamber volume

of first molars from cone-beam computed tomography images. Forensic Sci Int.

2015 Aug;253:133.e1-7

25. Pinchi V, Pradella F, Buti J, Baldinotti C, Focardi M, Norelli GA. A new age estimation

procedure based on the 3D CBCT study of the pulp cavity and hard tissues of the

teeth for forensic purposes: A pilot study. J Forensic Leg Med. 2015 Nov;36:150-7.

26. Biuki N, Razi T, Faramarzi M. Relationship between pulp-tooth volume ratios and

chronological age in different anterior teeth on CBCT. J Clin Exp Dent. 2017 May

1;9(5):e688-e693. doi: 10.4317/jced.53654. eCollection 2017 May.

196

GENERAL DISCUSSION.

This thesis emerges as an answer to some observed challenges faced by forensic dentistry

for dental age estimation in adults. The age indicator to do this analysis was the formation

of secondary dentine on the walls of the root canal, which generates a negative correlation

between pulp/tooth ratio with age. As there is a demand for non-invasive methods, this

thesis analyses radiological-based methods for dental age estimation in adults as those

proposed by Kvaal et al., (linear measurements) Cameriere et al., (area measurements),

and the different approaches using volume calculation from cone beam computed

tomography (CBCT).

The above-mentioned methods have been tested in many studies. However, there is no

available systematic review which compares their results and helps forensic dentists to

select a method that can be systematically applied, which was the first observed challenge

in this thesis. To face this challenge, Chapter 1 presents what can be considered the first

systematic review for dental age estimation in adults, following the established parameters

in the Cochrane handbook.

This systematic review discloses some of the weaknesses observed in previous studies as

the lack of consensus for minimum standards for the methodological design of different

studies regarding the materials, methods, minimum sample size, statistical analysis, and

data analysis. Based on this fact, it is possible to state that there is a need to generate

proper methodological guidelines to be followed by the international community, taking in

197

account the different requirements to avoid age mimicry, and the over-estimation of age

for young individuals or the under-estimation of age for older individuals, when regression

models are applied, which would guarantee the proposal of more accurate methods as well

as the uniform use of the existent ones. In this way, a proper meta-analysis could be

performed in future studies.

The second observed challenge was the need to test the available methods on individuals

whose dental conditions were different to “totally sound teeth”; to this end, one of the

mentioned methods was used to estimate the age in individuals whose dental conditions

had been affected by the same external factor.

The selected method to do this analysis was that proposed by Kvaal et al. in 1995, owing to

the linear measurements of length and width of tooth and root canal that this method

demands, and the external factor was orthodontic treatment which affects tooth length as

well as the production of secondary dentine. This analysis was perfomed in two papers

following different methodological designs (Chapters 1 and 2) presented in section one.

Section one also defies one of the established parameters for the use and proposal of

dental age estimation methods: the exclusive inclusion of sound teeth, a status that

changes once the individual is exposed to orthodontic treatment, because even though,

these teeth are free from other pathologies, such as caries or periodontal disease, their

integrity is altered once the orthodontic treatment starts.

198

No previous study in forensic dentistry, had considered the evaluation of the effects that

orthodontic treatment has on dental anatomy and how these associated dental changes

could affect the accuracy of the available methods.

The hypothesis for this section was that in individuals with orthodontic treatment, the age

estimates would show a considerable over-estimation of age, owing to the root shortening

and the production of dentine associated to this treatment. Before obtaining the results, it

was believed that it would be necessary to propose different standards for dental age

estimation in individuals with orthodontic treatment.

In Chapter 2 different linear regression models were proposed for individuals before and

after orthodontic treatment. The methodological design of this chapter facilitated the

comparison of correlation coefficients of different pulp/tooth ratios, and the standard error

of estimation (SEE) obtained from linear regression models obtained before and after

orthodontic treatment.

Although, for some teeth a higher SEE after orthodontic treatment, in general terms, there

was no significant difference between both observations (Pearson correlation test r=0.58),

with a difference that ranged between -0.69 to 2 years. The results of this chapter also

reassure some previous concepts that are important for the following chapters of this

thesis, as the non-significant correlation between age and the length ratios (Chapter 2

table 2.1), as well as the concept that the more teeth are included to estimate the age of

the individual, the better the results are going to be (Chapter 2, tables 2.3 and 2.4).

199

Alike previous studies, that evaluated the root resorption associated to orthodontic

treatment, Chapter 2 also demonstrated that this dental change is evident in dental

radiographs, and affects in a higher percentage upper teeth. Contrary to the expected,

although there are some changes in the morphology of tooth and pulp chamber, these

changes do not alter the accuracy of the linear regression models.

However, Chapter 2 had a debatable point in its methodology: this was a longitudinal study

that observed the same group of individuals before and after orthodontic treatment. To do

a more detailed analysis, that confirmed the findings of Chapter 2, the study presented in

Chapter 3 was designed following the methodology of epidemiological studies for risk

assessment, to test whether orthodontic treatment could be considered a risk factor able

to affect the accuracy of the Kvaal et al method, causing more over-estimation of age.

This methodological design has no precedents in forensic dentistry, making this chapter a

relevant piece for research of dental age estimation in adults. This analysis included data

from different individuals, who were grouped in two cohorts, with the same chronological

age and sex distribution. After age estimation using previously published formulae, a

comparison between age estimation obtained from individuals from both groups,

distributed in equal pairs, was done at the tooth level.

After obtaining the different age estimates, and compare the percentage of observations

where the result showed over- or under-estimation, no significant difference was observed

between both groups (Chapter 3, table 3.1). The respective contingency table was

200

elaborated (Chapter 3, table 3.2) revealing that orthodontic treatment could not be

considered as a real risk factor for obtaining over-estimation of age, when the Kvaal et al.

method is used for age estimation in adults (RR=1.0071). It is necessary to mention that

none of the teeth analyzed in this chapter presented severe apical root resorption.

The analysis presented in Chapter 3, is an outstanding contribution to dental age

estimation in adults, as it presents a study design that could be applied to analyze the

effect of other dental conditions on different methods for age estimation. This

methodology could be replicated, including teeth with and without cavities, dental fillings,

or periodontal diseases, to establish whether those teeth are non-suitable for age

estimation of adults.

The fact that this section did not find significant differences among both groups, could

represent two different scenarios: 1. the Kvaal method could be used regardless the history

of orthodontic treatment exposure, or 2. the instrument used to measure the tooth and

pulp dimensions, as well as the statistical analysis are not sensitive enough to detect the

differences between both groups. To clarify which one of the two scenarios is the

applicable, it would be necessary to perform a similar analysis using periapical radiographs,

and data from a larger sample.

Nevertheless, there are some obstacles: the use of periapical radiographs for orthodontic

treatment is not as popular as the use of panoramic radiographs, as they would provide

limited information about the different anatomic structures, for example maxillary sinuses,

201

and osseous structures. For that reason, the analysis presented in this section used only

panoramic radiographs.

As the third listed challenge for dental age estimation in this thesis, was the lack of

methods for age estimation in young individuals with absence of third molars, the Kvaal et

al method was tested in a younger population, as presented in Chapter 4. For the analyzed

sample, the Kvaal et al method seems to be a useful tool for younger individuals. However,

owing to the SEE (±4.5 years), it would be necessary to analyse other age indicators, to

establish whether an individual is an adult, using methods with higher accuracy.

To cover the fourth listed challenge in this thesis, Chapter 5 presents a study that explores

new uses and applications of one the existent methods for age estimation in adults. This

chapter presents the first adaptation of the Kvaal et al method on CBCT. For this case,

width measurements only were collected, in sagittal and coronal views. Although the

exclusive analysis of width measurements of root and pulp canal on CBCT also showed

negative correlation between age and pulp/tooth ratio, the accuracy (SEE>10 years) does

not surpass the analysis of dental radiographs with mesio-distal pulp/tooth width

measurements.

Even though the results of the papers presented in Chapters 4 and 5 (section 3), are not as

accurate as expected, the analysis was deemed necessary to be done, exploring different

applications of this method for dental age estimation, especially in the case of the use of

CBCT doing linear measurements. It is necessary to mention that for future studies

202

following the presented methodology, it would be necessary to analyze larger samples and

test different statistical methods, not only linear regression models as it was done for these

studies.

On the other hand, the constant advances in imaging analysis and telecommunications,

demand forensic dentistry to evolve and to adapt to new sources of information and

emerging technologies. In this way, Chapters 6 and 7 are clear examples of the application

of one the most recent programs for imaging segmentation on dental CBCT, and illustrate,

how, counting with the support of the international academic community it is possible to

design innovative studies.

The use of CBCT and volume reconstruction of different anatomical structures has led to

the proposal of different methods for identification of individuals. However, when it comes

to their use in forensic dentistry for dental age estimation in adults, the results are limited.

The analysis is more complex and demands more time, than when using dental

radiographs, aspects that could be considered as disadvantages of previously proposed

methods.

To help forensic dentists to face the fifth listed challenge in this thesis: the need to

understand the different technological aspects that affect those methods based on

pulp/tooth volume reconstruction, the study presented in Chapter 6 compares the

previously proposed methodologies, and analyses the possible circumstances that affect

their accuracy.

203

Chapter 6, demonstrated that although there are different approaches available to quantify

the volume of root canal, it is difficult to obtain the real volume of this space, also that

regardless the technique used to obtain volume measurements from pulp chamber, there

are more implicit aspects in the software that forensic dentists need to understand and be

aware of, to maintain the objectivity of the proposed methods. In this study, it was also

observed that the method which proved to be more reliable to measure volume was

semiautomatic segmentation using the software ITK-SNAP version 3.4 (open source

software, www.itksnap.org), as it was less likely to be affected by the subjectivity of the

observer.

Based on the findings of the study presented in Chapter 6, the software was used to

analyze a sample that included data from two population groups with different ethnic

backgrounds.

The paper presented in Chapter 7, is a homage to the international community that

collaborated in this research, and an example of how the use of telecommunications is

promoting the generation of innovative methods for forensic sciences, helping researchers

to overcome the geographical barriers, that previously limited the design of studies using

large scale radiological data from different population groups.

This chapter also discusses one of the most recent findings in forensic dentistry as it is a

selection bias called age mimicry, which probably has affected the results of previous

studies for age estimation. This selection bias could be the explanation to the wrong

http://www.itksnap.org/

204

assumption that there are population and ethnic differences that affect the accuracy of

past studies.

To avoid age mimicry, the sample needs to meet rigorous conditions that guarantee the

reliability of the results. Although, the paper presented in Chapter 7 could not fulfill all the

required conditions, it presents the pathway forward in proposing a model for inter-

population studies for age estimation in adults.

GENERAL CONCLUSIONS

This thesis confronts some established paradigms for the use of methods for dental age

estimation in adults. Likewise, each one of the chapters that this thesis encompasses,

reveals new concepts, and presents new proposals that might be useful for future studies,

which makes each one of the sections and chapters of this thesis a valuable contribution

for forensic dentistry.

Likely, the existent literature has elaborated some parameters for the design and

application of methods for age estimation. These parameters specify: the study of

individuals from the same population; and the exclusion of teeth with pathologies or dental

procedures. Evidence presented in this thesis challenges these rather limited parameters

which are too restrictive to be used as guidelines. Rather it has been shown that samples

from widely different populations and from teeth which are in various conditions may be

used to generate meaningful data in forensic analyses.

205

Despite, the similarities of the study design of previous publications, that followed or

slightly modified the three mentioned methods in this thesis, it was possible to observe

that there are no established protocols for data analysis and the report of the results.

More importantly, there are no parameters that consider the chracteristics of the

individuals included in the study as well as a sample size, which might be big enough, to

simultaneously include individuals with different dental conditions and the uniqueness of

each subject.

Under this frame, it is possible that the main conclusion from section one is that an ideal

research scenario, the study design should ensure the inclusion of individuals with a dental

status different to “totally sound teeth” representing the variability existent in any

population, taking in account the parameters menioned in Chapter 7 to avoid age mimicry

bias.

Section two sends some clear messages:

1. It is necessary to test the available methods for dental age estimation in adults in

individuals with different dental characteristics.

2. Orthodontic treatment is not an obstacle to the application of the Kvaal et al

method.

3. It is necessary to perform similar analyses using periapical radiographs.

206

4. The use of CBCT for dental age estimation needs to be thoroughly tested, for linear,

area and volume measurements, and if necessary, discourage its use, due to the

simplicity of those methods using dental radiographs.

5. It is recommended to conduct studies that include individuals from different groups

of population.

After analyzing the results of Section three it is possible to state:

1. The Kvaal et al. method is a validated tool for age estimation, and could be an

alternative for those individuals with agenesis of third molars. However, for the

determination of the age of majority of an individual, it would be advisable to use

the Kvaal et al. method in association with an analysis of other age indicators, such

as skeletal age-related changes.

2. Minor dental changes are not an obstacle to the use of the Kvaal et al method, as

long as they do not affect the root of the tooth.

3. Although CBCT provides more information, the accuracy for age estimation in adults

using the Kvaal et al. method is not better than that provided by dental radiographs.

The main conclusions from Section four are:

207

1. Before proposing a method for dental age estimation using volume measurements,

it is necessary to understand the functioning of the software to be used.

2. Although there are different approaches for tooth volume measurements from

CBCT, the more reliable are those based on semi-automatic or automatic

segmentation.

3. With the new technological advances in telecommunications and imaging analysis,

it is possible to do studies which include data from individuals with different ethnic

backgrounds.

4. Future studies need to follow the stated conditions to avoid the selection bias

called age mimicry, which are:

a. Even number of participants in each age category.

b. Large samples.

c. Well stablished age ranges.

d. Equal numbers of individuals per age range.

e. Individuals from different populations.

5. It is recommended to test the methodology proposed in Chapter7 in a larger

sample, which includes at least two groups of individuals with different ethnic

backgrounds but, with the same age and sex distribution, to ensure the reliability of

the results.

208

6. To do a reliable meta-analysis in the future, it is necessary to establish appropriate

standards for the publication of the results and accuracy of different studies.

And finally, the different new approaches and statements presented by this thesis

deserved to be tested, debated, and confronted by future researchers, using innovative

techniques and more developed tools that will be available in the coming years for forensic

dentistry.

Final note: Some of the chapters have been slightly modified after the publication of the

most recent findings about age mimicry (Chapter 1, Chapter 5, Chapter 7).

209

APPENDIX 1. Age estimation in adults by dental imaging assessment systematic review.

Marroquin, T. Y., Karkhanis, S., Kvaal, S. I., Vasudavan, S., Kruger, E., & Tennant, M. Age

estimation in adults by dental imaging assessment systematic review. Forensic Science

International Journal. Accepted for publication, March 2017.

Age estimation in adults by dental imaging assessment systematicreview

T.Y. Marroquina, S. Karkhanisa, S.I. Kvaalb, S. Vasudavanc, E. Krugera,*, M. Tennanta

aUniversity of Western Australia, School of Anatomy Physiology and Human Biology, 35 Stirling Hwy, Crawley WA 6009, AustraliabUniversity of Oslo, Department of Oral Pathology and Section of Forensic Odontology, P.O. Box 1109, Blindern, Oslo, NorwaycDepartment of Developmental Biology, Harvard School of Dental Medicine, 188 Longwood Ave, Boston, MA 02115, United States

A R T I C L E I N F O

Article history:Received 26 October 2016Received in revised form 28 January 2017Accepted 17 March 2017Available online 27 March 2017

Keywords:Forensic dentistryAdultsAge estimationDental imagingSystematic review

A B S T R A C T

Importance: The need to rely on proper, simple, and accurate methods for age estimation in adults is still aworld-wide issue. It has been well documented that teeth are more resistant than bones to thetaphonomic processes, and that the use of methods for age estimation based on dental imagingassessment are not only less invasive than those based on osseous analysis, but also have shown similaror superior accuracy in adults.Objectives: To summarise the results of some of the recently most recently cited methods for dental ageestimation in adults, based on odontometric dental imaging analysis, to establish which is more accurate,accessible, and simple.Evidence review: A literature search from several databases was conducted from January 1995 to July2016 with previously defined inclusion criteria.Conclusion: Based on the findings of this review, it could be possible to suggest pulp/tooth area ratiocalculation from first, upper canines and other single rooted teeth (lower premolars, upper centralincisors), and a specific statistical analysis that considers the non-linear production of secondary dentinewith age, as a reliable, easy, faster, and predictable method for dental age estimation in adults. The secondrecommended method is the pulp/tooth width–length ratio calculation. The use of specific populationformulae is recommended, but to include data of individuals from different groups of population in thesame analysis is not discouraged. A minimum sample size of at least 120 participants is recommended toobtain more reliable results. Methods based on volume calculation are time consuming and still needimprovement.

© 2017 Elsevier B.V. All rights reserved.

1. Introduction

Age estimation is one of the most important characteristicsused to establish the identity of any individual in different legal,forensic, or anthropological research context [1]. To this end,forensic teams depend on osseous analysis based methods, whichhave acceptable results for young individuals or in their earlyadulthood [2], and dental development based methods, which arehighly reliable in individuals under 21 years of age [3]. However,these methods have some disadvantages: the poor resistance ofbones to the taphonomic process [4], and once the individual

reaches the threshold of 21 years of age, and the third molarsdevelopment concludes [3], the currently available dental devel-opment based methods are no applicable. In individuals with thecongenital absence of third molar teeth, this threshold falls downup to 14–15 years of age.

To respond to the need of an ageing population, and with theevident resistance of teeth to the taphonomic process, alternativemethods for dental age estimation in adults have been proposed.Primarily, these are based on the formation of secondary dentine,studied since 1950 [5] and the subsequent narrowing of the pulpcavity, which can be observed in dental radiographs, leading to theproposal of minimally invasive methods. This systematic reviewfocuses on three methods based on odontometric analysis of thepulp cavity, performing length and with measurements [6], areameasurements [7] and lastly volume calculation [8]. The objectiveof this review is to summarise the results of these recently mostcited methods for dental age estimation in adults, to establishwhich method is more accurate, accessible, and simple.

* Corresponding author.E-mail addresses: [email protected]

(T.Y. Marroquin), [email protected] (S. Karkhanis),[email protected] (S.I. Kvaal), [email protected](S. Vasudavan), [email protected] (E. Kruger), [email protected](M. Tennant).

http://dx.doi.org/10.1016/j.forsciint.2017.03.0070379-0738/© 2017 Elsevier B.V. All rights reserved.

Forensic Science International 275 (2017) 203–211

Contents lists available at ScienceDirect

Forensic Science International

journal homepage: www.elsevier .com/ locat e/ f orsc i in t

undefined

undefined

http://dx.doi.org/10.1016/j.forsciint.2017.03.007

undefined

www.elsevier.com/locate/forsciint

http://dx.doi.org/10.1016/j.forsciint.2017.03.007

undefined

undefined

http://crossmark.crossref.org/dialog/?doi=10.1016/j.forsciint.2017.03.007&domain=pdf

undefined

undefined

http://www.sciencedirect.com/science/journal/03790738

1.1. Description of the problem or issue

Different methods have been published for dental age estima-tion in adults, based on the pulp/tooth dimensions’ ratios.Nevertheless, the obtained results of the application of some ofthese methods for dental age estimation, in adults from differentpopulation groups, surpass the accepted threshold in forensicsciences which says that the standard deviation of a method foradult’s age estimation should preferable be below a standarddeviation (SD) of years �10 years [9].

1.2. Description of the methods being investigated

The methods for dental age estimation in adults analysed in thispaper were selected based on their minimally invasive nature, notrequirement for the extraction of teeth to be performed, and pulp/tooth ratio calculation which have been applied in individuals fromdifferent populations. Kvaal et al. [6] method is based on theanalysis of linear measurements of the pulp, tooth, and root lengthas well as root and pulp width measurements at three differentroot levels, initially applied on periapical radiographs and later onpanoramic radiographs and tomographs. Cameriere et al. [7,10]method is based on the analysis of pulp and tooth area measure-ments on periapical and panoramic radiographs. Finally, thedifferent methods for dental age estimation in adults based onpulp/tooth volume ratio from cone beam computer tomography[8,11–14] were also included in this systematic review.

1.3. How these methods might work

The included methods in this study are based on a negativecorrelation between age and the pulp chamber size, as well as onthe tooth/pulp ratio calculation, regardless the nature of the usedmeasurements: length/width, area, or volume. In other words, allof them look at the association of one age related phenomenon, asit is the formation of secondary dentine and the decrease of pulpchamber size with age, which has been accepted as an ageindicator, observable and measurable with different dentalimaging techniques. As an ideal, the accuracy of the studiedmethods should not exceed the threshold of a SD �10 years [4,1].

1.4. Why it is important to do this review

The relevance of this review is grounded on the need torecommend a method for dental age estimation with the followcharacteristics: simple, fast, non-invasive, non-expensive, repro-ducible and over all, accurate, that can be systematically used indifferent academic and forensic scenarios. Helping the reconstruc-tion of identity profiles of unidentified deceased individuals oralive individuals with doubtful identity documents.

2. Methods

2.1. Criteria for considering studies for this review

Qualitative analysis of the information: original studies, inhumans, reporting the use of any of the listed methods for dentalage estimation, based on pulp/tooth ratio calculation (length/width, area, volume) that preferably reported intra-inter observercalibration, generating population specific formulae or in case thatdid not, that reported if the obtained results were obtained byusing the method’s author’s original formulae. English or Spanishlanguage that expressed the results in terms of accuracy for dentalage estimation.

Quantitative analysis of the information: same criteria thanqualitative analysis plus the exclusion of studies which sample

included individuals younger than 14 years of age, studies withsmall samples (n < 50), studies using extracted teeth, and studiesthat did not report the use of specific population formulae.

2.2. Search methods for identification of studies

The information was searched though the data-base available atthe University of Western Australia which included the collectionsof:

Directory of open access journals (DOAJ), Medline/Pubmed(NLM), OneFile (GALE), ProQuest, collection, (Web of Science),Science Direct Journals (Elsevier), Social Sciences Citation Index(Web of Science, Scopus (Elsevier)), SocialSciences Citation Index(Web of Science), Wiley (CrossRef), Wiley Online Library. Also,google scholar, by looking at the papers that reference the originalstudy performed by Kvaal et al. Cameriere et al., and those methodsfor age estimation that referred in their methodology the use ofCBCT and volume reconstruction of pulp chamber and tooth.

The search key words were as follow: Kvaal and dental ageestimation; Cameriere and dental age estimation; age; and toothvolume.

The literature search included papers published after thepublication of the original papers of the authors to July 2016, thesearch was conducted during the years 2014 to 2016. In the lack of amanual for systematic review in forensic dentistry, this systematicreview follows the Cochrane handbook for systematic reviews-methodology review, and when possible the RevMan softwarerecommended by the Cochrane handbook (Figs. 1–3 ).

2.3. Data collection and analysis

2.3.1. Selection of studiesThe initial selection was based on the title and then abstract.

Papers with titles that referred the inclusion of only minors wereexcluded (children). Studies reporting also the use of third molarsor developing teeth, were excluded. Studies that included in thetittle also the use of other methods for dental age estimation inchildren were excluded (Demirjian) or any other invasive methodfor dental age estimation were excluded (Diagrams 1–3,Tables 1–3 ).

2.3.2. Data extraction and managementThe collected information was organized in an excel spread-

sheet as follow: Author, year, country, number of participants(male and female), age, intra and inter-observer agreementassessment, imaging technique to obtain the images, measuringinstrument, best correlation coefficient between age and thedifferent age predictors, best result in terms of accuracy perindividual tooth or per set of teeth, when possible, as well asthe highest error recorded by set of teeth or individual tooth(Tables 4–6 ).

2.3.3. Assessment of risk of bias in included studiesTo avoid bias in this systematic review, and to avoid false

positive (declare that a method is more accurate than other whenit is not) or false negative conclusions (declare that a method isless accurate than other when it is not), it was necessary toanalyse the possibility of author bias. This owed to theparticipation of the same authors in repeated publications. Tothis end, the results were analysed comparing individual papers,and then grouping them per author. Additionally, in certain caseswere there were doubts in regards to the origin of the sample,which means the likelihood to find studies that had used the samesample, the authors were contacted to confirm the origin of thesample, and in case that two studies had the same sample, theauthors were asked to suggest which study should be included in

204 T.Y. Marroquin et al. / Forensic Science International 275 (2017) 203–211

the meta-analysis, in case that the study met the criteria for thequantitative analysis. In the same way, to deal with missing data,the authors were contacted via e-mail.

There was another possible source of bias observed in thisstudy: the use of non-specific population regression models andequations, which could cause the judgement of a method as no orless accurate. To overcome this issue, only studies using specificpopulation formulae were used in the quantitative analysis.

3. Results

3.1. Kvaal et al. [6] (29 Papers, n = 3254)

3.1.1. Qualitative analysisThe original paper about Kvaal et al. [6] method was published

in 1995. Since then, this method has been applied to a globalsample of 3254 individuals from 12 different countries (Graph 1,Table 1). The initial methodology suggested the use of periapicalradiographs from six different teeth, using vernier callipers tomeasure the maximum tooth length, the pulp length and rootlength, and a stereomicroscope with a measuring eyepiece to thenearest 0.1 mm to measure pulp and root width at three differentlevels previously described by the author of this method [6]. Justfew studies have followed the original methodology with slightadaptations [1,6,9,16–19]. In 2005, this method is applied onpanoramic radiographs (20), which became popular [21–35]. Theuse of digital imaging in dentistry also introduced the use ofdifferent software to perform the odontometric measurements.The most commonly used software are: Adobe Photoshop(different versions n = 5), Image J (n = 4) and Kodak dental imagingsoftware (n = 2). From those studies that tested the effect of sex onthe accuracy of the age estimated, 3 found that sex does not affectthe models [16,20,31] and 2 that it does [19,24]. In the original

Fig. 1. Flow chart of the study selection for Kvaal et al. method.

Fig. 2. Flow chart of the study selection for Cameriere et al. method.

T.Y. Marroquin et al. / Forensic Science International 275 (2017) 203–211 205

study, sex was included as a factor for the mandibular lateralincisor, indicating that pulp cavity size changes occurred faster inmales, as females need another 6 years to get the same age asfemales for this tooth [6].

3.1.2. Quantitative analysisFrom the total sample, only 16 papers met the inclusion criteria

for the quantitative analysis (Table 4). The most accurate result(SD = 5.6, r = �0.95) was obtained from panoramic radiographsusing Hipax program (version 3.01) software, 168 participants(95 female and 102 male) in a Caucasian group [20] The larges errorreported error was SEE > 13 years r2 = 0.1 [16,36] obtained from thelower canine, following original Kvaal methodology on periapicalradiographs [36] and the Adobe Photoshop 6.0 software inpanoramic radiographs [16].

Kvaal et al. measurements have been reported to have a highdegree of intra and inter-observer agreement indicating thereproducibility of the measurements, even in those studies thatdid not follow the original methodology. Only one study reportedsignificant differences in regards to the root length and pulp widthat the level B and root width at the level A measurement, althoughthis study reported a SEE = �13.8 (r2 = 0.38), the average error ofage estimation was �18–21 years using periapical radiographs[36].

Note: the studies that did not meet the inclusion criteria for thequantitative analysis also reported high intra and inter-observer

agreement [32–34], One study reported that the lower lateralincisor produced lower intra-observer correlation [33], and one ofthem affirmed that the use of stereomicroscope improve theaccuracy of the age estimates [9].

Fig. 3. Flow chart of the study selection for methods based on volume calculation.

Table 1Total studies reporting the use of Kvaal et al. method.

Author Year Country Totala Age

Kvaal et al. [6] 1995 Norway 100 20–87Kvaal et al. [17] 1999 Sweden 21 20–60Willems et al. [9] 2002 Belgic 29 26–85Soomer et al. [1] 2003 Caucasian 20 14–95Bosmans et al. [21] 2005 Belgic 197 19–75Paewinsky et al. [20] 2005 Germany 168 18–81Meinl A. [32] 2007 Austria 44 13–24Avendaño et al. [19] 2009 Colombia 107 21–50Landa M. [33] 2009 Portugal 100 14–60Shetty et al. [16] 2010 India 100 20–70Sharma et al. [72] 2010 India 50 15–60Saxena. [31] 2011 India 120 21–60Saxena et al. [41] 2011 India 160 21–60Chandramala et al. [22] 2012 India 100 20–70Kanchan et al. [36] 2012 India 100 25–77Agarwal. [18] 2012 India 50 20–70Erbudak et al. [30] 2012 Turkey 123 15–57Thevissen et al. [34] 2012 Belgic 450 15–23Limdiwala et al. [23] 2013 India 150 20–55Parkin et al. [29] 2013 India 30 15–60Kostenko. [73] 2013 Ukraine 64 –

Karkhanis et al. [25] 2014 Australia 279 20–62Misrlioglu et al. [26] 2014 Turkey 114 17–72Patil et al. [70] 2014 India 200 20–50Ayad et al. [24] 2014 Sudan 99 15–30Muszynska et al. [35] 2015 Poland 3 –

Marroquin et al. [27] 2016 Australia 74 12–28Mittal et al. [28] 2016 India 152 14–56Rajpal et al. [71] 2016 India 50 15–5729 papers 12 countries 3254 12–95

a Number of individuals reported in each study. Sample size.

Table 2Total studies reporting the use of Cameriere et al. method.


Cameriere et al. [7] 2004 Italy 100 18–72Cameriere et al. [53] 2006 Italy 33 –

Cameriere et al. [10] 2007 Italy 100 20–79Cameriere et al. [37] 2007 Italy 100 20–79Cattaneo et al. [65] 2008 Ethiopia 1 52Cameriere et al. [38] 2009 Portugal and Italy 229 20–84Singaraju et al. [39] 2009 India 200 18–72Babshed. [4] 2010 India 178 20–70De luca et al. [2] 2010 Spain–Italy 73 –

Babshed et al. [51] 2011 India 61 21–71Cameriere et al. [74] 2011 Italy 90 50–79Jeevan et al. [40] 2011 India 228 16–72Saxena [31] 2011 India 120 21–60Vodanovic et al. [66] 2011 Croatia 192 –

Zaher et al. [42] 2011 Egypt 144 12–60De luca et al. [52] 2011 Mexico 85 18–60Cameriere et al. [43] 2012 Spain 606 18–75Cameriere et al. [47] 2013 Portugal 116 18–74Charis et al. [44] 2013 India 120 20–70Cameriere et al. [47] 2013 Portugal 116 18–74Azevedo et al. [45] 2014 Italy 81 19–74Misirlioglu et al. [26] 2014 Turkey 114 17–72Azevedo et al. [45] 2015 Brazil 443 20–78Ravindra et al. [48] 2015 India 308 9–68Cameriere et al. [75] 2015 Italy 70 20–70De Angelis et al. [68] 2015 – 1 –

Fabbri et al. [67] 2015 Italy 18 –

Sakhdari et al. [49] 2015 Iran 120 >12Torkian [76] 2015 Iran 120 >1229 papers 11 countries 4167 12–79



3.2. Cameriere et al. [7] (29 papers n = 4167)

3.2.1. Qualitative analysisThe original paper documenting Cameriere et al. method [7]

was published in 2004. In this systematic review, 29 papers usingthis method were found, having a global sample of 4167 individualsfrom 11 countries (Graph 2, Table 2). This method was initiallyapplied to digital panoramic radiographs, using (AutoCAD2000,Install Shield3.0, 1997) and different versions of this software havebeen used in 11 studies. The most used software to perform thepulp/tooth area measurements is Adobe Photoshop (13 studies).This method was designed to be applied in single rooted teeth,especially canines, but it has also been tested in premolars andcentral and lateral incisors. From those studies that assessed theeffect of sex on the regression models, the vast majority reportedthat sex does not affect the regression model [4,7,26,31,37–46].Only 4 studies reported that sex affect the regression model [47–50].

3.2.2. Quantitative analysisTwelve studies met the inclusion criteria for the quantitative

analysis (Table 5). The use of this method has shown a standarderror of estimation (SEE) from �1.2 [42] to �12–13 years (r = 0.2–0.4) [51] From these studies, 2 reported a median error < � 5 years[7,38], 6 studies reported an error from �5 to �8.5 years [26,40,42–44,46], and three studies reported an error from �10 to �13 years[4,47,51]. From the included studies in the qualitative analysis, onlyone reported significant intra-observer differences in lateralincisors and first premolars (p < 0.05) [51].

Note: all the studies that did not meet the inclusion criteria forthe qualitative analysis and that did intra and inter-observercalibration, reported high intra inter-observer agreement[37,49,52,53].

3.3. Volume (16 papers n = 2444)

3.3.1. Qualitative analysisOwed to the low number of studies doing pulp/tooth volume

reconstruction, in the qualitative analysis those studies usingextracted teeth were included. The first study doing pulp/toothvolume reconstruction reported the use of micro-focus X-ray in2004 [8]. In this systematic review, 16 studies were found using thepulp/toot volume reconstruction, by means of micro-focuscomputer tomography, CT scan or CBCT, and the use of differentsoftware to do manual or semiautomatic volume reconstruction,

such as Tri 3D Bon, Mimics1, ITK-SNAP 2.4, Osirix, Amira amongothers, in a global sample of 2444 individuals from 7 countries(Graph 3, Table 3). However, the methodology of some of themrequire the use of extracted teeth [8,12,13,54,55]. In regards to theeffect of sex on the regression models, 9 studies reported that sexhas no effect [8,14,26,54,56–60]. 3 studies reported that the modelsare more accurate in female than in male, having a higherdetermination coefficient for women than form men [12,61,62], asfollow: R2 0.6 for males and R2 0.7 for females [12], or R2 = 0.29 formales and R2 = 0.15 for men [61] and R2 = 0.67 for male andR2 = 0.75 for female when the mandibular central incisor is used orR2 = 0.56 for male and R2 = 0.58 for female when the secondpremolar is used. Two studies, which only reconstructed the pulpcavity volume, also reported significant difference between toothtype and sex [63,64], One of them found a significant difference inthe volume between the volume of both genders (p = 0.028 < 0.05)and a stronger relation between pulp chamber and age for female(R2 0.6) than for male (R2 0.5) [63]. The other one reported asignifficant difference in volumebetween genders for 12 types ofteeth (p = 0.000) [64].

3.3.2. Quantitative analysisFour studies meet the inclusion criteria for the qualitative

analysis. The error of these studies variates between �3.47 years[58] to �28 years [59]. It was reported high intra-inter-observeragreement [60]. Among those studies that did not meet theinclusion criteria for the quantitative analysis, seven studies alsoreported high intra-inter-observer agreement [8,11–14,57,63,64].

4. Discussion

Age estimation in adults is a challenge in all forensic contexts,especially in cases that require the use of non-invasive methods.The formation of secondary dentine and the non-linear narrowingof the root canal with age [11], is one age predictor measurable indental radiographs and tomographs, leading to the proposal ofdifferent methods for age estimation in adults as an alternative tomore invasive methods and as a complement to osseous analysis[2,17,35,65–68].

In the same way, many methods have been suggested for dentalage estimation in adults. Nevertheless, there are few availablestudies that compare their accuracy [69]. In the lack of a consensusto uniformly apply a method for age estimation in adults, the needto perform a systematic review seemed to be evident. Thissystematic review summarises and compares the results of some ofthe most used methods for dental age estimation in adults,performing a qualitative and quantitative analysis.

In terms of the qualitative analysis, it has been reported thatthere are no significant differences between right or left teeth[14,57]. Another aspect to test among the different papers was theeffect of sex of the different methods. Although not all the includedstudies assessed the effect of sex on age estimation, those that did(n = 38), 71% (n = 27) affirm that sex has not statistically significanteffect on age estimation, when pulp/tooth ratio is calculated.However, if only pulp volume is measured and used as ageindicator, there is a significant difference between the accuracy formales and females [63,64].

In the light of the evidence one could suggest that ratiocalculation not only diminishes the effects of tooth magnificationin dental radiographs, as reported by Kvaal et al. [6], but alsoreduces the effect of sexual dimorphism related to tooth size. Theage of the participants is also another relevant aspect to include, inthis systematic review those studies including participants under14 years of age where excluded for two main reasons. First, thereare other methods more reliable for aging teenagers and infants,and second, before 14 years of age, not all the teeth have completed

Table 3Total studies reporting the use of volume calculation.


Vandervoort et al. [8] 2004 Belgic 52 24–66Yang et al. [11] 2006 Belgic 19 23–70Someda et al. [12] 2009 Japan 155 12–79Agematsu et al. [62] 2010 japan 258 teeth 20–79Aboshi et al. [13] 2010 Japan 50 20–78Star et al. [56] 2011 Belgic 111 10–65Tardivo et al. [57] 2011 France 58 14–74Jagannathan et al. [14] 2011 India 188 10–70Sakuma et al. [54] 2013 Japan 136 14–79Tardivo et al. [58] 2014 France 210 15–85Sasaki et al. [55] 2014 Japan 363 15–70Mendonca et al. [61] 2015 Brazil 72 22–70Ge et al. [63] 2015 China 403 12–69De Angelis et al. [59] 2015 Italy 91 17–80Pinchi et al. [60] 2015 Italy 148 10–80Ge et al. [64] 2016 China 240 16–6316 papers 7 countries 2296 10–85



the apex closure, which by definition, is a requirement for theformation of the secondary dentine [77].

For the qualitative analysis, characteristics related with samplesize and inclusion criteria stablished in different studies werecompared. As compared to previous efforts, we did not find asignificant correlation between sample size and the accuracy of themethod, when all the data were analysed together (r2 = 0.1, p-value<0.05, sample size of the analysed studies n = 50–604) or when thedata were analysed individually for each method. It is necessary tohighlight that from the studies included to perform this analysis(n = 30) only 4 (13.3%) had a sample smaller than 100 individuals(n = 50–91 individuals), which may suggest that in any study forage estimation the minimum sample should include data from atleast 100 participants, or over 100 teeth from different individualsin those studies using only one tooth type. Some of the excludedstudies for the quantitative analysis, fail in this aspect[56,57,61,62], including data from several teeth of the sameindividual and counting them as different samples, which mayaffect the reported results.

A previous study recommended a minimum of 120 participantsto achieve 80% of power and 5% of significance [31]. One almostuniversal inclusion criteria, among the papers, was the use of onlytotally sound teeth. Only two studies did not follow this parameter,using teeth with cervical lessons to do volume calculation [55]. or

using panoramic radiographs that did not meet the inclusioncriteria suggested by Limdiwala et al. [23]. Both of them found notsignificant effect of these situations on the accuracy of the results.

Another aspect to consider, is the description of the differentmethods, in terms of the procedure to measure the different pulp/tooth dimensions, as well as the statistical processing of the data.In certain cases, it was necessary to re-read the papers severaltimes to understand the proposed methodology and how theaccuracy was reported.

In the quantitative analysis two main characteristics werecompared among the studies: repeatability of the measurements,in terms of intra and inter-observer agreement and their accuracy,reported as standard error of estimation (SEE), mean absolute error(MAE) or standard deviation (SD). It is necessary to mention thatowed to this variety in the approach to evaluate and to report theobtained accuracy, it was not possible to perform a deeperstatistical analysis, a meta-analysis, to confirm that one methodwas superior to another, however in this review we found that inaverage, the studies using pulp/tooth area ratio reported a lowererror (Tables 4–6).

All the studies reported significant intra- and inter-observeragreement, regardless the used method and the measuringinstrument, which suggests that after adequate training any ofthese methods is reproducible. However, in regards of the

Table 4Kvaal et al. method. Studies included in the quantitative analysis.

Study Sample Measuring instrument SEE � years per tooth or group of teeth (FDI)

6_teeth upper lower 11/21

12/22

15/25

32/42

33/43

34/44

11/21_32/42

13/23

Kvaal et al. [6] 1995, Norway 100 PER20–87 years

Stereomicroscope andVernier callipers

8.6 8.9 9.4 9.5 10 11 10.5 11.5 11.5

Bosman et al. [21] 2005,Belgic

197 OPG19–75 years

Adobe Photoshopsoftware

9.5 9.2 9.9 9.7 9.8 9.3 11.6 8.2 8.1

Paewinsky et al. [20]a 2005,Germany


Hipax program 5.6 6.4

García et al. [19] 2009Colombia

107 PER21–50 years

Scion Image software 7.1

Ranjani et al. [16] 2010, India 100 PER20–70 years

SV- 4 mini slide viewer,digital Vernier callipers,Stereomicroscope

10.5 10.2 11.8 11.5 12.4 11.8 12.7 13.3 11.8

Erbudak et al. [30] 2012,Turkey


Image J software 10.01 10.12 8.73 8.82

Saxena et al.b [41] 2011, India 120 OPG21–60 years

AutoCAD2005 software 3.63

Kanchan-Talreja et al. [36]2012, India

Group A47 PER25–75 years

Adobe Photoshop software 12.08 12.08 12.4 12.79 13.25 11.87 13.3 13.28

Group B43 PER25–77 years

11.9 11.27 12.46 11.17 13.4 12.62 13.43 13.89 12.75

Limdawala and Shah [23]2013, India

Group A100 PER25–50 years

Kodak Dental ImagingSoftware

8.3 8.21 9.09

Group B50 PER25-50 years

9.45 9.5 9.88

Misirlioglu et al. [26] 2014Turkey


Easy-Dent PC software 5.88 7.36 7.52 6.94

Patil et al. [70] 2014, India 200 PER20–50 years

Image-Pro Plus II software 6.5

Karkhanis et al. [25] 2015,Australia


Image J software 8.99 9.6 8.36 9.36 9.64 9.52 10.22 10.9 10.53

Mittal et al. [28] 2016, India 152 OPG14–60 years

VistaScan DBSWIN software 7.97 8.59 7.51 8.15 8.53 7.89 8.85 7.95 7.58

Rajpal et al. [71] 2016, India 50 PER15–57 years

Kodak Dental ImagingSoftware

6.42 7.3 7.84

PER = periapical radiographs. OPG = panoramic radiographs. Tooth numeration FDI (Federation Dentaire International).a Error reported as standard deviation.b Error reported as the difference between the chronological age and the estimated.


simplicity of the technique, and the access to the required tool tomeasure and to calculate the respective pulp/tooth ratio dimen-sions, Kvaal et al. and Cameriere et al. are more convenient, usingperiapical radiographs or panoramic radiographs, as the radiologi-cal technique does not affect the accuracy of the measurements, aslong as the quality of the image allows the observer to clearlyobserve the boundaries on the root canal and tooth surface [23].The volumetric reconstruction of anatomic structures involves amore technical and time consuming training as well as the use ofmore complex software, that are not always free access. This couldbe considered as an obstacle for the big scale application of some ofthese methods. Likewise, the user spends more time doing thevolumetric reconstruction per tooth, which variates from 4 h [8] to15 min [61], when reported.

Another important aspect in regards of accuracy, is the use ofthe specific population formulae, which also improves the resultsof the age estimates, independently of the used method[14,30,42,46,64,70]. In those studies, using non-specific populationformulae the reported error was notably larger (error > �20 years)[32,33]. However, the use of data from different population in thesame statistical analysis, to generate a unique age estimationequation, is not discouraged [2,21,47]. Furthermore, it has beenreported that the pulp chamber size variation can be detected only

each 10 years, which opens a question mark about the reliability ofthose methods reporting a lower error.

In the specific analysis of Kvaal et al. method, although thismethod required the inclusion of six different single rooted teethper idividual, the use of only one tooth [19], only upper centralincisor [70], canines [26], mandibular teeth [33], or differentcombination of teeth to apply the odontometric analysisdescribed by Kvaal and co-workers [25,27], has also beenreported, with acceptable results (SEE < �10 years). It has alsobeen reported that its accuracy depends on the quality of theimage and on the precision of the measurements [23]. Severalstudies reported that tooth length is not strongly correlated withage [16,18,20,23] as tooth length would depend on tooth wear,bruxism, and food habits, rather than a physiological agingrelated process [16].

In regards to Cameriere et al. method, it only requires the use ofone teeth, and the majority of the studies report the use of canines.However, one study using three different lower teeth (Lateralincisor, canines, first premolars) found that lower canine had thepoorest correlation coefficient with age (r = �0.2) (51). Anotherstudy found more accurate estimations from upper lateral incisors,claiming that the narrowing velocity in the pulp chamber of thistooth is twice faster than in the lower incisors [47].

Table 5Cameriere et al. method. Studies included in the quantitative analysis.

Study Sample Measuring instrumet SEE � years per tooth (FDI)

32/42 11/21 12/22 31/41 32/42 33/43 34/44 35/45 13/23

Cameriere et al. [7]a

Italy, 2004100 OPG18–72 years

AutoCAD2000 3.27

Cameriere et al. [38]Portugal and Italy, 2009


Adobe Photoshop 4.33 4.24

Babshet et al. [4]India, 2010


Adobe PhotoshopAutoCAD 2004

10 10

Babshet et al. [51]India, 2011

61 PER21–71 years

Adobe PhotoshopAutoCAD 2004

12.22 12.28 13.08 12.45

Jeevan et al. [40]a

India, 2011228 PER16–72 years


Zaher et al. [42]Egypt, 2011


AutoCAD2008 2.63 1.94

Cameriere et al. [43]Spain, 2012



Cameriere et al. [47]Portugal, 2013


Adobe Photoshop 7.03 6.64 10.8 10.9

Charis et al. [44]a

India, 2013120 PER20–70 years

Jenoptic ProgRessVersion ss2.7

5.4

Azevedo et al. [45]b

Italy, 201481 PER19–74 years

Jenoptic ProgRessVersion ss2.7

3.05

Misirlioglu et al. [26]Turkey, 2014


Adobe Photoshop 6.75

Azevedo et al. [45]Brazil, 2015



PER = periapical radiographs. OPG = panoramic radiographs. Tooth numeration FDI (Federation Dentaire International).***Error reported as mean error.

a Error reported as mean absolute error (MAE).b Error reported as the module of the differences between the chronological and estimated age.

Table 6Volume calculation based methods.

Author Sample Image segmentation method Measuring instrument Accuracy

Star et al. [56]2011, Belgic

111 CBCT10–65 years

Automatic segmentation (threshold),and manual correction

Simplant1 11/21 SD = 12.813/23 SD = 13.115/25 SD = 8.4

Tardivo et al. [58]2014, France

210 CT15–85 years

Semiautomatic Mimics1 13/23 MAE = 3.433/43 MAE = 4.6

De Angelis et al. [59]2015, Italy


Manual Osirix software prediction interval = �28 years

Pinchi et al. [60]2015, Italy


Cone shape approximation Osirix1 software SEE = �11.45

CBCT: cone beam computed tomography; SD: standard deviation; MAE: mean absolute error; SEE: standard error of estimation.


Although, the majority of the studies reporting the use of thismethod also mention its creator as part of the authors, which couldgenerate certain bias in the results. It was observed that in thosestudies that did not count with his participation, the results interms of the repeatability of the method and accuracy were similar[26]. In average, Cameriere et al. method reported the lowest error(total average �5.6 years), and a sample of at least 120 individualswas observed as adequate to obtain more accurate age estimates[31].

The combined use of Kvaal et al. width ratios and Camerierearea ratios has been reported on canines from digital panoramicradiographs, finding more accurate age estimates when the pulp/tooth width ratio at level C and pulp/tooth area ratio were usingtogether [31] or when area ratio calculation from canines is usedindividually [26].

In this systematic review we found the following advantagesabout the use of pulpt/tooth area calculation over other methods:area measurement on digital radiographs with different software isfaster, and the most recent study using this method proposes asoftware to do and automatic selection of the borders of the pulpand tooth, which minimizes the required time to obtain the area oftooth and pulp chamber, and reduces the error associated with theobserver, when performing the area selection [75]. Linearmeasurements can also be performed on digital radiographs,but, the observer needs to take nine measurements per tooth, in sixteeth, and calculate several ratios. In the method using pulp/tootharea measurements, only one ratio needs to be calculated andallows the researcher to develop the respective statisticalregression model. In the same way, the use of canines, especiallyupper canines to calculate pulp/tooth area ratio has certainadvantages, as their longer survival, in comparison with otherteeth, less wear, and the big size of the pulp chamber [59]. Also, asobserved in Table 5, there are not only more studies reporting theuse of upper canines to do pulp/tooth area calculation but also,these studies present an acceptable accuracy, in contrast to whatwas reported by Kvaal and Solheim, when doing pulp/tooth length/with ratio calculation who observed that upper canines had thelowest correlation with age, when using dental radiographs ofextracted teeth [77], which was an exclusion criteria in thissystematic review.

Finally, although the use of pulp/tooth volume ratio calculationfrom non-extracted teeth still requires more research. The initialresults of the analysis of pulp volume among individuals ofdifferent age groups, bring important and detailed information tounderstand of the formation of secondary dentine. It has beenreported that the pulp/tooth volume ratios in the cervical areawere more correlated with age, and that this correlation decreasetowards the apex [13]. Also, that the most marked reduction involume ratio was observed between the second and the fifthdecades of life in lower first and second premolars, and betweenthe second and the third decades of life in lower first premolars.[13] This clear findings may enlighten the proposal of future andmore accurate methods for age estimation in adults.

5. Conclusion

The narrowing of root canal caused by the formation ofsecondary dentine is a well-accepted age indicator in adults. Thissystematic review observed that age estimation methods based onpulp/tooth area ratio calculation reported more accurate results,even when one tooth is analysed per individual. However certainconditions need to be fulfilled: a sample of minimum 120 individ-uals, older than 14 years of age, assessment of the accuracy of theobserver, and generation of specific population equation. Theinclusion of data of individuals from different population group inthe same analysis is not discouraged. These studies must consider

in their statistical analysis the non-linear deposition of secondarydentine through life. This systematic review also recommends theuse of dental age estimation methods, firstly pulp/tooth area ratiocalculation of single first, upper canines and other single rootedteeth (lower premolars, upper central incisors) and secondly pulp/tooth length/with ratio calculation, as reported by Kvaal et al., incombination with other methods that include diverse ageindicators to produce a more reliable age estimates. The authorsof this systematic review also recommend future studies to reporttheir results in terms of standard deviation, mean absolute errorand standard error of estimation simultaneously, with this, therewill be more be more homogeneity. Thereby, to perform a propermeta-analysis will be more likely.

References

[1] H. Soomer, H. Ranta, M.J. Lincoln, A. Penttila, E. Leibur, Reliability and validityof eight dental age estimation methods for adults, J. Forensic Sci. 48 (1) (2003)149–152.

[2] S. De Luca, I. Alemán, F. Bertoldi, L. Ferrante, P. Mastrangelo, M. Cingolani, et al.,Age estimation by tooth/pulp ratio in canines by peri-apical X-rays: reliabilityin age determination of Spanish and Italian medieval skeletal remains, J.Archaeol. Sci. 37 (12) (2010) 3048–3058.

[3] K.A. Shahin, L. Chatra, P. Shenai, Dental and craniofacial imaging in forensics,JOFRI 1 (2) (2013) 56–62.

[4] M. Babshet, A.B. Acharya, V.G. Naikmasur, Age estimation in Indians from pulp/tooth area ratio of mandibular canines, Forensic Sci. Int. 197 (1–3) (2010) 125e1–e4.

[5] G. Gustafson, Age determination on teeth, J. Am. Dent. Assoc. 41 (1) (1950) 45–54.

[6] S.I. Kvaal, K.M. Kolltveit, I.O. Thomsen, T. Solheim, Age estimation of adultsfrom dental radiographs, Forensic Sci. Int. 74 (3) (1995) 175–185.

[7] R. Cameriere, L. Ferrante, M. Cingolani, Variations in pulp/tooth area ratio asan indicator of age: a preliminary study, J. Forensic Sci. 49 (2) (2004) 317–319.

[8] F.M. Vandevoort, L. Bergmans, J. Van Cleynenbreugel, D.J. Bielen, P. Lambrechts,M. Wevers, et al., Age calculation using X-ray microfocus computed tomo-graphical scanning of teeth: a pilot study, J. Forensic Sci. 49 (4) (2004) 787.

[9] G. Willems, C. Moulin-Romsee, T. Solheim, Non-destructive dental-agecalculation methods in adults: intra- and inter-observer effects, ForensicSci. Int. 126 (3) (2002) 221–226.

[10] R. Cameriere, L. Ferrante, M.G. Belcastro, B. Bonfiglioli, E. Rastelli, M. Cingolani,Age estimation by pulp/tooth ratio in canines by mesial and vestibular peri-apical X-rays, J. Forensic Sci. 52 (5) (2007) 1151–1155.

[11] F. Yang, R. Jacobs, G. Willems, Dental age estimation through volume matchingof teeth imaged by cone-beam CT, Forensic Sci. Int. 159 (Supplement (1))(2006) S78–S83.

[12] H. Someda, H. Saka, S. Matsunaga, Y. Ide, K. Nakahara, S. Hirata, et al., Ageestimation based on three-dimensional measurement of mandibular centralincisors in Japanese, Forensic Sci. Int. 185 (1–3) (2009) 110–114.

[13] H. Aboshi, T. Takahashi, T. Komuro, Age estimation using microfocus X-raycomputed tomography of lower premolars, Forensic Sci. Int. 200 (1–3) (2010)35–40.

[14] N. Jagannathan, P. Neelakantan, C. Thiruvengadam, P. Ramani, P. Premkumar, A.Natesan, et al., Age estimation in an Indian population using pulp/toothvolume ratio of mandibular canines obtained from cone beam computedtomography, J. Forensic Odontostomatol. 29 (1) (2011) 1–6.

[16] S. Ranjani, L. Ashok, G.P. Sujatha, Age estimation in adults using intra oralperiapical radiographs in Indian population using Kvaal’s method, Medico LegUpdate 10 (2) (2010) 73–77.

[17] S.I. Kvaal, E.M. During, A dental study comparing age estimations of the humanremains from the Swedish warship Vasa, Int. J. Osteoarchaeol. 9 (3) (1999)170–181.

[18] N. Agarwal, P. Ahuja, A. Sinha, A. Singh, Age estimation using maxillary centralincisors: a radiographic study, J. Forensic Dent. Sci. 4 (2) (2012) 97.

[19] G.A. García, Y.M.R. García, L.D.E. Velásquez, Estimación de la edad poraposición de dentina secundaria en una muestra de la población de Bogotáentre 21 y 50 años de edad, Univ. Odontol. 28 (60) (2009) 29–38.

[20] E. Paewinsky, H. Pfeiffer, B. Brinkmann, Quantification of secondary dentineformation from orthopantomograms—a contribution to forensic age estima-tion methods in adults, Int. J. Legal Med. 119 (1) (2005) 27–30.

[21] N. Bosmans, P. Ann, M. Aly, G. Willems, The application of Kvaal’s dental agecalculation technique on panoramic dental radiographs, Forensic Sci Int. 153(2–3) (2005) 208–212.

[22] R. Chandramala, R. Sharma, M. Khan, A. Srivastava, Application of Kvaal’stechnique of age estimation on digital panoramic radiographs, Dentistry 2012(2012).

[23] P.G. Limdiwala, J.S. Shah, Age estimation by using dental radiographs, J.Forensic Dent. Sci. 5 (2) (2013) 118–122.

[24] E. Ayad, H.M. Hamid, E.A. Abdalla, S.A. Kajoak, Estimation of age for Sudaneseadults using orthopantomographs, GJMR 14 (1) (2014).


http://refhub.elsevier.com/S0379-0738(17)30108-1/sbref0010

































































[25] S. Karkhanis, P. Mack, D. Franklin, Age estimation standards for a WesternAustralian population using the dental age estimation technique developed byKvaal et al, Forensic Sci. Int. 235 (2014) 104 e1–e6.

[26] M. Misirlioglu, R. Nalcaci, M.Z. Adisen, S. Yilmaz, S. Yorubulut, Age estimationusing maxillary canine pulp/tooth area ratio, with an application of Kvaal’smethods on digital orthopantomographs in a Turkish sample, Aust. J. ForensicSci. 46 (1) (2014) 27–38.

[27] T. Marroquin, S. Karkhanis, S. Kvaal, S. Vasudavan, E. Castelblanco, E. Kruger,et al., Determining the effectiveness of adult measures of standardised ageestimation on juveniles in a Western Australian population, Aust. J. ForensicSci. (2016) 1–9.

[28] S. Mittal, S.G. Nagendrareddy, M.L. Sharma, P. Agnihotri, S. Chaudhary, M.Dhillon, Age estimation based on Kvaal’s technique using digital panoramicradiographs, J. Forensic Dent. Sci. 8 (2) (2016) 115.

[29] N. Parikh, G. Dave, Application of Kvaal’s dental age estimation technique onorthopantomographs on a population of Gujarat—a short study, BUJOD 8 (3)(2013) 18–24.

[30] H.O. Erbudak, M. Ozbek, S. Uysal, E. Karabulut, Application of Kvaal et al.’s ageestimation method to panoramic radiographs from Turkish individuals,Forensic Sci. Int. 219 (1–3) (2012) 141–146.

[31] S. Saxena, Age estimation of Indian adults from orthopantomographs, Braz.Oral Res. 25 (2011) 225–229.

[32] A. Meinl, S. Tangl, E. Pernicka, C. Fenes, G. Watzek, On the applicability ofsecondary dentin formation to radiological age estimation in young adults, J.Forensic Sci. 52 (2) (2007) 438–441.

[33] M.I. Landa, P.M. Garamendi, M.C. Botella, I. Aleman, Application of the methodof Kvaal et al. to digital orthopantomograms, Int. J. Legal Med. 123 (2) (2009)123–128.

[34] P. Thevissen, D. Galiti, G. Willems, Human dental age estimation combiningthird molar(s) development and tooth morphological age predictors, Int. J.Legal Med. 126 (6) (2012) 883–887.

[35] D. Lorkiewicz-Muszy�nska, A. Przysta�nska, T. Kulczyk, A. Hyrchała, B. Bartecki,W. Kociemba, et al., Application of X-rays to dental age estimation in medico-legal practice, Arch. Med. Sadowej Kryminol. 65 (1) (2015) 1.

[36] P. Kanchan-Talreja, A.B. Acharya, V.G. Naikmasur, An assessment of theversatility of Kvaal’s method of adult dental age estimation in Indians, Arch.Oral Biol. 57 (3) (2012) 277–284.

[37] R. Cameriere, L. Ferrante, M.G. Belcastro, B. Bonfiglioli, E. Rastelli, M. Cingolani,Age estimation by pulp/tooth ratio in canines by peri-apical X-rays, J. ForensicSci. 52 (1) (2007) 166–170.

[38] R. Cameriere, E. Cunha, E. Sassaroli, E. Nuzzolese, L. Ferrante, Age estimation bypulp/tooth area ratio in canines: study of a portuguese sample to testCameriere’s method, Forensic Sci. Int. 193 (1–3) (2009) 128 e1–e6.

[39] S. Singaraju, P. Sharada, Age estimation using pulp/tooth area ratio: a digitalimage analysis, J. Forensic Dent. Sci. 1 (1) (2009) 37–41.

[40] M.B. Jeevan, A.D. Kale, P.V. Angadi, S. Hallikerimath, Age estimation by pulp/tooth area ratio in canines: Cameriere’s method assessed in an Indian sampleusing radiovisiography, Forensic Sci. Int. 204 (1–3) (2011) 209 e1–e5.

[41] S. Saxena, S. Tiwari, A. Bhambal, Variations in morphological variables ofcanine as indicators of age among Indian adults, Indian J. Stomatol. 2 (1) (2011)18–20.

[42] J.F. Zaher, I.A. Fawzy, S.R. Habib, M.M. Ali, Age estimation from pulp/tooth arearatio in maxillary incisors among Egyptians using dental radiographic images,J. Forensic Leg. Med. 18 (2) (2011) 62–65.

[43] R. Cameriere, S. De Luca, I. Alemán, L. Ferrante, M. Cingolani, Age estimation bypulp/tooth ratio in lower premolars by orthopantomography, Forensic Sci. Int.214 (1–3) (2012) 105–112.

[44] C.C. Joseph, B.S. Reddy, N.M. Cherian, S.K. Kannan, G. George, S. Jose, Intraoraldigital radiography for adult age estimation: a reliable technique, JIAOMR 25(4) (2013) 287–290.

[45] A.C. Azevedo, E. Michel-Crosato, M.G.H. Biazevic, I. Gali�c, V. Merelli, S. De Luca,et al., Accuracy and reliability of pulp/tooth area ratio in upper canines by peri-apical X-rays, Leg. Med. 16 (6) (2014) 337–343.

[46] A.d.C.S. Azevedo, N.Z. Alves, E. Michel-Crosato, M. Rocha, R. Cameriere, M.G.H.Biazevic, Dental age estimation in a Brazilian adult population usingCameriere’s method, Braz. Oral Res. (2015) 29.

[47] R. Cameriere, E. Cunha, S.N. Wasterlain, S. De Luca, E. Sassaroli, F. Pagliara, et al.,Age estimation by pulp/tooth ratio in lateral and central incisors by peri-apicalX-ray, J. Forensic Leg. Med. 20 (5) (2013) 530–536.

[48] S.V. Ravindra, G.P. Mamatha, J.D. Sunita, A.Y. Balappanavar, V. Sardana,Morphometric analysis of pulp size in maxillary permanent central incisorscorrelated with age: an indirect digital study, J. Forensic Dent. Sci. 7 (3) (2015)208–214.

[49] S. Sakhdari, S. Mehralizadeh, M. Zolfaghari, M. Madadi, Age estimation frompulp/tooth area ratio using digital panoramic radiography, JIDAI 27 (1) (2015)1.

[50] N.N. Kumar, M.G. Panchaksharappa, R.G. Annigeri, Digitized morphometricanalysis of dental pulp of permanent mandibular second molar for ageestimation of Davangere population, J. Forensic Leg. Med. 39 (2016) 85–90.

[51] M. Babshet, A.B. Acharya, V.G. Naikmasur, Age estimation from pulp/tooth arearatio (PTR) in an Indian sample: a preliminary comparison of three mandibularteeth used alone and in combination, J. Forensic Leg. Med. 18 (8) (2011) 350–354.

[52] S. De Luca, J. Bautista, I. Alemán, R. Cameriere, Age-at-death estimation bypulp/tooth area ratio in canines: study of a 20th-century Mexican sample ofprisoners to test Cameriere’s method, J. Forensic Sci. 56 (5) (2011) 1302–1309.

[53] R. Cameriere, G. Brogi, L. Ferrante, D. Mirtella, C. Vultaggio, M. Cingolani, et al.,Reliability in age determination by pulp/tooth ratio in upper canines inskeletal remains, J. Forensic Sci. 51 (4) (2006) 861–864.

[54] A. Sakuma, H. Saitoh, Y. Suzuki, Y. Makino, G. Inokuchi, M. Hayakawa, et al., Ageestimation based on pulp cavity to tooth volume ratio using postmortemcomputed tomography images, J. Forensic Sci. 58 (6) (2013) 1531–1535.

[55] T. Sasaki, O. Kondo, Human age estimation from lower-canine pulp volumeratio based on Bayes’ theorem with modern Japanese population as priordistribution, Anthropol. Sci. 122 (1) (2014) 23–35.

[56] H. Star, P. Thevissen, R. Jacobs, S. Fieuws, T. Solheim, G. Willems, Human dentalage estimation by calculation of pulp-tooth volume ratios yielded on clinicallyacquired cone beam computed tomography images of monoradicular teeth, J.Forensic Sci. 56 (s1) (2011) S77–S82.

[57] D. Tardivo, J. Sastre, M. Ruquet, L. Thollon, P. Adalian, G. Leonetti, et al., Three-dimensional modeling of the various volumes of canines to determine age andsex: a preliminary study, J. Forensic Sci. 56 (3) (2011) 766–770.

[58] D. Tardivo, J. Sastre, J.-H. Catherine, G. Leonetti, P. Adalian, B. Foti, Agedetermination of adult individuals by three-dimensional modelling of canines,Int. J. Leg. Med. 128 (1) (2014) 161–169.

[59] D. De Angelis, D. Gaudio, N. Guercini, F. Cipriani, D. Gibelli, S. Caputi, et al., Ageestimation from canine volumes, Radiol. Med. 120 (8) (2015) 731–736.

[60] V. Pinchi, F. Pradella, J. Buti, C. Baldinotti, M. Focardi, G.-A. Norelli, A new ageestimation procedure based on the 3D CBCT study of the pulp cavity and hardtissues of the teeth for forensic purposes: a pilot study, J. Forensic Leg. Med. 36(2015) 150–157.

[61] L.V.M.G. Porto, J. Celestino da Silva Neto, A.d. Anjos Pontual, R.Q. Catunda,Evaluation of volumetric changes of teeth in a Brazilian population by usingcone beam computed tomography, J. Forensic Leg. Med. 36 (2015) 4–9.

[62] H. Agematsu, H. Someda, M. Hashimoto, S. Matsunaga, S. Abe, H.-J. Kim, et al.,Three-dimensional observation of decrease in pulp cavity volume using micro-CT: age-related change, Bull. Tokyo Dent. Coll. 51 (1) (2010) 1–6.

[63] Z.-p. Ge, R.-h. Ma, G. Li, J.-z. Zhang, X.-c. Ma, Age estimation based on pulpchamber volume of first molars from cone-beam computed tomographyimages, Forensic Sci. Int. 253 (2015) 133 e1–e7.

[64] Z.-p. Ge, P. Yang, G. Li, J.-z. Zhang, X.-c. Ma, Age estimation based on pulpcavity/chamber volume of 13 types of tooth from cone beam computedtomography images, Int. J. Leg. Med. 130 (4) (2016) 1159–1167.

[65] C. Cattaneo, D. De Angelis, M. Ruspa, D. Gibelli, R. Cameriere, M. Grandi, Howold am I? Age estimation in living adults: a case report, J. ForensicOdontostomatol. 26 (2) (2008) 39–43.

[66] M. Vodanovi�c, J. Duman9ci�c, I. Gali�c, I. Savi�c Pavi9cin, M. Petrove9cki, R. Cameriere,et al., Age estimation in archaeological skeletal remains: evaluation of fournon-destructive age calculation methods, J. Forensic Odontostomatol. 29 (2)(2011) 14.

[67] P.F. Fabbri, S. Viva, L. Ferrante, N. Lonoce, I. Tiberi, R. Cameriere, Radiologicaltooth/pulp ratio in canines and individual age estimation in a sample of adultneolithic skeletons from Italy, Am. J. Phys. Anthropol. 158 (3) (2015) 423–430.

[68] D. De Angelis, D. Gibelli, P. Fabbri, C. Cattaneo, Dental age estimation helpscreate a new identity, Am. J. Forensic Med. Pathol. 36 (3) (2015) 219–220.

[69] H.-M. Jeon, S.-M. Jang, K.-H. Kim, J.-Y. Heo, S.-M. Ok, S.-H. Jeong, et al., Dentalage estimation in adults, JOMP 39 (4) (2014) 119–126.

[70] S. Patil, K. Mohankumar, M. Donoghue, Estimation of age by Kvaal’s techniquein sample Indian population to establish the need for local Indian-basedformulae, J. Forensic Dent. Sci. 6 (3) (2014) 171 (Original Article) (Report).

[71] P.S. Rajpal, V. Krishnamurthy, S.S. Pagare, G.D. Sachdev, Age estimation usingintraoral periapical radiographs, J. Forensic Dent. Sci. 8 (1) (2016) 56.

[72] R. Sharma, A. Srivastava, Radiographic evaluation of dental age of adults usingKvaal’s method, J. Forensic Dent. Sci. 2 (1) (2010) 22–26.

[73] Y.Y. Kostenko, M.Y. Goncharuk-Khomyn, Clinical and experimrntal study forimproving methods of determining the age of adults by dental status,Morphologia 7 (1) (2013) 85–88.

[74] R. Cameriere, L. Ferrante, Canine pulp ratios in estimating pensionable age insubjects with questionable documents of identification, Forensic Sci. Int. 206(1–3) (2011) 132–135.

[75] R. Cameriere, S. De Luca, N. Egidi, M. Bacaloni, P. Maponi, L. Ferrante, et al.,Automatic age estimation in adults by analysis of canine pulp/tooth ratio:preliminary results, JOFRI 3 (1) (2015) 61–66.

[76] Torkian, Age estimation using digital panoramic radiography, IJBPAS 4 (9)(2015) Special Issue: 124–129.

[77] S. Kvaal, T. Solheim, A non-destructive method for age estimation, J. ForensicOdontostomatol. 12 (1994) 6–11.































































































































































219

APPENDIX 2. Changes in age estimation standards secondary to full fixed edgewise

orthodontic treatment.


Tennant, M. (2016). Changes in age estimation standards secondary to full fixed edgewise

orthodontic treatment. European Journal of Forensic Science, Accepted for publication.

June 2016.

Eur J Forensic Sci ● Jan-Mar 2017 ● Vol 4 ● Issue 1 1

European Journal of Forensic Sciences

DOI: 10.5455/ejfs.225119www.ejfs.co.uk

INTRODUCTION

Kvaal et al. [1] developed a non-invasive method for the estimation of age in adults with the measurement of dimensions of the pulp chamber and tooth length to formulate regression models in the estimation of age. Originally, this method was proposed to be used on periapical radiographs. However, more recent work has focused on the use of the Kvaal et al. [1] method in panoramic radiographs [2-4]. During tooth development, secondary dentin deposition commences once the root formation is complete, and the tooth erupts to the level of the occlusal plane [5]. Annually, the rate of secondary dentin formation has been documented to be 6.5 μm/year for the crown and 10 μm/year for the root [6]. The rate of secondary dentin formation is influenced by external stimuli such as occlusal forces, trauma, and caries [7]. These external factors also generate the production of tertiary dentin [5], which is indistinguishable to secondary dentin on dental radiographs [8].

The estimation of dental age has been a valid tool in forensic and anthropological research [1,9,10]. For younger persons,

including newborns, infants, and adolescents, there are dental age estimation methods, generally based on tooth maturation, which are considered the most accurate to estimate the chronological age in subadults [3,11-13]. In adults, once the root formation is deemed complete, methods for dental age estimation are broadly based on the analysis of secondary dentin deposition and the subsequent narrowing of the pulp chamber [1,6,14]. A potential confounding factor in the estimation of dental age is the high incidence of orthodontic treatment in developed countries [15]. The judicious application of forces is central to successful orthodontic treatment and is well known to induce biological changes in dental hard tissues, including the formation of secondary dentin [7] and root shortening [16]. The corollary being a change in the dental morphological features analyzed in the estimation of dental age in adults [7,17].

Given the potential of the development of tooth structure in adolescents and adult patients to be influenced by the application of orthodontic force, and these structures being the foundation of age estimation methods, and it is important

Changes in age estimation standards secondary to full fixed edgewise orthodontic treatmentTalia Yolanda Marroquin1, Shalmira Karkhanis1, Sigrid Kvaal2, Sivabalan Vasudavan1, Edwin Castelblanco1, Estie Kruger1, Marc Tennant1

Original Research

1School of Anatomy, Physiology and Human Biology, The University of Western Australia, Australia, 2Department of Oral Pathology and Section of Forensic Odontology, University of Oslo, Norway

Address for correspondence: Estie Kruger, School of Anatomy, Physiology and Human Biology, The University of Western Australia. E-mail: [email protected]

Received: April 20, 2016

Accepted: June 15, 2016

Published: March 12, 2017

ABSTRACTObjective: The aim of this cohort study was to evaluate if the side effects of orthodontic treatment have any significant impact when the Kvaal et al., method is used for dental age estimation. The objective of this study was to observe the potential effects of orthodontic treatment on the accuracy of the age estimation, as performed using the Kvaal et al. Standards when it was applied on individuals before and after orthodontic treatment. Materials and Methods: Following the methodological approach of Kvaal et al., odontometric measurements were acquired, and the data were statistically analyzed to develop age estimation regression models. The total number of radiographs analyzed was 182 (644%, n = 58) female and 36% (n= 33) male. The ages ranged from 12 to 50 years for females (mean age 22 years) and 12-52 years for males (mean age 22 years) before starting the treatment. The average length of the treatment was 2.2 years for both females and males. Results: It was observed that the standard error of estimate for the regression models did not change dramatically for the pre- and post-treatment data. Conclusions: It is recommended that similar analyses be performed for other methods of dental age estimation, for example, methods based on cone-beam computer tomography or microfocus computer tomography. These novel approaches, though more accurate, are also potentially more sensitive to dental changes caused by orthodontic treatment.

KEY WORDS: Forensic sciences, forensic odontolgy, adults, dental age estimation, orthodontic treatment, root resorption, secondary dentin formation

Marroquin, et al.: Changes in age estimation secondary to orthodontic

2 Eur J Forensic Sci ● Jan-Mar 2017 ● Vol 4 ● Issue 1

to recognize and quantify the effect. The first aim of this study was to evaluate if the side effects of orthodontic treatment had any significant impact on dental age estimation when using the Kvaal et al. [1] method. The objective of this study was to observe the potential effects of orthodontic treatment on the accuracy of the age estimation as performed using the Kvaal et al. [1] standards when applied on individuals before and after orthodontic treatment.


Ethics approval was obtained from the Human Research Ethics Committee of The University of Western Australia (Ref: RA/4/1/6797) before commencement. In the design of the study, data collection and analysis have been performed in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki.

Sample

This cohort study consisted of a series of subjects who consecutively presented for orthodontic treatment at a private specialist orthodontic clinic (SV). All panoramic radiographs were acquired from the same machine and in digital format, with the informed consent of the participants. All of the panoramic radiographs were unidentified, and the only information known was the sex and age of the participants at the starting and finishing time of the orthodontic treatment.

Inclusion Criteria

All subjects who had a recorded panoramic radiograph of high quality with respect to factors of image brightness, contrast, and sharpness were included. In addition, all the analyzed teeth were clinically sound with completed root formation and in functional occlusion.

Exclusion Criteria

Any radiographs that did not meet the inclusion criteria or had observable failings (e.g. image distortion, poor contrast, overlap of tooth structure, or improper positioning) were excluded. Teeth with rotations, incomplete root formation, dilacerations, pulpal pathologies, and endodontic treatment and/or restorations were excluded. Teeth with developmental abnormalities in size, shape, and tooth structure, as well as previous pulpal and periodontal pathologies and endodontic treatment, were excluded. Teeth with large areas of enamel overlap between neighboring teeth were excluded in this study.

Sample Population

This study analyzed a total of 182 pre- and post-orthodontic treatment panoramic radiographs from 91 participants, 64% (n = 58) female and 36% (n = 33) male, from a Western Australian population. All the participants presented for orthodontic treatment that, after diagnosis and treatment planning, was completed with fixed appliances. The ages ranged

from 12 to 50 years for females (mean age 22) and 12-52 years for males (mean age 22) before starting the treatment. The average length of treatment was 2.2 years for both females and males. Digital panoramic radiographs were taken as part of routine clinical orthodontic treatment to evaluate the initial condition of the participants and the final treatment results.

Teeth Selection

As established by Kvaal et al. [1], three single rooted upper teeth (with Federation Dentaire International [FDI] notation 11/21, 12/22, 15/35) and three single rooted lower teeth (FDI 32/42, 33/43, 34/44) were measured. Kvaal et al. [1] reported that there are not significant differences between teeth from the left or the right side of the arch. In accordance with this, if one tooth was absent, the contralateral was used, as long as this was observable in a straight position.

Measurement

The measurements were performed using Image J software (version 1.48 19 April 2014 - National Institute of Health, USA). As proposed by Kvaal et al [1], different odontometric measurements were recorded: tooth length, pulp length and root length as well as pulp width and tooth width at three different levels: A (cemento-enamel junction in the mesial surface of the roots); B (midpoint between the points A and C); C (midpoint between the cemento-enamel junction and root apex). After the collection of the figures in an Excel (2013) spreadsheet, a serial of ratios was calculated: (T) tooth/root length, (R) pulp/tooth length, (P) pulp/root length, and root width/pulp width ratio at levels A, B, and C. The ratio calculation and use were proposed by Kvaal et al. [1] with the aim to reduce the effect of magnification and angulation inherent in most radiographs [1,4]. The different mean values calculated with these ratios provided the age estimation predictors (M: Mean value all ratios, W: Mean value width ratios B and C, L: Mean value ratios P and R, W-L: Difference between W-L). The predictors M and W-L were later used to formulate the age estimation models using multiple regression analysis as established by Kvaal et al. [1]. The measurements were completed by a single observer (TYM). All data were collected using Excel (Version 2013 Microsoft Redmont, USA), and statistical analysis was completed using the program R Core Team version 3.1.3 (2015) (R: A language and environment for statistical computing and R Foundation for Statistical Computing, Vienna, Austria. URL http://www.R-project.org/).

Calibration

Intra- and inter-observer calibration was estimated using five randomly selected panoramic radiographs from the final study sample. These panoramic radiographs were measured on five different days with at least 1 day interval. With the aim to avoid the recall of the measurements and landmarks, five new panoramic radiographs were also measured on each occasion. Data were recorded using the software Image J by TYM and SK. To determine intra-observer precision, three estimates of



precision were calculated: The technical error of measurement (TEM <1.0), relative TEM (rTEM <5%), and coefficient of reliability were calculated to quantify the inter- and intra-observer measurement error [18].


The strength of linear association between chronological age and the Kvaal et al. [1] dental ratios was determined by Pearson correlation coefficient (r). Following this stage, regression analysis was applied to evaluate the relationship between the ratios and the chronological age. The final results were later used to generate the statistical models for age estimation. To quantify the predictive accuracy of the models, the standard error of estimation (SEE) was used in the cross-validation sample. Multivariable linear regression analyses were used to examine the association between the independent and outcome variables. The outcome variables included chronological age (dependent variable) and the predictors M and W-L (independent variables). Different equations were generated to estimate the chronological age, by individual tooth and by group of teeth, as follows: Three maxillary teeth with M and W-L values (6 predictors) individually calculated for each tooth, and the average of M and W-L for the same teeth (2 predictors). Three mandibular teeth with M and W-L values (6 predictors) individually calculated for each tooth, and the average of M and W-L for the same teeth (2 predictors). All six teeth, with M and W-L values (12 predictors) individually calculated for each tooth, and the average of M and W-L for the same teeth (2 predictors). All regression models were built using the ordinary least squares approach. R2 values were computed to examine the amount of variance explained by the predictor variables. The formula proposed by Preoteasa et al. [19] was used to calculate the percentage of teeth with root shortening.

RESULTS

Measurement Precision

After the estimation of intra- and inter-observer error, the obtained values were as follow: Intra-observer error: TEM <1.0 (0.92), rTEM<5% (2.99%), and R > 0.75 (0.95). Inter-observer error (between SK and TYM) was: TEM <1.0 (0.80), rTEM <5 (2.37), and R > 0.75 (0.98) showing comparable intra- and inter-observer measurement precision.

Age Distribution

The age distribution before orthodontic treatment for females was: 12-19 years (62%, n = 36), 20-39 years (28%, n = 16), >40 years (10%, n = 6), and after treatment, it was: 15-19 years (57%, n = 33), 20-39 years (15%, n = 19), and >40 years (10%, n = 6). For males, before orthodontic treatment was: 12-19 years (52%, n = 17), 20-39 years (42%, n = 14), and >40 years (6%, n = 2), and after orthodontic treatment, it was: 13-19 years (48%, n = 16), 20-39 years (45%, n = 15), and >40 years (6%, n = 2).

The correlation coefficient between chronological age and

the dental ratios was calculated for the data before and after orthodontic intervention [Table 1]. In both cases, the correlation coefficient related with the width ratios (A, B, and C) was more significant than those related with length (P, T, R, and L). As established by Kvaal et al. [1], the predictors M and W-L are required to be included in the final equation for dental age estimation. Based on the Pearson correlation coefficients, it could be observed that after the orthodontic treatment, the significance of the values for the predictor W-L decreased.

Individual tooth regression models and multiple teeth regression models were proposed for individuals that had not received orthodontic treatment [Tables 2 and 3, respectively] and for individuals after finishing the treatment [Tables 3 and 4]. In regards to the models for age estimation using individual teeth, the SEE tends to increase after orthodontic treatment, especially for the mandibular canine, where the increase was ±2 years. Furthermore, while the results for the SEE provided by the mandibular canine showed a lower SEE (±8.26 years) in the data analysis before treatment, the same tooth showed the greatest value of SEE (±10.23 years) for individual tooth analysis after orthodontic treatment.

When multiple regression models were formulated using different combinations of teeth, it is also noticeable that in a few

Table 1: Correlation coefficients between chronological age before(1) and after(2) starting orthodontic treatment and the Kvaal et al.[1] dental ratios

Correlation coefficients before treatment

Tooth number (FDI tooth numeration)

Ratio 11/21 12/22 15/25 32/42 33/43 34/44

P1 0.156 0.079 −0.080 0.085 −0.083 0.205P2 −0.261 −0.137 −0.065 −0.026 −0.220 −0.125T1 0.032 0.034 −0.134 −0.186 −0.237 −0.110T2 −0.162 −0.127 −0.148 −0.098 −0.085 −0.159R1 0.163 0.041 0.076 0.321* 0.174 0.383**R2 −0.099 −0.014 0.130 0.070 −0.143 0.037A1 −0.442** −0.419** −0.422** −0.418** −0.325** −0.168A2 −0.306* −0.388** −0.475** −0.172 −0.209 −0.194B1 −0.439** −0.407** 0.026 −0.211* −0.628** −0.595**B2 −0.440** −0.494** −0.449** −0.221* −0.488** −0.522**C1 −0.347** −0.345* −0.322* −0.185 −0.517** −0.459**C2 −0.349** −0.439** −0.364* −0.320* −0.332* −0.436**M1 −0.130 −0.338* −0.283* −0.194 −0.566** −0.378**M2 −0.308* −0.409** −0.412** −0.227* −0.382* −0.443**W1 −0.448** −0.416** −0.142 −0.237* −0.522** −0.573**W2 −0.450** −0.511** −0.469** −0.299* −0.433** −0.522**L1 0.183 0.074 −0.134 0.193 0.019 0.332L2 −0.223* −0.109 −0.148 0.015 −0.218 −0.076W‑L1 −0.365 −0.329 0.027 −0.327* −0.481** −0.574**W‑L2 −0.013 −0.234* −0.058 −0.227* −0.261* −0.412**

*P<0.05, **P<0.001, NS non‑significant. P: Pulp length/root length, T: Tooth length/root length, R: Pulp length/tooth length, A: Pulp width/root width at level A, B: Pulp width/root width at level B, C: Pulp width/root width at level C and predictors, M: Mean value all ratios, W: Mean value width ratios B and C, L: Mean value ratios P and R, W‑L: Difference between W‑L



cases the SEE decreased after orthodontic treatment: Maxillary lateral incisor (±0.437), maxillary first premolar (±0.698), the multiple tooth regression models using three maxillary teeth with two (±0.574) and six (±0.613) predictors, and in the multiple regression models using six teeth and two predictors (±0.237). In both circumstances, the multiple regression models using the different teeth combinations improved the accuracy of estimation of chronological age, consistent with previous studies [2].

Root shortening was detected in 62% of the upper central incisors, 53% of upper lateral incisors, 48% of upper second

premolars, 48% of lower lateral incisors, 50% of lower canine, and 51% of lower second premolars, with a formula previously proposed [19].

DISCUSSION

This is the first study to specifically apply the Kvaal et al. method to a group of orthodontically treated subjects, and also the first evaluating the impact of the side effects of orthodontic treatment on one of the proposed methods for dental age estimation in adults based on the formation of secondary dentine and the pulp/tooth dimension ratios. Previous

Table 2: Multiple regression for estimation of chronological age (in years) for individual maxillary and mandibular teeth in subjects before(1) and after(2) orthodontic treatment

Teeth FDI n R R2 Equation SEE±years

11/21(1) 83 0.229 0.21 Age=39.87−76.79(M)−51.75(W‑L) 8.57411/21(2) 83 0.143 0.122 Age=54.50−73.67(M)−30.64(W‑L) 9.06412/22(1) 81 0.186 0.165 Age=68.21−99.54(M)−33.52(W‑L) 8.92312/22(2) 81 0.255 0.236 Age=83.07−128.42(M)−42.68(W‑L) 8.48615/25(1) 58 0.08 0.046 Age=79.8105−81.5881(M)−0.5444(W‑L) 9.99315/25(2) 58 0.213 0.184 Age=90.22−129.66(M)−18.62(W‑L) 9.29532/42(1) 87 0.124 0.103 Age=15.40−37.51(M)−43.05(W‑L) 9.73132/42(2) 87 0.095 0.74 Age=49.10−77.55(M)−37.65(W‑L) 9.89333/43(1) 63 0.476 0.458 Age=111.80−238.35(M)−132.7(W‑L) 8.26033/43 2) 63 0.196 0.169 Age=76.60−159.75(M)−105.79(W‑L) 10.2634/44(1) 68 0.356 0.336 Age=21.67−69.84(M)−69.59(W‑L) 8.40434/44(2) 68 0.312 0.291 73.52−137.96(M)−64.43(W‑L) 8.716

R2: Coefficient of determination. SEE: Standard error of estimation in years. See Table 1 for abbreviations

Table 3: Multiple regression for estimation of chronological age (in years) for the combination of different teeth before orthodontic treatmentTeeth FDI n R R2 Equation SEE±years

3 max 2 pds 50 0.278 0.247 Age=80.40−145.51(M)−47.69(W‑L) 7.7893 max 6 pds 50 0.314 0.219 Age=63.704−6.412(11/21M)−7.216(11/21W‑L)−32.812(12/22M)

−19.332(12/22W‑L)−77.969(15/25M)+10.594(15/25W‑L)7.932

3 mdb 2 pds 46 0.4 0.372 Age=13.11−36.64(M)−19.83(W‑L) 9.1243md 6 pds 46 0.601 0.54 Age=124.528+81.230(32/42M)−7.784(32/42W‑L)

−328.184(33/43M)−121.021(33/43W‑L) −37.503(34/44M)−9.855(34/44W‑L)7.805

6 teeth 2 pds 36 0.45 0.417 Age=133.86−238.98(M)−67.88(W‑L) 7.6096 teeth 12 pds 36 0.729 0.5889 Age=130.727+(15.058(11/21M)+(23.906(11/21W‑L)+61.249(12/22M)−29.091(12/22W‑L)

−100.481(15/25M)−2.441(15/25W‑L)+54.517(32/42M)+3.089(32/42W‑L)−268.438(33/43M)−89.805(33/43W‑L)−41.009(34/44M)‑25.32(34/44W‑L)

6.392


Table 4: Multiple regression for estimation of chronological age (in years) for the combination of different teeth after orthodontic treatmentTeeth n R R2 Equation SEE±years

3 max 2 pds 50 0.389 0.363 Age=102.46−102.16(M)−63.13(W‑L) 7.2153 max 6 pds 50 0.425 0.344 Age=114.631+(−0.876(11/21M))+(2.218(11/21W‑L))+(−115.748(12/22M))

+(−38.699(12/22W‑L))+(−91.406(15/25M))+(−20.963(15/25W‑L))7.319

3 mdb 2 pds 46 0.337 0.306 Age=98.68−213.10(M)−105.05(W‑L) 9.6323md 6 pds 46 0.447 0.362 Age=42.761(81.109(32/42M))+(1.905(32/42W‑L)+(−111.346(33/43M))

+(−173.252(33/43W‑L))+(−151.769(34/44M))+(−43.162(34/44W‑L))9.232

6 teeth 2 pds 36 0.490 0.460 Age=174.12−288.42(M)−59.22(W‑L) 7.3726 teeth 12 pds 36 0.602 0.395 Age=154.569+9.383(11/21M)+19.718(11/21W‑L)−70.675(12/22M)−39.284(12/22W‑L)

−39.932(15/25M)−4.881(15/25W‑L)+47.114(32/42M)−17.071(32/42W‑L)−58.568(33/43M)−29.183(33/43W‑L)−149.635(34/44M)−0.842(34/44W‑L)

7.803




investigations failed to disclose whether subjects had previously received orthodontic treatment. There are well-known biological changes that occur secondary to the application of orthodontic forces including apical root resorption and secondary dentin formation.

External root resorption is one unavoidable, unpredictable, and undesirable sequela [20-22] of orthodontic treatment. The reported frequency of external root resorption is about 100% when it is examined under microscopy. When it is quantified on periapical or panoramic radiographs, the frequency falls to 70%, with a mean reported value of root shortening of 1.42 ± 0.44 mm and 1% to 5% of the cases with a loss of 4mm or one-third of root length [19]. In addition, maxillary teeth are more sensitive than mandibular teeth to root resorption. The most frequently affected teeth, according to severity, are the maxillary laterals, maxillary centrals, mandibular incisors, the distal root of mandibular first molars, second premolars, and maxillary second premolars [21,23], and all these teeth have been used in different methods for dental age estimation in adults [1,2,24]. In this study, a higher percentage of root resorption for upper teeth was also observed.

Confirming the findings of Karkhanis et al. [2], the length ratios (P, T, R, and L ratios) in this study have a non-significant correlation coefficient with age (P > 0.05), and the percentage of negative correlation coefficients [Table 1], closer to −1, is higher in the observed values after orthodontic treatment (23% before vs. 66% after). As the Kvaal et al. method is based on the negative correlation between age and the mentioned ratios, it is possible to observe that after the orthodontic treatment this negative correlation is more evident but not statistically significant. In this study, there was no significant variation in the correlation coefficients of the calculated length ratios with age among the different teeth. Furthermore, as the Kvaal et al. method is based on length ratio calculations, not the linear measurements per se, it is possible that the effect of root shortening on the estimated ages had been diminished when the ratios are calculated.

In terms of secondary dentin formation owed to orthodontic treatment (a phenomenon that has not been as widely investigated as root resorption), there are reports of complete obliteration of the canals in maxillary incisors [25], and in one study analyzing, the tooth changes in terms of root length and pulp chamber within maxillary central incisors (tooth 11 and 21), statistically significant changes in the width of the pulp chamber at the midpoint of the dental root [26] were found, equivalent to the point C in the Kvaal et al. method. [1] Another study using cone-beam computer tomography (CBCT) in maxillary teeth found a statistically significant difference between the volumes before and after orthodontic treatment, with the highest mean volume loss observed in the upper left lateral incisor (3.86 mm3) and the least for the upper right central incisor (3.04 mm3) [7].

In this study, the correlation coefficients of the pulp/root width ratios (A, B, C, and W) maintained their negative correlation with age before and after the treatment, without showing the same variation observed for the length ratios. It would still

be necessary to assess if the root shortening after orthodontic treatment, displaced the location of the reference points B and C toward the crown (where the pulp chamber is wider), and its relation with the observed variation of the SEE before and after orthodontic treatment, which in both cases is acceptable in forensic terms (SEE =±10 years).

CONCLUSION

Our investigation demonstrated an alteration in tooth and pulp chamber morphology and dimension following orthodontic treatment. However, these changes did not serve to significantly influence the validity and accuracy (SEE) of the Kvaal et al. [1] method when applied in the assessment of panoramic radiographs to estimate chronological age. We recommend further analysis of other methods for dental age estimation in adults based on the formation of secondary dentin, such as Cameriere et al. [3], or the more recently proposed methods based on volumetric reconstructions of the tooth and pulp chamber in CBCT [27,28] which could be more susceptible to be affected by orthodontic treatment sequels.

REFERENCES

1. Kvaal SI, Kolltveit KM, Thomsen IO, Solheim T. Age estimation of adults from dental radiographs. Forensic Sci Int 1995;74:175-85.

2. Karkhanis S, Mack P, Franklin D. Age estimation standards for a Western Australian population using the dental age estimation technique developed by Kvaal et al. Forensic Sci Int 2014;235:104.e1-6.

3. Panchbhai AS. Dental radiographic indicators, a key to age estimation. Dentomaxillofac Radiol 2011;40:199-212.

4. Kanchan-Talreja P, Acharya AB, Naikmasur VG. An assessment of the versatility of Kvaal’s method of adult dental age estimation in Indians. Arch Oral Biol 2012;57:277-84.

5. Hargreaves KM. Cohen’s Pathways of the Pulp Expert Consult. St. Louis: Mosby; 2010.

6. Cameriere R, De Luca S, Alemán I, Ferrante L, Cingolani M. Age estimation by pulp/tooth ratio in lower premolars by orthopantomography. Forensic Sci Int 2012;214:105-12.

7. Venkatesh S, Ajmera S, Ganeshkar SV. Volumetric pulp changes after orthodontic treatment determined by cone-beam computed tomography. J Endod 2014;40:1758-63.

8. Nanci A. Ten Cate’s Oral Histology-Pageburst on VitalSource: Development, Structure, and Function. St. Louis, MO: Elsevier Health Sciences; 2007.

9. Sarkar S, Kailasam S, Mahesh Kumar P. Accuracy of estimation of dental age in comparison with chronological age in Indian population – a comparative analysis of two formulas. J Forensic Leg Med 2013;20:230-3.

10. Senn DR, Weems RA, editors. Manual of Forensic Odontology. 5th ed. Boca Raton, FL: CRC Press; 2013.

11. Yusof MY, Thevissen PW, Fieuws S, Willems G. Dental age estimation in Malay children based on all permanent teeth types. Int J Legal Med 2014;128:329-33.

12. Ebrahim E, Rao RK, Chatra L, Shenai P, Veena KM, Prabhu RV, et al. Dental age estimation using schour and Massler method in South Indian Children. Sch J Appl Med Sci 2014;2:1669-74.

13. Demirjian A, Goldstein H, Tanner JM. A new system of dental age assessment. Hum Biol 1973;45:211-27.

14. Gustafson G. Age determination on teeth. J Am Dent Assoc 1950;41:45-54.

15. Pretty IA, Sweet D. A look at forensic dentistry – Part 1: The role of teeth in the determination of human identity. Br Dent J 2001;190:359-66.

16. Baumrind S, Korn EL, Boyd RL. Apical root resorption in orthodontically treated adults. Am J Orthod Dentofacial Orthop 1996;110:311-20.

17. Ramazanzadeh BA, Sahhafian AA, Mohtasham N, Hassanzadeh N, Jahanbin A, Shakeri MT. Histological changes in human dental pulp



following application of intrusive and extrusive orthodontic forces. J Oral Sci 2009;51:109-15.

18. Weinberg SM, Scott NM, Neiswanger K, Marazita ML. Intraobserver error associated with measurements of the hand. Am J Hum Biol 2005;17:368-71.

19. Preoteasa CT, Ionescu E, Preoteasa E, Machado JA, Baleanu MC, Baleanu D. Multidimensional scaling for orthodontic root resorption. Math Probl Eng 2013;2013:1-6.

20. Ponder SN, Benavides E, Kapila S, Hatch NE. Quantification of external root resorption by low - Vs high-resolution cone-beam computed tomography and periapical radiography: A volumetric and linear analysis. Am J Orthod Dentofacial Orthop 2013;143:77-91.

21. Campos MJ, Silva KS, Gravina MA, Fraga MR, Vitral RW. Apical root resorption: The dark side of the root. Am J Orthod Dentofacial Orthop 2013;143:492-8.

22. Harris DA, Jones AS, Darendeliler MA. Physical properties of root cementum: Part 8. Volumetric analysis of root resorption craters after application of controlled intrusive light and heavy orthodontic forces: A microcomputed tomography scan study. Am J Orthod Dentofacial Orthop 2006;130:639-47.

23. Brezniak N, Wasserstein A. Root resorption after orthodontic treatment: Part 2. Literature review. Am J Orthod Dentofacial Orthop 1993;103:138-46.

24. Mathew DG, Rajesh S, Koshi E, Priya LE, Nair AS, Mohan A. Adult

forensic age estimation using mandibular first molar radiographs: A novel technique. J Forensic Dent Sci 2013;5:56-9.

25. Delivanis HP, Sauer GJ. Incidence of canal calcification in the orthodontic patient. Am J Orthod 1982;82:58-61.

26. Tobón D, Aristizabal D, Álvarez C, Urrea J. Cambios radiculares en pacientes tratados ortodoncicamente (Spanish). Root changes in patients treated orthodontically (English). CES Odontol 2014;27:37-46.

27. Ge ZP, Ma RH, Li G, Zhang JZ, Ma XC. Age estimation based on pulp chamber volume of first molars from cone-beam computed tomography images. Forensic Sci Int 2015;253:133.e1-7.

28. Sakuma A, Saitoh H, Suzuki Y, Makino Y, Inokuchi G, Hayakawa M, et al. Age estimation based on pulp cavity to tooth volume ratio using postmortem computed tomography images. J Forensic Sci 2013;58:1531-5.

Source of Support: Nil, Conflict of Interest: None declared.

226

APPENDIX 3. Orthodontic treatment: Real risk for dental age estimation in adults?


Tennant, M. (2016). Orthodontic Treatment: Real Risk for Dental Age Estimation in Adults?

Journal of Forensic Science. Accepted for publication. October 2016.

PAPER

ODONTOLOGY

Talia Yolanda Marroquin Penaloza,1 D.M.D.; Shalmira Karkhanis,1 Ph.D.; Sigrid Ingeborg Kvaal,2

Ph.D.; Sivabalan Vasudavan,1 D.M.D.; Edwin Castelblanco,1 BEcon.; Estie Kruger,1 D.M.D., Ph.D.; andMarc Tennant,1 D.M.D., Ph.D.

Orthodontic Treatment: Real Risk for DentalAge Estimation in Adults?

ABSTRACT: Dental age estimation becomes a challenge once the root formation is concluded. In living adults, one dental age indicator isthe formation of secondary dentine, also associated with orthodontic treatment as well as root shortening. The aim of this study was to establishwhether these secondary effects of orthodontic treatment could generate a statistically significant difference in dental age estimations whenusing Kvaal’s method. The study sample included 34 pairs of pre- and postorthodontic panoramic radiographs, from different individuals withexactly the same age and sex distribution. Females 65%, median age 17.5 years, and males 35%, median age 22.5 years, were included. Afterdata collection, dental age was estimated per tooth using formulae previously published. The risk of obtaining over-estimation of age was calcu-lated. (RR = 1.007). The changes caused by orthodontic treatment do not have any significant effect on age estimation when Kvaal et al.’smethod is applied on panoramic radiographs.

KEYWORDS: forensic science, dental age estimation, secondary dentin formation, adults, orthodontic treatment, root resorption

The analysis of teeth for age estimation has been scientificallyreported since the early 1800’s (1). Methods based on tooth for-mation in juveniles have shown high reliability (2–4). Once theroot formation has finished (generally at the age of 14 years,excluding the third molar) (5), dental age estimation becomes achallenge, partly as a result greater variability in the develop-ment of third molar (6). The most reported noninvasive methodsfor dental age estimation are based on the formation of sec-ondary dentine and the decrease in pulp chamber dimensions.These features are measurable in periapical radiographs (7,8),panoramic radiographs (9,10), micro-focus-computed tomographs(11), computed tomographs (12), and cone beam-computedtomographs (13). These methods proposed different formulae tobe used in specific populations. An important characteristic ofthese studies is that in their analysis, they only included totallysound teeth.It is well known that orthodontic forces generate irreversible

changes on tooth structure, such as root shortening (14) and sec-ondary dentine formation (15). These two biological changes intooth structure may directly affect the features used for dentalage estimation in adults, especially the method proposed byKvaal et al (7). This method is based on the measurement oftooth/pulp length, root canal, and root width at different levels,followed by ratio calculations and linear regression analysis fordental age estimation. It would be expected that with the

mentioned changes, secondary to orthodontic treatment, age esti-mation in participants’ postorthodontic treatment would show ahigher estimation error and higher over-estimation, compared tothat of nontreated participants. If that were to be the case, itwould be necessary to develop specific standards for orthodonti-cally treated participants, not only for the Kvaal et al.’s method(7), but for any method based on secondary dentine formationand the variation of pulp/tooth dimensions with age. In the eventof the results being different to the expected, it would mean thatthe Kvaal’s method could be used despite the evidence of ana-tomic changes related to orthodontic treatment. The aim of thisstudy was to establish whether orthodontic treatment would gen-erate changes in dental age estimations at the tooth level whenthe Kvaal et al.’s (7) method is applied per individual tooth.

Materials and Methods

Ethics approval was obtained from the Human ResearchEthics Committee of The University of Western Australia(Ref: RA/4/1/6797) prior to commencement.

Panoramic Radiographs, Orthodontics, and Dental AgeEstimation

Initially, the Kvaal et al.’s (7) method was proposed to beused on periapical radiographs. However, recent studies havealso applied this method on panoramic radiographs (9,10).Panoramic radiographs have more image distortion than periapi-cal radiographs, but it has been reported that when root resorp-tion associated with orthodontic treatment is measured onpanoramic radiographs, it is significantly higher than when mea-sured on periapical radiographs (16). This study presents the first

1School of Anatomy Physiology and Human Biology, University ofWestern Australia, 35 Stirling Hwy, Crawley, WA 6009, Australia.

2Faculty of Dentistry, Institute of Clinical Dentistry, University of Oslo,P.O. Box 1109, Blindern, N-0317, Oslo, Norway.

Received 4 May 2016; and in revised form 14 Aug. 2016; accepted 18Oct. 2016.

1© 2017 American Academy of Forensic Sciences

J Forensic Sci, 2017doi: 10.1111/1556-4029.13371

Available online at: onlinelibrary.wiley.com

http://orcid.org/0000-0001-8405-1822

http://orcid.org/0000-0001-8405-1822

http://orcid.org/0000-0001-8405-1822

study to see the effect of orthodontic-related changes on age esti-mation, and therefore, no power calculation was achievable.However, based on the size of previous research assessing pulpalchanges associated with orthodontic treatment (17), a samplesize in a similar range was deemed appropriate.

Sample Selection

The study cohort consisted of a series of participants who con-secutively presented for orthodontic treatment at a private special-ist orthodontic clinic (SV). The initial sample for this study was 91participants from a Western Australia population, who had a pre-and post-treatment panoramic radiograph. From the initial sample,a total of 34 pre- and 34 postorthodontic treatment panoramicradiographs were selected, resulting in a final sample of 34 pairsof panoramic radiographs (n = 68). Sample size reduction corre-sponded to the need of age and sex matching between the differentindividuals in each pair, one of them without orthodontic treatmentand the other one with a concluded treatment. Female 65%(n = 22, for both groups), age range 15–50 years old, median17.5. Male 35% (n = 12, for both groups) age range 16–37 yearsold, median 22.5. The minimum length treatment was 1.2 years,the maximum 3.6 years, and median 2.1 years.Intra-observer calibration was performed to test the repeatabil-

ity and reliability of the main observer (TM). All the measure-ments were completed by a single observer (TM), and theanalysis was completed with Image J software (version 1.48 19April 2014—National Institute of Health, Bethesda, Maryland).All data were collated using Excel (version 2013 Microsoft,Redmond, WA), and statistical analysis was completed using theprogram R Core Team version 3.1.3 (2015) (R: A language andenvironment for statistical computing. R Foundation for Statisti-cal Computing, Vienna, Austria. URL http://www.R-project.org/).

Inclusion Criteria

All participants who had a recorded panoramic radiograph ofhigh quality with respect to factors of image brightness, contrast,and sharpness were included. In addition, all teeth included inthe analysis were clinically sound with completed root formationand in functional occlusion.

Exclusion Criteria

Radiographs with observable failings (image distortion, poorcontrast, superposition of tooth structure, or improper position-ing) were excluded. Teeth with pulpal or periodontal pathologiesand endodontic treatment and/or restorations, incomplete rootformation, dilacerations, or rotations were excluded. Teeth withdevelopmental abnormalities in size, shape and tooth structure,or large areas of enamel overlap between neighboring teeth werealso excluded in this study.

Teeth Analyzed

The purpose of this study was to assess the biologic variationeffects of orthodontic treatment at the tooth level. Therefore, thetooth-by-tooth formulae for the Kvaal method were applied(10,18).Following the parameters as provided by Kvaal et al. (7),

three upper and three lower single-rooted teeth were included:one maxillary central incisor (with Federation Dentaire

Internationale (FDI) notation 11 or 21), one maxillary lateralincisors (FDI notation 12 or 22), and one maxillary second pre-molar (FDI notation 15 or 25), mandibular lateral incisors (FDInotation 32 or 42), one mandibular canine tooth (FDI notation33 or 43), and one first premolar teeth (FDI notation 34 or 44).In accordance with the study by Karkhanis et al. (10), panoramicradiographs that presented any combination of the required teethwere included. There is no evidence claiming that the accuracyof the results had varied depending on the use of left or rightteeth. Consequently, in the absence of one of the teeth the con-tralateral was used, as long as they met the inclusion criteria.

Measurements

All the panoramic radiographs were obtained in digital format,and the measurements were performed using the software ImageJ software (developed by the National Institute of Health, USA).According to Kvaal et al. (7), the following measurements wererecorded: tooth, pulp, and root length, as well as the pulp andtooth width at three different levels: A (cemento–enamel junctionon the mesial surface of the roots), B (midpoint between thepoints A and C), and C (midpoint between the cemento–enameljunction and root apex). All the measurements were recordedbefore the sample was categorized into two groups: group prefor panoramic radiographs obtained from nontreated participantsand group post for treated participants, to avoid observation bias.Following the recording of measurements, a series of length andwidth ratios were calculated: tooth/root length; pulp/tooth length,and root width/ pulp width ratio at levels A, B, and C. Theseratio calculations were proposed by Kvaal et al. (7) as it reducesthe effect of magnification and angulation inherent in mostradiographs (7,19). The different mean values from the ratioswere used to obtain age estimation predictors. (M: mean valueall five ratios, W: mean value of width ratios at level B and C,L: mean value of length ratios P and R, W-L: differencebetween W-L) (7). The predictors M and W-L were later used toestimate the age of all the participants.Five randomly selected panoramic radiographs from the final

study sample were used to estimate intra-observer calibration.These panoramic radiographs were measured on five differentdays with at least one-day interval. With the aim to avoid therecall of the measurements and landmarks, five new panoramicradiographs were also measured on each occasion. Three esti-mates of precision were calculated to quantify intra-observermeasurement error and precision: the technical error of measure-ment (TEM), relative technical error of measurement (rTEM),and coefficient of reliability (20).


After the completion of the measurements, dental age estima-tions were calculated using the age estimation equations fromindividual teeth. In participants aged 30 years or older, the equa-tions used were those reported by Karkhanis et al. (10) And forparticipants 29 years or younger, the used equations wereobtained from a different group of nontreated participants(n = 74 aged 12–28 years). Both sets of equations were obtainedfrom a Western Australian population.

Results

For the estimation of intra-observer error, the obtained valueswere within acceptable standards for all the measurements. The

2 JOURNAL OF FORENSIC SCIENCES

http://www.R-project.org/

http://www.R-project.org/

intra-observer results were as follows: TEM < 1.0 (0.92),rTEM < 5% (2.99%), and R > 0.75 (0.95).After age estimation per individual tooth, there were a total of

189 age estimates for group pre and 196 estimates for grouppost. The age estimates obtained were then compared betweenparticipants with exactly the same age and sex in group pre andgroup post. In this way, 183 pairs of age estimates wereobtained and compared 1 to 1.For 79% (n = 145) of the total observations, the fluctuation of

the data was the same, regardless whether there was over- orunder-estimation of the age. In terms of over-estimation, in 47%of these observations, the over-estimation was higher after theorthodontic treatment.The fluctuation of the data with regard to the chronological

age of the individual was analyzed per individual tooth(Table 1), showing that there was no significant difference in thepercentages of over- and under-estimation between both groups(Pearson’s correlation coefficient r = 0.78).Although the fluctuation of the data was the same in both

groups, pre and post, there was a clear difference between theages. It was observed that before the age of 25, the large major-ity of results showed a slight over-estimation of age which didnot exceed the reported SEE. In contrast, the majority of ageestimation data obtained from older participants showed under-estimation of the age that notably exceeded the reported SEE.A contingency table (Table 2) was developed to test whether

the secondary effects of orthodontic treatment were a real risk toproduce over-estimation of the age when the Kvaal et al.’smethod is applied. The relative risk calculated was RR = 1.0071(low risk)

Discussion

The reliability of dental age estimation in adults, based on theformation of secondary dentine, has shown superior accuracy toother methods, based on the analysis of other age-related dentalor osseous changes (8,10). However, teeth are subjected to

different changes through life related to pathology as well asbiological, chemical, or mechanical trauma or dental procedures.It has been reported that orthodontic treatment causes mechan-

ical trauma to the periodontal ligament and induces pulpal reac-tions (21). The most frequently reported side effect is externalroot resorption, a process characterized by the destruction of rootstructure (21), with the subsequent diminution of root length(mean reported value of 1.42 � 0.44 mm) (22). Another sideeffect is the reduction in pulp chamber dimensions, owing tosecondary dentine formation (15). With these changes, it wasexpected to obtain significant differences between group pre andpost, with higher percentages of over-estimates of age in treatedparticipants. The analysis of the obtained data facilitated the cal-culation of the potential risk of having over-estimation in partici-pants after finishing the orthodontic treatment (Table 1). Aftercalculating the incidence of over-estimation in groups pre andpost, there was no evidence of association between over-estima-tion of age and orthodontic treatment (RR = 1.0071). However,it is necessary to mention that in this study none of the testedteeth showed signs of severe apical root resorption.According to previous studies, it was found that maxillary

teeth are more affected by root shortening due to orthodontictreatment, specially lateral and central incisors (23,24). In thisstudy, they also presented the higher percentage of under-esti-mates of age for both groups, followed by the mandibular lateralincisor. In the case of under-estimation, it was observed in ahigher percentage in lower canines for group pre and grouppost, 54% and 61%, respectively.Previous studies using the Kvaal et al.’s method and her pro-

posed formulae reported a constant under-estimation of age,from 18 to 20 years (19) up to 47.10 years (25). These studiesdid not use population-specific formulae. In this study, the for-mulae used were obtained from the same Western Australianpopulation. There was a clear-cut line (24 years for both groups)where the estimates would notably exceed the reported SEE (1–5 years), having statistically significant under-estimates. It isnecessary to determine in future studies whether this findingcould be related to the fact that the apposition of secondary den-tine does not occur in a linear manner through life (25), causinga larger decrease in pulp dimensions between 20 and 40 years ofage, than between 40 and 60 years of age (26). A limiting factorto clearly examine this relation in the current study was the lackof individual over 40 years of age.As the main objective of this study was to establish whether

orthodontic treatment would generate changes in dental age esti-mation, and as each type of tooth is affected by different degreesof severity, in this study we used previously published formulaefor dental age estimation for individual teeth rather than per setof teeth or per individual. The use of Kvaal et al.’s (7) individ-ual teeth formulae to estimate dental age has been reported onextracted teeth (27). In the same way, there are other methodsbased on the assessment of a single tooth per individual to

TABLE 1––Comparison of over- and under-estimates obtained per tooth before and after orthodontic treatment. And reported standard estimation error SEEper individual tooth, for participants younger than 30 years (SEE � years (<30 years)) and over 30 years of age (SEE � years (>30 years)).

Tooth 11/21 12/22 15/25 32/42 33/43 34/44

Over-estimation Pretreatment 39% (n = 13) 50% (n = 16) 52% (n = 17) 47% (n = 16) 54% (n = 13) 47% (n = 15)Post-treatment 38% (n = 13) 45% (n = 15) 48% (n = 16) 44% (n = 15) 61% (n = 20) 52% (n = 15)

Under-Estimation Pretreatment 61% (n = 20) 50% (n = 16) 48% (n = 16) 53% (n = 18) 46% (n = 11) 53% (n = 17)Post-treatment 62% (n = 21) 55% (n = 18) 52% (n = 17) 56% (n = 19) 39% (n = 13) 48% (n = 14)

SEE � years (<30 years) 4.344 4.258 4.536 4.465 3.708 3.709SEE � years (>30 years) 9.367 9.648 9.525 10.222 10.903 10.534

TABLE 2––Contingency table to estimate whether orthodontic treatment wasa causal of over-estimation, when the Kvaal et al.’s method is used to age

estimation.

Orthodontic treatment Over-estimation Under-estimation Total

YESPost-treatment

48%n = 94

52%n = 102

n = 196

NOPretreatment

48%n = 184

52%n = 99

n = 189

Total n = 184 n = 201 n = 385

Incidence over-estimation group A = 47.9%. Incidence under-estimationgroup B = 47.6%. Relative risk of presenting over-estimation owed toorthodontic treatment secondary effects when Kvaal et al.’s method is usedfor dental age estimation: RR = 1.0071 (No or low risk).

MARROQUIN PENALOZA ET AL. . EFFECT OF ORTHODONTIC TREATMENT AGE ESTIMATION 3

generate dental age estimation models with acceptable results ina forensic framework (5,7,13).

Conclusion

In our study, orthodontic treatment did not affect the finalresults when the Kvaal et al.’s method was used for dental ageestimation. Although it has been previously established that, touse the pulp complex as a biomarker for general aging, the ana-lyzed teeth have to fulfill the requirements of being in normaland functional occlusion, totally sound and free from dental pro-cedures (7), the real effect of this conditions on methods basedon the formation of secondary dentine, for dental age estimation,has not been tested. The results of our study allow forensic den-tists to use the Kvaal et al.’s method in participants who havehad previous orthodontic treatment, when there are no signs ofsevere apical root resorption.

References

1. Saunders E. The teeth a test of age. Lancet 1838;30(774):492–6.2. Ebrahim E, Prasanna KR, Laxmikanth C, Prashanth S, Veena K, Rachanna

V, et al. Dental age estimation using Schour and Massler method in SouthIndian children. Sch J App Med Sci 2014;2(5C):1669–74.

3. Demirjian A, Goldstein H, Tanner JM. A new system of dental ageassessment. Hum Biol 1973;45(2):211–17.

4. Panchbhai AS. Dental radiographic indicators, a key to age estimation.Dentomaxillofac Radiol 2011;40(4):199–212.

5. Tardivo D, Sastre J, Ruquet M, Thollon L, Adalian P, Leonetti G, et al.Three-dimensional modeling of the various volumes of canines to deter-mine age and sex: a preliminary study. J Forensic Sci 2011;56(3):766–70.

6. Celikoglu M, Kamak H. Patterns of third-molar agenesis in an orthodon-tic patient population with different skeletal malocclusions. Angle Orthod2012;82(1):165–9.

7. Kvaal SI, Kolltveit KM, Thomsen IO, Solheim T. Age estimation ofadults from dental radiographs. Forensic Sci Int 1995;74(3):175–85.

8. Cameriere R, Ferrante L, Cingolani M. Variations in pulp/tooth area ratioas an indicator of age: a preliminary study. J Forensic Sci 2004;49(2):317–19.

9. Paewinsky E, Pfeiffer H, Brinkmann B. Quantification of secondary den-tine formation from orthopantomograms – a contribution to forensic ageestimation methods in adults. Int J Legal Med 2005;119(1):27–30.

10. Karkhanis S, Mack P, Franklin D. Age estimation standards for a Wes-tern Australian population using the dental age estimation techniquedeveloped by Kvaal et al. Forensic Sci Int 2014;235:104.e1–e6.

11. Vandevoort FM, Bergmans L, Van Cleynenbreugel J, Bielen DJ, Lam-brechts P, Wevers M, et al. Age calculation using x-ray microfocus com-puted tomographical scanning of teeth: a pilot study. J Forensic Sci2004;49(4):787–90.

12. Sakuma A, Saitoh H, Suzuki Y, Makino Y, Inokuchi G, Hayakawa M,et al. Age estimation based on pulp cavity to tooth volume ratio using

postmortem computed tomography images. J Forensic Sci 2013;58(6):1531–5.

13. Ge Z-p, Ma R-h, Li G, Zhang J-z, Ma X-c. Age estimation based onpulp chamber volume of first molars from cone-beam computed tomog-raphy images. Forensic Sci Int 2015;253:133.e1–e7.

14. Baumrind S, Korn EL, Boyd RL. Apical root resorption in orthodonti-cally treated adults. Am J Orthod Dentofacial Orthop 1996;110(3):311–20.

15. Venkatesh S, Ajmera S, Ganeshkar SV. Volumetric pulp changes afterorthodontic treatment determined by cone-beam computed tomography. JEndod 2014;40(11):1758–63.

16. Sameshima G, Asgarifar K. Assessment of root resorption and rootshape: periapical vs panoramic films. Angle Orthod 2001;71(3):185–9.

17. Lazzaretti DN, Bortoluzzi GS, Torres Fernandes LF, Rodriguez R, GrehsRA, Martins Hartmann MS. Histologic evaluation of human pulp tissueafter orthodontic intrusion. J Endod 2014;40(10):1537–40.

18. Marroquin TY, Karkhanis S, Kvaal SI, Vasudavan S, Castelblanco E,Kruger E, et al. Determining the effectiveness of adult measures of stan-dardised age estimation on juveniles in a Western Australian population.Aust J Forensic 2017;49(4):459–467.

19. Kanchan-Talreja P, Acharya AB, Naikmasur VG. An assessment of theversatility of Kvaal’s method of adult dental age estimation in Indians.Arch Oral Biol 2012;57(3):277–84.

20. Weinberg SM, Scott NM, Neiswanger K, Marazita ML. Intraobservererror associated with measurements of the hand. Am J Hum Biol2005;17(3):368–71.

21. Krishnan V, Davidovitch Z. Cellular, molecular, and tissue-level reac-tions to orthodontic force. Am J Orthod Dentofacial Orthop 2006;129(4):469.e1–e32.

22. Preoteasa CT, Ionescu E, Preoteasa E, Tenreiro MJ, Baleanu MC,Baleanu D. Multidimensional scaling for orthodontic root resorption.Math Prob Eng 2013;2013(201):Article ID 383698.

23. Brezniak N, Wasserstein A. Root resorption after orthodontic treatment:part 2. Literature review. Am J Orthod Dentofacial Orthop 1993;103(2):138–46.

24. da Silva Campos MJ, Silva KS, Gravina MA, Fraga MR, Vitral RWF.Apical root resorption: the dark side of the root. Am J Orthod Dentofa-cial Orthop 2013;143(4):492–8.

25. Meinl A, Tangl S, Pernicka E, Fenes C, Watzek G. On the applicabilityof secondary dentin formation to radiological age estimation in youngadults. J Forensic Sci 2007;52(2):438–41.

26. Oi T, Saka H, Ide Y. Three-dimensional observation of pulp cavities inthe maxillary first premolar tooth using micro-CT. Int Endod J 2004;37(1):46–51.

27. Willems G, Moulin-Romsee C, Solheim T. Non-destructive dental-agecalculation methods in adults:intra-and inter-observer effects. ForensicSci Int 2002;126:221–6.

Additional information and reprint requests:Estie Kruger, D.M.D., Ph.D.International Research Collaborative – Oral Health and EquityThe University of Western AustraliaNedlands, 6009 WAAustraliaE-mail: [email protected]

4 JOURNAL OF FORENSIC SCIENCES

231

APPENDIX 4. Determining the effectiveness of adult measures of standardized age estimation

on juveniles in a Western Australian population.


Tennant, M. (2016). Determining the effectiveness of adult measures of standardised age

estimation on juveniles in a Western Australian population. Australian Journal of Forensic

Sciences. Accepted for publication. March 2016.

Full Terms & Conditions of access and use can be found athttp://www.tandfonline.com/action/journalInformation?journalCode=tajf20

Download by: [1.132.96.27] Date: 23 May 2016, At: 08:36

Australian Journal of Forensic Sciences

ISSN: 0045-0618 (Print) 1834-562X (Online) Journal homepage: http://www.tandfonline.com/loi/tajf20

Determining the effectiveness of adult measuresof standardised age estimation on juveniles in aWestern Australian population

T.Y. Marroquin, S. Karkhanis, S.I. Kvaal, S. Vasudavan, E. Castelblanco, E.Kruger & M. Tennant

To cite this article: T.Y. Marroquin, S. Karkhanis, S.I. Kvaal, S. Vasudavan, E. Castelblanco, E.Kruger & M. Tennant (2016): Determining the effectiveness of adult measures of standardisedage estimation on juveniles in a Western Australian population, Australian Journal of ForensicSciences

To link to this article: http://dx.doi.org/10.1080/00450618.2016.1177593

Published online: 22 May 2016.

Submit your article to this journal

View related articles

View Crossmark data

Determining the effectiveness of adult measures of standardised ageestimation on juveniles in a Western Australian population

T.Y. Marroquina* , S. Karkhanisa, S.I. Kvaalb, S. Vasudavana,c, E. Castelblancoa,E. Krugera and M. Tennanta

aSchool of Anatomy Physiology and Human Biology, University of Western Australia, CrawleyWA, Australia; bDepartment of Oral Pathology and Section of Forensic Odontology, Universityof Oslo, Oslo, Norway; cDepartment of Developmental Biology, Harvard School of Dental

Medicine, Boston, MA, USA

(Received 20 December 2015; accepted 15 March 2016)

The estimation of chronological age through assessment of dental radiographs iswell-established and a useful method to assist in the identification of persons inforensic and anthropological scenarios. The objective of our investigation was two-fold: (i) to validate the Kvaal et al. age-estimation method on a sample of WesternAustralian subjects, and (ii) to increase the range of chronological ages to which theKvaal et al. method can be applied. Our sample size included panoramic radio-graphs from 74 subjects (aged 12–28 years). A set of ratios were calculated and thenused to apply different statistical models of linear regression, in order to generate afinal formula to estimate age. The most accurate estimations were obtained from themodels generated by the mandibular canine measurement (SEE ± 3.708 years), andfor the three mandibular teeth (SEE ± 3.388 years). The results indicate that inclu-sion of juveniles did not affect final results, and the method still produced estimatesacceptable in a forensic framework.

Keywords: forensic dentistry; dental age estimation; secondary dentine formation;young adults

Introduction

The accurate approximation of chronological age, secondary to sex determination, is anintegral factor in constructing a biological profile for forensic1 and anthropological pur-poses, including the confirmation of identification at times of mass disasters, crimes,accidents and of unknown remains2−4. Further, the estimation of the chronological agein the living is also needed in situations such as immigration and refugee determina-tions4. Saunders originally proposed the estimation of chronological age based on dif-ferent stages of tooth eruption5; since then, several more methods have been suggestedto estimate dental age in children. Amongst these methods, those that are based on theradiological examination of permanent teeth development, are found to be the mostaccurate6,7 The Demirjian et al.8,9 classification is reported as the best method for den-tal age estimation in children and adolescents thanks to its high observer agreementand correlation between the defined stages and age10. This method has been adapted to

*Corresponding author. Email: [email protected]

© 2016 Australian Academy of Forensic Sciences

Australian Journal of Forensic Sciences, 2016http://dx.doi.org/10.1080/00450618.2016.1177593

Dow

nloa

ded

by [

1.13

2.96

.27]

at 0

8:36

23

May

201

6

different populations, showing a mean difference between the chronological age andthe dental age of around one year in different studies6,9,11.

The completion of tooth development is marked with the closure of the root apex,with the third molar being the last tooth to complete its eruption and root developmentat approximately 17–21 years of age. Following this stage, the accurate estimation ofchronological age based on dental formation becomes challenging12, leading to thedevelopment of alternative methods of dental age estimation in adults. Bodecker13,identified the relation of the continued apposition of secondary dentine and the subse-quent change in the morphology of the pulp chamber with chronological age. Accord-ingly, several dental age estimation methods have been developed, based on therelation between secondary dentine formation, and the subsequent narrowing andchange in pulp chamber dimensions and shape with age1,2,14,15.

Dental analysis using radiographs for the purpose of age estimation presents numer-ous advantages over other histological and biochemical methods: it is relatively simple,non-invasive and economically viable16. The method developed by Kvaal et al.1 is arelatively non-invasive method and has been validated in diverse populations 17−20.Karkhanis et al.17 applied the Kvaal et al.1 method in a Western Australian populationto develop age estimation standards with acceptable results in a forensic framework(±10 years)2. Originally, the Kvaal et al.1 method was proposed to be applied in anadult population. More recently, the method has been applied in younger populationswith varying degrees of success. The level of error in estimates of ages remains highin younger populations, with Landa et al. reporting a standard deviation approaching15 years20.

The aim of the present study is to validate the applicability of the Kvaal et al.1

method in a younger population (Western Australia sample) and to provide furtherrefinement of the level of variation in younger people. It will also assess the potentialof this method as an additional tool for dental age estimation in juveniles, where meth-ods based on the analysis of tooth development cannot be used.

Materials and methods

Sample selection

The study cohort consisted of a series of subjects who consecutively presented fororthodontic treatment at a private specialist orthodontic office. From the initial sample,74 Western Australians were aged less than 30 years, of those 63.5% (n=47) werefemale and 36.5% (n=27) were male, with a median age of 16 year for both genders.All panoramic radiographs were acquired from the same machine in digital format.Analysis was completed with Image J software (version 1.48 19 April 2014 – NationalInstitute of Health, USA) and the measurements were completed by a single observer(TM). All data was collated using Excel (version 2013 Microsoft, Redmont, USA) andstatistical analysis was completed using R Core Team version 3.1.3 (2015). (R: A lan-guage and environment for statistical computing. R Foundation for Statistical Comput-ing, Vienna, Austria. URL http://www.R-project.org/.)

All subjects with a panoramic radiograph of high quality with respect to factors ofimage brightness, contrast and sharpness were included. Any radiographs that did notmeet this requirement or had observable failings (e.g. image distortion, poor contrast,superposition of tooth structure, or improper positioning) were excluded. All teethincluded in the analysis were clinically sound with completed root formation and in

2 T.Y. Marroquin et al.

Dow

nloa

ded

by [

1.13

2.96

.27]

at 0

8:36

23

May

201

6

functional occlusion. Teeth with rotations, incomplete root formation, dilacerations,pulpal pathologies and endodontic treatment and/or restorations were excluded. Finally,teeth with large areas of enamel overlap between neighbouring teeth were excluded inthis study.

Teeth analysed

Following the parameters as provided by Kvaal et al.1, the teeth analysed were themaxillary central incisors (with Federation Dentaire International (FDI) notation 11 and21) lateral incisors (FDI notation 12 and 22), second premolars (FDI notation 15 and25), mandibular lateral incisors (FDI notation 32 and 42), mandibular canines (FDInotation 33, and 43) and first mandibular premolars (FDI notation 34 and 44). In theoriginal study, Kvaal et al.1 analysed only those periapical radiographs from individualswhere all the six teeth were present and suitable for examination. In the present study,panoramic radiographs with any combination of the required teeth available for mea-surement were included. Previous research1,17,19 has demonstrated the absence of bilat-eral asymmetry in the deposition of secondary dentine. Consequently, teeth from eitherside that fulfilled the inclusion criteria were analysed for the purpose of developing ageestimation standards.

Measurements

Odontometric data were acquired following the methodological approach of Kvaalet al. 1. In this way, the data were obtained from measuring: the maximum tooth length(from the incisal border or cusp tip to the root apex); the maximum pulp length (from thepulp chamber roof to the root apex); and the maximum root length (from the cemento-enamel junction – CEJ – on the mesial surface of each tooth to the root apex). The pulpchamber and root width measurements were collected at the points A (CEJ), B (mid-pointbetween the points A and C) and C (mid-point between the CEJ and root apex).


A pilot set of 30 panoramic radiographs, which were not included in the final analysis,was used to perform initial training and benchmarking to an expert in the field (SK).Six of the 30 radiographs were measured on six different days, in addition to four ran-domly selected panoramic radiographs, thus resulting in the analysis of ten panoramicradiographs per day. Based on this, intra- and inter-observer error calculations were per-formed17. The measurements were acquired from all the panoramic radiographs with aminimum of one day between each session to minimise the possibility of memorisingreference points and/or measurements in each observer. A second intra- and inter-observer calibration was performed using five randomly selected panoramic radiographsfrom the final study sample, and recording the measurements on five different days,also using four additional panoramic radiographs per day, to avoid measurements mem-orising, with at least one day between each evaluation.


Simple descriptive statistics (mean, standard deviation, and frequency distributions)were used to summarise the data. Initial tests for normality (assessment for skewness,

Australian Journal of Forensic Sciences 3

Dow

nloa

ded

by [

1.13

2.96

.27]

at 0

8:36

23

May

201

6

kurtosis and Shapiro-Wilk) were performed to determine, where appropriate, parametricand non-parametric univariate analysis testing for the continuous variables (Table 1).Pearson’s correlation coefficient (r) was calculated to assess the strength of correlationbetween age and dental ratios based on the Kvaal et al.1 method (Table 1). The correla-tion coefficients calculated (using the width ratios) presented significant correlation val-ues. The most representative were observed for the ratio calculated between pulp androot width at the B point of the root, and for the predictor W-L. The mandibularcanines showed the highest correlation coefficient for the predictors B (–0.570) andW-L (–0.625).

To examine the independent association of several factors with chronological age,multivariable linear regression analysis was used to look at the association between theindependent and outcome variables. The outcome variables included chronological age.The predictor variables M and W-L were used as the independent variables. Age esti-mation models were developed individually for each tooth (Table 2) and the followingtooth combinations (Table 3): three maxillary teeth with the predictors M and W-L indi-vidually introduced for each tooth (six predictors), and the average of M and W-L (twopredictors); three mandibular teeth with the predictors M and W-L individually intro-duced for each tooth, having six predictors in the equation, and the average of M andW-L resulting in two predictors; and mandibular and maxillary teeth M and W-L pre-dictors, introduced individually in the equation (12 predictors) and the average of Mand W-L (two predictors). All regression models were built using the ordinary leastsquares approach. R-square values were computed to examine the amount of varianceexplained by the predictor variables.

All statistical tests were two-sided and a p value of less than 0.01 was consideredto be statistically significant. Corrections were made for multiplicity using a modifiedBonferonni method to reduce the likelihood of Type I errors; an alpha threshold forstatistical significance for all comparisons was set at 0.01. Univariate analyses and

Table 1. Correlation coefficients between chronological age and Kvaal et al.1 dental measure-ments and ratios, for individuals under 30 years of age.

Tooth number (FDI numbering system)

Ratio 11/21 12/22 15/25 32/42 33/43 34/44P 0.129NS 0.236 NS 0.133 NS 0.164 NS 0.455* 0.152 NS

T 0.041 NS −0.071 NS −0.002 NS −0.130 NS 0.171 NS −0.117 NS

R 0.287* 0.287* 0.151 NS 0.327* 0.279* 0.397**A −0.130 NS −0.281* −0.004 NS −0.263* −0.254 NS −0.137 NS

B −0.378* −0.305* −0.280* −0.354* −0.570** −0.381*C −0.345* −0.377* 0.009 NS −0.088 NS −0.605** −0.504**M −0.049 NS −0.202 NS 0.0001 NS −0.123 NS −0.108 NS −0.265*W −0.415** −0.397** −0.150 NS −0.237* −0.123 NS −0.497**L 0.151 NS 0.295* −0.002 NS 0.255* 0.467** 0.297*W – L −0.310* −0.421** −0.052 NS −0.360* −0.625** −0.547**

*= p<0.05; **=p<0.001.NS=Not significantNote: P is the ratio between length of pulp and root; T is the ratio between length of tooth and root; R is theratio between length of pulp and root; A is the ratio between width of pulp and root at CEJ (Level A); B isthe ratio between width of the pulp and root at mid-point between level C and A (level B); C is the ratiobetween width of pulp and root at mid-root level (level C); M is the mean value of all ratios (first predictor);W is the mean value of width ratios from levels B and C; L is the mean value of the length ratios P and R;W – L is the difference between W and L (second predictor) 1.


Dow

nloa

ded

by [

1.13

2.96

.27]

at 0

8:36

23

May

201

6

multivariable linear regression analyses were performed using R Core Team version3.1.3 (2015) (R: A language and environment for statistical computing. R Foundationfor Statistical Computing, Vienna, Austria. URL http://www.R-project.org/).

Results


The technical error of measurement (TEM), relative technical error of measurement(rTEM) and coefficient of reliability were within acceptable standards for the intra- andinter-observer measurement precision (TEM<1.0, rTEM<5%, R>0.75). These valueswere 0.92, 2.99% and 0.95 respectively for intra-observer precision analysis. Similarly,the inter-observer calibration (between TM and SK) showed comparable precision21

(TEM 0.80, rTEM 2.37 and R 0.98).

Table 2. Multiple regression for estimation of chronological age (in years) from individualmaxillary and mandibular teeth for individuals under 30 years of age.

Tooth (FDI) n R R2 Equation SEE ± years

11/21 69 0.152 0.1271 Age= 22.915 – 28.78 (M) −22.376 (W – L) 4.34412/22 67 0.197 0.1727 Age= 19.724 – 25.935 (M) – 23.631 (W – L) 4.25815/25 68 0.003 −0.273 Age= 17.78 – 3.962 (M) – 2.204 (W – L) 4.53632/42 71 0.131 0.106 Age= 4.644 – 5.363 (M) – 22.558 (W – L) 4.46533/43 48 0.417 0.391 Age= 11.26 – 41.33 (M) – 86.06 (W – L) 3.70834/44 71 0.327 0.307 Age= 10.513 – 27.075 (M)–38.176 (W – L) 3.709

Note. *R2, coefficient of determination. SEE, standard error of estimation in years. See Table 1 forabbreviations.

Table 3. Multiple regression for estimation of chronological age (in years) from the combinedmaxillary and mandibular teeth, the respective predictors (pds) for individuals under 30 years ofage.

Teeth (FDI) n R R2 EquationSEE ±years

3 mx (2 pds) 60 0.154 0.125 Age= 28.702 – 46.338 (M) – 24.233 (W – L) 4.1413 mx (6 pds) 60 0.237 0.150 Age=24.394 – 30.751 (11/21 M) –12.090 (11/

21 W–L) – 29.2257 (12/22 M) – 12.894 (12/22 W– L)+ 2.611 (15/25 M) – 0.631 (15/25 W – L)

4.08

3 mdb (2 pds) 45 0.409 0.381 Age= −27.44 + 10.48(M) – 63.82 (W–L) 3.6483 mdb (6 pds) 45 0.539 0.466 Age= 1.894 + 13.428 (32/42 M)–5.931 (32/42 W

– L) – 51.067 (33/43 M) – 66.495 (33/43 W – L)– 5.114 (34/44 M) – 16.806 (34/44 W – L)1

3.388

6 teeth (2 pds) 40 0.252 0.211 Age= 31.91 – 63.34 (M) – 63.34 (W – L) 4.1386 teeth (12 pds) 30 0.542 0.339 Age= 2.627 −58.42 (11/21 M) + 4.526 (11/21 W

– L) + 32.169 (12/22 M) – 12.913 (12/22 W – L)+ 25.460 (15/25 M) + 1.098 (15/25 W – L) +5.270 (32/42 M) + 0.096 (32/42 W – L) – 28.692(33/43 M) + 1.686 (33/43 W – L) – 13.857 (34/44 M) – 49.336 (34/44 W – L)

3.788

*R2, coefficient of determination. SEE, standard error of estimation in years. See Table 1 for abbreviations.


Dow

nloa

ded

by [

1.13

2.96

.27]

at 0

8:36

23

May

201

6

In this way, individual (Table 2) and multiple (Table 3) tooth regression modelswere obtained. The accuracy to predict the age was quantified by the standard error ofestimation (SEE ± years). The most accurate age estimation model based on the analy-sis of individual tooth was for mandibular canines (SEE ± 3.708 years) (Figure 1). Pre-diction accuracy improved with the combined analysis of teeth; it was highest for themodel developed with the three mandibular teeth (six predictors, SEE ± 3.388 years).

Discussion

The estimation of chronological age has been a very relevant topic of discussionthrough human history. For this reason, diverse skeletal and dental methods have beendeveloped, showing variable levels of reliability in different populations. These meth-ods, based on dental changes, are well accepted and have shown their potential forforensic and anthropological applications22. Once the root formation has been com-pleted, progressive dental changes, such as secondary dentine formation, can be radio-graphically assessed based on the narrowing on the pulp chamber. Odontometric andmorphometric measurements can thus be acquired to quantify secondary dentine depo-sition and is the foundation for non-invasive adult age estimation methods such asKvaal et al.1 and Cameriere et al.23.

The Kvaal et al.1 method has been widely validated in different populations becauseof its numerous benefits, such as, among other benefits, its relatively non-invasiveapproach, applicability in live individuals and low cost. Previous research has appliedthis method to a Western Australian population17 and has shown significant results thatare valid under forensic standards. Although the Kvaal et al.1 method was initiallydeveloped to estimate age in individuals over 20 years old, previous studies have testedthis method in younger populations19. The present study applied the Kvaal et al.method1 in Western Australian participants under 30 years of age. The primary aimwas to assess the applicability of this method in a younger population to broaden theage range that the method can be applied to, without compromising the age estimationaccuracy.

Estim

ated

age

toot

h 33

/43

Chronological age tooth 33/43

Figure 1. Scatter plot showing a positive association between the estimated age and the chrono-logical age for the teeth 33/43.


Dow

nloa

ded

by [

1.13

2.96

.27]

at 0

8:36

23

May

201

6

To this end, regression models were developed, based on the data acquired fromindividual teeth (Table 2) and a combination of teeth (Table 3). The accuracy to predictage was quantified by the standard error of estimation (SEE ± years). The most accuratemodel for individual teeth was for the mandibular canines (±3.708 years), in contrast tothe study of Karkhanis et al.17, where the same tooth presented the highest SEE(±10.903 years).

Prediction accuracy improved when multiple teeth were included in the regressionmodels (Table 3). In this study, the highest level of accuracy was obtained when theequation included three mandibular teeth and six predictors (SEE ± 3.388 years), incomparison with the study of Karkhanis et al.17, where the most accurate model wasfor the combined analysis of the six teeth and 12 predictors (SEE ± 7.963).

It has been observed that age estimation in individuals older than 14 years is diffi-cult, as all permanent teeth, except the third molars (when present), have completedtheir development19. Some examples of previous studies amongst people under 30 yearsof age, using the Kvaal et al. method 1 on panoramic radiographs, are: Erbudak et al.24

in a Turkish population, including in their sample 75 participants between 14 to35 years of age, obtaining a SEE= ± 8.73 years at best, using the regression formulasof Paewinsky et al.25 and Kvaal et al.1; Landa et al.20 in a Spanish population with anage range of 14 to 60 (n=100,) from which 40 were aged younger than 31 years ofage, and with an underestimation of age when using the regression formulae of Kvaalet al.1 and Paewinsky et al.25 and a standard deviation of 12.53 at best when the Kvaalet al1. regression equation was used. The last example is the study of Meinl et al.19,which was conducted in an Austrian population (n=44) with an age range of 13 to24 years, and resulted in a mean underestimation of 31.44 years when the Kvaal et al.1

regression models were applied, or a mean overestimation of 20.88 years when thePaewinsky et al.25 formula was applied. In contrast, the present study (n=74) showsthat the Kvaal et al.1 method provides acceptable results2 in a sample composed ofsub-adults and young adults. Further research is warranted, in other populations withsimilar age ranges and using larger samples.

It is also worth comparing the results of this study with other methods based onthird molar development as an estimator of chronological age. One of the most remark-able studies is the research performed by Lewis and Senn10. In their study, theyreported a standard deviation of no more than 3 years, when different methods areapplied in a North-American population. Another method based on third molars is theexamination of the periodontal membrane in lower third molars in a German popula-tion26, which found a standard deviation of between 1.9 to 4.8 years. However, thirdmolar agenesis has been reported to be between 14% up to 51% in different studies27.The Kvaal et al.1 method can be considered as an alternative in these cases.

Conclusion

Panoramic radiographs are a unique diagnostic tool, but also provide useful informationvalid for forensic purposes. It has been well established that dental records are a highlyprecise instrument for establishing an individual’s identity. The use of the Kvaal et al.1

method was initially purposed on periapical radiographs, obtained from adults, butlately this method was applied using panoramic radiographs, and included juvenile par-ticipants. The validation of the Kvaal et al.1 method for age estimation, using panora-mic radiographs obtained from a population including juvenile individuals, enlarges therange of ages where the ratio between secondary dentine production and the decrease


Dow

nloa

ded

by [

1.13

2.96

.27]

at 0

8:36

23

May

201

6

of pulp chamber can be applied. This also presents the opportunity to examine theapplication of this method in other juvenile population groups.

Disclosure statement

No potential conflict of interest was reported by the authors.

Funding

This study received ethics approval from the Human Research Ethics Committee ofThe University of Western Australia (Ref: RA/4/1/6797).

ORCID

T.Y. Marroquin http://orcid.org/0000-0002-3335-5305

References1. Kvaal SI, Kolltveit KM, Thomsen IO, Solheim T. Age estimation of adults from dental radio-

graphs. Forensic Sci Int. 1995;74(3):175–185.2. Mathew DG, Rajesh S, Koshi E, Priya LE, Nair AS, Mohan A. Adult forensic age estimation

using mandibular first molar radiographs: A novel technique. J Forensic Dent Sci. 2013;5(1):56–59.

3. Sheelavant SS, Chandra GYP, Harish S. Estimation of age from the physiological changes ofteeth in adults between 25–60 years. An autopsy study. Indian. J Forensic Med Toxicol.2014;8(1):202–207.

4. Bosmans N, Ann P, Aly M, Willems G. The application of Kvaal’s dental age calculationtechnique on panoramic dental radiographs. Forensic Sci Int. 2005;153(2):208–212.

5. Saunders E. The teeth a test of age. The Lancet. 1838;30(774):492–496.6. Yusof MYPM, Thevissen PW, Fieuws S, Willems G. Dental age estimation in Malay children

based on all permanent teeth types. Int J Legal Med. 2014;128(2):329–333.7. Franklin D. Forensic age estimation in human skeletal remains: Current concepts and future

directions. Legal Med. 2010;12(1):1–7.8. Demirjian A, Goldstein H, Tanner JM. A new system of dental age assessment. Human Biol.

1973;45(2):211.9. Feijóo G, Barbería E, De Nova J, Prieto JL. Dental age estimation in Spanish children.

Forensic Sci Int. 2012;223(1–3):371.e1–.e5.10. Lewis JM, Senn DR. Dental age estimation utilizing third molar development: A review of

principles, methods, and population studies used in the United States. Forensic Sci Int.2010;201(1–3):79–83.

11. Kırzıoğlu Z, Ceyhan D. Accuracy of different dental age estimation methods on Turkishchildren. Forensic Sci Int. 2012;216(1–3):61–67.

12. Panchbhai AS. Dental radiographic indicators, a key to age estimation. Dento Maxillo FacialRadiol. 2011;40(4):199.

13. Bodecker CF. A consideration of some of the changes in the teeth from young to old age.Dental Cosmos. 1925;67:543–549.

14. Cameriere R, Ferrante L, Cingolani M. Variations in pulp/tooth area ratio as an indicator ofage: a preliminary study. J Forensic Sci. 2004;49(2):317–319.

15. Drusini AG, Toso O, Ranzato C. The coronal pulp cavity index: a biomarker for age determi-nation in human adults. Am J Phys Anthropology. 1997;103(3):353–363.

16. Sarkar S, Kailasam S, Mahesh Kumar P. Accuracy of estimation of dental age in comparisonwith chronological age in Indian population – A comparative analysis of two formulas.J Forensic Legal Med. 2013;20(4):230–233.


Dow

nloa

ded

by [

1.13

2.96

.27]

at 0

8:36

23

May

201

6

17. Karkhanis S, Mack P, Franklin D. Age estimation standards for a Western Australian popula-tion using the dental age estimation technique developed by Kvaal et al. Forensic Sci Int.2014;235:104.e1–104.e6.

18. Kanchan-Talreja P, Acharya AB, Naikmasur VG. An assessment of the versatility of Kvaal’smethod of adult dental age estimation in Indians. Arch Oral Biol. 2012;57(3):277–284.

19. Meinl A, Tangl S, Pernicka E, Fenes C, Watzek G. On the applicability of secondary dentinformation to radiological age estimation in young adults. J Forensic Sci. 2007;52(2):438–441.

20. Landa MI, Garamendi PM, Botella MC, Aleman I. Application of the method of Kvaal et alto digital orthopantomograms. Int J Legal Med. 2009;123(2):123–128.

21. Lewis SJ. Quantifying measurement error. Oxford: Oxbow Books; 1999. Current and recentresearch in osteoarchaeology 2: Proceedings of the 4th, 5th and 6th meetings of the Osteoar-chaeological Research Group, Anderson IS, editor; 54–55.

22. Manual of Forensic Odontology. 5th ed. Edited by David R, Senn R, Weems A, Senn DR,Weems RA. Boca Raton, FL: CRC Press; 2013.

23. Cameriere R, De Luca S, Alemán I, Ferrante L, Cingolani M. Age estimation by pulp/toothratio in lower premolars by orthopantomography. Forensic Sci Int. 2012;214(1–3):105–112.

24. Erbudak HO, Ozbek M, Uysal S, Karabulut E. Application of Kvaal’s et al. age estimationmethod to panoramic radiographs from Turkish individuals. Forensic Sci Int. 2012;219(1–3):141–146.

25. Paewinsky E, Pfeiffer H, Brinkmann B. Quantification of secondary dentine formation fromorthopantomograms—a contribution to forensic age estimation methods in adults. Int J LegalMed. 2005;119(1):27–30.

26. Olze A, Solheim T, Schulz R, Kupfer M, Pfeiffer H, Schmeling A. Assessment of the radio-graphic visibility of the periodontal ligament in the lower third molars for the purpose offorensic age estimation in living individuals. Int J Legal Med. 2010;124(5):445–448.

27. Celikoglu M, Kamak H. Patterns of third-molar agenesis in an orthodontic patient populationwith different skeletal malocclusions. The Angle Orthodontist. 2012;82(1):165–169.


Dow

nloa

ded

by [

1.13

2.96

.27]

at 0

8:36

23

May

201

6

242

APPENDIX 5. Application of the Kvaal method for adult dental age estimation using Cone

Beam Computed Tomography (CBCT).



dental age estimation using Cone Beam Computed Tomography (CBCT). Journal of Forensic

and Legal Medicine, Accepted for publication. October 2016.

Research Paper

Application of the Kvaal method for adult dental age estimation usingCone Beam Computed Tomography (CBCT)

Talia Yolanda Marroquin Penaloza a, Shalmira Karkhanis a, Sigrid Ingeborg Kvaal b,Firdaus Nurul c, Shalini Kanagasingam d, Daniel Franklin a, Sivabalan Vasudavan a,Estie Kruger a, *, Marc Tennant a

a University of Western Australia, School of Anatomy Physiology and Human Biology, 35 Stirling Hwy, Crawley, WA 6009, Australiab University of Oslo, Faculty of Dentistry, Institute of Clinical Dentistry, P.O. Box 1109, Blindern, N-0317 Oslo, Norwayc Radiology Unit, Dental Faculty University of Kebangsaan Malaysia, Jalan Raja Muda Abdul Aziz, 50300 Kuala Lumpur, Malaysiad Faculty of Dentistry, University of Kebangsaan Malaysia, Jalan Raja Muda Abdul Aziz, 50300 Kuala Lumpur, Malaysia

a r t i c l e i n f o

Article history:Received 11 August 2016Received in revised form13 October 2016Accepted 22 October 2016Available online 24 October 2016

Keywords:AdultsAge estimationTomographyKvaal method

a b s t r a c t

Different non-invasive methods have been proposed for dental age estimation in adults, with the Kvaalet al. method as one of the more frequently tested in different populations.

The purpose of this study was to apply the Kvaal et al. method for dental age estimation on modernvolumetric data from 3D digital systems. To this end, 101 CBCT images from a Malaysian population wereused. Fifty-five per cent were female (n ¼ 55), and forty-five percent were male (n ¼ 46), with a medianage of 31 years for both sexes. As tomographs allow the observer to obtain a sagittal and coronal view ofthe teeth, the Kvaal pulp/root width measurements and ratios were calculated in the bucco-lingual andmesio-distal aspects of the tooth. From these data different linear regression models and formulae werebuilt. The most accurate models for estimating age were obtained from a diverse combination of mea-surements (SEE ±10.58 years), and for the mesio-distal measurements of the central incisor at level A(SEE ±12.84 years). This accuracy, however is outside an acceptable range in for forensic application (SEE±10.00 years), and is also more time consuming than the original approach based on dental radiographs.

© 2016 Elsevier Ltd and Faculty of Forensic and Legal Medicine. All rights reserved.

1. Introduction

Currently, the need for developing more accurate and non-invasive methods for age estimation, as part of the identificationof adult individuals in situations of forensic scenarios is increasingglobally.1 Gustafson first proposed a method for dental age esti-mation in adults,2 since then, the analysis of dental changes inadults relative to chronological age has continued. The adultdentition is relatively resistant to environmental and chemical in-fluences, and methods based on the assessment of teeth areconsidered advantageous for this reason.3,4 Furthermore, methods

based on odontometric measurement of pulp and tooth structuresfrom dental radiographs or tomographs are non-invasive/non-destructive and can be applied to both living or deceased in-dividuals.5 Kvaal et al., proposed a method for age estimation inadults which has been tested in different populations around theworld since 1995, but all reported a high standard error (±8.5 to±13 years).1,5,6 The basis of this method is the analysis of the nar-rowing of the pulp chamber with age, which is observable andmeasurable in dental radiographs. However, there is scope forimprovement in the accuracy of the age estimations.

With the emergence of computer tomography (CT) and conebeam computer tomography (CBCT), new methods based on volu-metric reconstruction of tooth and pulp volume and ratio calcula-tions have been proposed for dental age estimation in adults.7,8

However, those techniques have not provided better accuracyrelative to the previously described Kvaal method, using dentalradiographs. Moreover, pulp/tooth volume calculation is labourintensive,9 and can require the use of complex, and often costlycomputer software.10 Clearly, these more technical multi-

* Corresponding author.E-mail addresses: [email protected] (T.Y. Marroquin Penaloza),

[email protected] (S. Karkhanis), [email protected] (S.I. Kvaal),[email protected] (F. Nurul), [email protected] (S. Kanagasingam),[email protected] (D. Franklin), [email protected](S. Vasudavan), [email protected] (E. Kruger), [email protected](M. Tennant).

Contents lists available at ScienceDirect

Journal of Forensic and Legal Medicine

journal homepage: www.elsevier .com/locate/ jflm

http://dx.doi.org/10.1016/j.jflm.2016.10.0131752-928X/© 2016 Elsevier Ltd and Faculty of Forensic and Legal Medicine. All rights reserved.

Journal of Forensic and Legal Medicine 44 (2016) 178e182

dimensional systems provide considerably more data, at a sub-stantially more detailed level, and with a reduction in many of thecomplications of simple plane film images. The method describedby Kvaal et al.5 was initially proposed, with measuring dimensionson periapical radiographs with a microscope and a pair of callig-raphers. The applications of this method have been expanded toinclude panoramic radiographs using different analytical soft-ware.6,11 The purpose of the present study is to apply the Kvaalet al.5 method for dental age estimation using modern computertomographic data from 3D digital systems. The null hypothesis isthat the substantially increased quantity (and quality) of the datafrom these systems will provide greater accuracy of dental ageestimation in adults.

2. Materials and methods

2.1. Sample

The study sample included a total of 101 tomographs obtainedas part of normal treatment from the radiology department of theUniversity of Keibangsaan Malaysia. Almost half (n ¼ 47) wereobtained with a Kodak device, reference K9000-3D (exposure pa-rameters: scanning time: 10.8 s, 3e15 mA, 60e85kVp, field of viewFOV: 5.2 cm � 5.2 cme5.5 cm to 5.5 cm, 180� rotation, slice thick-ness 0.076e0.2 mm), and 54 were obtained with an i-CAT device(Dental Imaging Technologies Corporation. i-CAT® FLX™) (expo-sure parameters: scanning time 4 s, 5 mA, 120 kVp, FOV:16 cm � 16 cm, Slice thickness 0.3 mme0.4 mm). Of these tomo-graphs, 55% were from female (n ¼ 55), and 45% were from male(n ¼ 46) patients, with a median age of 31 years for both sexes(Table 1). All the images were received in DICOM format, andanalysed using the Osirix® software package (OnDemand 3D soft-ware CyberMed Inc, Seoul, South Korea). The study sampleincluded teeth with very small coronal restorations, as long as theydid not compromise the pulp chamber and secondary caries wasnot present. Teeth with shape abnormalities, active caries, rootcanal treatment, prosthetic restorations, signs of pulp or periapicalpathologies, or open root apex, were duly excluded from thesample. Multi-rooted upper second premolars were also excluded.

2.2. Teeth analysed

Initially, Kvaal et al.5 proposed a set of linear measurements,including the length of the tooth, pulp and root, in addition to thewidth of the root and pulp chamber at different levels, in sixdifferent teeth: upper central and lateral incisors, and the secondpremolar; and the lower lateral incisor, canine and first premolar. Inthe present study, due to the unclear definition of the pulp chamberand tooth boundaries for the lower teeth (most likely related totheir small dimensions) only upper teeth were included: centralincisor, lateral incisor, canine, and second upper premolar (when

possible). As the dental images are tomographs, it is possible toassess the teeth in sagittal and coronal views; accordingly, the rootlength, and pulp/tooth width measurements proposed by Kvaalet al.5 were acquired in the mesio-distal and bucco-lingual aspectsof the teeth (see below). Significant differences between teeth ofthe left and right side of the maxilla have not been reported pre-viously,5 and therefore in this study, teeth from either the left orright side was measured.

2.3. Measurements

As proposed by Kvaal et al.,5 the mesio-distal measurements ofthe teeth (coronal view) were performed as follows: first, the levelA, was located at the CEJ (cemento-enamel junction) on the mesialside of the tooth, then level C was located halfway between point Aand the root apex, and finally point B was located halfway betweenpoint A and C. As this is the first study applying the Kvaal et al.method exclusively in tomographs, bucco-lingual measurementswere also acquired (sagittal view). Point A was located at thevestibular CEJ, point C halfway between the vestibular CEJ and theroot apex, and point B, halfway between point A and point C. In thisway, there were six pulp/tooth width measurements per tooth. Allthe measurements were recorded by one observer (TYM) using theOsirix® software.

2.4. Statistical analysis

Intra-observer error was tested, four tomographs wererandomly selected (two Kodak and two i-CAT) and all the mea-surements were recorded by one observer (TYM). Upper central,lateral and canine were measured on four separate occasions, witha minimum of one day between repetitions ensuring that theobserver did not remember the previously recorded measure-ments. Measurement error was assessed by calculating the tech-nical error of measurement (TEM), coefficient of reliability (R) andrelative technical error of measurement (rTEM).12

All data were collected in an Excel spreadsheet, Excel (version2013 Microsoft, Redmont, USA), and statistical analysis wascompleted using R Core Team version 3.1.3 (2015). (R: A languageand environment for statistical computing. R Foundation for Sta-tistical Computing, Vienna, Austria. URL http://www.R-project.org/). Descriptive statistics were used to summarise the data (mean,standard deviation, and frequency distributions). Normality wasalso tested (skewness, kurtosis and Shapiro-Wilk) to establishwhether parametric or non-parametric analysis was required.Pearson correlation coefficient (r) was calculated to assess therelationship between the dental ratios and age, for the data ob-tained from the Kodak CBCT and for the data obtained from i-CATCBCT individually, and also when both data sets were combined.This was done with the aim of testing if the difference between theresolutions of both devices could cause any statistically significantdifferences (Table 2). Linear regression models were built with agedesignated as the dependant variable. Different age estimationequations were then formulated using different groupings of agepredictors: 1) the tooth/pulp ratios at the levels A, B or C individ-ually; 2) calculating the average of the pulp/tooth ratios of all theteeth at the different levels (A, B or C) obtaining a formula for eachlevel; 3) calculating the average of the pulp/tooth ratios of thelateral (Lat) and central (Cent) upper incisor at the different levels(A, B and C); 4) Average of the pulp/tooth ratios from all the teeth atthe levels A, B and C in the same formula (Aþ Bþ C); and 5) same as4, but excluding the ratios obtained from the canines (Cent-Lat). Atotal of 17 linear regression models were thus generated per data-set: sagittal (s) view (bucco-lingual measurements); coronal (c)view; and the average of both (sc). The accuracy of the models is

Table 1Age and sex distribution for CBCT and i-CAT tomographs.

Sex Female Male Total

Age range (years) CBCTn ¼ 23

i-CATn ¼ 32

CBCTn ¼ 24

i-CATn ¼ 22

15e20 2 1 7 2 1220e29 12 9 5 8 3430e39 3 5 2 2 1240e49 4 6 8 5 2350e59 2 6 2 0 1060e75 0 5 0 5 10TOTAL 55 46 101

T.Y. Marroquin Penaloza et al. / Journal of Forensic and Legal Medicine 44 (2016) 178e182 179

reported as the standard error of estimation (SEE).11,13

3. Results

Intra-observer error results were considered to be withinacceptable range. (TEM <1.0 and the coefficient of reliabilityR>0.80%), regardless the tomograph source (Kodak or i-CAT) orwhether the measurements were recorded in the bucco-lingual ormesio-distal view of the teeth. In terms of the relative technicalerror of measurement (rTEM), 72% of the observations scored <5%and 28% had a rTEM < 10%. It was also observed that the accuracy ofthe observer was better for the bucco-lingual measurements(sagittal view of the teeth) regardless the source of the tomography.

Pearson correlation coefficient test (r) demonstrated a strongercorrelation between the pulp/tooth ratios and age for the bucco-lingual measurements (sagittal (s)) than for the mesio-distal mea-surements (coronal view (c)) (Table 2). Although there were dif-ferences between the correlation coefficients from the Kodak and i-CAT CBCT's, scatter-plots including the results of both sources donot show outliers associated with the specific use. There was a clearand significant negative correlation between age and pulp/toothratios in all the data sets. Correlation coefficient were calculated asfollows: i) for the measurements obtained from the Kodak tomo-graphs; ii) for the measurements recorded from the i-CAT tomo-graphs; and iii) for the combination of both. The i-CAT tomographsshowed the strongest correlation to age, especially for the uppercentral incisor (r ¼ �0.65), followed by the combination of theKodak & i-CAT data sets (r ¼ �0.61). The correlation coefficientsobtained from the Kodak, although statistically significant, werelower than that obtained from the i-CAT and combined Kodak & i-CAT (Table 2).

In relation to the accuracy of the different linear regressionmodels, it was observed that all the estimates were over thethreshold generally accepted for forensic application (SEE ±10years),14 however the SEE is reduced when both data-sets arecombined into a linear regression model (Table 3). There was onlyone equation with an acceptable level of predictive accuracy (SEE10.58 years (n ¼ 72, r ¼ 0.59, r2 ¼ 0.56)):

Age ¼ 101.46 þ 734.74 (Ac) �374.81 (Ac(Cent-Lat)) e 1075.18 (Csc) þ 327.07 (Bc) þ 310 (Cc(Cent-Lat)) e 159.25 (11/21 As)

Abbreviations: Ac: average of the pulp/tooth ratio in the coronalview at level A from the three teeth. Ac(Cent-Lat): average of the pulp/tooth ratios from the lateral and central incisors. Csc: average of thepulp/tooth ratio at level C from the sagittal and coronal views (sc)from all teeth. Bc: average of the pulp tooth ratio at the level B fromall the teeth in the coronal view. (Cc(Cent-Lat)): average of the pulp/tooth ratios at the level C from the lateral and central incisor in the

coronal view. 11/21 As: pulp/tooth ratio at the level A from thecentral incisor in the sagittal view.

The most accurate linear regression models obtained from theindividual analysis of the bucco-lingual measurements were asfollows: the central incisor at level A (Age¼ 84.059e191.003 (11/21As), r2 ¼ 0.37, SEE ¼ ±12.84), the average of the bucco-lingualmeasurements from the central incisor and lateral incisor at thelevel A, B and C (Age ¼ 102.08e157.94 (As) �73.0 (Bs) �36.81 (Cs),r2 ¼ 0.39, SEE ¼ ±12.73 years).

When both data-sets were combined (Table 3) the highest ac-curacy was achieved from: the average of the bucco-lingual (s) plusthe average of the mesio-distal (c) measurements at level A, (SEE±12.17 years); the average of the bucco-lingual and mesio-distalmeasurements at level A from the central and lateral incisors,(SEE ±12.76); and the average of the bucco-lingual andmesio-distalmeasurements from the central and lateral incisors at level A, pluslevel B, plus level C, (SEE ±12.7). The linear regression modelsgenerated by the inclusion from only the mesio-distal (c) mea-surements did not result in an SEE lower than ±14 years, andtherefore these results are not included.

4. Discussion

This is the first study testing the Kvaal et al. method exclusivelyin tomographs, and also the first one applying Kvaal et al.'s. pulp/tooth width ratio calculations in the bucco-lingual aspect of theteeth. Computer tomography is increasingly being used in dentalpractice and provides three-dimensional information about anyarea of interest, in a relatively quick and cost-effective manner.15 Inthis study it was expected that application of the Kvaal et al.method using CBCT diagnostic images would increase the array ofmethods available for the forensic profiling of living and deceasedindividuals. Until now, there has only been one other study pro-posing a similar method for adult dental age estimation, based onthe assessment of the pulp/tooth bucco-lingual dimensions (area)in canines.16 Although the results of the latter study were satis-factory (Mean prediction error, ME ±2.8 years at best), this methodrequired extracted teeth, and thus has restricted application in aforensic context. The use of tomographs overcomes issues relatedto the requirement for tooth extraction, and also permits theanalysis of multiple teeth with less radiation exposure.

Although, in this study, the number of individuals was small incertain age intervals, there are well documented biological agerelated changes in the human dentition, which have allowedforensic dentists to propose and use different methods for dentalage estimation.17 In previous publications applying the Kvaal et al.method for dental age estimation in adults, it has been noted thatone of the principal limitations is the absence of a clear limit be-tween the dentine and the pulp camber,11 which is observed as a

Table 2Correlation coefficient measurements CBCT.

Tooth number CBCT i-CAT CBCT & i-CAT

Coronal Sagittal Coronal Sagittal Coronal Sagittal

11/21 A �0.05 NS �0.36* �0.634** �0.65** �0.45** �0.61**11/21 B �0.26 NS �0.37* �0.421** �0.59** �0.32* �0.5**11/21 C �0.29 NS �0.45** �0.288* �0.36* �0.23* �0.41**12/22 A �0.53** �0.08 NS �0.326* �0.62** �0.31* �0.41**12/22 B �0.35* �0.5** �0.3* �0.60** �0.22* �0.46**12/22 C �0.24 NS �0.57** �0.013 NS �0.41** �0.05 NS �0.41**13/23 A �0.25 NS �0.13 NS �0.153 NS �0.50** �0.04 NS �0.17 NS

13/23 B �0.59** �0.56** �0.329* �0.41* �0.28** �0.39**13/23 C �0.52** �0.34** �0.348* �0.24 NS �0.28** �0.21*

*p < 0.05, **p < 0.001, NS ¼ non-significant.

T.Y. Marroquin Penaloza et al. / Journal of Forensic and Legal Medicine 44 (2016) 178e182180

grey zone instead of a distinct line.6 This affects the accuracy of themeasurements and intra/inter-observer agreement. In the presentstudy it was observed that neither Kodak nor i-CAT tomographswere exempt from this issue. In certain cases, the border betweenthe pulp chamber and dentine, or tooth and bone are also moreblurred than in periapical radiographs. It is true that dental Kodaktomographs may have higher resolution than i-CAT tomographs.However, in the present study it was observed that the data ob-tained from the i-CAT had a higher correlation to age compared tothe Kodak scans (Table 2). Similarly, the bucco-lingual width ratioshad a higher correlation to age than the mesio-distal width ratios(with the latter measurable in periapical radiographs).

It has previously been reported that Kvaal et al. length mea-surements are both difficult to register and do not provide signifi-cant information to the final equation.5,18 It is important to notethat in the tomographs analysed in the present study, it was evenmore difficult to observe the root apex compared to periapical ra-diographs. The only length measurement registered was the rootlength from the CEJ towards the apex to locate points A, B and C. Inthis way only pulp/toothwidth and ratios were used as the only agepredictors. The exclusive use of pulp/tooth ratios proposed by Kvaalet al. has previously been documented18 on panoramic radiographs,showing more accurate results (SEE ¼ ±10.02 years) than thoseobtained in this study. (SEE ¼ ±10.58 years at best). In a previousstudy that analysed the odontometric age-related changes indifferent teeth, it was found that the rate of secondary dentineformation varied depending on tooth type: in canines the increaseof coronal dentinal thickness was not as high as it was for theincisor and premolars.19 In the present study it was similarlyobserved that although the correlation coefficient between pulp/tooth ratios and age were significant, the exclusion of canine ratioswhen the linear regression models were generated, improved theiraccuracy.

On the other hand, this is also the first study documenting theinclusion of teeth with small crown fillings in the sample togenerate linear regression formulae, in contrast to previous publi-cations using the Kvaal et al. method,5 where one of the inclusioncriteria was to use only totally clinically sound teeth. It wasobserved that there were no statistically significant differencescaused by inclusion of these teeth. This may relate to the nature of

the formation of tertiary dentine or reactionary dentine to externalthreats. It has been reported that tertiary dentine formation ishighly dependent on the stimulus, and that caries mediators inducea focal dentine production,19e21 with a rate formation of about 3 mmper day.22

Regression analysis, has been chosen for age estimation due toits simplicity.23 Nevertheless, it is necessary to note the limitationsof the use of linear regression models for dental age estimation; forexample, the higher standard deviations when the originalformulae of Kvaal et al. are applied in populations of differentethnic backgrounds.24 The latter necessitates the need for popula-tion specific formulae. Also, it assumes that formation of secondarydentine is a linear process, when it is more similar to a curve.25

Additionally, it was observed that, although using tomographyresult in obtaining more odontometric data from the same teeth(mesio-distal and bucco-lingual measurements), this apparentadvantage over the use of dental radiographs, did not improve thefinal outcome. Also, the need to properly align the teeth in thesagittal and coronal plane demands more time than the traditionalKvaal et al. approach in dental radiographs.

Finally, it is necessary to mention that CBCT is not exempt ofartefacts intrinsic to the process of image acquisition.26 This couldexplain, not only the high SEE obtained in this study, but also thepoor accuracy of those methods based on pulp/tooth volumecalculation,9,10 when compared with more simple approaches suchas the measurement of pulp/tooth area on dental radiographs(Cameriere et al. method, with a SEE ¼ ±3.27 years).16

5. Conclusion

Radiological assessment of dental age-related changes is a validtool for age estimation. It is true that new diagnostic technologiesprovide the practitioners and forensic researchers new opportu-nities to use and propose new methods for dental age estimation.However, the use of tomographs in this study, applying the Kvaalet al. method, did not improve the results. It is recommended toalso include teeth with small fillings in studies based on pulp/toothdimension ratios, as long as the pulp chamber is not compromised.The aim of this is to further investigate the application of thesedifferent methods on those individuals who do not have totally

Table 3Linear regression analysis, linear regression formulae, correlation coefficient (r) between age and the ratios of bucco-lingual and mesio-distal width measurements from CBCT& i-CAT.

Sagittal and coronal

Age predictors n Linear regression formulae r r2 SEE

11/21 A sc 86 Age ¼ 99.74�72.07(11Ac)�161.63(11As) 0.4 0.4 12.211/21 B sc 86 Age ¼ 81.82�35.031(11Bc)�147.14(11As) 0.26 0.24 1411/21 C sc 86 Age ¼ 71.92�29.16(11Cc)�134.69(11Cs) 0.17 0.15 14.912/22 A sc 89 Age ¼ 87.43�68.05(12Ac)�129.37(12As) 0.21 0.19 14.512/22 B sc 89 Age ¼ 73.98�24.34(12Cb)�127.46(12Bs) 0.21 0.19 14.512/22 C sc 89 Age ¼ 57.56 þ 37.84(12Cc)�148.22(12 cs) 0.18 0.17 14.713/23 A sc 95 Age ¼ 51.14�5.4(13Ac)�47.192(13As) 0.03 0.01 15.413/23 B sc 95 Age ¼ 69.51�47.15(13Bc)�70.44(13Bs) 0.17 0.15 14.213/23 C sc 95 Age ¼ 60.81�78.34(13Cc)�34.37(13Sc) 0.09 0.08 14.9A sc 78 Age ¼ 100.41�33.11(Ac)�206.40(As) 0.34 0.33 13.5B sc 78 Age ¼ 89.22�20.54(Bc)�178.29(Bs) 0.31 0.29 13.9C sc 78 Age ¼ 69.46 þ 16.4(Cc)�174.42(Cs) 0.17 0.15 15.2A sc Cent/Lat 83 Age ¼ 105.42�72.96(Ac1/2)�187.22(As1/2) 0.4 0.38 12.8B sc Cent/Lat 83 Age ¼ 86.31�19.86(Bc1/2)�180.19(Bs1/2) 0.29 0.27 13.9C sc Cent/Lat 83 Age ¼ 70.41 þ 42.09(Cc1/2)-206.7(Cs1/2) 0.23 0.21 14.4Asc þ Bsc þ Csc 74 Age ¼ 104.09�17.65 (Ac)�135.48 (AS)�38.18(Bc)�105.54(Bs)þ40.87(Cc)þ10.93 (Cs) 0.4 0.35 13.3Asc þ Bsc þ Csc Lat/Cent 82 Age ¼ 106.25�180.14(Asc1/2)�135.83(Bsc1/2)þ51.73(Ccs1/2) 0.4 0.39 12.7

Abbreviations: Tooth number (11/21, 12/22 and 13/23) with Asc, Bsc or Csc ¼ pulp/tooth ratio at the levels A, B or C in the sagittal and coronal view (sc) calculated per tooth.Asc, Bsc or Csc: average of the pulp/tooth ratio at the respective level, from all the teeth. Asc, Bsc or Csc with Lat/Cent: average of the pulp/tooth ratio only from central andlateral incisors. Asc þ Bsc þ Csc ¼ average from the pulp/tooth ratio from all the teeth at the respective level. Asc þ Bsc þ Csc ¼ average from the pulp/tooth ratio from all theteeth at the respective level including only lateral and central teeth. SEE: Standard error of estimation in ± years. Best results in bold.

T.Y. Marroquin Penaloza et al. / Journal of Forensic and Legal Medicine 44 (2016) 178e182 181

sound dentitions.

Conflict of interest

None of the authors have any conflict of interest associated withthis study.

Funding

None.

Ethics approval

Ethics approval for this study was obtained from the HumanResearch Ethics Committee of the University of Western Australia(Ref: RA/4/1/6797). Permission was also granted by the Universityof Keibangsaan Malaysia. The design of the study, data collection,and analyses have been performed in accordance with the ethicalstandards set forth in the 1964 Declaration of Helsinki.

Acknowledgments

The authors would like to thank the department of radiology ofthe dental faculty of the University of Kebangsaan Malaysia, and toDr. Shahida Mohd Said also to Ambika Flavel, from the Centre forForensic Science at the University of Western Australia (UWA) toMartin Firth form the statistics clinic at UWA.

References

1. Bosmans N, Ann P, Aly M, Willems G. The application of Kvaal's dental agecalculation technique on panoramic dental radiographs. Forensic Sci Int.2005;153(2e3):208e212.

2. Gustafson G. Age determination on teeth. J Am Dent Assoc. 1950;41(1):45e54.3. Soomer H, Ranta H, Lincoln MJ, Penttila A, Leibur E. Reliability and validity of

eight dental age estimation methods for adults. J Forensic Sci. 2003;48(1):149e152.

4. Limdiwala PG, Shah JS. Age estimation by using dental radiographs. J ForensicDent Sci. 2013;5(2):118e122.

5. Kvaal SI, Kolltveit KM, Thomsen IO, Solheim T. Age estimation of adults fromdental radiographs. Forensic Sci Int. 1995;74(3):175e185.

6. Kanchan-Talreja P, Acharya AB, Naikmasur VG. An assessment of the versatilityof Kvaal's method of adult dental age estimation in Indians. Arch Oral Biol.2012;57(3):277e284.


estimation procedure based on the 3D CBCT study of the pulp cavity and hardtissues of the teeth for forensic purposes: a pilot study. J Forensic Leg Med.2015;36:150e157.

8. Ge Z-p, Ma R-h, Li G, Zhang J-z, Ma X-c. Age estimation based on pulp chambervolume of first molars from cone-beam computed tomography images. ForensicSci Int. 2015;253:133.e1e133.e7.

9. De Angelis D, Gaudio D, Guercini N, et al. Age estimation from canine volumes.Radiol Med. 2015:1e6.

10. Venkatesh S, Ajmera S, Ganeshkar SV. Volumetric pulp changes after ortho-dontic treatment determined by cone-beam computed tomography. J Endod.2014;40(11):1758e1763.

11. Karkhanis S, Mack P, Franklin D. Age estimation standards for a WesternAustralian population using the dental age estimation technique developed byKvaal et al. Forensic Sci Int. 2014;235(0):104.e1e104.e6.

12. Lewis SJ. Quantifying measurement error. In: Anderson IS, ed. Current andRecent Research in Osteoarchaeology 2: Proceedings of the 4th, 5th and 6thmeetings of the Osteoarchaeological Research Group. Oxford: Oxbow Books;1999:54e55.

13. Marroquin T, Karkhanis S, Kvaal S, et al. Determining the effectiveness of adultmeasures of standardised age estimation on juveniles in a Western Australianpopulation. Aust J Forensic Sci. 2016:1e9.

14. Solheim T, Sundnes PK. Dental age estimation of Norwegian adults - a com-parison of different methods. Forensic Sci Int. 1980;16(1):7e17.

15. Oi T, Saka H, Ide Y. Three-dimensional observation of pulp cavities in themaxillary first premolar tooth using micro-CT. Int Endod J. 2004;37(1):46e51.

16. Cameriere R, Ferrante L, Belcastro MG, Bonfiglioli B, Rastelli E, Cingolani M. Ageestimation by pulp/tooth ratio in canines by mesial and vestibular peri-apicalx-rays. J Forensic Sci. 2007;52(5):1151e1155.

17. Panchbhai AS. Dental radiographic indicators, a key to age estimation. Dento-maxillofac Radiol. 2011;40(4):199.

18. Paewinsky E, Pfeiffer H, Brinkmann B. Quantification of secondary dentineformation from orthopantomogramsda contribution to forensic age estima-tion methods in adults. Int J Leg Med. 2005;119(1):27e30.

19. Murray PE, Stanley HR, Matthews JB, Sloan AJ, Smith AJ. Age-related odonto-metric changes of human teeth. Oral Surg Oral Med Oral Pathol Oral RadiolEndod. 2002;93(4):474e482.

20. Hargreaves KM, Berman LH. Cohen's Pathways of the Pulp Expert Consult. ElsevierHealth Sciences. 2015.

21. Bergenholtz G, H€orsted-Bindslev P, Reit C, eds. Textbook of Endodontology.second ed. Hoboken: Hoboken. Wiley; 2009.

22. The principles of endodontics. In: Patel S, Barnes JJ, eds. Ovid Technologies I.second ed. Oxford: Oxford University Press; 2013.

23. Darmawan MF, Yusuf SM, Abdul Kadir MR, Haron H. Age estimation based onbone length using 12 regression models of left hand X-ray images for Asianchildren below 19 years old. Leg Med. 2015;17(2):71e78.

24. Parikh N, Dave G. Application of kvaal's dental age estimation technique onOrthopantomographs on A Population of GujarateA Short Study. BUJOD.2013;3(3):18e24.

25. Woods MA, Robinson QC, Harris EF. Age-progressive changes in pulp widthsand root lengths during adulthood: a study of American blacks and whites.Gerodontology. 1990;9(2):41e50.

26. Schulze R, Heil U, Grob D, et al. Artefacts in CBCT: a review. DentomaxillofacRadiol. 2011;40(5):265e273.

T.Y. Marroquin Penaloza et al. / Journal of Forensic and Legal Medicine 44 (2016) 178e182182

248

APPENDIX 6. Reliability and repeatability of pulp volumetric reconstruction through three

different volume calculations


Tennant, M. Reliability and repeatability of pulp volumetric reconstruction through three

different volume calculations. Journal of Forensic Odontostomatology. Accepted for

publication. March 2017.

JOURNAL of FORENSIC ODONTO-

STOMATOLOGY VOLUME 34 Number 2 December 2016

35

SECTION AGE ESTIMATION

Reliability and repeatability of pulp volume

reconstruction through three different volume

calculations

Talia Yolanda Marroquin Penaloza1, Shalmira Karkhanis

1, Sigrid Ingeborg Kvaal

2, Sivabalan Vasudavan

1, Edwin

Castelblanco1, Estie Kruger

1, Marc Tennant

1

1School of Anatomy, Physiology and Human Biology University of Western Australia, Crawley, Australia. 2Faculty of Dentistry, Institute of Clinical Dentistry, Oslo, Norway. Corresponding author: [email protected]

The authors declare that they have no conflict of interest.

ABSTRACT

Objective: To test the variability of the volume measurements when different segmentation methods

are applied in pulp volume reconstruction. Materials and methods: Osirix® and ITK-SNAP software

were used. Different segmentation methods (Part A) and volume approaches (Part B) were tested in a

sample of 21 dental CBCT’s from upper canines. Different combinations of the data set were also

tested on one lower molar and one upper canine (Part C) to determine the variability of the results

when automatic segmentation is performed. Results: Although the obtained results show correlation

among them(r>0.75), there is no evidence that these methods are sensitive enough to detect small

volume changes in structures such as the dental pulp canal (Part A and Part B). Automatic

segmentation is highly susceptible to be affected by small variations in the setting parameters (Part

C). Conclusions: Although the volumetric reconstruction and pulp/tooth volume ratio has not shown

better results than methods based on dental radiographs, it is worth to persevere with the research in

this area with new development in imaging techniques.

KEYWORDS: age estimation, pulp volume calculation, image segmentation.

JFOS. July 2017, Vol.34, No.2 Pag 35 xx-xx

ISSN :2219-6749

Reliability and repeatability of pulp volume reconstruction through three different volume calculations. Penaloza et al.

36

INTRODUCTION

Since 1950 different methods have been

proposed to estimate dental age in adults.1

Most of them are based on the formation of

secondary dentine and the decrease of the

pulp chamber size with age.2,3

Methods

that have an invasive approach, (for

example aspartic acid racemization and

cementum annulation), are not of

preference, as they often require the

physical damage of the sample.4 With the

evolution of different diagnostic imaging

techniques, non-invasive methods for

dental age estimation have become the

preferred methods,5 starting with the use of

periapical dental radiographs,2,3

then

panoramic radiographs6 and more recently

cone beam computed tomography

(CBCT).7

Micro-Computer Tomography (µCT) was

introduced in the early 20th century8 to

determine age related three-dimensional

changes in pulp cavities, from maxillary

first premolar teeth. Consistent with early

histological research, µCT study’s findings

indicate that decrease in pulp volume is not

linear. It demonstrates a quicker reduction

between the 20 and 40 years and then it

slows down thereafter. Further research

correlating the ratio between pulp and

tooth volume with the chronological age,

using µCT and linear regression models to

estimate dental age in adults, found

promising outcomes.8 However, in light of

the high radiation doses associated with

µCT, only extracted teeth can be measured

in this way.9 More recently CBCT has

been used for dental age estimation

calculating the volumes from tooth and

pulp chamber. Since then, volumetric

reconstruction from CBCT with different

software have been reported in various

studies for dental age estimation in

adults.10-13

The most recent studies

document the use of CBCT from single- 5,13

and multi-rooted teeth.14

. However,

these new methods have not shown

superior accuracy to the methods based on

dental radiographs.2,15

It is evident that

within each type of software there are a

series of sensitivity settings that influences

the mathematical approach to measure the

baseline volumetric data. These settings

clearly have consequences for the outcome

and may well be in part the cause of the

substantial variation. The primary aim of

this study was to test the variability of the

volume measurements when different

segmentation methods were applied to pulp

volume reconstruction. This study is

designed to ensure that the underpinning

assumptions of a commonly used age

estimation approach based on pulp volume

calculation are as accurate as possible.


Ethics approval for this study was obtained

from the Human Research Ethics

Committee of the University of Western

Australia (Ref: RA/4/1/6797). This study

has been performed in accordance with the

ethical standards laid down in the 1964

Declaration of Helsinki.

Study sample

A total of 22 anonymized dental CBCT’s

were used in this study. All of them were

recorded with therapeutic purposes, with

the consent of the participants, and there

was no unnecessary or repeated radiation

exposure. All the images were previously

unidentified, and age and sex of each

individual was the only known

information. From these, 57% (n=12) were

female, age range 17-63 years, mean age

33.7, and 43% (n=9) were male, age

range=15-52 years, mean age 36.7. The

CBCT’s were obtained from the radiology

department of the University of

Kebangsaan, Malaysia, and the research

group INVIENDO from the National

University of Colombia, In both cases the

images were obtained using the 9000 3D

Extraoral Imaging System (Carestream


37

Dental, Atlanta, GA, USA) which had a

180˚ rotation and a field of view (FOV) of

50 mm by 37 mm. The radiation exposure

parameters were 8-10 mA, 70 Kv, with a

slice thickness of 0.076mm. The images

were saved in DICOM format. The

selected tooth to apply the different

segmentation approaches was the upper

right canine, or the left when the right was

absent (with Federation Dentaire

International (FDI) notation 13 and 23)

reconstructing one tooth per subject. Only

sound teeth, free of any restorations with

completed apex formation were included.

Measurement software ITK-SNAP version

3.4 (open source software,

www.itksnap.org) and Osirix®

(OnDemand 3D software CyberMed Inc,

Seoul, South Korea) were used in this

study. All the segmentations were

completed by a single observer (TYM).

All data was collated using Excel (version

2013 Microsoft Redmont, USA) and

statistical analysis was completed using R

Core Team version 3.1.3 (2015). (R: A

language and environment for statistical

computing. R Foundation for Statistical

Computing, Vienna, Austria. URL

http://www.R-project.org/.)

Description of the method

This study tested the comparative outcome

of three different ways of estimating the

pulp volume,5,13,14

dental age indicator used

in adults. The tests were carried out on the

root pulp chamber of upper canines (n=21),

to calculate the correlation between

automatic14

and manual segmentation5

(Part A)(Table 1) and cone shape approach

volume calculation13

(Part B)(Table 2).

The final part of the study (Part C)(Table

3) tested the variability of the results

obtained with different setting parameter

values combinations for automatic

segmentation of the pulp canal of one

multi-rooted tooth, as previously reported

in the literature14

and one single rooted

tooth. (Figure 1)

Part A. Automatic and manual

segmentation

Automatic segmentation was performed

using the software ITK-SNAP. (Figure 1.

A). Manual segmentation used software

Osirix ® and its 2D viewer, (Figure 1. B)

which allows the observer to go slice by

slice and manually select the boundaries of

the space to reconstruct, using the tool

polygon. The automatic segmentation

process in ITK-SNAP is called seed

region-growing. To apply it, the observer

sets up a “seed” inside the structure to

reconstruct, in this case the pulp chamber.

This seed grows and a volume is finally

obtained.

Automatic segmentation of the pulp was

generated using the same combination of

setting parameters through the entire

sample: Scale of Gaussian blurring: 3.0,

edge transformation contrast: 0.2, edge

transformation exponent: 4.0, expansion

(balloon) force: 1.0, smoothing (curvature)

force: 0.2, edge attraction (advection) force

2.0. In three cases the edge transformation

exponent needed to be adjusted to 0.05.

Regression models were run to establish

which of the setting parameters would

have the most significant influence on the

final results. Little changes in the Scale of

Gaussian generated the bigger changes in

the final obtained values (p< 0.001).

With manual segmentation two approaches

were used. Firstly, every fourth slice was

analysed as for the published method.

Secondly, the first and last slice was used

as the sample. To analyse the same length

of root in the automatic and manual

approaches, a length defined in the

automatic approach was set for the manual

approach.

Part B. Cone shape geometric

approximation



38

The geometric approximation of the root

canal to cone is a simple conservative

method and has been used for dental age

estimation in adults. Using Osirix®, ovals

of best fit were formed in the root canal at

the cemento-enamel junction (CEJ) level,

following the reported method.13

Secondly,

instead of best fit ovals, the boundaries of

the pulp canal were drawn using the tool

polygon. (Figure 1, C) The root length was

measured from the CEJ to the apex,

(Figure 1, D) and finally two different

volumes of cones were calculated per tooth

with a mathematical formula.

Part C. Automatic segmentation and

different setting parameters

combinations

As observed in part A, it is not possible to

apply the same combination of setting

parameters values to all teeth with

automatic segmentation. Taking into

account the advantage of this automatic

process16

twenty combinations of different

values for all the setting parameters (Table

3) were used to reconstruct the pulp

chamber of one lower first molar (FDI 36,

Male 23 year old Malaysian origin, slice

thickness 0.2mm) and an upper canine

(FDI 23, Male 31 year old Colombian

origin). The volumetric reconstruction with

each combination was completed twice per

tooth. The variation among the obtained

volume from the same tooth was

calculated.

Statistical analysis Shapiro Wilk normality test, mean,

standard deviation (SD), paired t-Test and

Pearson correlation coefficient, were

calculated to compare the results obtained

using the different segmentation methods

(Part A), cone shape volume approach

(Part B), and the obtained results when

different parameters are used to do the

automatic segmentation of the same tooth

(Part C).

Fig.1: Volumetric reconstructions, upper right canine using different methods Volumetric reconstruction of tooth 13 from a female individual, 36 years old, using the different

methodologies mentioned in this paper. A: Semiautomatic recontruction using ITK-SNAP software. B:

Manual segmentation using Osirix ® software. C and D: Cone shape geometric approximation, following the methodology reported by the author doing area measurement of he root canal at the

level of the CEJ (D) and length measurement of the root from the CEJ up to the root apex (C), using

Osirix ® software.

RESULTS

Part A. Automatic and manual

segmentation

Three data-sets were obtained from the

volumetric reconstruction of the root canal


39

of the canines using the different methods

(Table 1).

When automatic and manual segmentation

using each fourth slice were compared, the

Pearson’s correlation coefficient (r=0.83),

shows a greater correlation between them,

than between automatic and manual

segmentation using only the first and the

last slice (r=0.75) or between the two

manual segmentation methods (r=0.79).

However, this does not mean that manual

segmentation using every fourth slice and

automatic segmentation produce

comparable results. The paired t test shows

that there is a statistically significant

difference between the means of the results

of both manual segmentations and

automatic segmentation (p<0.001).

When comparing the obtained volumes, it

was observed that manual segmentation

produced volumes that are larger than

those obtained with automatic

segmentation. To facilitate the

understanding of this difference, a simple

mathematical analysis was used, taking the

obtained volume with automatic

segmentation as 100% of the volume.

Then the difference between the volumes

obtained with automatic segmentation and

both manual segmentation, were used in a

rule of three to calculate the percentage

this difference represents.

The results of this calculation showed that

the volumes obtained using every fourth

slice are in average 25% (1%-53%) larger

than those obtained with automatic

segmentation. Also, that using the first and

the last slice are in average 57% (12% to

210%) larger than those obtained with

automatic segmentation. The average of

the obtained volumes from both manual

segmentation methods is in general 41%

larger than automatic segmentation.

The volumes obtained doing manual

segmentation of each fourth slice were in

average 28% larger than those obtained

using only the first and the last slice.

However, some of them were in average

16% smaller (p>0.05). The mathematical

analysis was like the previously described,

but in this case, the volume obtained with

manual segmentation using only the first

and last slice was taken as 100% in the

rule of three.

Part B. Cone shape geometric

approximation

It was expected to obtain similar area and

volume values when using ovals of best fit

and the tool polygon following the outline

of the pulp chamber. Different than

expected, it was observed that with the tool

polygon the obtained areas and volumes

were noticeably larger (50%) than with the

ovals (Table 2). However, there is a strong

correlation coefficient between both

measured areas using the tools oval and

polygon (r=0.84) at the CEJ and the

obtained volumes (r=0.88).

When the obtained areas using the two

different tools are compared the data show

a big variability and a high SD, this effect

is diminished once the volume is

calculated. In this way, the length of the

root is what really determines the final

obtained volume.

Part C. Automatic segmentation and

different setting parameters

combinations

Repeatability of the automatic

segmentation for the first and the second

repetition for both teeth was evaluated

with the paired t test (r=0.7, p=0.99)

showing that there is no statistically

difference between both repetitions with a

95% level of confidence. The average

between the obtained volumes in both


40

repetitions was calculated. The volumes

obtained for the molar pulp chamber (FDI

36), ranged from 31.7mm3 to 53.74mm

3,

with a SD of 5.7and a variance of 33.

For the volumetric reconstruction of the

canine (FDI 23), the volume value ranged

between 29.8 mm3 to 50.8 mm

3, SD=7.12

and variance of 50.75. The SD for the

different parameters was not bigger than

0.6 and the sample variance was not

greater than 0.3

Table 1. Volume results of using automatic segmentation (Vol 1), manual segmentation using

just the first and last slice (Vol 2) and manual segmentation each forth slice (Vol3), of cone

beam computer tomography from upper canines (FDI 13, 23)

Individual Vol 1 (mm3) Vol 2 (mm3) Vol 3 (mm3)

1 10.87 10.6 13.2

2 16.03 17 19.7

3 16.4 20.9 19.8

4 9.309 12.9 14.8

5 5.478 7.1385 8.04

6 11.25 12.5 12.6

7 13.26 25.9 16.3

8 16.72 17 23.2

9 37.01 32.6 41.9

10 16.98 23.4 28.9

11 11.77 19.2 17.7

12 5.857 8.7131 11.7

13 9.107 13.3 28.3

14 17.24 33 29.8

15 12.76 27.5 21.1

16 15.87 16.1 27

17 13.2 19.7 22.6

18 11.61 19.7 16.5

19 24.52 35.3 40

20 15.51 22.1 22.1

21 17.6 29.3 35.5

Mean 14.68 20.18 22.41

DISCUSSION

With the introduction of computer vision

techniques for medical imaging and their

application in dentistry and forensic

sciences, the possibilities to propose

methods for age and sex estimation are

almost unlimited. Owed to the apposition

of secondary dentine, the anthropometric

assessment of the variation in the ratio

between size of the pulp chamber and size


41

of the tooth is still an acceptable dental age

predictor.14

The aim of this study was to

test the variability of the volume

measurements when different

segmentation methods are applied in pulp

volume reconstruction, because although it

would be expected that with volume

measurements the results for age

estimation were more reliable, the

literature shows the opposite. To find out

an explanation to the lower accuracy of the

methods based on volume reconstruction,

in this study we tested three different

methods doing only pulp volume

reconstruction.

Table 2. Cone shape approach results using the tool polygon (Pol) and oval (Oval). Pulp area

(in mm2) at the CEJ level and volume (in mm

3) calculation with the measured root length (in

mm).

Area Area Pol Area Oval Root length Vol Pol Vol Oval

1 22 20 1.55 11.2 10.2

2 29 30.5 1.46 14.1 14.8

3 28 24.5 1.46 13.6 11.9

4 20 9.85 1.51 10.0 4.9

5 21 16 1.04 7.3 5.5

6 16 13.5 1.56 8.3 7.0

7 46 29 1.86 28.6 18.0

8 37 20.5 1.64 20.3 11.2

9 52 41 1.38 24.0 18.9

10 32 26.5 1.63 17.4 14.4

11 30 17 1.28 12.8 7.2

12 34 16 1.38 15.8 7.4

13 32 22 2.10 22.4 15.4

14 28 14 1.71 16.1 8.0

15 42 28.5 1.79 24.9 16.9

16 27 16 1.71 15.4 9.1

17 25 17.5 1.96 16.2 11.3

18 35 25.5 1.47 17.2 12.5

19 58 44.5 1.73 33.9 26.0

20 39 26.5 1.52 19.8 13.4

21 46 24.5 2.17 33.3 17.7

SD 8.7 10.86 7.45 5.2

The methodology analysed in this study,

using CBCT, seems to be no more

accurate than those methods based on

radiographs for dental age estimation in

adults, regardless the sample size of the

study which ranges from 19 individuals,

analysing only 12 canines7 to 403

individuals analysing only the pulp

chamber of first molars.14

This is clear

when the reported accuracy of the different


42

methods is compared. The reported mean

absolute error using automatic

segmentation of the first molar’s pulp

chamber is 6.26 years at best.14

The cone

shape geometric approach of pulp and

tooth has a standard error of estimation of

± 11.45 years,13

and finally, manual

segmentation of pulp and tooth, has a

prediction interval of ± 12 years.5

Table 3. Setting parameter combinations used in part C

Repetition

Scale of

Gaussian

Edge

contrast

Edge

transformation

exponent.

Smoothing

force

Average

molar volume

mm3

Average

canine

volume mm3

1 2.14 0.09 4 0.2 38.415 36.58

2 3 0.05 3 0.2 45.775 49.7

3 2.05 0.05 3 0.2 34.525 49.7

4 2.5 0.05 3 0.2 46.075 41.81

5 2.5 0.1 2.5 0.2 49.38 48.87

6 2.5 0.2 2.05 0.2 53.74 48.94

7 2 0.1 2.5 0.2 43.55 39.27

8 2.15 0.09 4 0.2 36.53 40.26

9 2.15 0.09 4 0.3 38.94 39.76

10 2.15 0.09 4 0.35 42 39.38

11 3 0.05 3 0.3 47.29 50.89

12 3 0.05 3 0.35 43.555 48.92

13 3 0.05 3 0.4 44.69 48.9

14 2.05 0.05 3 0.3 38.695 35.98

15 2.05 0.05 3 0.35 38.79 35.89

16 2.05 0.05 3 0.4 38.46 36.04

17 1.5 0.05 2.51 0.2 34.3 31.92

18 1.5 0.05 2.5 0.2 34.955 30.08

19 1 0.1 3 0.2 31.73 29.83

20 2 0.05 4 0.2 36.405 35.10

SD 0.53 0.03 0.59 0.07 5.74 7.12

Variance 0.28 0.001 0.35 0.005 32.94 50.75

There are 3 parameters available in ITK-SNAP software that are not listed, as they were constant for all the

reconstructions: Edge attraction force (=2) and step size (=1) and smoothing force which was changed from 1 to 0.5 only for the 18th combination.

Other methods also using automatic

segmentation of canines reported a

prediction interval of 3.47 years at best.17

When these reported results are compared

with some of the most commonly used

adult age estimation methods utilising

dental radiographs, such as Cameriere et

al.3 (with reported mean predictor errors of

2.37 years18

to 11.01 years19

(tooth/pulp

area ratio)), or the method developed by

Kvaal et al.2 (pulp/tooth-linear

measurements ratios with reported

standard estimation errors of ±9.367

years,20

) it is possible to observe that the


43

use of CBCT and dental structures volume

reconstruction, do not improve the final

results for adult age estimation. The

relation between the lack of the accuracy

of the mentioned methods for age

estimation and the different methods for

volume reconstruction are presented as

follow:

Part A. Manual and automatic

segmentation may produce similar results,

as observed in this study (r=0.83). Manual

segmentation is time consuming, difficult

to perform and prone to be influenced by

the subjectivity of the observer, and

personnel trained in this aspect is

necessary.21

Additionally, it is not accurate

enough to estimate the variation of the

volume in small anatomic structures, such

as the root canal, as observed in this study.

Although there are different methods for

automatic image segmentation, the first

problem faced in the bio-medical area, is

the exclusive dependence on gradient or

intensity analysis without using anatomical

information.21

In the case of automatic

segmentation, with ITK-SNAP software,

although the software chooses which

pixels will be included in the

segmentation, the observer finally decides

how to adjust the parameters to control the

algorithms in the segmentation, making

this method prone to a certain degree of

subjectivity from the observers. These

parameters may change from one CBCT to

another. Unfortunately, in the previous

study using this software the values of the

setting parameters were not mentioned.14

In this study it was also observed that

although all the CBCT were obtained from

machines with the same reference, and

under similar exposure conditions (slice

thickness, Kv, mA and time of exposure),

and the same parameters for automatic

segmentation, the interaction of the seed

region-growing with the boundaries of the

structure to reconstruct may also differ

from one CBCT to another. For this

reason, it is not possible to follow the

recommendations of the ITK-SNAP

manufacturer, who recommends to keep

the same parameters among different

CBCT, and it is neither possible to

guarantee the uniformity of the

segmentations obtained.14

In this study, we also observe that small

changes in the values of the setting

parameters can generate significant

variation in the final obtained volumes

with automatic segmentation, even though

the segmentation was made on the same

CBCT, same tooth and same anatomic

region.

Which means that the final volume will be

always the same under the same

parameters in the same CBCT and

anatomic region, but these parameters

generally need to be modified from one

CBCT to another, and in the same way,

small changes in the parameters alter the

final volume.

Part B. With the aim of simplifying the

measurement of the tooth structure

volume, a cone shape geometric approach

was proposed (Part B).13

The developers of

this method were aware of the main

disadvantages of this proposal:

measurements inaccuracy and subjectivity.

They also reported that this method tends

to underestimate the real volume of the

tooth structures (56% to 67%), which in

theory would not affect the final pulp/tooth

ratio that was calculated to estimate the

age.

To test this method, we used two differ

approaches for doing the volume

calculation. The first, following the

published methodology13

displaying ovals


44

at the CEJ, and a second one selecting

manually the outline of the root canal at

the CEJ, using the tool polygon. As

observed in the results section, despite the

difference between both area

measurements, at the CEJ, once the

volume is calculated the difference

decreases. Neither of these two

approaches is recommended to carry

outpulp volume measurements, , firstly

because of the subjectivity of the observer,

who only counts with the naked eye to

display one or another and secondly,

because both bring a distorted

representation of the root canal, ignoring

the different variations in its internal

shape.

Part C. In this study, it has been shown,

that there are different parameters that

affect the final volumetric reconstruction

of any anatomical region. The scale of

Gaussian, or blurrier factor22

was found to

be the most relevant when doing automatic

segmentation. To increase the value of the

Gaussian scale, allows the observer to

eliminate certain noises in the image,

which affect the “growing of the seed”.

However, the bigger the value the

Gaussian scale is, the more principal

structures and details will be lost.23

This

would mean that the initial volume of the

pulp chamber will be magnified.16

Although automatic segmentation is less

prone to the subjectivity of the observer

than manual segmentation, in this study we

observed that it is not possible to keep the

same volume reconstruction parameters for

all the CBCT, and that minimum changes

may produce significant variations even

though they are done on the same tooth.

The volume obtained with this tool must

be understood as an approximation to the

real volume of the structure being

measured

Taking into consideration the reported rate

of secondary dentin formation is 6.5

μm/year for the crown and 10 μm/year for

the root,24

the analysed methods in this

study are not sensitive enough to detect

such small dimension changes, generating

a big variation of the measurements, which

could explain the greater error of

pulp/tooth volume based methods for age

estimation when compared with simpler

methods based on linear2 or area

measurements3 on dental radiographs.

Previous research in the biomedical area

have highlighted the problems related to

semi-automatic segmentation, and

specifically, to the process that the

software uses to select which pixels or

voxels to be included in the final volume

reconstruction. To overcome this problem,

they have elaborated the probabilistic

anatomic atlas, which facilitates the

segmentation of an organ of interest.21

When doing the semi-autographic

segmentation of teeth and pulp chambers,

different obstacles were observed. Firstly,

owing to the similar density of dentine and

bone and the irregularity of the periodontal

ligament, it was not possible to obtain a

volumetric reconstruction of the tooth as it

is in for the pulp chamber. This is the

reason why the automatic tooth volume

reconstruction was not analysed.

Secondly, owing to the large variety in the

shape of the pulp chamber in the apical

third of the tooth, there was always a leak

of the seed region -growing in to the

dentine. To create a probabilistic atlas for

dentistry could help to overcome these

issues, and help the dentist and forensic

dentist to implement more accurate and

easier methods for automatic segmentation

and volume reconstruction. Thirdly,

although the automatic reconstruction of

the pulp chamber of molars may be

understood as a simple process, it may be


45

more reliable to use other methods for age

estimation when using molars.

CONCLUSIONS

Based on the results of this study, it is

possible to affirm that the obtained volume

measurements with any segmentation

method must be understood as an

approximation of the measured structure

and not as the real volume. In the same

way, manual segmentation could produce

volume values that are even more

inaccurate, and should be avoided in future

studies.

Although the volumetric reconstruction

and P/T volume ratio has not shown better

results than methods based on dental

radiographs, it is worth to persevere with

the research in this area with an

interdisciplinary team to build software

that can reliably reconstruct pulp and tooth

volumes for forensic purposes.

ACKNOWLEDGEMENTS

The authors of the study want to thank to

Nurul Firdaus, Dr Shahida Mohd Said and

Dr Shalini Kanagasingam from the

Radiology Unit, Dental Faculty University

of Kebangsaan Malaysia. To the Forensic

Science Centre from The University of

Western Australia and to the research

group INVIENDO from the National

University of Colombia.

REFERENCES


2. Kvaal SI, Kolltveit KM, Thomsen IO, Solheim T. Age estimation of adults from dental radiographs. Forensic Sci Int. 1995;74(3):175-85. 3. Cameriere R, Ferrante L, Cingolani M. Variations in pulp/tooth area ratio as an indicator of age: a preliminary study. J Forensic Sci. 2004;49(2):317-9. 4. Vandevoort FM, Bergmans L, Van Cleynenbreugel J, Bielen DJ, Lambrechts P, Wevers M, et al. Age calculation using X- ray microfocus computed tomographical scanning of teeth: a pilot study. J Forensic Sci. 2004;49(4):787. 5. De Angelis D, Gaudio D, Guercini N, Cipriani F, Gibelli D, Caputi S, et al. Age estimation from canine volumes. Radiol Med. 2015:1-6.

6. Paewinsky E, Pfeiffer H, Brinkmann B. Quantification of secondary dentine formation from orthopantomograms—a contribution to forensic age estimation methods in adults. Int J Legal Med 2005;119(1):27-30. 7. Yang F, Jacobs R, Willems G. Dental age estimation through volume matching of teeth imaged by cone-beam CT. Forensic Sci Int. 2006;159, Supplement(0):S78-S83. 8. Oi T, Saka H, Ide Y. Three-dimensional observation of pulp cavities in the maxillary first premolar tooth using micro-CT. Int Endod J. 2004;37(1):46-51. 9. Agematsu H, Someda H, Hashimoto M, Matsunaga S, Abe S, Kim H-J, et al. Three-dimensional Observation of Decrease in Pulp Cavity Volume Using Micro-CT: Age-related Change. Bull Tokyo Dent Coll. 2010;51(1):1-6.

10. Someda H, Saka H, Matsunaga S, Ide Y, Nakahara K, Hirata S, et al. Age estimation based on three-dimensional measurement of mandibular central incisors in Japanese. Forensic Sci Int. 2009;185(1–3):110-4. 11. Jagannathan N, Neelakantan P, Thiruvengadam C, Ramani P, Premkumar P, Natesan A, et al. Age estimation in an Indian population using pulp/tooth volume ratio of mandibular canines obtained from cone beam computed tomography. J Forensic Odontostomatol. 2011;29(1):1-6. 12. Star H, Thevissen P, Jacobs R, Fieuws S, Solheim T, Willems G. Human Dental Age Estimation by Calculation of Pulp–Tooth Volume Ratios Yielded on Clinically Acquired Cone Beam Computed Tomography Images of Monoradicular Teeth. J Forensic Sci. 2011;56(s1):S77-S82.

13. Pinchi V, Pradella F, Buti J, Baldinotti C, Focardi M, Norelli G-A. A new age estimation procedure based on the 3D CBCT study of the pulp cavity and hard tissues of the teeth for forensic purposes: A pilot study. J Forensic Legal Med. 2015;36:150-7. 14. Ge Z-p, Ma R-h, Li G, Zhang J-z, Ma X-c. Age estimation based on pulp chamber volume of first molars from cone-beam computed tomography images. Forensic Sci Int. 2015;253:133.e1-.e7. 15. Cameriere R, Ferrante L, Belcastro MG, Bonfiglioli B, Rastelli E, Cingolani M. Age Estimation by Pulp/Tooth Ratio in Canines by Mesial and Vestibular Peri-Apical X-Rays. J Forensic Sci. 2007;52(5):1151-5. 16. Yushkevich PA, Piven J, Hazlett HC, Smith RG, Ho S, Gee JC, et al. User-guided 3D active contour segmentation of anatomical structures: Significantly improved efficiency and reliability. NeuroImage. 2006;31(3):1116-28.


46

17. Tardivo D, Sastre J, Catherine J-H, Leonetti G, Adalian P, Foti B. Age determination of adult individuals by three-dimensional modelling of canines. Int J Legal Med. 2014;128(1):161-9. 18. Cameriere R, Cunha E, Sassaroli E, Nuzzolese E, Ferrante L. Age estimation by pulp/tooth area ratio in canines: Study of a Portuguese sample to test Cameriere's method. Forensic Sci Int. 2009;193(1–3):128.e1-.e6. 19. Babshet M, Acharya AB, Naikmasur VG. Age estimation in Indians from pulp/tooth area ratio of mandibular canines. Forensic Sci Int. 2010;197(1–3):125.e1-.e4. 20. Karkhanis S, Mack P, Franklin D. Age estimation standards for a Western Australian population using the dental age estimation technique developed by Kvaal et al. Forensic Sci Int. 2014;235(0):104.e1-.e6.

21. Dong C, Chen Y-w, Foruzan AH, Lin L, Han X-h, Tateyama T, et al. Segmentation of liver and spleen based on computational anatomy models. Comput Biol Med. 2015;67:146-60. 22. Tschirsich M, Kuijper A. Notes on Discrete Gaussian Scale Space. J Math Imaging Vis. 2015;51(1):106-23. 23. Guerrero González N. Segmentación de imágenes empleando el espacio de escala Gaussiano. Diss. Universidad Nacional de Colombia-Sede Manizales. 24. Murray PE, Stanley HR, Matthews JB, Sloan AJ, Smith AJ. Age-related odontometric changes of human teeth. Oral Surg Oral Med Oral Pathol Oral Radiol Endod. 2002;93(4):474-82.

*****

261

APPENDIX 7. A new proposal to overtake population group differences for dental age

estimation in adults through pulp/tooth volume calculation. Pilot Study

In press version. Under revision.



differences for dental age estimation in adults through pulp/tooth volume calculation. Pilot

Study. Submitted to Journal of Forensic Sciences.

A new proposal to overtake population group differences for dental age estimation in adults

through pulp/tooth volume calculation.

Abstract

Objectives.

The formation of secondary dentine on the walls of the root canal, has been used as an age

indicator in adults. The aim of this study is to propose a model for dental age estimation in

adults, regardless of their nationality or ethnicity by volume reconstruction of only a fraction of

the tooth to obtain linear regression models and an equation for dental age estimation.

Materials and Methods

Eighty-one anonymized cone beam computed tomography images obtained from two different

population groups were used. Only sex and age information was known. One group had a Latin-

American background (Colombian individuals aged 23 to 71 years) and the other had an Asian

background (Malaysian individuals aged 15 to 58 years) The analyzed tooth was the maxillary

canine. The used software was ITK-SNAP version 3.4 (open source software, www.itksnap.org),

to do automatic volume reconstruction of the cervical third of root and root canal.

Results

Analysis of the samples showed that the ages were unequally distributed in the two samples, but

by combining them a more equal age distribution was obtained. The correlation coefficient

between pulp/pulp+tooth volume ratio increased when data from individuals of both

populations were included in the same statistical analysis (R2=0.42). A formula for age estimation

regardless the origin of the individual was proposed: Age=67.104+(-434.741 x p/pt) (Standard

error of estimation=±11.4 years)

Conclusion

It has been established that methods for age estimation must be population specific. This study

presents an analysis that included data from individuals that are ethnically different and

geographically separated, obtaining promising results.

Key words:Computed tomography, secondary dentine, age estimation, ethnic variation, Forensic

Anthropology, Population Data

*Manuscript (without author details)

Click here to view linked References

Introduction

Much has been said in biological sciences about race and ethnicity since 1775 when Johann

Friedric Blumenbach defined five different races: Caucasians (whites), Mongolians (East Asians)

Malayans (South Asians), Negroids (Black Africans) and Americans (First Nations) (1), since then,

different studies followed this structure to analyze and interpret data and present results,

phenomena that has been named as scientific racism (2). Owed to the different nomadic and

migration processes in human history, this racial classification is day by day less accurate and

practical (2). Nevertheless, in forensic sciences it is still stipulated that the ethnic background of

an individual is of paramount relevance, to formulate a biological profile, and to clarify the

identity of living or deceased individuals, next to sex, stature and age (3). In terms of age

estimation, many methods have been proposed for individuals of different ages, based on bone

and tooth formation or degenerative changes with age (4).

Methods founded on the radiological analysis of dental development, have shown acceptable

accuracy, owed to the genetically controlled process of tooth formation, which seems to be

similar for individuals from different group of populations (5). The most recent systematic

review about age estimation based on tooth and bone formation (6,7) have stated that those

studies claiming that there are differences among group of populations may be incurring on a

selection bias inducing a phenomenon called age mimicry (6, 7), which has caused the erroneous

need to generate specific population formulae for age estimation, fact that could be also

affecting the results of studies for dental age estimation in adults

Simultaneously, age estimation in adults, by means of non-invasive analysis, is still a challenge for

scientists of different areas, specially, those who are involved in forensic investigations. Although

different methods have been proposed, there is no consensus about a unique method to be

applied to individuals from different countries, and even more, it has been claimed that methods

for age estimation must be population specific (8). Additionally, the few changes that bones and

teeth have after body maturation, are a setback for the accuracy of the existing methods.

One of the first methods for dental age estimation in adults, stablished that the formation of

secondary dentine on the wall of the root canal could be considered an age predictor in adults

(9). Based on the negative correlation of pulp/tooth ratio size with age, different non-invasive

methods were proposed, doing first, linear measurements (10), then area measurements, in

dental radiographs (11) and lastly volume measurements from dental tomography (12).

The aim of this study is to generate a model for dental age estimation in adults, regardless their

nationality or ethnic background, by reconstructing only a fraction of the tooth to obtain linear

regression models and an equation for age estimation.

Materials and methods.

The sample for this cohort study encompasses 81 cone beam computed tomography (CBCT)

obtained from two different population groups. Latin-American (Colombian individuals aged 23

to 71 years, 60% female (n=23) and 40% male (n=15)) and Asian (Malaysian sample individuals

aged 15 to 58 years 46% female (n=20) and 54% male (n=23)) previously collected for different

dental diagnosis and treatment procedures, with the respective informed consent. All the images

were anonymized keeping information about sex and age.

Table 1 shows the age and country distribution. Although the CBCT were obtained in different

institutions, in both cases the reference of the used device was the same: Kodak device,

reference K9000-3D (exposure parameters: scanning time: 10.8 s, 3-15 mA, 60-85kVp, field of

view FOV: 5.2 cm x 5.2cm to 5.5cm to 5.5cm, 180° rotation, slice thickness 0.076-0.2 mm) The

images were shared via online under the strictest security parameters to warrantee the integrity

and safety of the data.

The analyzed tooth was the maxillary canine, left or right, one per individual, as previous studies

have shown no right/left difference in tooth secondary dentine production (13). The

reconstructed root fraction was in all the cases 5 mm from the highest point of the vestibular

cemento-enamel junction, towards the root apex, with the aim to include only the cervical third

of the root.

Inclusion criteria: The included roots portions were from sound teeth, teeth with no evidence of

physical trauma, free of active caries, free of evidence of periapical disease, without signs of pulp

pathology and free of cervical lesson of any nature. The analysis also included teeth with small

coronal fillings, with at least 2 mm of dentine between the tooth restoration and the pulp

chamber, which is assumed to have negligible effect on the root portion.

The used software was ITK-SNAP version 3.4 (open source software, www.itksnap.org), doing

automatic reconstruction of the root canal and root. In certain cases, it was necessary to

manually correct the root segmentation. The measurements were collected by only one

observer, who had been previously trained in the use of this software. All the data were

collected in an excel spreadsheet, using Excel (version 2013 Microsoft Redmont, USA), before

and after manual correction of the image segmentation. Data analysis was performed with R

Core Team version 3.1.3 (2015). (R: A language and environment for statistical computing. R

Foundation for Statistical Computing, Vienna, Austria. URL http://www.R-project.org/).

Results

Pearson correlation coefficient test was calculated to observe if there was any difference

between the volume values obtained for pulp and tooth doing automatic segmentation and after

manual correction slide by slide. It was found that there is not significant difference between

both measurements (r=0.98). Two different proceedings were tested to obtain the pup/tooth

ratio. First dividing the volume of the pulp by the volume of the tooth, and second, dividing the

volume of the pulp by the sum of the volume of the tooth and pulp. It was observed that the

second approach generated a higher correlation coefficient (R2=0.34 vs R

2=0.42). The

determination coefficient between age and the pulp/pulp+tooth ratio (p/pt) for the individuals

of each group was R2=0.26 for the Colombian sample and R

2=0.38 for the Malaysian sample and

for the total sample using the data of both groups in the same analysis, R2=0.42.

In regards to tooth condition, it was observer that those teeth with small fillings did not generate

any outlier that could affect the linear regression model (Figure 1).

The linear regression models were generated to produce a formula for dental age estimation

from the combined sample. The formula was:

Age=67.104+(-434.741 x p/pt)

where p/pt is the ratio between the volume of the pulp divided by the sum of the volume of pulp

and tooth. The standard error of estimation (SEE) is ±11.4 years (R2=0.42)

Discussion

It is undeniable that dentistry has done a valuable contribution to anthropological sciences

bringing insights about human evolution. In the same way, dental analysis has become an

important tool for human identification in forensic sciences, in both fields, anthropology and

forensics, the concepts about racial and ethnic differences have deep roots and a significant

influence in research (14). However, in a changing world, where day by day the biological

boundaries between individuals of different backgrounds are constantly vanishing, to think

about methods for age estimation, that are regionally, racially, or ethnically focused must be re-

evaluated.

The last systematic review about Demirjian’s developmental stages of third molars (6) and the

Greulich and Pyle’s bone atlas (7) for age estimation in children and adolescents, exposes a

selection bias called age mimicry, present in numerous publications. This bias explains the

observed results variation, which has been wrongly associated to the population origin and

ethnicity of the participants, instead to the intrinsic characteristic of the used sample.

To avoid age mimicry, studies must have a study population with an even age distribution, large

samples, same number of individuals in each age group, appropriate upper and lower age limits

facilitating the observation of the different age related changes, and ideally samples that include

data of individuals from different regions, to really conclude whether regional differences are

important for age estimation in children (6) and probably in adults. Subsequently, it is necessary

to perform multi-population studies in adults. This research, is a good start to disclose the

population related differences in age estimation in adults when secondary dentine production is

analyzed as an age predictor.

In this study, it was found that the production of secondary dentine and the narrowing of the

root canal with age, which is one of the used parameters for dental age estimation in adults, has

a similar behavior in individuals that are not only ethnically different but also geographically

separated. Also, that the separate generation of regression models, for each one of the

population groups, would have generated age mimicry, owed to the unequal age distribution of

both samples, as in the Malaysian group the younger group dominates, while for the Colombian

group the middle age dominates (Table 1). By adding both samples it is possible to obtain a more

even age distribution and also a higher correlation coefficient between age and p/pt ratio

(R2=0.42). However, for future studies it is recommended to have a larger sample with the same

age distribution for the different populations to analyse, and same number of individuals among

the different age groups, avoiding age mimicry, which would generate more reliable results.

Up to now different methods and formulae have been published to estimate dental age in adults

(15). Nevertheless, few studies have tested the proposed formulae in groups of population

different from those included in the original studies. In the few studies that have done this

analysis, it has been suggested that local formulae bring more accurate results. (16) With the

results of the present study it could be possible to affirm that it is necessary to include the data

of individuals of different populations with equal age distribution in the same statistical analysis

to observe that there is not significant difference between population or ethnical groups, in

regards to the production of secondary dentine, and also to generate an equation that can be

used regardless the origin of the individual.

Furthermore, there are other factors that can affect the length, area and the total volume of the

tooth, as bruxism and attrition (17), among others, that also affect the production of dentine in

the crown and apical third of the root. These factors are not ethnically related but individual

dependent. By analyzing sound cervical root third of canines, or any other teeth, the problems

related with tooth attrition are diminished. In this study, the maxillary canine was selected as the

tooth to analyze, owed to the reported longer survival, compared to other teeth (18). It was also

selected to measure only the volume of the cervical third, as it has been reported that

pulp/tooth ratios, had a higher correlation with age at the cervical root level and that the

correlation with age decreased towards the apex, for different types of teeth, when doing linear

(19) or volume (20) measurements.

In the first studies, that analyzed the production of secondary dentine with age, (9) the proposed

methods required not only the use of extracted teeth but also the destruction of the sample,

which is unethical in a modern context. The use of different radiological techniques has allowed

researchers to analyze age indicators in a non-invasive and more conservative way. The use of

CBCT to analyze and obtain the volume of any anatomical structure is much easier, cheaper,

faster and has great potential for applications in anthropological and forensic studies. In this

study, this radiological technique allowed the observer to separate only a fraction of the tooth,

calculate its volume, and warrantee not only the integrity of the individual but also the

uniformity of the tooth sample size.

In terms of accuracy, although the obtained error exceeds the threshold that must accomplish a

method for age estimation in adults (±10 years) (21), the results of this study (SEE=±11.4 years)

are not so different to previous studies that used data of individuals from the same country or

belonging to the ethnic group (±12.5 years at best). (22)

Conclusion

It has been claimed that the modern world would testify the end of ethnicity, and also that it

changes according to circumstances, (14) This study proposes the use of data of individuals not

only from different countries but that also have been historically cataloged as members of

different racial and ethnical groups. It was found that by calculating the pulp/tooth+pulp volume

of only one part of the tooth it is possible propose a formula to estimate the age for adult

individuals regardless their origin.

Ethics

This study received ethics approval from the Human Research Ethics Committee of The

University of Western Australia (Ref: RA/4/1/6797).

References

1. Meer N. Key Concepts in Race and Ethnicity. London: SAGE Publications; 2014.

2. Wade P. Race and Ethnicity in Latin America. London: Pluto Press; 2010.

3. Franklin D. Forensic age estimation in human skeletal remains: Current concepts and

future directions. Legal Medicine. 2010;12(1):1-7.

4. Schmitt A, Cunha E, Pinheiro JES. Forensic Anthropology and Medicine. Totowa, NJ,

United States: Humana Press; 2007.

5. Liversidge HM, Chaillet N, Mörnstad H, Nyström M, Rowlings K, Taylor J, et al. Timing of

Demirjian's tooth formation stages. Annals of human biology. 2006;33(4):454

6. Rolseth V MA, Dahlberg PS, Ding KY, Bleka Ø, Skjerven-Martinsen M, Straumann, GH DG,

Vist GE. Demirjians utviklingsstadier på visdomstenner for estimering av kronologisk

alder: en systematisk oversikt. [Demirjian’s development stages on wisdom teeth for

estimation of chronological age: a systematic review.]. Oslo: Folkehelseinstituttet, 2017.

7. Dahlberg PS MA, Ding KY, Bleka Ø, Rolseth V, Straumann GH, Skjerven-Martinsen M,

Delaveris GJM, Vist GE. Samsvar mellom kronologisk alder og skjelettalder basert på

Greulich og Pyle-atlaset for aldersestimering: en systematisk oversikt. [Agreement

between chronological age and bone age based on the Greulich and Pyle-atlas for

ageestimation: a systematic review]. Oslo: Folkehelseinstituttet, 2017.

8. Senn DR. Senn, Weems R. Manual of forensic odontology. 5th ed. Boca Raton, FL: Boca

Raton, Press; 2013.



radiographs. Forensic Science International. 1995;74(3):175-85.

11. Cameriere R, Ferrante L, Cingolani M. Variations in pulp/tooth area ratio as an indicator

of age: a preliminary study. Journal of forensic sciences. 2004;49(2):317-9.

12. Star H, Thevissen P, Jacobs R, Fieuws S, Solheim T, Willems G. Human Dental Age

Estimation by Calculation of Pulp–Tooth Volume Ratios Yielded on Clinically Acquired

Cone Beam Computed Tomography Images of Monoradicular Teeth. Journal of forensic

sciences. 2011;56(s1):S77-S82.

13. Tardivo D, Sastre J, Ruquet M, Thollon L, Adalian P, Leonetti G, et al. Three-dimensional

Modeling of the Various Volumes of Canines to Determine Age and Sex: A Preliminary

Study. Journal of Forensic Sciences. 2011;56(3):766-70.

14. Ansell A. Race and Ethnicity: The Key Concepts. Abingdon, Oxon, Taylor and Francis;

2013.

15. Jeon H-M, Jang S-M, Kim K-H, Heo J-Y, Ok S-M, Jeong S-H, et al. Dental Age Estimation in

Adults: A Review of the Commonly Used Radiological methods. Journal of Oral Medicine

and Pain. 2014;39(4):119-26.

16. Patil S, Mohankumar K, Donoghue M. Estimation of age by Kvaal's technique in sample

Indian population to establish the need for local Indian-based formulae. Journal of

forensic dental sciences. 2014;6(3):171.

17. Vandevoort FM, Bergmans L, Van Cleynenbreugel J, Bielen DJ, Lambrechts P, Wevers M,

et al. Age calculation using X- ray microfocus computed tomographical scanning of teeth:

a pilot study. Journal of forensic sciences. 2004;49(4):787.

18. De Angelis D, Gaudio D, Guercini N, Cipriani F, Gibelli D, Caputi S, et al. Age estimation

from canine volumes. La radiologia medica. 2015:1-6.

19. Solheim T. Amount of secondary dentin as an indicator of age. European Journal of Oral

Sciences. 1992;100(4):193-9.

20. Aboshi H, Takahashi T, Komuro T. Age estimation using microfocus X-ray computed

tomography of lower premolars. Forensic Science International. 2010;200(1–3):35-40.

21. Willems G, Moulin-Romsee C, Solheim T. Non-destructive dental-age calculation methods

in adults: intra- and inter-observer effects. Forensic Science International.

2002;126(3):221-6.

22. Landa MI, Garamendi PM, Botella MC, Aleman I. Application of the method of Kvaal et al.

to digital orthopantomograms. International journal of legal medicine. 2009;123(2):123-

8.

Figure 1. Scatter plot showing the correlation between age and the pulp/pulp+tooth (p/pt) volume

ratio. (P/PT ratio) R2=0.42.

0.01

0.03

0.05

0.07

0.09

0.11

0.13

10 30 50 70

p/p

t ra

tio

Age

Figure 1

Age range Mal Col Total %

15-20 12 0 12 15

21-30 13 2 15 18

31-40 6 11 17 20

41-50 9 12 21 26

51-60 3 7 10 12.3

61-71 0 6 6 0.7

Total 43 38 81 100%

Table 1. Distribution of the individuals for age in years and country; Malaysia (Mal) and

Colombia (Col)

Table 1

Dental Age Estimation in Adults. An in-depth analysis of ...

Documents

Transcript of Dental Age Estimation in Adults. An in-depth analysis of ...