Assessment of normal tissue complications following prostate cancerirradiation: Comparison of radiation treatment modalitiesusing NTCP models

Rungdham Takam and Eva Bezaka�

School of Chemistry and Physics, The University of Adelaide, Adelaide SA 5000, Australiaand Department of Medical Physics, Royal Adelaide Hospital, Adelaide SA 5000, Australia

Eric E. YeohSchool of Medicine, The University of Adelaide, Adelaide SA 5000, Australia and Department of RadiationOncology, Royal Adelaide Hospital, Adelaide SA 5000, Australia

Loredana MarcuSchool of Chemistry and Physics, The University of Adelaide, Adelaide SA 5000, Australia and Faculty ofScience, University of Oradea, Oradea 410086, Romania

�Received 13 August 2009; revised 20 July 2010; accepted for publication 21 July 2010;published 31 August 2010�

Purpose: Normal tissue complication probability �NTCP� of the rectum, bladder, urethra, andfemoral heads following several techniques for radiation treatment of prostate cancer were evalu-ated applying the relative seriality and Lyman models.Methods: Model parameters from literature were used in this evaluation. The treatment techniquesincluded external �standard fractionated, hypofractionated, and dose-escalated� three-dimensionalconformal radiotherapy �3D-CRT�, low-dose-rate �LDR� brachytherapy �I-125 seeds�, and high-dose-rate �HDR� brachytherapy �Ir-192 source�. Dose-volume histograms �DVHs� of the rectum,bladder, and urethra retrieved from corresponding treatment planning systems were converted tobiological effective dose-based and equivalent dose-based DVHs, respectively, in order to accountfor differences in radiation treatment modality and fractionation schedule.Results: Results indicated that with hypofractionated 3D-CRT �20 fractions of 2.75 Gy/fractiondelivered five times/week to total dose of 55 Gy�, NTCP of the rectum, bladder, and urethra wereless than those for standard fractionated 3D-CRT using a four-field technique �32 fractions of 2Gy/fraction delivered five times/week to total dose of 64 Gy� and dose-escalated 3D-CRT. Rectaland bladder NTCPs �5.2% and 6.6%, respectively� following the dose-escalated four-field 3D-CRT�2 Gy/fraction to total dose of 74 Gy� were the highest among analyzed treatment techniques. Theaverage NTCP for the rectum and urethra were 0.6% and 24.7% for LDR-BT and 0.5% and 11.2%for HDR-BT.Conclusions: Although brachytherapy techniques resulted in delivering larger equivalent doses tonormal tissues, the corresponding NTCPs were lower than those of external beam techniques otherthan the urethra because of much smaller volumes irradiated to higher doses. Among analyzednormal tissues, the femoral heads were found to have the lowest probability of complications asmost of their volume was irradiated to lower equivalent doses compared to other tissues. © 2010American Association of Physicists in Medicine. �DOI: 10.1118/1.3481514�

Key words: NTCP models, prostate radiotherapy

I. INTRODUCTION

The main therapeutic aim of all radiotherapy treatment tech-niques including those for prostate cancer is to maximizedamage to the tumor while, at the same time, keeping dam-age to the surrounding normal tissues as small as possible.During treatment planning, normal tissue complication prob-ability �NTCP� as well as tumor control probability �TCP�should be assessed, so as to optimize the therapeutic ratio ofany particular radiotherapy modality. Among plans whichhave similar TCP, the one with the lowest NTCP should beconsidered superior.

Many groups have published tumor control results follow-

ing various radiotherapy techniques �external beam and

5126 Med. Phys. 37 „9…, September 2010 0094-2405/2010/37

brachytherapy� based on biochemical and other clinical out-comes. For instance, Livsey et al.1 reported 5 yr overall sur-vival and disease-specific survival rate in patients with pros-tate cancer who received hypofractionated �3.13 Gy/fraction�four-field conformal radiotherapy of 83.1% and 91%, respec-tively. Kupelian et al.2 analyzed the long term relapse-freesurvival rates in the patients treated with hypofractionated�2.5 Gy/fraction for 70 Gy� radiotherapy using the intensitymodulated radiation therapy technique and observed 5 yroverall American Society for Therapeutic Radiology and On-cology biochemical relapse-free survival and Houston�nadir+2� biochemical relapse-free survival rates of 85%

and 88%, respectively.

5126„9…/5126/12/$30.00 © 2010 Am. Assoc. Phys. Med.

5127 Takam et al.: Normal tissue complications following prostate cancer irradiation 5127

Blasko et al.3 reported the 9 yr overall biochemical con-trol rate of 83.5% in a group of patients treated with low-dose-rate brachytherapy �LDR-BT� using palladium-103�Pd-103� for a minimum dose of 115 Gy. Twelve-year overalland disease-specific survival rates of 84% and 93%, respec-tively, were observed among patients treated with LDR-BTusing I-125 or Pd-103.4 Similarly, Zelefsky et al.5 reportedthe 8 yr nadir+2 prostate specific antigen-disease-free sur-vival �PSA-DFS� rates of 73%, 60%, and 41% for low, in-termediate, and high-risk prostate cancer patients, respec-tively, treated with I-125 LDR-BT �median dose of 160 Gy�.In addition, for those patients who received Pd-103 LDR-BT�median dose of 120 Gy�, the 8 yr nadir+2 PSA-DFS ratesfor low, intermediate, and high-risk patients were 73%, 64%,and 38%, respectively. Poor implant quality as reflected byD90 value �the dose received by 90% of target volume� mayhave contributed to slightly lower tumor control rates in thisstudy compared with previous reports.3,4

Mark et al.6 investigated the treatment outcomes of Ir-192high-dose-rate brachytherapy �HDR-BT� �45 Gy in six frac-tions� in localized prostate cancer patients and found that thePSA disease-free survival rate was 90.3%. For a comprehen-sive list of studies, see the review by Nilsson et al.7

For the treatment of prostate cancer, while TCP increaseswith increasing dose, the total radiation dose which can begiven to the prostate is limited by the tolerance of surround-ing normal tissues such as the bladder, rectum, urethra, andbowel. As shown above, although differences in dose levels,fractionation, and quality of treatment delivery can affect theefficacy of radiation treatment, clinical studies indicate thatcurrently used treatment techniques generally report similartumor control.7 As a result, assessment of NTCP values fororgans-at-risk �OARs� in association with each treatmentplan or technique would assist clinicians and patients in theselection of suitable treatment modality and dose per fractionfor a given treatment.

The purpose of the current study is to assess NTCP of therectum, bladder, urethra, and femoral heads, following radia-tion treatment of prostate carcinoma using the relative seri-ality and Lyman models. Real patient plans for externalbeam radiotherapy �EBRT�, LDR, and HDR brachytherapytechniques, which had evolved approximately over a 7 yrperiod at our center, were evaluated. Currently, EBRT andLDR brachytherapy are used as monotherapy but HDR iscombined with EBRT.

II. MATERIALS AND METHODS

II.A. Prostate treatment techniques and differentialdose-volume histograms

Contouring of normal tissues in all plans was carried outby one radiation oncologist only to ensure that all organswere contoured consistently for all patients and treatmenttechniques analyzed. While this does not eliminate intraob-server variability in contouring, this variability is minimizedby using absolute rather than percent volumes of the con-toured OARs in the dose-volume histograms �DVHs�.8 The

full extent of the rectum and bladder were contoured based

Medical Physics, Vol. 37, No. 9, September 2010

on CT slices of the entire pelvis obtained at 2–3 mm inter-vals in the axial plane. The rectum was defined as extendingfrom the anal canal to the rectosigmoid junction. Intravenouscontrast was used to assist in the definition of the bladder forcontouring purposes. The urethra was not contoured for theEBRT techniques as this normal tissue would have receivedthe same homogenous radiation dose as the prostate. Follow-ing the dose calculation, DVHs of the rectum, bladder, andurethra were exported from the corresponding treatmentplanning systems.

In total, 215 DVHs from 101 patients were analyzed inthis study. Real treatment plans of treated patients were usedin the current study. As a result, different groups of patientsare compared when analyzing individual radiotherapy tech-niques. While acknowledging that this introduces anothervariable into the study, it allows our risk estimates to becorrelated with patient data in the future. Details about eachtreatment technique are briefly described in Table I. Whilethe heterogeneity of the techniques is acknowledged, this hasresulted from the development and validation of treatmenttechniques reported in the medical literature. For example,HDR-BT is now acknowledged to have an emerging role inthe management of prostate cancer.9

Although HDR-BT as monotherapy is currently not yetavailable at our center, the promising results in terms oftreatment efficacy and low normal tissue toxicities ofHDR-BT as monotherapy �often prescribed as four fractionsof 9.5 Gy� for prostate cancer have been recentlyreported.10–16 In order to simulate the effect on NTCP usingHDR-BT as monotherapy, the original HDR-BT live treat-ment plans used for the combined modality treatment wereused as monotherapy plans by increasing the number of frac-tions �of 9.5 Gy� from two to four �same dose distributionwas assumed for each fraction�.

II.B. Biologically effective dose and equivalent doseconversion techniques

The probabilities of normal tissue complications were cal-culated from differential DVHs of organs-at-risk. The physi-cal dose-based differential DVHs from hypofractionatedthree-dimensional conformal radiotherapy �3D-CRT�,HDR-BT �live planning�, and LDR-BT �live planning� werefirst converted to biologically effective dose �BEffD�-baseddifferential DVHs �BEffDVHs� in order to normalize thedoses in the DVHs to the same biological end-point. Thisconversion was performed using the formalism developed byDale17

BEffD = D�RE, �1�

where D is the delivered physical dose �Gy� and RE is afunction of dose called relative effectiveness as defined inEqs. �2�–�4� in the following sections.

II.B.1. Dose conversion for hypofractionated 3D-CRT, HDR-BT, and LDR-BT

Relative effectiveness for hypofractionated 3D-CRT with17

dose per fraction d can be written as

T � � � � � � a

5128 Takam et al.: Normal tissue complications following prostate cancer irradiation 5128

Medical Physics, Vol. 37, No. 9, September 2010

RE = 1 +d

��/��, �2�

where � /� is the dose when the linear and quadratic com-ponents of cell killing are equal �using the linear-quadraticdose-response model�.

Relative effectiveness for a nonpermanent implant with adecaying radioactive source as used in HDR-BT is definedas17

RE = 1 + � 2R0�

� − ���

��� � �1 − e−��T��−1 � 1

2��1 − e−�2�T��

−1

� + ��1 − e−T��+��� , �3�

where R0 is the initial dose rate ��94.86 Gy /h�, � is thesource decay constant �0.000 39 h−1�, � is the rate of suble-thal damage repair �0.46 h−1�,18 and T is the total treatmenttime ��25 min per fraction�.

For a permanent implant with a �infinite� decaying sourceas used in LDR-BT, assuming that repair rates never exceedrates of double-strand break induction and without cell pro-liferation, the relative effectiveness in this case is definedas17

RE = 1 +R0

� + ���

�� , �4�

where, R0 is the initial dose rate �0.0704 Gy/h�, � is thesource decay constant �0.000 48 h−1�, and � is the rate ofsublethal damage repair.18

Following the BEffD conversion of DVHs, in order toaccount for the differences in dose fractionation schemessuch as between standard fractionated �2 Gy/fraction for 64Gy� and hypofractionated �2.75 Gy/fraction for 55 Gy�, andalso between HDR-BT and LDR-BT, BEffDVHs were sub-sequently converted to equivalent dose �Deq�-based differen-tial DVHs �DeqVHs�. Equivalent dose Deq for a particulardose delivering scheme is the dose which would be givenusing conventionally fractionated �2 Gy/fraction� irradiationfor the same biological effect. It can be calculated using theformalism developed by Nag and Gupta19

Deq =BEffD

�1 + dref/�/��, �5�

where dref is the reference dose per fraction for a convention-ally fractionated EBRT. In this study, the dref was 2 Gy/fraction.

The aim of conversion of physical doses in DVHs toBEffD and Deq in this study was to normalize the physicaldose from individual radiation treatment techniques to thedose which would produce the same biological end-point�BEffD� as that of the standard fractionated �2 Gy/fraction�dose schedule �Deq�.

For standard fractionated 3D-CRT or other EBRT tech-niques based on 2 Gy fraction delivering scheme, these doseconversions were not needed because the final Deq obtained

AB

LE

I.R

adia

tion

trea

tmen

tte

chni

ques

for

pros

tate

carc

inom

aat

Roy

alA

dela

ide

Hos

pita

l,R

adia

tion

Onc

olog

yD

epar

tmen

t,So

uth

Aus

tral

ia,

whi

chw

ere

invo

lved

inth

isst

udy.

Tre

atm

ent

tech

niqu

e�n

=no

.of

DV

H�

Pres

crip

tion

dose

�Gy�

Dos

e/fr

actio

nor

dose

rate

�Gy�

Bea

ms

arra

ngem

ent

orim

plan

tatio

nM

argi

n

Ave

rage

plan

ning

trea

tmen

tvo

lum

e�P

TV

�in

cm3

�ran

ge�

Tre

atm

ent

plan

ning

syst

em

1�St

anda

rdfr

actio

nate

d3D

-CR

T�n

=21

�64

218

MV

phot

ons

�Var

ian

2100

EX

�,fo

ur-fi

eld

�AP/

PA,

late

rals

�T

hepr

osta

tegl

and

with

a1.

5cm

95%

isod

ose

mar

gina

275.

0�

24.5

�249

.1–3

14.6

�Pi

nnac

le3

6.2b

�Phi

llips

Med

ical

Syst

em�

2�H

ypof

ract

iona

ted

3D-C

RT

�n=

30�

552.

75Sa

me

as�1

�Sa

me

as�1

�29

7.9

�55

.6�2

53.6

–429

.2�

Sam

eas

�1�

3�D

ose-

esca

late

d3D

-CR

T�n

=44

�70

or74

2Sa

me

as�1

�Sa

me

as�1

�19

8.0

�55

.7�1

12.9

–283

.8�

Sam

eas

�1�

4�Fi

ve-fi

eld

3D-C

RT

�n=

42�

702

18M

Vph

oton

s�V

aria

n21

00E

X�,

five-

field

�AP/

two

late

rals

,tw

oob

lique

s�Sa

me

as�1

�fo

rth

efir

st64

Gy,

then

pros

tate

glan

dw

ithno

mar

gin

201.

3�

85.7

�93.

0–39

6.9�

Sam

eas

�1�

5�L

DR

-BT

mon

othe

rapy

�I-1

25�

�n=

73�

145

7cG

yh−

1A

vera

gene

edle

sus

ed:

24;

aver

age

seed

sim

plan

ted:

70N

/AN

ucle

tron

SPO

T-PR

O™

�live

plan

ning

�

6�H

DR

-BT

mon

othe

rapy

�Ir-

192�

�n=

19�

389.

5N

ucle

tron

Mic

roSe

lect

ron

HD

RN

/AN

ucle

tron

SWIF

T™

�live

plan

ning

�

Ref

eren

ce8.

from BEffD conversions will be equal to the original physical

5129 Takam et al.: Normal tissue complications following prostate cancer irradiation 5129

doses. The original differential DVHs obtained from treat-ment planning system were used directly in these cases.

II.C. NTCP calculations

The relative seriality model was applied to most of theDeqVHs in this study, with the model parameters of the or-gans of interest for specified end points obtained from litera-ture. This model was chosen mainly because it accounts forthe architecture of the organ through parameter “s,” which isderived from the ratio of serial subunits to all subunits in theorgan.20 In this scheme, an organ where the substructures areorganized in series becomes nonfunctional when one sub-structure is damaged, while for a parallel organ, the probabil-ity of complication depends on the fraction of substructuresdamaged.21 Hence, the magnitude of volume irradiated to acertain radiation dose will strongly affect the final outcomeof irradiated normal organ. This is particularly important forbrachytherapy techniques where small volumes are exposedto high doses.

The following logistic function was used to estimate theNTCP:22–24

NTCP = 1 − �i�1 − � 1

1 + �D50/Deq,i�k�s��i/V1/s. �6�

D50 in the above equation represents the dose required toproduce 50% probability of specific tissue complications, �i

is the normal organ subvolume which received the equivalentdose Deq,i. Parameters s and “k” are empirically determinedNTCP parameters which dictate the seriality of the organstructural architecture and steepness of dose-response curve,respectively.

Model parameters used for calculations of NTCP for eachOAR are summarized in Table II.

The model parameters for calculation of the NTCP of theurethra are not readily available despite extensive reports ofurethral toxicity following various prostate cancer radio-

26

TABLE II. The default parameter values of the relainvolved in this study.

Parameters Rectum B

�1� � /� ratio 5.4 Gya 7.�2� s 0.75c

�3� k 10.64e 1�4� m –�5� n –

�6� D50

80 Gy for severeproctitis/necrosis/stenosis/fistulab

80sym

bladderand vo

aReference 25.bReference 26.cReference 24.dReference 27.eReference 28.

therapy techniques. For example, Burman et al. lists sev-

Medical Physics, Vol. 37, No. 9, September 2010

eral OARs end points and tolerance parameters for use inestimating NTCP following radiotherapy but not the urethra.As the urethra has similar anatomical structures to OARssuch as the colon, esophagus, and small intestine, and stric-tures leading to the obstruction of the passage of the luminalcontents are common end points following radiotherapy, itwas decided to use the end points and tolerance parametersof the esophagus �Table II� to estimate the urethral NTCP inthis work. In addition, in case of standard fractionated andhypofractionated 3D-CRT techniques, the urethra was notcontoured and, as a result, differential DVHs of urethra forthese techniques were not available. However, assuming thatequivalent doses, Deq,i are the same as the target dose andwere uniformly delivered to the urethral volume within theprostate, urethral NTCP was calculated using the followingequation:23

NTCP =1

1 + �D50/Deq,i�k , �7�

where D50 and k have the same definition as described pre-viously.

In case of femoral heads, the relative seriality model pa-rameters were not available. Therefore, the Lyman NTCPmodel with effective volume DVH reduction scheme29–31

was used instead. The Lyman NTCP model may be definedas follows:

NTCP�D,V� =1

2��

�

t

e−�t�2/2�dt�, �8�

where

t =Dmax − D50��eff�

m�D50��eff�. �9�

The normal deviate t represents the number of standard de-viations the point �Dmax,�eff� is away from D50��eff�, the 50%tolerance dose for the effective volume ��eff�. D50��eff� is

32

seriality and the Lyman models for organs-at-risk

efault values

r Urethra Femoral heads

a 7.5 Gy�estimated� 6 Gyb

1.0�estimated� –14.5e –

– 0.12b

– 0.25b

rticacturelossb

68 Gy forclinical stricture/

perforationb65 Gy forNecrosisb

tive

D

ladde

5 Gy1.3d

4.5e

––

Gy foptomacontrlume

taken to vary with �eff as

5130 Takam et al.: Normal tissue complications following prostate cancer irradiation 5130

D50��eff� = D50�1�/��eff�n. �10�

The effective volume ��eff� is calculated using the followingequation:

�eff =1

�ref�

i

�i� Di

Dmax�1/n

, �11�

where �ref can be either a volume of the whole organ orreference volume of that organ and n is the volume depen-dence of the complication probability.26

Model parameters used for femoral heads are shown inTable II.

II.D. Assessment of NTCP values dependence on therelative seriality model parameters

It is clear from Eq. �6� described above that the relativeseriality NTCP model contains several variable parameterssuch as Deq,i, s, and k. The dose Deq,i �derived in this studyby first converting physical dose to BEffD and then to Deq asdescribed earlier� depends on the � /� ratio of the OAR. Toinvestigate the sensitivity of the NTCP values obtained de-pending on the model parameters, values of � /� ratio, aswell as s and k parameters were varied and rectal NTCPs forhypofractionated 3D-CRT and HDR-BT treatment planswere calculated. The parameter k was calculated applying thefollowing equation:

k =4

2�m, �12�

showing that it is related to the value of parameter “m” �theslope of the complication probability vs dose curve�. Hence,testing of sensitivity on the NTCP model associated with theparameter k can be done either by varying the value of pa-rameter k directly or by varying the value of parameter m.The latter approach was used in this study by varying thevalue of one parameter at a time while keeping others con-stant by using their default value. Typical rectal BEffDVHs

FIG. 1. A plot shows differential DVHs of rectum obtained from a four-fieldhypofractionated 3D-CRT treatment plan for prostate. Differential volume�cm3� of rectum was plotted against original physical doses and correspond-

ing converted biological effective doses and equivalent doses.

Medical Physics, Vol. 37, No. 9, September 2010

for the various treatment modalities were used to demon-strate the results of this sensitivity testing in the followingsubsections.

III. RESULTS

III.A. DVHs of organs-at-risk

Changes in OAR DVHs as a result of physical doses con-version are demonstrated in Figs. 1–3, showing examples oftypical differential DVHs of the rectum obtained from four-field hypofractionated 3D-CRT, HDR-BT, and LDR-BT asmonotherapy treatment plans for the prostate. In these fig-ures, the normalized cumulative volume �%� of rectum wasplotted against the original physical doses.

Calculated NTCP values were statistically analyzed usingone-way ANOVA and t-tests for their significance. NTCP forstandard four-field 3D-CRT technique and 64 Gy total dosewas used as reference.

FIG. 2. A plot shows differential DVHs of rectum obtained from a HDR-BTas monotherapy treatment plan for prostate.

FIG. 3. A plot shows differential DVHs of rectum obtained from a LDR-BT

as monotherapy treatment plan for prostate.

5131 Takam et al.: Normal tissue complications following prostate cancer irradiation 5131

III.B. NTCP of OARs

III.B.1. Rectal NTCP

Table III shows the volumetric, radiation dosimetric data,and NTCP of rectum for prostate radiation treatment tech-niques investigated. Calculations based on the relative seri-ality model indicate that the risk of rectal complications wasthe highest following dose-escalated 3D-CRT to a total doseof 74 Gy being approximately 5.2�1.0%. Average rectalNTCP were smaller for HDR-BT �0.5�0.4%� and LDR-BT�0.6�0.4%� treatment plans.

The combination of large irradiated volume and high ra-diation dose exposure led to higher probability of rectal com-plications in standard fractionated and dose-escalated 3D-CRT compared to other techniques. The NTCP in four-field3D-CRT increases with the increasing total dose and doseescalation can only be recommended if PTV margin can bereduced.

With five-field 3D-CRT, radiation beams were arranged insuch a way that irradiation of critical organs such as rectumand bladder was minimized. Therefore, the average rectalNTCP following five-field 3D-CRT was smaller than that ofdose-escalated four-field 3D-CRT with the same total dose.For HDR-BT and LDR-BT monotherapy, only approxi-mately 1% and 0.1% of rectal tissues were exposed to theprescribed doses, hence, calculated probabilities of rectalcomplications for these techniques were the lowest. These

TABLE III. Average calculated rectal NTCP following various prostate canparameters �equivalent dose was used in calculation�.

Treatment techniqueNo. of

DVH/patient

Average ofequivalent doGy�S.D �ra

Standard fractionated 3D-CRT�64 Gy at 2 Gy/fraction� 7

48.5�4.�41.6–53.

Hypofractionated 3D-CRT�55 Gy at 2.75 Gy/fraction�

10 43.9�2.�39.6–46.

Dose-escalated 3D-CRTA. Total dose of 70 Gy 13 46.6�5.

�38.1–55.

B. Total dose of 74 Gy 3 51.6�0.�50.8–52.

Five-field 3D-CRT�70 Gy at 2 Gy/fraction�

14 38.6�5.�30.2–51.

HDR-BT �Ir-192�monotherapy �49.5 Gy�

9 59.8�8.�49.6–78.

LDR-BT �I-125� 37 61.9�5.�50.5–73.

techniques offer better dose conformality and less peripheral

Medical Physics, Vol. 37, No. 9, September 2010

radiation dose exposure of surrounding normal tissues. Thelower NTCP values for LDR and HDR brachytherapy werefound to be statistically significant.

III.B.2. Bladder NTCP

Table IV shows the volumetric, radiation dosimetric data,and NTCP of the bladder for various prostate radiation treat-ment techniques. Similar to rectal complications, it was ob-served that severe bladder complications are most likely fol-lowing dose-escalated four-field 3D-CRT �to a total dose of74 Gy� for prostate cancer. Average bladder NTCP followingthis technique was 6.6% �range 5.8%–7.4%�. Following thesame technique with a smaller total dose of 70 Gy deliveredto the prostate, the average bladder NTCP was reduced to5.0% despite a larger irradiated volume.

For standard fractionated four-field 3D-CRT, the maxi-mum bladder irradiated volume receiving equivalent dosearound 63 Gy was approximately 10%, resulting in an aver-age 1.9% NTCP for bladder. Similar fractions of bladder�9%� were irradiated to lower equivalent dose of 59 Gy fromhypofractionated 3D-CRT which led to smaller average blad-der NTCP at 0.7%.

Similar to the discussion for rectal DVHs analysis, blad-der NTCP for hypofractionated four-field 3D-CRT techniqueresulted in lesser equivalent dose given and smaller volumeof bladder irradiated, thus the probability of severe bladder

eatment techniques calculated with relative seriality model and dosimetric

Average irradiatedvolume in

cm3�S.D �range�

AverageNTCP in

%�S.D �range�

P-valuet-value

Statisticalsignificance

93.9�44.4�54.6–186.6�

2.8�1.0�1.1–4.1� Reference

83.8�29.6�45.9–142.5�

1.3�0.2�1.1–1.6�

0.000131.32Yes

72.0�31.1�25.3–141.7�

3.3�1.6�1.2–5.5�

0.26371.173

No62.7�9.8

�51.8–70.9�5.2�1.0�4.1–6.1�

0.05534.072

No98.5�51.9

�36.6–204.5�2.7�0.9�1.3–4.1�

0.56320.5932

No5.4�2.6�2.1–8.1�

0.5�0.4�0.0–1.1�

0.000117.55Yes

3.4�1.0�1.5–5.3�

0.6�0.4�0.0–1.8�

0.000129.27Yes

cer tr

meanse innge�

16�

02�

58�

60�

76�

35�

83�

complications was able to be reduced.

5132 Takam et al.: Normal tissue complications following prostate cancer irradiation 5132

III.B.3. Urethral NTCP

Table V shows the volumetric, radiation dosimetric data,and NTCP of urethra for radiation treatment techniques dis-cussed. As expected, urethral NTCPs following standardfractionated and hypofractionated four-field 3D-CRT tech-niques were higher than those for other organs due to uni-form high-dose exposure. Following standard fractionated3D-CRT, average urethral NTCP was found to be approxi-mately 9% �range 8.2%–11.2%�. Similar high average ure-thral NTCP was also predicted for HDR-BT �18.4%, range12.2%–31.1%� as well as LDR-BT �24.7%, range 12.0%–55.1%�. The corresponding average urethral NTCP for hy-pofractionated four-field 3D-CRT was lower at 3.6%.

III.B.4. Femoral heads NTCP

The volumetric, radiation dosimetric data, and NTCP offemoral heads for individual treatment techniques discussed

TABLE IV. Average calculated bladder NTCP following various prostate canparameters �equivalent dose was used in calculation�.

Treatment techniqueNo. of

DVH/patient

Averagemean equiv

dose inGy�S.D �ra

Standard fractionated 3D-CRT�64 Gy at 2 Gy/fraction� 7

53.4�4.�44.6–56.

Hypofractionated 3D-CRT�55 Gy at 2.75 Gy/fraction�

10 50.8�4.�42.8–54.

Five-field 3D-CRT�70 Gy at 2 Gy/fraction�

14 43.0�12�20.4–63

Dose-escalated 3D-CRTA. Total dose of 70 Gy 13 48.3�13

�20.6–65

B. Total dose of 74 Gy 3 44.2�6.�37.9–50.

TABLE V. Average urethral NTCP in various prostate cancer treatment teccalculation�.

Treatment techniqueNo. of

DVH/patient

m

Gy

Standard fractionated 3D-CRT�64 Gy at 2 Gy/fraction� 7Hypofractionated 3D-CRT�55 Gy at 2.75 Gy/fraction� 10HDR-BT �Ir-192� monotherapy�49.5 Gy� 10 �

LDR-BT �I-125� 36 �

Medical Physics, Vol. 37, No. 9, September 2010

are shown in Table VI. Necrosis of femoral heads may be aconsequence of excessive exposure to radiation from prostateradiotherapy. Assessment of femoral heads DVHs retrievedfrom treatment plans for dose-escalated four-field 3D-CRTtreatment for prostate cancer indicated that approximately11% and 14% of the femoral heads volume was irradiated todoses of 70 and 74 Gy, respectively. The mean equivalentdose received was lower than other OARs partly because oftheir distance from the treated volume. Accordingly, an av-erage NTCP for femoral heads was observed to be as low as0.02% for dose-escalated four-field 3D-CRT �to total dose of70 Gy� and 0.06% for dose-escalated four-field 3D-CRT �tototal dose of 74 Gy�.

III.C. Assessment of NTCP values dependence on therelative seriality model parameters

Figure 4 shows the effect of varying � /� ratio on rectalNTCP calculated with the relative seriality model. For hy-

reatment techniques calculated with relative seriality model and dosimetric

Averageirradiatedvolume in

cm3�S.D �range�

AverageNTCP in

%�S.D �range�

P-valuet-value

Statisticalsignificance

133.4�32.9�90.7–181.0�

1.9�0.2�1.6–2.3� Reference

119.6�42.3�56.0–184.6�

0.7�0.2�0.4–0.9�

0.000123.24Yes

162.4�99.2�46.6–456.8�

3.3�1.0�1.4–4.8�

0.00025.146Yes

161.7�72.6�81.5–306.0�

5.0�2.4�1.3–9.1�

0.00015.784Yes

199.4�147.9�72.4–361.8�

6.6�0.8�5.8–7.4�

0.009510.18Yes

ues calculated with relative seriality model �equivalent dose was used in

age ofquivalente in

D �range�

Averageirradiated volume incm3�S.D �range�

AverageNTCP in

%�S.D �range�

�0.6–65.3�

5.2�0.5�4.6–5.9�

9.4�1.1�8.2–11.2�

�0.1–59.4�

5.5�1.1�4.3–7.5�

3.6�0.7�2.8–5.0�

�5.7103.4�

0.8�0.3�0.5–1.5�

11.2�3.9�6.5–19.3�

�5.1–139.2�

0.6�0.2�0.2–1.6�

24.7�8.0�12.0–55.1�

cer t

ofalent

nge�

14�

39�

.2.7�

.3.5�

35�

hniq

Averean e

dos�S.

64.2�63.859.3

�59.293.583.7–130.4118.0

5133 Takam et al.: Normal tissue complications following prostate cancer irradiation 5133

pofractionated 3D-CRT, variation of � /� ratio from 1 to 10Gy causes around 5% change in rectal NTCP. However, asmaller change ��2%� in rectal NTCP was observed eitherfor hypofractionated 3D-CRT and HDR-BT considering � /�ratio for rectum �5 Gy �typically assumed for normal tis-sues�. If � /� is less than 3 Gy then the NTCP differencebetween the techniques will only be greater, with EBRT re-sulting in worse NTCP.

In case of varying the value of the s parameter, the rectalNTCP for hypofractionated 3D-CRT appears to have a linearrelationship with the s parameter. The rectal NTCP variesfrom approximately 0.3%–1.6%, i.e., around 1.3% change,for the whole range of the s parameter values from 0 to 1�Fig. 5�. The relationship between the s parameter and therectal NTCP appears to be exponential for HDR-BT. How-ever, small changes only in rectal NTCP �1.2%� were ob-served with increasing of the s parameter value from 0.1 to1.0. This full extent of the s parameter values is unrealistic asrectum is considered a serial organ �s values closer to 1�rather than parallel �s values closer to 0�. The variations inNTCPs between s values of 0.5 and 1 are less then 0.5% for

TABLE VI. Average femoral heads NTCP and in various treatment techniqu

Treatment techniqueNo. of

DVH/patient

Amean

Gy�

Five-field 3D-CRT�70 Gy at 2 Gy/fraction� 14

3�2

Dose-escalated 3D-CRT

A. Total dose of 70 Gy 103

�1

B. Total dose of 74 Gy 23

�3

FIG. 4. A plot shows changes in rectal NTCP �%� in fractionated 3D-CRT��� and HDR-BT monotherapy ��� corresponding to variation in the value

of rectal � /� ratio.

Medical Physics, Vol. 37, No. 9, September 2010

hypofractionated EBRT and less than a percent for HDR-BT,therefore not contributing to the final error bar significantly.

Finally, Fig. 6 shows the relationship between value ofparameter k and rectal NTCP as predicted by the relativeseriality model. Variation of the value of this parameter from1 to 20 causes considerable change in rectal NTCP for hy-pofractionated 3D-CRT especially when the value of param-eter k is smaller than 10. Contrarily, varying the value ofparameter k in the same range has a much smaller effect inrectal NTCP for HDR-BT with approximately 1%–2%change. For values of k less than 10, the difference in NTCPsbetween HDR-BT and hypofractionated EBRT will only beaccentuated/increased. As a result, we believe that the NTCPdifferences between modalities �EBRT and BT� are valideven within the model uncertainties.

IV. DISCUSSION

From all rectal differential DVHs evaluated for standardfour-field 3D-CRT, it was observed that some DVHs con-tained a dose peak at the Deq around 30–40 Gy while someof them had a peak at the Deq of 60–65 Gy. For those DVHs

prostate cancer �equivalent dose was used in calculation�.

e ofivalentin�range�

Averageirradiatedvolume in

cm3�S.D �range�

AverageNTCP in

%�S.D �range�

7.24.0�

204.0�68.9�101.5–372.8�

0.2�0.4�0.0–1.3�

7.19.0�

121.9�55.7�38.6–217.2�

0.02�0.02�0.0–0.05�

1.40.3�

117.7�7.3�112.6–122.8�

0.06�0.04�0.04–0.09�

FIG. 5. A plot shows changes in rectal NTCP �%� in fractionated 3D-CRT��� and HDR-BT monotherapy ��� corresponding to variation in the value

es for

veragequ

doseS.D

0.2�

0.4–4

3.5�

7.3–39.4�

8.4–4

of parameter s.

5134 Takam et al.: Normal tissue complications following prostate cancer irradiation 5134

which had the peak around intermediate Deq range, the aver-age rectal NTCP was around 1%, while those rectal DVHswhich had the peak around the prescribed tumor doses haveaverage rectal NTCPs in a range of 2%–4%. Similarly, mostof the bladder DVHs for standard fractionated 3D-CRT hadthe dose peak around the prescribed tumor doses of 60–65Gy which resulted in a bladder NTCP of approximately 2%,while those DVHs having a peak around 30–40 Gy resultedin NTCP of less than 2%. In addition, reduction of normaltissue complication risk following external beam radio-therapy can also be achieved by decreasing the normal tissuevolume that might be exposed to therapeutic radiation dose,i.e., reducing the treatment margin.

With hypofractionated four-field 3D-CRT, dose-volumedistributions of the rectum, bladder, and urethra were similarto that of standard fractionated four-field 3D-CRT, but theirradiated volumes were approximately 11% smaller. Addi-tionally, lower equivalent doses �maximum of 55 Gy� wereused to irradiate the prostate which, accordingly, resulted inlower dose exposure of the surrounding normal tissues,which ultimately led to lower estimated probability of com-plications following the treatment. Since the prostate hasbeen observed to have lower � /� ratio than normal tissues,this gives some advantages to hypofractionated EBRT orHDR-BT as treatment of choice for prostate cancer becauseboth techniques have a potential to yield increased tumorcontrol for a given level of late complications or decreasedlate complications for a given level of tumor control.33 Re-sults from this study partly confirmed this theory as the es-timated NTCP for a particular organ from hypofractionated3D-CRT was lower than that from standard fractionated 3D-CRT. However, it should be noted that there are few reportsof higher values of � /� ratio for prostate.34,35 Possible rea-sons for the lower NTCP estimated in this study includelower average irradiated volumes particularly for the bladderand the use of an � /� ratio of 7.5 for bladder and of 5.4 for

FIG. 6. A plot shows changes in rectal NTCP �%� in fractionated 3D-CRT��� and HDR-BT monotherapy ��� corresponding to variation in the valueof parameter k.

rectum. Based on the � /� ratio considerations, the linear-

Medical Physics, Vol. 37, No. 9, September 2010

quadratic model would predict lower 2 Gy equivalent dosesfor the hypofractionated schedule used in this study. Out-come analysis of latest randomized trials comparing hypof-ractionated schedules with conventional fractionation forprostate show quite inconclusive results regarding NTCP.While some trials show no difference between the twoschedules in regards to NTCP,36,37 others show better NTCPwith hypofractionation38 or only a small increase in certaintoxicities and not necessarily on genitourinary ones.39,40 Ob-viously, NTCP results depend on fraction size too, as reallylarge fraction sizes can induce toxicity. However, more stud-ies and longer follow-up are needed, especially for the recenttrials, to fully validate the efficacy of hypofractionation bothNTCP and TCP wise.

HDR-BT, either as monotherapy or as a boost to EBRT,has been reported to cause very low rates of severe late tox-icity to surrounding normal tissues.41–43 Mean Deq receivedby rectum from HDR-BT ranged between 50 and 78 Gy,while the mean Deq for urethra ranged similarly between 50and 73 Gy. DVHs obtained from prostate treatment plansindicated that only small fractions �approximately 1.5% oftotal volume� of rectum were exposed to high therapeuticdoses during the treatment. Therefore, average rectal NTCPfollowing HDR-BT predicted was much smaller than thatfollowing EBRT. This conclusion is in agreement with clini-cal findings.

Data based on a review of clinical results followingHDR-BT as a boost to EBRT indicate a small prevalence ofsevere long term toxicity.7 Furthermore, chronic toxicitiesafter HDR-BT as monotherapy for prostate cancer have beenreported to less than LDR-BT after a median follow-up pe-riod of 35 months.11 Most complications observed in theHDR monotherapy patients were low grade toxicity andnone of the patients experienced grade 4 toxicities. With amedian follow-up of 4 yr, the most severe late complicationobserved in patients treated with HDR-BT was urethral stric-ture with a 5 yr actuarial risk of 7% and no patient experi-enced late severe rectal complications.41 The incidence ofurological complications observed in the previous report wasobvious when it is related to differential DVHs assessmentobserved in this study where average Deq received by theurethra from the treatment was as high as 120 Gy represent-ing the highest received by all normal tissues.

For LDR-BT using I-125 permanent radioactive seeds, ra-diotherapy parameters such as average BEffD and Deq weresimilar to HDR-BT although rectal and urethral irradiatedvolume were slightly smaller. Dose-volume distribution ofLDR-BT in the rectum appeared to be more inhomogeneouscompared to other techniques and ranged widely from 30 to130 Gy. Although a wide range of equivalent dose was de-livered to rectum, only small fractions �approximately 4% intotal� were irradiated to the prescription dose. Hence, a smallvalue of average rectal NTCP ensued.

For urethra, during prostate irradiation, a part of the ure-thra located inside the prostate �prostatic urethra� may re-ceive the same dose which was given to the prostate espe-cially with the EBRT. As a result, higher NTCP values were

observed from DVHs assessment for the urethra than other

5135 Takam et al.: Normal tissue complications following prostate cancer irradiation 5135

irradiated organs. For brachytherapy, planning was done insuch a way that dose to the urethra was minimized. However,some fractions of the urethra were still irradiated with theequivalent doses in a range of 120–140 Gy for LDR-BT and110–130 Gy for HDR-BT, which were considerably higherthan the doses that other organs received. Accordingly, it wasindicated by the NTCP model that severe complications ofurethra following prostate irradiation are more likely to beobserved than that of other surrounding healthy organs. WithLDR-BT, approximately 3% of urethral volume was irradi-ated to equivalent doses in the range of 100–150 Gy, clearlythe highest among the doses received by other OARs. Theaverage urethral NTCP of 24.7% predicted is higher thansevere complication rates reported clinically. The discrep-ancy is likely to be attributable to the lack of published ure-thral specific model parameters. The average urethral NTCPof 24.7% predicted by this model is, however, consistentwith reports of low grade �grade 0–grade 2� urinary toxicity�incontinence� after I-125 LDR-BT range widely between0%–40%.44,45

Dose-escalated four-field 3D-CRT to a total dose of 70and 74 Gy is currently used at our center for treatment of lowand intermediate/high-risk prostate cancer. The planning tar-get volume is reduced after 64 Gy to avoid exposure of sur-rounding normal tissues to the full prescription dose. Distri-bution of equivalent doses over the volume of rectum andbladder was therefore similar to that of standard fractionatedand hypofractionated 3D-CRT. Although PTV was reducedin order to minimize the healthy tissue damage, some por-tions of rectum and bladder volume were still exposed tohigh doses resulting in a higher prediction of rectal and blad-der NTCP.

Zelefsky et al.5 reported 5 yr actuarial likelihood of de-velopment of grade 2 and grade 3 late GI toxicities of 11%and 0.75%, respectively, following the prostate treatment us-ing dose-escalated 3D-CRT up to 81 Gy. In addition, the 5 yractuarial likelihood of development of grade 2 and grade 3late GU toxicities was 10% and 3%, respectively. Our differ-ential DVHs assessment showed NTCP ranges of 1.2%–6.1% and 1.3%–9.1% for the rectum and bladder, respec-tively, which are consistent with the rates of severe late GIand GU toxicities.

Michalski et al.46 recently investigated dose-volume ef-fects in radiation induced rectal injury. They reviewed sev-eral published data on rectal injury and estimated parametersfor the Lyman–Kutcher–Burman NTCP model. While the svalue of the relative seriality model used in this study forestimation of NTCP for rectal complications is slightly lowerthan the n parameter of the Lyman–Kutcher–Burman modelin their work, it yields NTCPs within a similar range fordose-escalated 3D-CRT.

Severe complications of femoral heads have been rarelyreported. Corresponding differential DVHs from our dose-escalated 3D-CRT treatment plans indicated that they wouldnormally receive equivalent doses in a range of 30–40 Gyexplaining the lack of reports of severe complications. Inaddition, the complication rates for femoral heads are lower

than those observed compared to other OARs because the

Medical Physics, Vol. 37, No. 9, September 2010

DVHs show that only 11%–14% of the femoral head wereirradiated to the prescription doses, the remainder receivinglower physical and therefore equivalent doses. While theNTCP increases with the dose �from 0.02% to 0.06% for 70and 74 Gy, respectively�, a larger increase in NTCP has beenobserved for larger volumes of the femoral heads irradiatedto the prescribed dose of 70 Gy �0.2% for 204 cm3 and0.02% for 122 cm3�.

Borghede and Hedelin47 reported the estimated femoralheads dose of 49 Gy from 3D-CRT treatment technique �totaldose of 70 Gy with standard fractionation� and 64.8 Gy withhypofractionated �2.4 Gy/fraction� for prostate cancer. Out of184 patients involved in their study, only one patient �0.5%�experienced osteonecrosis of the hip joint 18 months afterthe treatment and was suspected as a result of three-fieldtreatment technique which increased the dose to femoralheads compared to four-field technique. A range of meandoses similar to ours, received by femur head and neck dur-ing prostate treatment, was reported by Gershkevitsh et al.48

For a dose prescribed to target of 64 Gy with different plans,the mean doses to this OAR were in the range of 3–34 Gy,the maximum mean dose falling within the range reported inthis study. Bedford et al.49 reported the use of the Lymanmodel to estimate a complication probability of femoralheads for different radiation treatment plans of conformalradiotherapy for prostate cancer. In his report, the NTCP offemoral heads was generally small �0.1%� except for a fewplans with NTCP of up to 5.5%. Luxton et al.50 also reportedvery small NTCP probability �up to 0.05%� of femoral headsas a result of 3D-CRT for prostate carcinoma.

Testing for the sensitivity of NTCP predictions using therelative seriality model, the values of � /� ratio and s and kparameters for the rectum were varied within the total rangeand changes in estimated rectal NTCP in hypofractionated3D-CRT and HDR-BT were investigated. In case of � /�ratio for rectum which is typically assumed to be larger thanthat for the target organ �prostate�, increasing from its defaultvalue �5.4 Gy� to the maximum value �10 Gy� had littleeffect on estimated rectal NTCP in both hypofractionatedEBRT and HDR-BT. It may be assumed that within the typi-cal range of � /� ratio �5–10 Gy� for rectum, the estimatedrectal NTCP is virtually independent to rectal � /� ratio. Themagnitude of variation in rectal NTCP in both techniques asa result of changes in the s parameter was not substantial asonly around a 1.2% increase in rectal NTCP was observed.

Variations of the k parameter within the expected range�1–20� had greater effect on estimated rectal NTCPs in hy-pofractionated 3D-CRT than HDR-BT, especially for k �10.The effect of this dependence on rectal NTCP in HDR-BTwas far less pronounced and seems to be virtually indepen-dent of the value. For values of k less than 10, the differencein NTCPs between HDR-BT and hypofractionated EBRTwill only be increased.

V. SUMMARY

Results form this study are based on theoretical predic-

tions using available radiobiological models and are intended

5136 Takam et al.: Normal tissue complications following prostate cancer irradiation 5136

to provide clinicians with additional information to assist inthe selection of a radiation treatment technique and plan forradiotherapy of prostate cancer. Validation of the predictedNTCP awaits long term toxicity data of the different tech-niques and fractionation studies preferably from randomizedclinical trials.

a�Electronic mail: eva.bezak@health.sa.gov.au1J. E. Livsey, R. A. Cowan, J. P. Wylie, R. Swindell, G. Read, V. S. Khoo,and J. P. Logue, “Hypofractionated conformal radiotherapy in carcinomaof the prostate: Five-year outcome analysis,” Int. J. Radiat. Oncol., Biol.,Phys. 57�5�, 1254–1259 �2003�.

2P. A. Kupelian, V. V. Thakkar, D. Khuntia, C. Reddy, E. A. Klein, and A.Mahadevan, “Hypofractionated intensity-modulated radiotherapy �70 Gyat 2.5 Gy per fraction� for localized prostate cancer: Long-term out-comes,” Int. J. Radiat. Oncol., Biol., Phys. 63�5�, 1463–1468 �2005�.

3J. C. Blasko, P. D. Grimm, J. E. Sylvester, K. R. Badiozamani, D. Hoak,and W. Cavanagh, “Palladium-103 brachytherapy for prostate carci-noma,” Int. J. Radiat. Oncol., Biol., Phys. 46�4�, 839–850 �2000�.

4L. Potters, C. Morgenstern, E. E. Mullen, P. Fearn, A. Jassal, and M.Kattan, “Twelve year outcomes following permanent brachytherapy inpatients with clinically localized prostate cancer,” Int. J. Radiat. Oncol.,Biol., Phys. 60�1�, S183–S184 �2004�.

5M. Zelefsky, D. A. Kuban, L. Levy, L. Potters, D. C. Beyer, J. C. Blasko,B. J. Moran, J. P. Ciezki, A. L. Zietman, T. Pisansky, M. Elshaikh, and E.M. Horwitz, “Long-term multi-institutional analysis of stage T1–T2 pros-tate cancer treated with permanent brachytherapy,” Int. J. Radiat. Oncol.,Biol., Phys. 63�3�, S33–S34 �2005�.

6R. Mark, G. Vallabhan, A. Akins, P. Anderson, M. Nair, T. Neumann, D.White, and S. Gurley, “Interstitial high dose rate �HDR� brachytherapyfor early stage prostate cancer,” Int. J. Radiat. Oncol., Biol., Phys. 63�2�,S304 �2005�.

7S. Nilsson, B. J. Norlen, and A. Widmark, “A systematic overview ofradiation therapy effects in prostate cancer,” Acta Oncol. �Madr� 43�4�,316–381 �2004�.

8P. A. Kupelian, C. A. Reddy, T. P. Carlson, and T. R. Willoughby, “Dose/volume relationship of late rectal bleeding after external beam radio-therapy for localized prostate cancer: Absolute or relative rectal vol-ume?,” Cancer J. 8�1�, 62–66 �2002�.

9G. C. Morton, “The emerging role of high-dose-rate brachytherapy forprostate cancer,” Clin. Oncol. 17, 219–227 �2005�.

10E. Panitsa, J. C. Rosenwald, and C. Kappas, “Quality control of dosevolume histogram computation characteristics of 3D treatment planningsystems,” Phys. Med. Biol. 43�10�, 2807–2816 �1998�.

11A. A. Martinez, I. Pataki, G. Edmundson, E. Sebastian, D. Brabbins, andG. Gustafson, “Phase II prospective study of the use of conformal high-dose-rate brachytherapy as monotherapy for the treatment of favorablestage prostate cancer: A feasibility report,” Int. J. Radiat. Oncol., Biol.,Phys. 49�1�, 61–69 �2001�.

12Y. Yoshioka, T. Nose, K. Yoshida, R.-J. Oh, Y. Yamada, E. Tanaka, H.Yamazaki, T. Inoue, and T. Inoue, “High-dose-rate brachytherapy asmonotherapy for localized prostate cancer: A retrospective analysis withspecial focus on tolerance and chronic toxicity,” Int. J. Radiat. Oncol.,Biol., Phys. 56�1�, 213–220 �2003�.

13I. S. Grills, A. A. Martinez, M. Hollander, R. Huang, K. Goldman, P. Y.Chen, and G. S. Gustafson, “High dose rate brachytherapy as prostatecancer monotherapy reduces toxicity compared to low dose rate palla-dium seeds,” J. Urol. �Baltimore� 171�3�, 1098–1104 �2004�.

14T. Martin, D. Baltas, R. Kurek, S. Röddiger, M. Kontova, G. Anagnosto-poulos, T. Dannenberg, T. Buhleier, G. Skazikis, U. Tunn, and N. Zam-boglou, “3-D conformal HDR brachytherapy as monotherapy for local-ized prostate cancer,” Strahlenther. Onkol. 180�4�, 225–232 �2004�.

15C. D. Springer, R. N. Boyd, T. C. Elsdon, L. S. Skeates, A. A. McKenzie,and S. Ozard, “Outcome analysis of HDR prostate brachytherapy in themanagement of patients with localized prostate cancer in a communitycancer center,” Brachytherapy 6�2�, 106 �2007�.

16D. J. Demanes, M. Gilhezan, L. Schour, G. Gustafson, D. R. Hill, K.Marvin, I. Bernstein, and A. A. Martinez, “High dose rate brachytherapy�HDR-BT� as monotherapy for favorable prostate cancer: Excellent5-year control rates and low toxicity,” Int. J. Radiat. Oncol., Biol., Phys.

69�3�, S83 �2007�.

Medical Physics, Vol. 37, No. 9, September 2010

17R. G. Dale, “The application of the linear-quadratic dose-effect equationof fractionated and protracted radiotherapy,” Br. J. Radiol. 58, 515–528�1985�.

18R. G. Dale, “The application of the linear-quadratic model to fractionatedradiotherapy when there is incomplete normal tissue recovery betweenfractions, and possible implications for treatments involving multiplefractions per day,” Br. J. Radiol. 59, 919–927 �1986�.

19S. Nag and N. Gupta, “A simple method of obtaining equivalent doses foruse in HDR brachytherapy,” Int. J. Radiat. Oncol., Biol., Phys. 46, 507–513 �2000�.

20T. Rancati, C. Fiorino, G. Gagliardi, G. M. Cattaneo, G. Sanguineti, V. C.Borca, C. Cozzarini, G. Fellin, F. Foppiano, G. Girelli, L. Menegotti, A.Piazzolla, V. Vavassori, and R. Valdagni, “Early clinical and radiologicalpulmonary complications following breast cancer radiation therapy:NTCP fit with four different models,” Radiother. Oncol. 82, 308–316�2007�.

21G. Gagliardi, J. Bjöhle, I. Lax, A. Ottolenghi, F. Eriksson, A. Liedberg, P.Lind, and L. E. Rutqvist, “Radiation pneumonitis after breast cancer irra-diation: Analysis of the complication probability using the relative serial-ity model,” Int. J. Radiat. Oncol., Biol., Phys. 46�2�, 373–381 �2000�.

22P. Kallman, A. Agren, and A. Brahme, “Tumour and normal tissue re-sponse to fractionated non-uniform dose delivery,” Int. J. Radiat. Biol.62�2�, 249–262 �1992�.

23R. I. MacKay, J. H. Hendry, C. J. Moore, P. C. Williams, and G. Read,“Predicting late rectal complications following prostate conformal radio-therapy using biologically effective doses and normalized dose-surfacehistograms,” Br. J. Radiol. 70, 517–526 �1997�.

24M. Zaider, M. J. Zelefsky, G. N. Cohen, C.-S. Chui, E. D. Yorke, L.Ben-Porat, and L. Happersett, “Methodology for biologically-based treat-ment planning for combined low-dose-rate �permanent implant� and high-dose-rate �fractionated� treatment of prostate cancer,” Int. J. Radiat. On-col., Biol., Phys. 61�3�, 702–713 �2005�.

25A. Daşu, I. Toma-Daşu, J. Olofsson, and M. Karlsson, “The use of riskestimation models for the induction of secondary cancers following ra-diotherapy,” Acta Oncol. �Madr� 44�4�, 339–347 �2005�.

26C. Burman, G. J. Kutcher, B. Emami, and M. Goitein, “Fitting of normaltissue tolerance data to an analytic function,” Int. J. Radiat. Oncol., Biol.,Phys. 21, 123–135 �1991�.

27M. Mattia, P. D. Giudice, and B. Cacciac, “IMRT optimization: Variabil-ity of solutions and its radiobiological impact,” Med. Phys. 31�5�, 1052–1060 �2004�.

28A. Niemierko and M. Goitein, “Calculation of normal tissue complicationprobability and dose-volume histogram reduction schemes for tissueswith a critical element architecture,” Radiother. Oncol. 20, 166–176�1991�.

29J. T. Lyman, “Complication probability as assessed from dose-volume-histogram,” Radiat. Res. 104, S13–S19 �1985�.

30G. Kutcher, C. Burman, L. Brewster, M. Goitein, and R. Mohan, “Histo-gram reduction method for calculating complication probabilities forthree-dimensional treatment planning evaluations,” Int. J. Radiat. Oncol.,Biol., Phys. 21�1�, 137–146 �1991�.

31G. J. Kutcher and C. M. Burman, “Calculation of complication probabil-ity factors for non-uniform normal tissue irradiation: The effective vol-ume method,” Int. J. Radiat. Oncol., Biol., Phys. 16�6�, 1623–1630�1989�.

32J. T. Lyman and A. B. Wolbarst, “Optimization of radiation therapy, III: Amethod of assessing complication probabilities from dose-volume histo-grams,” Int. J. Radiat. Oncol., Biol., Phys. 13, 103–109 �1987�.

33D. J. Brenner, “Hypofractionation for prostate cancer radiotherapy—What are the issues?,” Int. J. Radiat. Oncol., Biol., Phys. 57�4�, 912–914�2003�.

34A. E. Nahum, B. Movsas, E. M. Horwitz, C. C. Stobbe, and J. D. Chap-man, “Incorporating clinical measurements of hypoxia into tumor localcontrol modeling of prostate cancer: Implications for the � /� ratio,” Int.J. Radiat. Oncol., Biol., Phys. 57�2�, 391–401 �2003�.

35J. Z. Wang and X. A. Li, “Evaluation of external beam radiotherapy andbrachytherapy for localized prostate cancer using equivalent uniformdose,” Med. Phys. 30�1�, 34–40 �2003�.

36G. Arcangeli, B. Saracino, S. Gomellini, and M. G. Petrongari, “A pro-spective phase III randomized trial of hypofractionation versus conven-tional fractionation in patients with high-risk prostate cancer,” Int. J. Ra-diat. Oncol., Biol., Phys. 78�1�, 11–18 �2010�.

37

H. Lukka, C. Hayter, J. A. Julian, and P. Warde, “Randomized trial com-

5137 Takam et al.: Normal tissue complications following prostate cancer irradiation 5137

paring two fractionation schedules for patients with localized prostatecancer,” J. Clin. Oncol. 23, 6132–6138 �2005�.

38D. Norkus et al., “A randomized trial comparing hypofractionated andconventionally fractionated three-dimensional external-beam radiotherapyfor localized prostate adenocarcinoma: A report on acute toxicity,” Strahl-enther. Onkol. 185�11�, 715–721 �2009�.

39A. Pollack et al., “Dosimetry and preliminary acute toxicity in the first100 men treated for prostate cancer on a randomized hypo fractionationdose escalation trial,” Int. J. Radiat. Oncol., Biol., Phys. 64�2�, 518–526�2006�.

40E. Yeoh et al., “Hypofractionated versus conventionally fractionated ra-diation therapy for prostate carcinoma: Updated results of a phase IIIrandomized trial,” Int. J. Radiat. Oncol., Biol., Phys. 66, 1072–1083�2006�.

41L. Astrom, D. Pedersen, C. Mercke, S. Holmang, and K. A. Johansson,“Long-term outcome of high dose rate brachytherapy in radiotherapy forlocalized prostate cancer,” Radiother. Oncol. 74, 157–161 �2005�.

42R. M. Galalae, G. Kovacs, J. Schultze, T. Loch, P. Rzehak, R. Wilhelm,H. Bertermann, B. Buschbeck, P. Kohr, and B. Kimmig, “Long-term out-come after elective irradiation of the pelvic lymphatics and local doseescalation using high-dose-rate brachytherapy for locally advanced pros-tate cancer,” Int. J. Radiat. Oncol., Biol., Phys. 52�1�, 81–90 �2002�.

43T. Akimoto, H. Katoh, S.-E. Noda, K. Ito, T. Yamamoto, B. Kashiwagi,and T. Nakano, “Acute genitourinary toxicity after high dose rate �HDR�brachytherapy combined with hypofractionated external-beam radiationtherapy for localized prostate cancer: Second analysis to determine the

Medical Physics, Vol. 37, No. 9, September 2010

correlation between the urethral dose in HDR brachytherapy and the se-verity of acute genitourinary toxicity,” Int. J. Radiat. Oncol., Biol., Phys.63�2�, 472–478 �2005�.

44T. L. McElveen, F. M. Waterman, H. Kim, and A. P. Dicker, “Factorspredicting for urinary incontinence after prostate brachytherapy,” Int. J.Radiat. Oncol., Biol., Phys. 59�5�, 1395–1404 �2004�.

45S. Dinges, S. Deger, S. Koswig, D. Boehmer, D. Schnorr, T. Wiegel, S.A. Loening, M. Dietel, W. Hinkelbein, and V. Budach, “High-dose rateinterstitial with external beam irradiation for localized prostate cancer—Results of a prospective trial,” Radiother. Oncol. 48�2�, 197–202 �1998�.

46J. Michalski, H. Gay, A. Jackson, S. Tucker, and J. Deasy, “Radiationdose-volume effects in radiation-induced rectal injury,” Int. J. Radiat.Oncol., Biol., Phys. 76�3�, S123–S129 �2010�.

47G. Borghede and H. Hedelin, “Radiotherapy of localised prostate cancer.Analysis of late treatment complications. A prospective study,” Radiother.Oncol. 43�2�, 139–146 �1997�.

48E. Gershkevitsh, I. Rosenberg, P. D. Dearnaley, and K.-R. Trott, “Bonemarrow doses and leukemia risk in radiotherapy of prostate cancer,” Ra-diother. Oncol. 53, 189–197 �1999�.

49J. L. Bedford, V. S. Khoo, M. Oldham, D. P. Dearnaley, and S. Webb, “Acomparison of coplanar four-field techniques for conformal radiotherapyof the prostate,” Radiother. Oncol. 51, 225–235 �1999�.

50G. Luxton, S. L. Hancock, and A. L. Boyer, “Dosimetry and radiobio-logic model comparison of IMRT and 3D conformal radiotherapy in treat-ment of carcinoma of the prostate,” Int. J. Radiat. Oncol., Biol., Phys.59�1�, 267–284 �2004�.