Topographic Asymmetry Indices: Correlation...

Research ArticleTopographic Asymmetry Indices Correlation betweenInferior Minus Superior Value and Index of Height Decentration

Sherine S Wahba 12 Maged M Roshdy 12 Ramy R Fikry 23 Mona K Abdellatif 1

and Abdulrahman M Abodarahim 4

1Faculty of Medicine Ain Shams University Cairo Egypt2Consultant of Ophthalmology Al Watany Eye Hospital Cairo Egypt3Faculty of Medicine Cairo University Cairo Egypt4Staff Physician Paediatric Ophthalmology King Abdullah Specialised Childrenrsquos Hospital Riyadh Saudi Arabia

Correspondence should be addressed to Mona K Abdellatif monakkkamalyahoocom

Received 9 April 2018 Revised 16 July 2018 Accepted 2 August 2018 Published 2 October 2018

Academic Editor Nora Szentmary

Copyright copy 2018 Sherine S Wahba et al -is is an open access article distributed under the Creative Commons AttributionLicense which permits unrestricted use distribution and reproduction in any medium provided the original work isproperly cited

Purpose To investigate the correlation between decentration index (index of height decentration IHD) automatically calculatedby the Pentacam HR software and the manually calculated inferior minus superior (I-S) value Setting Al Watany Eye HospitalCairo Egypt Methods In a retrospective study history taking clinical examination and rotating Scheimpflug camera scanning(by Oculyzer II equivalent to Pentacam HR) were done to 128 eyes 82 normal 24 forme fruste keratoconus FFKC (apparentlynormal cornea with evident keratoconus in the fellow eye) and 22 keratoconus (KC) All cases of corneal scars or previous cornealsurgeries were excluded -e (I-S) value was calculated manually from 10 points astride the horizontal meridian -e IHD iscalculated automatically by the device software 117r119 Results -e mean (plusmnSD) of (I-S) value in normal FFKC and KC eyeswere 030plusmn 093 011plusmn 203 and 462plusmn 389 respectively and those of IHD were 0008plusmn 0004 0011plusmn 0005 0066plusmn 0067respectively -e two indices were highly correlated (plt 00001) with a correlation coefficient (r2 0874) Deduced regressionformulae linking the two indices were calculated Conclusions -e two topographic decentration indices are highly correlatedDeduced formulae were proposed linking them

1 Introduction

Although clinical diagnosis of advanced keratoconus (KC) isrelatively easy with biomicroscopic and keratometric data itis rather complicated to rule out subclinical KC before re-fractive surgery Corneal topography provided a bettermeans of evaluating themorphologic change in patients withearly KC [1]

With the introduction of rotating Scheimpflug devicesfor corneal tomography many indices have been proposedfor discriminating corneas with subclinical KC from normalcorneas including curvature pachymetric elevation andmost recently biomechanical indices [2] -e most repu-table ectasia risk score system [3] uses the topographicasymmetry decentration index inferior minus superior (I-S)

value which is not calculated automatically by the PentacamHR software -erefore the topographic asymmetrydecentration indices have a special interest

-e aim of our study was to investigate the correlationbetween the manually calculated (I-S) value and the auto-matically calculated index of height decentration (IHD) byPentacam software and proposed deduced regression for-mulae linking the two indices

2 Patients and Methods

-is is a retrospective study that included 128 eyes 82of these eyes had normal corneas 24 of them were diagnosedas forme fruste KC (FFKC) (a cornea that has no abnormalfinding by clinical examination curvature elevation and

HindawiJournal of OphthalmologyVolume 2018 Article ID 7875148 4 pageshttpsdoiorg10115520187875148

pachymetry parameters with evident KC in the fellow eye)[4] and 22 as KC -e approach taken was through revisingthe charts looking for recorded history taking and clinicalexamination notes of candidates for refractive surgeries inour hospital All cases of corneal scars or previous cornealsurgeries were excluded

Every eye was scanned at least thrice by a rotatingScheimpflug camera (Allegro Oculyzer II equivalent toPentacam HR WaveLight Erlangen Germany) softwareversion 117r119 Each scan included 25 Scheimpflug im-ages Data were collected from only one scan which had thelargest analyzed area the highest percent of valid data andgood alignment

-e investigated indices included the following

(1) Inferior minus superior value (I-S value) [5] it is thekeratometric dioptric power difference between theaverage of five points of the inferior hemisphere andthe average of five points of the superior hemisphereIt was calculated manually astride the horizontalmeridian -e points lie on the 3mm radius circle atangular intervals of 30 degrees in the axial (sagittal)curvature display as described in Figure 1

(2) Index of height decentration (IHD) [6] it is theabsolute value of decentration of elevation data in thevertical direction (expressed in micrometres) It iscalculated automatically by the device

A validation sample of 78 eyes (52 normal 10 FFKC and16 KC) was then collected prospectively from patientscoming for screening or follow up -e right eyes only werechosen except in FFKC due to the paucity of these cases

3 Statistical Analysis

Data were collected and statistical analysis was performedusing MedCalc Statistics 12500 (MedCalc SoftwareOstend Belgium) Mean standard deviation (SD) MannndashWhitney test area under the receiver operating character-istic (AUROC) calculation and comparison by DeLong et almethod and Pearson correlation coefficient were calculatedLogistic regression using the enter technique was used todeduce formulae linking the two most correlated indices

Regarding the validation sample the AUROC medianand the 25 to 975 confidence interval of the absolutedifference between the actual and the deduced IHD werecalculated

-e study adhered to the Tenets of the Declaration ofHelsinki and to the local ethics committee (Watany ResearchDevelopment Center (WRDC)) standards

4 Results

-emean (plusmnSD) age of the patients was 285plusmn 75 years 41 eyesof males and 87 of females 65 were right eyes and 63 left eyes

-e mean (I-S) value in normal FFKC and KC eyes were030plusmn093 011plusmn203 and 462plusmn389 respectively and those ofIHD were 0008plusmn0004 0011plusmn0005 and 0066plusmn0067respectively

When differentiating normal from abnormal corneas(KC and FFKC collectively) the two indices gave accu-rate results (MannndashWhitney plt 00001) IHD had thehighest AUROC although no statistically significantsuperiority of one index over the other was confirmed-e same was found on differentiating normal fromFFKC (Table 1)

-e two indices were highly correlated (plt 00001) withcorrelation coefficient r2 0874

-e deduced regression formula linking the two indices(Figure 2) is as follows

IHD 0007334 + 0005676(I-S value)

+ 00007021(I-S value)2

middot r2

09282 residual SD 00094721113872 1113873

(1)

-e reverse formula (Figure 3) is as follows

I-S value minus04575 + 895389(IHD)minus 1118720(IHD)2

middot r2

07857 residual SD 119251113872 1113873

(2)

By applying the formula with the inferior half steeperthan superior (I-S) value of 05 1 and 14 correspond toIHD of 0010 0014 and 0017 taking into considerationthat IHD values gt0014 are considered abnormal andIHD values gt0016 are considered pathological [6](Table 2)

Regarding the validation sample the IHD (I-S) valueand the deduced IHD from the abovementioned formula arerepresented in Table 3

-e AUROC of IHD and (I-S) value were not statisticallysignificantly different from each other (Table 4)

-e absolute difference between the actual and the de-duced IHD is presented in Table 5

+2

+1

0

ndash1

ndash2

330deg

300deg240deg

210deg

180deg

150deg

120deg

270deg

90deg60deg

30deg

0deg

Figure 1 (I-S) value calculated astride the horizontal meridian

2 Journal of Ophthalmology

5 Discussion

Corneal topography has been found to be sensitive fordetecting subtle changes on the anterior corneal surface dueto corneal ectatic disorders before the appearance of clinicalsigns [2] Rabinowitz [7] had set useful topographic criteriato differentiate between normal and KC suspect eyes One ofthese criteria is the curvature asymmetric decentration (I-S)

value with a positive value indicating steeper curvature of theinferior half of the cornea Value of 14 is defined as a cutoffpoint for suspected KC [8]

Although elevation-based tomography is adding knowl-edge about the elevation and pachymetry indices of thecornea still the curvature analysis is very important as thetopographic abnormalities are very important evaluatedcriteria in post-LASIK ectasia and are involved in the ectasiarisk score system [3]-emembers of the American Academyof Ophthalmology (AAO)International Society of RefractiveSurgery (ISRS)American Society of Cataract and RefractiveSurgery (ASCRS) joint committee [9] recommended avoidingLASIK in patients with asymmetric inferior corneal steep-ening or asymmetric bowtie patterns with steep axes skewedabove and below the horizontal meridian

Moreover it is one of the four components involved inKISA used in KC identification which demonstratedsensitivity and specificity of 96 and 100 respectively interms of keratoconus diagnosis [10]

One of the currently used elevation-based tomographydevices is the Pentacam HR it provides analysis of topo-metric indices including the automatically calculateddecentration IHD index IHD is not only an index to detectcorneal asymmetry and consequently important in KCdetection but also it was reported as one of the mostsensitive and specific criteria for follow up of KC [6]

Our study revealed a high correlation between the (I-S)value and the IHD

As the Pentacam software does not include the (I-S)value and there is a high correlation between the (I-S) valueand IHD we deduced regression formulae between them tobe able to calculate one from the other with reasonableaccuracy -is could enable calculating the Randlemanectasia risk score system directly from Pentacam display

Table 1 -e AUROC of decentration indices

IndexNormal versus KC+FFKC Normal versus FFKC

AUROCAUROC compared to that of IHD (p value)


Mean SD Mean SD(I-S) value 0768 0048 0092 0582 0071 0135IHD 0840 0039 mdash 0700 0061 mdashKC keratoconus FFKC forme fruste keratoconus AUROC area under the receiver operating characteristic SD standard deviation (I-S) val-ue inferior minus superior value IHD index of height decentration

ndash10 ndash5 0 5 10 15 20ndash01

00

01

02

03

I-S astride horizontal

IHD

Figure 2 -e regression formula deducing the IHD from the (I-S)value with the 95 prediction lines

00 01 02 03ndash10

ndash5

0

5

10

15

20

IHD

IndashS

astr

ide h

oriz

onta

l

Figure 3-e regression formula deducing the (I-S) value from theIHD with the 95 prediction lines

Table 2 -e corresponding values of IHD(I-S) value 05 10 14IHD 0010 0014 0017(I-S) value inferior minus superior value IHD index of heightdecentration

Table 3 Validation sample data

Median 25ndash97 confidence intervalIHD 0017 0003ndash0151(I-S) value 101 0061ndash6475Deduced IHD 0014 0008ndash0074

Journal of Ophthalmology 3

which is currently not possible We listed the (I-S) value innormal and KC suspect with the corresponding IHD valuesas an example

-is regression formula was validated with a small ab-solute difference between the actual and the deduced IHD Itwas higher in KC eyes mostly due to relative higher verticalasymmetry in these eyes -erefore the difference wasunlikely to change the diagnosis

To our knowledge the current study is the first study toevaluate the correlation between these asymmetric decen-tration indices and to deduce the formulae linking them

6 Conclusion

-e two decentration asymmetric indices are highly cor-related and they can be deduced from each other

Data Availability

All data generated or analyzed during this study are includedwithin the article

Conflicts of Interest

Maged M Roshdy received travel support from AlconLaboratories Orchidia Pharma and Bayer Ramy R Fikryreceived travel support from Novartis -e authors declarethat they have no conflicts of interest regarding the publi-cation of this paper

Acknowledgments

Instruments used in the study belong to Al Watany EyeHospital and Watany Research Development Center(WRDC) Cairo Egypt

References

[1] S J Ahn M K Kim and W R Wee ldquoTopographic pro-gression of keratoconus in the Korean populationrdquo KoreanJournal of Ophthalmology vol 27 no 3 pp 162ndash166 2013

[2] R Ambrosio Jr and M W Belin ldquoImaging of the corneatopography vs tomographyrdquo Journal of Refractive Surgeryvol 26 no 11 pp 847ndash849 2010

[3] J B Randleman M Woodward M J Lynn andR D Stulting ldquoRisk assessment for ectasia after corneal re-fractive surgeryrdquo Ophthalmology vol 115 no 1 pp 37ndash502008

[4] R Ambrosio Jr M W Belin V L Perez J C Abad andJ A P Gomes ldquoDefinitions and concepts on keratoconus andectatic corneal diseases panamerican delphi consensusndashaPilot for the global consensus on ectasiasrdquo InternationalJournal of Keratoconus and Ectatic Corneal Diseases vol 3no 3 pp 99ndash106 2014

[5] R Martin ldquoCornea and anterior eye assessment with placido-disc keratoscopy slit scanning evaluation topography andscheimpflug imaging tomographyrdquo Indian Journal of Oph-thalmology vol 66 no 3 pp 360ndash366 2018

[6] A J Kanellopoulos and G Asimellis ldquoRevisiting keratoconusdiagnosis and progression classification based on evaluationof corneal asymmetry indices derived from scheimpflugimaging in keratoconic and suspect casesrdquo Clinical Oph-thalmology vol 7 pp 1539ndash1548 2013

[7] Y S Rabinowitz ldquoKeratoconusrdquo Survey of Ophthalmologyvol 42 no 4 pp 297ndash319 1998

[8] Y S Rabinowitz and P J McDonnell ldquoComputer-assistedcorneal topography in keratoconusrdquo Refractive and CornealSurgery vol 5 no 6 pp 400ndash408 1989

[9] P S Binder R L Lindstrom R D Stulting et al ldquoKerato-conus and corneal ectasia after LASIKrdquo Journal of Cataract ampRefractive Surgery vol 31 no 11 pp 2035ndash2038 2005

[10] M R Sedghipour A L Sadigh and B F Motlagh ldquoRevisitingcorneal topography for the diagnosis of keratoconus use ofRabinowitzrsquos KISA indexrdquo Clinical Ophthalmology vol 6p 181 2012

Table 4 Validation sample AUROC

IndexNormal versus KC+ FFKC Normal versus FFKC

AUROC AUROC compared to that ofIHD (p value)

AUROC AUROC compared to that ofIHD (p value)Mean SD Mean SD

(I-S) value 0858 0050 0710 0709 0104 0648IHD 0868 0048 mdash 0735 0100 mdash

Table 5 -e absolute difference between the actual and the deduced IHD

Number of eyes Median of theactual IHD

Median of the absolute differencebetween actual and deduced IHD

25ndash975 confidence interval of theabsolute difference between actual

and deduced IHDAll eyes 78 0017 0004 0003ndash0007Normal eyes 52 0015 0003 0002ndash0005FFKC 10 0024 0007 0001ndash0015KC 16 0074 0032 0012ndash0063


Stem Cells International

Hindawiwwwhindawicom Volume 2018


MEDIATORSINFLAMMATION

of

EndocrinologyInternational Journal of



Disease Markers


BioMed Research International

OncologyJournal of



Oxidative Medicine and Cellular Longevity


PPAR Research

Hindawi Publishing Corporation httpwwwhindawicom Volume 2013Hindawiwwwhindawicom

The Scientific World Journal

Volume 2018

Immunology ResearchHindawiwwwhindawicom Volume 2018

Journal of

ObesityJournal of



Computational and Mathematical Methods in Medicine


Behavioural Neurology

OphthalmologyJournal of


Diabetes ResearchJournal of



Research and TreatmentAIDS


Gastroenterology Research and Practice


Parkinsonrsquos Disease

Evidence-Based Complementary andAlternative Medicine

Volume 2018Hindawiwwwhindawicom

Submit your manuscripts atwwwhindawicom

pachymetry parameters with evident KC in the fellow eye)[4] and 22 as KC -e approach taken was through revisingthe charts looking for recorded history taking and clinicalexamination notes of candidates for refractive surgeries inour hospital All cases of corneal scars or previous cornealsurgeries were excluded

Every eye was scanned at least thrice by a rotatingScheimpflug camera (Allegro Oculyzer II equivalent toPentacam HR WaveLight Erlangen Germany) softwareversion 117r119 Each scan included 25 Scheimpflug im-ages Data were collected from only one scan which had thelargest analyzed area the highest percent of valid data andgood alignment

-e investigated indices included the following

(1) Inferior minus superior value (I-S value) [5] it is thekeratometric dioptric power difference between theaverage of five points of the inferior hemisphere andthe average of five points of the superior hemisphereIt was calculated manually astride the horizontalmeridian -e points lie on the 3mm radius circle atangular intervals of 30 degrees in the axial (sagittal)curvature display as described in Figure 1

(2) Index of height decentration (IHD) [6] it is theabsolute value of decentration of elevation data in thevertical direction (expressed in micrometres) It iscalculated automatically by the device

A validation sample of 78 eyes (52 normal 10 FFKC and16 KC) was then collected prospectively from patientscoming for screening or follow up -e right eyes only werechosen except in FFKC due to the paucity of these cases

3 Statistical Analysis

Data were collected and statistical analysis was performedusing MedCalc Statistics 12500 (MedCalc SoftwareOstend Belgium) Mean standard deviation (SD) MannndashWhitney test area under the receiver operating character-istic (AUROC) calculation and comparison by DeLong et almethod and Pearson correlation coefficient were calculatedLogistic regression using the enter technique was used todeduce formulae linking the two most correlated indices

Regarding the validation sample the AUROC medianand the 25 to 975 confidence interval of the absolutedifference between the actual and the deduced IHD werecalculated

-e study adhered to the Tenets of the Declaration ofHelsinki and to the local ethics committee (Watany ResearchDevelopment Center (WRDC)) standards

4 Results

-emean (plusmnSD) age of the patients was 285plusmn 75 years 41 eyesof males and 87 of females 65 were right eyes and 63 left eyes

-e mean (I-S) value in normal FFKC and KC eyes were030plusmn093 011plusmn203 and 462plusmn389 respectively and those ofIHD were 0008plusmn0004 0011plusmn0005 and 0066plusmn0067respectively

When differentiating normal from abnormal corneas(KC and FFKC collectively) the two indices gave accu-rate results (MannndashWhitney plt 00001) IHD had thehighest AUROC although no statistically significantsuperiority of one index over the other was confirmed-e same was found on differentiating normal fromFFKC (Table 1)

-e two indices were highly correlated (plt 00001) withcorrelation coefficient r2 0874

-e deduced regression formula linking the two indices(Figure 2) is as follows

IHD 0007334 + 0005676(I-S value)

+ 00007021(I-S value)2

middot r2

09282 residual SD 00094721113872 1113873

(1)

-e reverse formula (Figure 3) is as follows

I-S value minus04575 + 895389(IHD)minus 1118720(IHD)2

middot r2

07857 residual SD 119251113872 1113873

(2)

By applying the formula with the inferior half steeperthan superior (I-S) value of 05 1 and 14 correspond toIHD of 0010 0014 and 0017 taking into considerationthat IHD values gt0014 are considered abnormal andIHD values gt0016 are considered pathological [6](Table 2)

Regarding the validation sample the IHD (I-S) valueand the deduced IHD from the abovementioned formula arerepresented in Table 3

-e AUROC of IHD and (I-S) value were not statisticallysignificantly different from each other (Table 4)

-e absolute difference between the actual and the de-duced IHD is presented in Table 5

+2

+1

0

ndash1

ndash2

330deg

300deg240deg

210deg

180deg

150deg

120deg

270deg

90deg60deg

30deg

0deg

Figure 1 (I-S) value calculated astride the horizontal meridian


5 Discussion














00

01

02

03


IHD


00 01 02 03ndash10

ndash5

0

5

10

15

20

IHD

IndashS

astr

ide h

oriz

onta

l









6 Conclusion


Data Availability




Acknowledgments


References


























of




Disease Markers



OncologyJournal of





PPAR Research



Volume 2018


Journal of

ObesityJournal of



















5 Discussion














00

01

02

03


IHD


00 01 02 03ndash10

ndash5

0

5

10

15

20

IHD

IndashS

astr

ide h

oriz

onta

l









6 Conclusion


Data Availability




Acknowledgments


References


























of




Disease Markers



OncologyJournal of





PPAR Research



Volume 2018


Journal of

ObesityJournal of






















6 Conclusion


Data Availability




Acknowledgments


References


























of




Disease Markers



OncologyJournal of





PPAR Research



Volume 2018


Journal of

ObesityJournal of























of




Disease Markers



OncologyJournal of





PPAR Research



Volume 2018


Journal of

ObesityJournal of



















Topographic Asymmetry Indices: Correlation...

Documents

Transcript of Topographic Asymmetry Indices: Correlation...