Candela Juan2013

Calculated organ doses using Monte Carlo simulations in a reference malephantom undergoing HDR brachytherapy applied to localized prostatecarcinoma

Cristian Candela-Juana)

Radioprotection Department, La Fe University and Polytechnic Hospital, Valencia 46026, Spain

Jose Perez-CalatayudRadiotherapy Department, La Fe University and Polytechnic Hospital, Valencia 46026, Spain

Facundo BallesterDepartment of Atomic, Molecular and Nuclear Physics, University of Valencia, Burjassot 46100, Spain

Mark J. RivardDepartment of Radiation Oncology, Tufts University School of Medicine, Boston, Massachusetts 02111

(Received 23 October 2012; revised 27 January 2013; accepted for publication 28 January 2013;published 27 February 2013)

Purpose: The aim of this study was to obtain equivalent doses in radiosensitive organs (aside fromthe bladder and rectum) when applying high-dose-rate (HDR) brachytherapy to a localized prostatecarcinoma using 60Co or 192Ir sources. These data are compared with results in a water phantom andwith expected values in an infinite water medium. A comparison with reported values from protontherapy and intensity-modulated radiation therapy (IMRT) is also provided.Methods: Monte Carlo simulations in Geant4 were performed using a voxelized phantom describedin International Commission on Radiological Protection (ICRP) Publication 110, which reproducesmasses and shapes from an adult reference man defined in ICRP Publication 89. Point sources of60Co or 192Ir with photon energy spectra corresponding to those exiting their capsules were placed inthe center of the prostate, and equivalent doses per clinical absorbed dose in this target organ wereobtained in several radiosensitive organs. Values were corrected to account for clinical circumstanceswith the source located at various positions with differing dwell times throughout the prostate. Thiswas repeated for a homogeneous water phantom.Results: For the nearest organs considered (bladder, rectum, testes, small intestine, and colon), equiv-alent doses given by 60Co source were smaller (8%–19%) than from 192Ir. However, as the distanceincreases, the more penetrating gamma rays produced by 60Co deliver higher organ equivalent doses.The overall result is that effective dose per clinical absorbed dose from a 60Co source (11.1 mSv/Gy)is lower than from a 192Ir source (13.2 mSv/Gy). On the other hand, equivalent doses were the same inthe tissue and the homogeneous water phantom for those soft tissues closer to the prostate than about30 cm. As the distance increased, the differences of photoelectric effect in water and soft tissue, andappearance of other materials such as air, bone, or lungs, produced variations between both phantomswhich were at most 35% in the considered organ equivalent doses. Finally, effective doses per clini-cal absorbed dose from IMRT and proton therapy were comparable to those from both brachytherapysources, with brachytherapy being advantageous over external beam radiation therapy for the furthestorgans.Conclusions: A database of organ equivalent doses when applying HDR brachytherapy to the prostatewith either 60Co or 192Ir is provided. According to physical considerations, 192Ir is dosimetricallyadvantageous over 60Co sources at large distances, but not in the closest organs. Damage to dis-tant healthy organs per clinical absorbed dose is lower with brachytherapy than with IMRT or pro-tons, although the overall effective dose per Gy given to the prostate seems very similar. Given thatthere are several possible fractionation schemes, which result in different total amounts of thera-peutic absorbed dose, advantage of a radiation treatment (according to equivalent dose to healthyorgans) is treatment and facility dependent. © 2013 American Association of Physicists in Medicine.[http://dx.doi.org/10.1118/1.4791647]

Key words: HDR brachytherapy, prostate carcinoma, Monte Carlo methods, voxelized phantom,integral dose, organ equivalent dose

033901-1 Med. Phys. 40 (3), March 2013 © 2013 Am. Assoc. Phys. Med. 033901-10094-2405/2013/40(3)/033901/10/$30.00

033901-2 Candela-Juan et al.: Organ doses in HDR prostate brachytherapy 033901-2

I. INTRODUCTION

Prostate carcinoma is the second most frequently diagnosedtumor among men in economically developed countries (14%of cancer cases in 2008), being the sixth leading cause of can-cer death (6% of cancer deaths in males in 2008),1 althoughthere are variations among countries.2 If increasing life ex-pectancy is also considered, an even higher death rate due toprostate cancer is expected and it is thus worth directing ef-forts to improve treatment options for this disease.

Brachytherapy is a radiation treatment modality in whicha radiation source is placed near, in contact or inside (caseof the prostate) the tumor volume. The American Brachyther-apy Society has shown the success of high-dose-rate (HDR)brachytherapy applied to localized prostate cancer.3 In ad-dition to low toxicities (<5% for Grade 3 or higher), bio-chemical control rates of 85%–100%, 81%–100%, and 43%–93% have been reported for low-, intermediate-, and high-riskprostate tumors, respectively. From a population-based analy-sis, it is shown that mortality rates are reduced when applyingbrachytherapy alone or in combination with external beam ra-diation therapy (EBRT), even for high-risk cancers. In fact,besides HDR monotherapy being a treatment technique forthis disease, it also plays a fundamental tool as a boost (9 or15 Gy) after 60 or 46 Gy intensity modulated radiation ther-apy (IMRT), respectively.

Several manufacturers have traditionally offered HDR 192Iras a brachytherapy source. The new Eckert & Ziegler BEBIGGmbH MultiSource remote afterloader permits HDR 60Cobrachytherapy.4 The reasons why 60Co may replace in somecases the traditional 192Ir sources have already been reported.5

Physical differences between 60Co and 192Ir are less signifi-cant than prescription technique and the optimization param-eters. An advantage of 60Co over 192Ir is the significant costsavings due to source replacement every 2–6 years, whereas192Ir sources need replacement every 2–4 months. In compar-ison to 192Ir, equipment down-time and physics support timewith 60Co is also reduced by around 40%.6 A 60Co disadvan-tage is a need for more room shielding due to its higher energyphotons.

Tumor control is the main concern in a radiotherapy treat-ment plan. However, production of secondary cancers be-comes a criterion that can be used to establish the treatmentchoice when target coverage and dose sparing are compara-ble for different radiotherapy modalities. As stated in previ-ous works,7 new radiation treatment techniques such as IMRTor hadron therapy may increase cancer cure rates, althoughit is expected they can also increase secondary tumor inci-dence. This is because of the relatively high equivalent dosedeposition in organs at risk (OARs) that are within the pri-mary beam trajectory. In addition, a considerable increase inbeam-on time for IMRT gives to organs out of field a higherabsorbed dose, mainly from leakage radiation, since contribu-tion from scatter is equal or less than in 3D-conformal radio-therapy (3D-CRT) if proper collimation is used. In the caseof proton therapy or radiation therapy with high energy pho-tons, neutrons are created in the accelerator which, due totheir higher radiobiological effectiveness than photons, mayalso lead to an increase of secondary cancers.

Given the successful clinical results when usingbrachytherapy, IMRT, or proton therapy,7, 8 it is of interest todiscern the treatment modalities and compare probabilitiesfor secondary cancer occurrence. This comparison has beendone for 3D-CRT, IMRT, and proton therapy.7, 9–12 Thisassessment has also been done for brachytherapy appliedto the breast, using HDR 192Ir and electronic sources.13

However, it does not appear to have been done for HDRbrachytherapy with 60Co and 192Ir sources applied to prostatecancer, nor with comparison to EBRT techniques. Accordingto a summary on the topic done by Xu et al.,14 the AAPMTask Group 158 and the National Council on RadiationProtection and Measurements (NCRP) Scientific Committeeexcluded brachytherapy from their studies due to lack of data.This exclusion supports the study of organ equivalent dosesin a similar context to include brachytherapy.

Consequently, the aim of this work is to obtain equivalentdoses to a variety of organs when applying HDR brachyther-apy to the prostate using 60Co or 192Ir sources, and to provideEBRT comparisons. Given that absorbed doses to the bladderand rectum are dependent on patient-specific treatment plansand are already estimated during the planning process, the fo-cus is on farther organs. Due to the intrinsic difficulties andlimitations of absorbed dose measurements to organs, MonteCarlo (MC) simulated radiation transport has been selected asthe calculation method to estimate the needed data.

II. MATERIALS AND METHODS

II.A. Geometry definition

In order to reproduce a prostate brachytherapy treatment,a representative adult male phantom is needed. Since the1960s, more than 100 phantoms have been reported in theliterature,15–18 going from stylized phantoms formed by math-ematical shapes, to voxelized phantoms, and finally to meshphantoms, where their postures can be adjusted and the bodyorgans deformed. A recent and complete analysis of all thesebody phantoms can be found in Ref. 17.

In its Publication 110 (Ref. 19), the International Commis-sion on Radiological Protection (ICRP), jointly with the Inter-national Commission on Radiation Units and Measurements(ICRU), defines an official voxelized phantom that reproducesreference organ and tissue values given in ICRP Publica-tion 89 (Ref. 20). For design of the ICRP/ICRU 110 phan-tom, tomographic datasets from a real individual with phys-ical characteristics close to those from the reference phan-tom (176 cm tall and 73 kg weight) were selected, and voxelscaling was applied to adjust the body height and the skele-tal mass. Individual organs were then segmented and adjustedto reproduce reference masses within 0.01 g. The final phan-tom consists of over 140 organs and tissues, formed by almost1.95 million voxels (7.16 million voxels if exterior air is con-sidered). For the male phantom, the voxel height is 8 mm,whereas the voxel in-plane resolution is 2.137 mm with avoxel volume of 36.54 mm3. Since the physical posture ofa patient with prostate carcinoma who is being treated withHDR brachytherapy is not very different from the ICRP/ICRU110 phantom, this voxelized phantom was used in this work.

Medical Physics, Vol. 40, No. 3, March 2013

033901-3 Candela-Juan et al.: Organ doses in HDR prostate brachytherapy 033901-3

Some tissues/organs such as lymphatic nodes, blood, andbone marrow, among others, could not be perfectly segmenteddue to the limited voxel resolution.19 Lymphatic nodes wereincluded manually at specific locations, whereas a blood por-tion was incorporated in the elemental tissue composition ofeach organ. In addition, the mass percentage of bone marrowin the spongiosa part of skeletal targets is given in ICRP Pub-lication 110, which allows an estimation of absorbed dose tored bone marrow and endosteum tissue (bone surface), whichare part of the radiosensitive organs needed to obtain effectivedose.21 This is, however, an overestimation of absorbed dosein these two tissues given that secondary equilibrium betweenthe mineral bone and soft tissue components of the spongiosadoes not exist.19 Despite this, the voxelized phantom is con-sidered adequate for the purpose of this work given that vari-ation of absorbed doses due to exact organs resolutions is ex-pected to be small compared to variations between differentpatients. In addition, dealing with relatively high energy pho-tons, this voxel resolution should not be a concern.

In order to assess the adequacy of a homogeneous phantomfor HDR absorbed dose calculations, and to permit model val-idation against measurements by others, simulations were re-peated with the entire phantom composition replaced as liquidwater. In addition, an extra simulation with the heterogeneousvoxelized phantom immersed in water was later performed inorder to explain the behavior of organ doses as a function ofdistance when compared to an “infinite” medium. Additionalsimulations were performed in 1 m radii spheres of water(ρ = 1.00 g/cm3) and muscle (ρ = 1.05 g/cm3).

II.B. Brachytherapy sources

Energy spectra used in the MC simulations and their prob-abilities are the ones obtained by our group in previousstudies,22, 23 which provide spectra exiting real encapsulatedHDR 60Co and 192Ir sources. The photons were generatedat the prostate center of the voxelized phantom (i.e., pointsources) emitting photons isotropically. Source capsule ef-fects have been included in the simulations using those spec-tra. These spectra were also used for simulations at the centerof the spheres of water and muscle as explained above.

II.C. Monte Carlo simulation setup and organequivalent dose calculation

The simulation toolkit Geant4 version 9.4 (Ref. 24)was used to read ASCII data provided in ICRP/ICRU 110(Ref. 19) and to simulate the human phantom and the HDRbrachytherapy treatment. The low-energy electromagneticmodels of the Livermore physics package, which use detailedhandling, were employed. Standard multiple scattering crosssections were also utilized, with specialized processes foreach particle type. Detailed information about these modelscan be found in the Geant4 User Documentation webpage.25

This code has been widely validated for its use in brachyther-apy dosimetry.23, 26, 27 Secondary electrons were also tracked.For electrons and photons, the cutoff range was set to 0.1

mm, so secondary particles having a smaller range were notgenerated.

In order to hasten simulation time without extra memoryrequirement, the G4RegulatorNavigation algorithm was used,as implemented in the DICOM example of Geant4. It is basedon removing voxel frontiers when two voxels share the samematerial. An iterative algorithm is later applied to determineabsorbed dose in each individual voxel of a cluster. More de-tails about this technique, which has already been validated,can be found on the cited example of Geant4.25

A text file with absorbed dose in each voxel was obtainedas an output of the simulation. In-house developed softwarewas then used to convert this output to absorbed dose DT ineach organ or tissue T, by summing all voxel absorbed dosesin an organ and dividing by the number of voxels that form it.This is a mass weighted average considering that all voxels ofthe same tissue have the same mass. DT was then converted tomean absorbed dose per released photon dividing by the num-ber of events per simulation, N. Finally, organ equivalent doseper photon, HT/N, was obtained from absorbed dose througha radiation weighting factor wR:

HT

N=

∑

R

wR

DT,R

N. (1)

According to ICRP Publication 103,21 wR equals 1 for pho-tons and electrons. Since these are the only particle types inthe simulation environment, the equivalent dose HT in an or-gan was numerically equal to absorbed dose in that organ.Effective dose E was also obtained through a tissue weightingfactor wT:

E

N=

∑

T

wT

HT

N. (2)

The point source was located at the prostate center. Foreach source and phantom composition (heterogeneities or wa-ter), N = 109 initial photons were used, which provided statis-tical (Type A) uncertainties <1% for most of the organs con-sidered. As recommended in the joint AAPM/ESTRO TaskGroup No. 138 report,28 a coverage factor of k = 2 (con-fidence level of 95%) was used to express uncertainties inbrachytherapy dose.

The procedure by Pujades et al.29 was used to correlatethe MC simulations of the source at a single position at theprostate center with a realistic therapeutic absorbed dose tothe prostate for the source at various locations and with dif-ferent dwell times. They derived the relationship betweenprostate volume V , therapeutic absorbed dose DP, air-kermastrength SK, and total irradiation time t for 127 clinical cases.The correlation was presented as a nomogram fitted to a lin-ear function and allowed estimation of t for the source at thecenter of the prostate to deliver the required absorbed dose:

t × SK = DP (a × V + b), (3)

where a and b are fitting parameters obtained from real clin-ical cases.29 From previous MC studies,22, 23, 26, 27 the air-kerma strength per Bq, SK/A, was derived for both HDR 60Coand 192Ir sources. From all these clinical data, the number of

Medical Physics, Vol. 40, No. 3, March 2013

033901-4 Candela-Juan et al.: Organ doses in HDR prostate brachytherapy 033901-4

TABLE I. Organ equivalent dose per therapeutic absorbed dose to the prostate, (HT/DP), given to the heterogeneous voxelized phantom by the 60Co and 192Irsources. Distances correspond to those between the prostate and the organ center of masses. The statistical uncertainty (Type A) percentage (k = 2), ε, is alsoshown (but for those with ε < 0.05%), as well as the effective dose. The last column corresponds to the relative variation of absorbed dose between the waterand the heterogeneous phantom, �, being positive if absorbed dose in water is higher than in a tissue.

HT/DP (Sv/Gy)

60Co 192Ir � (%)

Organ name Mass (g) Distance (cm) Mean ε (%) Mean ε (%) 60Co 192Ir

Urinary bladder wall 50.0 5.0 1.30 × 10−1 0.1 1.60 × 10−1 0.1 0.6 − 0.4Rectum wall 30.0 5.3 1.51 × 10−1 0.1 1.83 × 10−1 0.1 0.6 0.5Testis, right 17.5 11.2 2.51 × 10−2 0.3 2.94 × 10−2 0.3 − 0.1 0.2Testis, left 17.5 11.3 2.42 × 10−2 0.2 2.82 × 10−2 0.2 0.2 0.2Small intestine wall 650.0 20.3 9.24 × 10−3 0.1 1.05 × 10−2 0.1 1.2 2.7Colon 340.0 24.0 6.21 × 10−3 0.1 6.78 × 10−3 0.1 0.9 2.7Kidney, right 157.0 29.9 1.75 × 10−3 0.3 1.40 × 10−3 0.3 5.1 11.5Kidney, left 153.0 31.1 1.56 × 10−3 0.5 1.20 × 10−3 0.3 4.9 10.8Pancreas 140.0 31.2 1.53 × 10−3 0.3 1.21 × 10−3 0.4 4.9 9.5Gall bladder wall 13.9 32.9 1.26 × 10−3 1.2 9.33 × 10−4 1.3 4.2 7.5Adrenal, right 7.0 33.6 1.11 × 10−3 1.9 7.63 × 10−4 1.4 10.5 19.5Adrenal, left 7.0 34.8 9.79 × 10−4 1.7 6.50 × 10−4 2.3 9.4 17.6Stomach wall 150.0 35.8 9.50 × 10−4 0.5 6.40 × 10−4 0.4 2.4 6.9Liver 1800.0 36.2 8.88 × 10−4 0.1 5.88 × 10−4 0.2 4.2 9.3Spleen 150.0 38.1 7.15 × 10−4 0.6 4.32 × 10−4 0.3 6.2 13.1Heart contents 510.0 44.3 4.03 × 10−4 0.3 2.09 × 10−4 0.7 9.3 17.9Heart wall 330.0 44.7 3.97 × 10−4 0.2 2.08 × 10−4 0.3 9.0 15.1Lung, right 471.7 47.3 3.37 × 10−4 0.4 1.60 × 10−4 0.4 1.0 8.8Lung, left 376.8 47.9 3.27 × 10−4 0.3 1.56 × 10−4 0.6 0.1 5.7Esophagus 40.0 50.0 3.01 × 10−4 1.8 1.55 × 10−4 1.9 14.7 23.3Spinal cord 36.6 50.9 4.35 × 10−4 1.3 2.51 × 10−4 1.7 13.1 34.4Thymus 25.0 54.6 1.53 × 10−4 2.5 6.16 × 10−5 4.5 13.6 6.3Trachea 10.0 57.0 1.28 × 10−4 4.1 4.74 × 10−5 6.1 11.4 18.5Thyroid 20.0 59.1 1.04 × 10−4 3.7 3.46 × 10−5 4.6 5.9 6.2Tongue (inner part) 42.3 70.0 5.62 × 10−5 1.8 1.50 × 10−5 6.7 0.6 0.8Salivary gland, right 42.5 70.3 5.88 × 10−5 2.3 1.45 × 10−5 6.7 − 16.7 − 14.1Salivary gland, left 42.5 70.3 4.71 × 10−5 2.5 1.20 × 10−5 6.3 − 0.9 − 4.0Brain 1450.0 79.5 2.08 × 10−5 1.1 4.17 × 10−6 2.0 2.6 4.3Active bone marrow . . . . . . 1.26 × 10−2 . . . 1.62 × 10−2 . . . 0.4 − 6.2Skin . . . . . . 3.21 × 10−3 . . . 3.23 × 10−3 0.1 2.5 4.7Lymphatic nodes . . . . . . 2.69 × 10−2 0.1 3.13 × 10−2 0.1 1.5 1.3

events N1 Gy necessary to provide a therapeutic absorbed doseof DP = 1 Gy to the prostate, as a function of V was obtained:

N1Gy

DP

= (a × V + b)

SK/A. (4)

Using N1 Gy with Eq. (1), equivalent dose per therapeu-tic absorbed dose to the prostate, HT/DP, was determined.Data correspond to those from a typical prostate volume ofV = 30 cm3, although variations with V were also analyzed.

III. RESULTS

III.A. Organ equivalent dose

N1 Gy = 8.4 × 1012 and N1 Gy = 3.0 × 1013 photon historiesfor 60Co and 192Ir, respectively, were needed to deliver 1 Gy oftherapeutic absorbed dose to a 30 cm3 prostate. If the methodby Pujades et al.29 had not been used, for that number of ini-

tial events, absorbed dose in the prostate would be 3.9 and4.9 Sv/Gy instead of 1 Sv/Gy for 60Co and 192Ir, respectively.They are larger than 1 Sv/Gy, which was expected given thatN1 Gy is used to account for a real case where the source islocated at various points inside the prostate. This source dis-tribution reduces overall absorbed dose in the prostate in ex-change for uniform dose deposition in the target volume. Thismethodology might underestimate absorbed doses to the near-est organs, mainly the ones in contact with the prostate, due toproximity of a specific source position. That increment mightincrease absorbed dose to rectum and urinary bladder by afactor that can be expected to be lower than an order of mag-nitude if full 3D source modeling was performed. For the restof the organs, the variations were expected to be negligible.

Table I shows mean equivalent doses (per therapeuticabsorbed dose to a 30 cm3 prostate), HT/DP, in severalradiosensitive organs in the heterogeneous phantom, for both

Medical Physics, Vol. 40, No. 3, March 2013

033901-5 Candela-Juan et al.: Organ doses in HDR prostate brachytherapy 033901-5

FIG. 1. Organ absorbed doses obtained in this work as a function of distanceto the prostate for: (a) 60Co source and (b) 192Ir source. Distances are betweenthe prostate CM and the organ CM. Absorbed doses are normalized such thatD(1 cm) = 1 (arbitrary units). The fitting curve obtained from Venselaar et al.(Ref. 30), which is valid for a point source in an infinite water medium, is alsoshown and compared with depth dose distribution in the 1 m radius watersphere simulated in this work, with real spectra exiting the source capsules.

the 60Co and 192Ir sources, all together with the Type A sta-tistical uncertainties. Effective dose is also included. In addi-tion, the relative variation of organ equivalent doses for thewater phantom in comparison to the heterogeneous phantomis shown. For completeness, organ masses and distances, r,from their center of mass to the center of mass of the prostateare also given.

Figure 1 shows organ absorbed doses as a function of r.The depth dose distribution in the 1 m radius water sphereand fitted curves obtained by Venselaar et al.30 from measure-ments of point sources (60Co and 192Ir) in water are plotted[see Figs. 1(a) and 1(b), respectively] for comparison. Simi-larity of the curves validates the simulations and physics pack-age used.

There is reasonable agreement between organ absorbeddoses and the fitted curves. There are three organs that re-ceived an absorbed dose higher than predicted by simulationsin water: spinal cord, small intestine wall, and the colon. Thelatter two are close to the prostate, and all three subtend vol-umes over a large range of distances from the prostate. Fororgans with d > 30 cm, organ absorbed dose is smaller thanexpected in water by 50%–125% for 192Ir, and by 17%–60%for 60Co. These discrepancies are also present for organ ab-sorbed doses in the homogeneous water phantom. There aretwo known radiological effects that contribute to these dif-ferences. First, there is dosimetric difference between wa-ter and soft tissue for d > 30 cm, as shown later. Second,there is lack of backscatter at the phantom skin comparedto an infinite medium. This was demonstrated by immers-ing the heterogeneous voxelized phantom in water. With thisconfiguration, organ absorbed doses increased ∼25% withvariations between different organs.

III.B. Influence of prostate size

Mean equivalent dose in each organ is proportional to thenumber of photons emitted by the source, N1 Gy. In addition,according to Eq. (4), N1 Gy is linear with V . Hence, if D0 isthe absorbed dose in a specific organ when the prostate vol-ume is V 0 = 30 cm3, when V = V0 (1 + X), i.e., when thevolume is X% higher or lower, then the relative variation oforgan absorbed dose is

DX − D0

D0= X

a × V0

a × V0 + b. (5)

Using real clinical values obtained by Pujades et al.29

(a = 0.0605 cm−1, b = 1.7 cm2), then the relative variationis 0.52 times X. Therefore, if prostate volume increases (de-creases) 10% relative to the reference volume, equivalent dosein each organ increases (decreases) by ∼5.2%.

III.C. 60Co vs 192Ir

Figure 2 represents the 60Co to 192Ir absorbed dose ratiofor each organ, for the heterogeneous and the water phan-tom. Presented error bars are statistical (Type A) uncertain-ties from MC simulations. Absorbed doses have been normal-ized such that the prostate receives the same absorbed dosefrom both sources. Absorbed doses in those organs near thetreated prostate, such as the bladder, rectum, testes, small in-testine, and colon are between 8% and 19% smaller whenusing 60Co. This can be rationalized considering that in areal treatment, as shown above, to give the same absorbeddose at 1 cm from the source, the required activity of 60Cois smaller and so is the number of emitted photons. This isnot the case for distant organs. As the distance between theprostate and the organ increases, so does the relative contri-bution of 60Co absorbed dose in comparison to 192Ir. Thiscan be understood from the mean energy of emitted photonsfor each radionuclide, which is higher for 60Co and has

Medical Physics, Vol. 40, No. 3, March 2013

033901-6 Candela-Juan et al.: Organ doses in HDR prostate brachytherapy 033901-6

FIG. 2. 60Co to 192Ir absorbed dose ratio for different body organs, using aheterogeneous and a water phantom.

greater penetration within the body. Given that the sametherapeutic absorbed dose is prescribed with both radionu-clides, and that estimated effective doses were 11.1 mSv/Gyfor 60Co and 13.2 mSv/Gy for 192Ir, an overall advantage of60Co seems to exist when organ doses are considered.

III.D. Heterogeneous vs water phantom

Currently, brachytherapy treatment planning systems con-sider the body to be composed of water. In this section, va-lidity of this assumption when obtaining organ absorbed dosein the whole body is analyzed. Figure 3 shows the organ ab-sorbed dose ratio of heterogeneous media to water for the twosources. For the closest organs considered, absorbed dose wasthe same for all organs having near-unity mass density.

However, as the distance increases the absorbed dose ra-tio decreases, reaching differences of almost 25% (except forthe spinal cord, which is surrounded by bone and so the ab-sorbed dose ratio is even lower), being lower for 192Ir thanfor 60Co (except for the thymus and the left salivary gland,where uncertainties are relatively high and so no conclusionscan be extracted). The reason for the absorbed dose rate re-

FIG. 3. Heterogeneous to water absorbed dose ratio, using 60Co and 192Irsources.

FIG. 4. Muscle to water absorbed dose ratio for 60Co and 192Ir encapsulatedsources at the center of 1 m radius spheres. Absorbed doses are averaged overshells of 0.5 cm width.

duction is the material composition. In order to prove this,absorbed dose as a function of distance was obtained in wa-ter and muscle spheres with the sources placed at the centers.The absorbed dose ratios when using both 60Co and 192Ir, withtheir real spectra exiting the sources, are shown in Fig. 4.

As an example, at 30 cm from the source, the muscle to wa-ter absorbed dose ratio for 192Ir is 0.92, being the differencebetween the kidneys and its corresponding water volume ofaround 10%, in agreement. Hence, the general decrease ofthe heterogeneous to water absorbed dose ratio can be ex-plained considering the differences between real tissue andwater, which are negligible for distances below 10–20 cm,but not after that. With 60Co, which has higher mean pho-ton energy, photoelectric effect is not dominant and so thedifference between water and soft tissue is smaller than for192Ir. The other factor which influences the curve from Fig. 3is the presence of different materials such as air inside someorgans, bone, or the lungs. An interesting case is the salivaryglands, where the heterogeneous to water absorbed dose ra-tio between the left and the right gland might seem consistentfor the 192Ir source considering their uncertainties, but not forthe 60Co source. This difference can be explained taking intoaccount that the prostate center of mass of the voxelized phan-tom is almost 1 cm at right from our geometrical center, andso it traverses more lung to reach the right salivary gland thanto reach the left one. In addition, the right lung is higher thanthe left one due to the presence of the heart, and so the effectis even higher.

IV. DISCUSSION

It is interesting to compare equivalent doses to organsfrom the different radiation modalities (brachytherapy, 3D-conformal radiotherapy, IMRT, and proton therapy).

There are some studies regarding equivalent dose fromproton therapy applied to prostate. Fontenot et al.31 esti-mated equivalent doses from stray radiation, i.e., neutrons and

Medical Physics, Vol. 40, No. 3, March 2013

033901-7 Candela-Juan et al.: Organ doses in HDR prostate brachytherapy 033901-7

photons generated (either inside or outside the body) in apassively scattered proton treatment to radiosensitive organs.The prescribed equivalent dose was 75.6 CGE (cobalt grayequivalent), i.e., 68.7 Gy × 1.1 RBE (radiobiological effec-tiveness of protons), to the clinical treatment volume. Simula-tions were run in the Monte Carlo N-Particle eXtended (MC-NPX) code, using an anatomically realistic male phantom de-veloped by Billings and Yucker.32 In their first work, they didnot consider absorbed dose from the therapeutic beam, whichmainly concerns additional absorbed dose to the bladder andrectum. In a follow-up paper,10 they considered primary ab-sorbed dose calculated using a commercial treatment planningsystem applied to three patients. Summing the contributionsfrom secondary particles, they obtained total equivalent doseand estimated the probability of secondary tumor inductionusing BEIR VII recommendations.33 Given that some datawere not explicitly written in the paper,10 the authors werecontacted and kindly provided the needed details. The threepatients were of different sizes (small, medium, and large).Their medium sized data are compared in the current study.In their work, equivalent dose to the colon was obtained as amass-weighted average of the equivalent doses to the colonand rectum:

Dtotal−colon = wcolonDcolon + wrectumDrectum, (6)

where the colon wcolon and rectum wrectum mass weights were20% and 80%, respectively, according to ICRP 89.20

Bednarz et al.11, 12 estimated organ equivalent doses andrisk probabilities from two 3D-CRT (a four-field box anda four-field box + six-field boost), and a seven-field IMRTtreatment applied to the prostate. Absorbed doses were es-timated through MCNPX using a voxelized male phantombased on the RPI-AM phantom,18 also in agreement with the

FIG. 5. Organ equivalent dose per therapeutic absorbed dose for differentradiation treatments applied to the prostate: brachytherapy (from this work,with 60Co and 192Ir), IMRT (Refs. 11 and 31), and proton therapy (Refs. 10and 31).

male reference phantom from ICRP Publication 89, and a pre-viously validated Varian Clinac 2100C linear accelerator.11

Their results are within an order of magnitude of those ob-tained by Kry et al.34 and Howell et al.35 Study discrepanciesare due in part to the different organ locations and methodsused to obtain absorbed doses. Organs within the therapeuticbeam (testes, bladder, skin, and prostate) were not yet con-sidered in those papers.11, 12 Fontenot et al.10 considered pri-mary radiation absorbed dose for IMRT to colon, bladder, andtestes. Equation (6) was also used to obtain equivalent dose tocolon from the IMRT treatment.

Figure 5 shows organ equivalent dose per therapeutic ab-sorbed dose for three radiation modalities: brachytherapy,IMRT, and proton therapy (numerical data in Table II include

TABLE II. Equivalent dose per therapeutic absorbed dose (HT/DP) for some body organs, comparing brachytherapy results from this work in a heterogeneousphantom and previously reported data for 3D-CRT (Refs. 11 and 12), IMRT (Refs. 11 and 31), and protontherapy (Refs. 10 and 31). Relative uncertainties ε

(Type A for brachytherapy (but for those with ε < 0.05%) and the ones published for EBRT) are given in brackets (%) with one decimal unit.

HT/DP (μSv/Gy)Organ name 60Co 192Ir Box + boost IMRT Protons

Bladder 1.30 × 10−1 (0.1) 1.60 × 10−1 (0.1) . . . 2.57 × 10−1 1.87 × 10−1 (14.5)Rectum 1.51 × 10−1 (0.1) 1.83 × 10−1 (0.1) . . . 3.65 × 10−1 2.75 × 10−1 (3.4)Testes 2.51 × 10−2 (0.3) 2.94 × 10−2 (0.3) . . . 1.26 × 10−2 7.60 × 10−3

Small intestine 9.24 × 10−3 (0.1) 1.05 × 10−2 (0.1) 2.36 × 10−2 (0.4) 1.16 × 10−2 (0.8) . . .Colon 6.21 × 10−3 (0.1) 6.78 × 10−3 (0.1) 9.07 × 10−3 (1.4) 6.01 × 10−3 (3.2) 7.80 × 10−3

Kidneys 1.75 × 10−3 (0.3) 1.40 × 10−3 (0.3) 5.88 × 10−3 (2.3) 3.14 × 10−3 (3.7) . . .Pancreas 1.53 × 10−3 (0.3) 1.21 × 10−3 (0.4) 3.43 × 10−3 (3.8) 2.27 × 10−3 (3.5) . . .Gall bladder 1.26 × 10−3 (1.0) 9.33 × 10−4 (1.3) 3.02 × 10−3 (3.4) 2.55 × 10−3 (2.5) . . .Stomach 9.50 × 10−4 (0.5) 6.40 × 10−4 (0.4) 3.50 × 10−3 (1.5) 2.87 × 10−3 (2.9) 3.40 × 10−3

Liver 8.88 × 10−4 (0.1) 5.88 × 10−4 (0.2) 3.68 × 10−3 (1.1) 2.65 × 10−3 (2.0) 3.40 × 10−3

Spleen 7.15 × 10−4 (0.6) 4.32 × 10−4 (0.3) 4.15 × 10−3 (2.3) 2.33 × 10−3 (3.6) . . .Heart 3.97 × 10−4 (0.2) 2.08 × 10−4 (0.3) 2.14 × 10−3 (2.0) 1.95 × 10−3 (3.6) . . .Lungs 3.37 × 10−4 (0.4) 1.60 × 10−4 (0.4) 2.76 × 10−3 (1.5) 1.57 × 10−3 (2.0) 2.50 × 10−3

Esophagus 3.01 × 10−4 (1.8) 1.55 × 10−4 (1.9) 1.60 × 10−3 (10.4) 9.75 × 10−4 (3.2) 1.90 × 10−3

Thyroid 1.04 × 10−4 (3.7) 3.46 × 10−5 (4.6) 2.12 × 10−3 (17.5) 8.67 × 10−4 (10.8) 1.90 × 10−3

Brain 2.08 × 10−5 (1.1) 4.17 × 10−6 (2.0) 1.21 × 10−3 (2.4) 6.50 × 10−4 (3.6) . . .Active bone marrow 1.26 × 10−2 . . . 1.62 × 10−2 . . . 2.28 × 10−3 (1.9) 2.00 × 10−3 (4.3) 6.40 × 10−3

Skin 3.21 × 10−3 . . . 3.23 × 10−3 (0.1) . . . . . . 6.40 × 10−3

Medical Physics, Vol. 40, No. 3, March 2013

033901-8 Candela-Juan et al.: Organ doses in HDR prostate brachytherapy 033901-8

3D-CRT). First, HT/DP in organs near the prostate (e.g., testes,urinary bladder, and colon) are about factor of 2 larger whenusing IMRT than when using proton therapy. A higher amountof beam time is needed for IMRT to conform absorbed doseto the prostate. In fact, besides results from Fontenot et al.for protons, which are based on a passive scanning system,there are other studies by Schneider7, 36 that have shown thatthe probability of secondary cancers could be decreased by afactor of 2 when using spot-scanned proton therapy instead ofIMRT. A spot-scanned system requires fewer beam conformalelements and results in a lower secondary neutron flux than ina passive system.

When considering HDR brachytherapy with 60Co and192Ir, equivalent doses (per prescribed absorbed dose) to or-gans up to the stomach (around 35 cm from the prostate)are the same within an order of magnitude when comparedto IMRT and proton therapy. For the testes, since a directbeam can be avoided when using IMRT and protons, equiv-alent doses are smaller for both EBRT. Actually, this pro-vides an effective dose of 12 mSv/CGE for proton therapy and13 mSv/Gy for IMRT, similar to 11.1 and 13.2 mSv/Gy ob-tained for both brachytherapy sources. Nevertheless, in theprevious figure, only stray radiation is considered for thetestes when using IMRT, and so real equivalent dose thereis expected to be higher. For all these near organs, there mightbe variations between patients within an order of magnitude.As an example, Fontenot et al.10 estimate an HT/DP = 0.07Sv/Gy for the colon when using IMRT, a factor 10 higherfrom what it is shown here. In addition, estimation done byFontenot et al. of equivalent dose to skin and bone marrow,which contributes considerably to effective dose, is averagevalues from equivalent dose in the other organs they consid-ered. Furthermore, in this work, equivalent dose to bone mar-row was overestimated, as explained above, although equiva-lent doses to rectum and bladder might be underestimated. Asa result, effective doses per therapeutic absorbed dose mightbe similar between all radiation therapy modalities includedin this comparison. These results are estimations for the con-sidered cases and exact conclusions for the nearest organs arepatient and equipment dependent. For 3D-CRT, dosimetriccontributions from the primary beam to those organs at riskwere not available and no conclusions may be drawn.

For the furthest organs, equivalent doses were approxi-mately the same (within an order of magnitude) throughoutthe whole body (Fig. 5) for 3D-CRT, proton therapy, andIMRT. Although equivalent doses were smaller for larger dis-tances, given a smaller contribution from scattering radiation,they were within the same order of magnitude. This was notthe case when using brachytherapy. As the distance increased,equivalent dose decreased rapidly, going from approximately1 mSv/Gy in the stomach to 10 μSv/Gy in the brain. For thefurthest organs, equivalent doses given by brachytherapy arebetween 1 and 2 orders of magnitude smaller to those deliv-ered by other radiation modalities. This gives brachytherapy aclear advantage if, for a given prostate absorbed dose, equiv-alent dose for distant organs needs to be minimized.

In this work, probabilities for secondary cancer inductionhave not been estimated given the low equivalent doses re-

ceived by organs far from the prostate and given the dose-risk model uncertainties in this dose range.37 Before doingthese calculations, a verified model from epidemiological datais needed, for which organ equivalent doses provided in thiswork can be used. This topic has been considered by Bed-narz et al.12 where these probabilities have been estimatedfor 3D-CRT and IMRT applying the BEIR VII model. Theyfound that probabilities for secondary cancer induction for theconsidered organs were about an order of magnitude lowerthan the baseline risks taken from the SEER Cancer StatisticsReview.38 It could be thought that given this low probability,one need not concern with secondary cancer induction follow-ing definitive radiation therapy. However, given the high rateof tumors diagnosed, in absolute terms, the number is consid-erable and so secondary cancer induction should be seriouslyaddressed.

Although this work provides results concerning radiationprotection applications in radiotherapy, the presented conclu-sions are based on certain assumptions. First, the point sourcewas placed in the prostate center. Although it was correctedfrom empirical clinical data (nomogram) to account for a clin-ical source with varying dwell positions within the prostate, itwould be closer to an OAR at some positions and would causesome discrepancies with results presented herein. This varia-tion, however, is lower than an order of magnitude and wouldonly apply to the nearest organs, being negligible for the rest.The aim of this work was to provide a general organ equiv-alent dose database for this particular treatment, comparing60Co and 192Ir sources in the same circumstances, bearingin mind that equivalent dose to the closest organs is patientand facility dependent, and so general values for these tis-sues cannot be established. Hence, the first assumption wasnot a limitation of the proposed aim. Second, the differentradiation treatments have been compared according to organequivalent dose per therapeutic absorbed dose. It has not beentaken into account that the therapeutic course is generallygiven in several fractions to take advantage of biological ef-fects like cell repair. The hypothesis made is that cell dam-age is proportional to overall absorbed dose, without consid-ering some biological effects which are dependent on param-eters that could be modality dependent. If this hypothesis iscorrect, then this work shows an advantage of brachyther-apy in comparison to IMRT and proton therapy whenconsidering equivalent doses to the farthest organs. If not, abiological model should be included. However, given that sev-eral possible fractionating plans can be applied for the differ-ent radiation treatments, with different total prostate absorbeddose in each case, no general conclusions can be extracted.As stated above, comparison herein between the different ra-diation modalities cannot be considered as representative ofall particular cases. Next, when comparing 60Co and 192Irsources, the same amount of absorbed dose was assumed inboth cases to produce the same biological effects, i.e., depen-dence of biological effects was assumed to be independentof source energy. A more precise estimation would accountfor radiobiological effectiveness (it was considered as unity inthis work) as a function of energy.39 However, this is expectedto be a proportionately small correction for these high energy

Medical Physics, Vol. 40, No. 3, March 2013

033901-9 Candela-Juan et al.: Organ doses in HDR prostate brachytherapy 033901-9

sources that would not change the overall findings. Therefore,after all these considerations, data presented in Table I can beconsidered as representative organ equivalent doses receivedby a patient with a body height similar to the reference malephantom, considering however a 30 cm3 prostate instead of16.5 cm3. The latter corresponds to the prostate size of thereference male phantom, which has many variations amongpeople and age.

V. CONCLUSION

A database of organ equivalent doses when applying HDRbrachytherapy to the prostate with either 60Co or 192Ir is pro-vided. Making only physical considerations, 192Ir seems to bea better choice than 60Co when considering damage to dis-tant organs, which have to be also considered because of theirradiosensitivity, but not to the farthest ones, which are theones that receive a considerably higher equivalent dose. Es-timated total effective doses per clinical absorbed dose were11.1 mSv/Gy for 60Co and 13.2 mSv/Gy for 192Ir.

Up to around 30 cm, organ variations due to differences be-tween water and soft tissue are negligible. However, for largerdistances this is not the case. More variations with an infinitewater medium are due to the lack of back-scattering, whichdepends on organ positions. Both effects together make thatan infinite water volume cannot be used to obtain organ ab-sorbed doses all over the body. Finally, the radiation treatment(considering HDR brachytherapy, IMRT, and proton therapy)to be used if considering organ equivalent doses are patient,treatment, and facility dependent. In any case, as the distanceincreases, brachytherapy shows a radiation protection advan-tage over all EBRT techniques.

ACKNOWLEDGMENTS

This study was supported in part by Generalitat Valen-ciana (Project No. PROMETEO2008/114) and Spanish Gov-ernment (Project No. FIS2010-17007).

a)Author to whom correspondence should be addressed. Electronic mail:[email protected]

1A. Jemal, F. Bray, M. M. Center, J. Ferlay, E. Ward, and D. Forman,“Global cancer statistics,” CA Cancer J. Clin. 61, 69–90 (2011).

2IARC, “Worldwide Cancer Incidence Statistics—Prostate,” in JNCI Can-cer Spectrum (Oxford University Press, 2001).

3Y. Yamada, L. Rogers, D. J. Demanes, G. Morton, B. R. Prestidge,J. Pouliot, G. N. Cohen, M. Zaider, M. Ghilezan, and I. C. Hsu, “AmericanBrachytherapy Society consensus guidelines for high-dose-rate prostatebrachytherapy,” Brachytherapy 11, 20–32 (2012).

4http://www.Bebig.eu. Last accessed January 24, 2013.5J. Richter, K. Baier, and M. Flentje, “Comparison of 60cobalt and 192irid-ium sources in high dose rate afterloading brachytherapy,” Strahlenther.Onkol. 184, 187–192 (2008).

6http://www.seedos.co.uk/co60_hdr_article.htm. Last accessed January 24,2013.

7U. Schneider, A. Lomax, P. Pemler, J. Besserer, D. Ross, N. Lombriser, andB. Kaser-Hotz, “The impact of IMRT and proton radiotherapy on secondarycancer incidence,” Strahlenther. Onkol. 182, 647–652 (2006).

8X. Shen, S. W. Keith, M. V. Mishra, A. P. Dicker, and T. N. Showalter, “Theimpact of brachytherapy on prostate cancer-specific mortality for definitiveradiation therapy of high-grade prostate cancer: A population-based analy-sis,” Int. J. Radiat. Oncol., Biol., Phys. 83, 1154–1159 (2012).

9S. Stathakis, J. Li, and C. C. Ma, “Monte Carlo determination of radiation-induced cancer risks for prostate patients undergoing intensity-modulatedradiation therapy,” J. Appl. Clin. Med. Phys. 8, 14–27 (2007).

10J. D. Fontenot, A. K. Lee, and W. D. Newhauser, “Risk of secondary malig-nant neoplasms from proton therapy and intensity-modulated x-ray therapyfor early-stage prostate cancer,” Int. J. Radiat. Oncol., Biol., Phys. 74, 616–622 (2009).

11B. Bednarz, C. Hancox, and X. G. Xu, “Calculated organ doses fromselected prostate treatment plans using Monte Carlo simulations and ananatomically realistic computational phantom,” Phys. Med. Biol. 54, 5271–5286 (2009).

12B. Bednarz, B. Athar, and X. G. Xu, “A comparative study on the risk ofsecond primary cancers in out-of-field organs associated with radiotherapyof localized prostate carcinoma using Monte Carlo-based accelerator andpatient models,” Med. Phys. 37, 1987–1994 (2010).

13M. M. Mille, X. G. Xu, and M. J. Rivard, “Comparison of organ doses forpatients undergoing balloon brachytherapy of the breast with HDR 192Iror electronic sources using Monte Carlo simulations in a heterogeneoushuman phantom,” Med. Phys. 37, 662–671 (2010).

14X. G. Xu, B. Bednarz, and H. Paganetti, “A review of dosimetry studies onexternal-beam radiation treatment with respect to second cancer induction,”Phys. Med. Biol. 53, R193–241 (2008).

15M. Caon, “Voxel-based computational models of real human anatomy: Areview,” Radiat. Environ. Biophys. 42, 229–235 (2004).

16H. Zaidi and X. G. Xu, “Computational anthropomorphic models of thehuman anatomy: The path to realistic Monte Carlo modeling in radiologicalsciences,” Annu. Rev. Biomed. Eng. 9, 471–500 (2007).

17X. G. Xu and K. F. Eckerman, Handbook of Anatomical Models for Radia-tion Dosimetry (CRC/Taylor & Francis, Boca Raton, FL, 2010).

18J. Zhang, Y. H. Na, P. F. Caracappa, and X. G. Xu, “RPI-AM and RPI-AF, a pair of mesh-based, size-adjustable adult male and female compu-tational phantoms using ICRP-89 parameters and their calculations for or-gan doses from monoenergetic photon beams,” Phys. Med. Biol. 54, 5885–5908 (2009).

19H. G. Menzel, C. Clement, and P. DeLuca, “ICRP Publication 110. Re-alistic reference phantoms: An ICRP/ICRU joint effort. A report of adultreference computational phantoms,” Ann. ICRP 39, 3–5 (2009).

20“Basic anatomical and physiological data for use in radiological protection:Reference values. A report of age- and gender-related differences in theanatomical and physiological characteristics of reference individuals. ICRPPublication 89,” Ann. ICRP 32, 5–265 (2002).

21“The 2007 Recommendations of the International Commission on Ra-diological Protection. ICRP publication 103,” Ann. ICRP 37, 3–5(2007).

22F. Ballester, D. Granero, J. Perez-Calatayud, C. S. Melhus, and M. J. Ri-vard, “Evaluation of high-energy brachytherapy source electronic dise-quilibrium and dose from emitted electrons,” Med. Phys. 36, 4250–4256(2009).

23M. J. Rivard, D. Granero, J. Perez-Calatayud, and F. Ballester, “Influenceof photon energy spectra from brachytherapy sources on Monte Carlo sim-ulations of kerma and dose rates in water and air,” Med. Phys. 37, 869–876(2010).

24GEANT4 Collaboration, “GEANT4-a simulation toolkit,” Nucl. Instrum.Methods Phys. Res. 506, 250–303 (2003).

25Geant4 user documentation, http://geant4.cern.ch/support/userdocuments.shtml. Last accessed January 24, 2013.

26D. Granero, J. Vijande, F. Ballester, and M. J. Rivard, “Dosimetry revisitedfor the HDR 192Ir brachytherapy source model mHDR-v2,” Med. Phys. 38,487–494 (2011).

27J. Vijande, D. Granero, J. Perez-Calatayud, and F. Ballester, “Monte Carlodosimetric study of the Flexisource Co-60 high dose rate source,” J. Con-temp. Brachyther. 4, 34–44 (2012).

28L. A. DeWerd, G. S. Ibbott, A. S. Meigooni, M. G. Mitch, M. J. Rivard,K. E. Stump, and B. R. Thomadsen, “A dosimetric uncertainty analysis forphoton-emitting brachytherapy sources: Report of AAPM Task Group No.138 and GEC-ESTRO,” Med. Phys. 38, 782–801 (2011).

29M. C. Pujades, C. Camacho, J. Perez-Calatayud, J. Richart, J. Gimeno,F. Lliso, V. Carmona, F. Ballester, V. Crispin, S. Rodriguez, and A. Tormo,“The use of nomograms in LDR-HDR prostate brachytherapy,” J. Con-temp. Brachyther. 3, 121–124 (2011).

30J. L. Venselaar, P. H. van der Giessen, and W. J. Dries, “Measurement andcalculation of the dose at large distances from brachytherapy sources: Cs-137, Ir-192, and Co-60,” Med. Phys. 23, 537–543 (1996).

Medical Physics, Vol. 40, No. 3, March 2013

033901-10 Candela-Juan et al.: Organ doses in HDR prostate brachytherapy 033901-10

31J. Fontenot, P. Taddei, Y. Zheng, D. Mirkovic, T. Jordan, andW. Newhauser, “Equivalent dose and effective dose from stray radiationduring passively scattered proton radiotherapy for prostate cancer,” Phys.Med. Biol. 53, 1677–1688 (2008).

32M. P. Billings, and W. R. Yucker, Summary final report: The ComputerizedAnatomical Man (CAM) Report, McDonnell Douglas Corporation MDCG4655, 1973.

33Committee to Assess Health Risks from Exposure to Low Levels of Ioniz-ing Radiation (2006) Health risks from exposure to low levels of ionizingradiation, BEIR VII phase 2 (National Academic Press, Washington, DC).

34S. F. Kry, M. Salehpour, D. S. Followill, M. Stovall, D. A. Kuban,R. A. White, and I. I. Rosen, “Out-of-field photon and neutron dose equiv-alents from step-and-shoot intensity-modulated radiation therapy,” Int. J.Radiat. Oncol., Biol., Phys. 62, 1204–1216 (2005).

35R. M. Howell, N. E. Hertel, Z. Wang, J. Hutchinson, and G. D. Fuller-ton, “Calculation of effective dose from measurements of secondary neu-

tron spectra and scattered photon dose from dynamic MLC IMRT for6 MV, 15 MV, and 18 MV beam energies,” Med. Phys. 33, 360–368(2006).

36U. Schneider, A. Lomax, J. Besserer, P. Pemler, N. Lombriser, andB. Kaser-Hotz, “The impact of dose escalation on secondary cancer riskafter radiotherapy of prostate cancer,” Int. J. Radiat. Oncol., Biol., Phys.68, 892–897 (2007).

37E. J. Hall, “Henry S. Kaplan Distinguished Scientist Award 2003. Thecrooked shall be made straight; dose-response relationships for carcino-genesis,” Int. J. Radiat. Biol. 80, 327–337 (2004).

38N. Howlader, A. M. Noone, M. Krapcho, and N. Neyman, SEER CancerStatistics Review 1975-2009 (Vintage 2009 Populations) (National CancerInstitute, Bethesda, MD, 2012).

39B. Reniers, D. Liu, T. Rusch, and F. Verhaegen, “Calculation of relativebiological effectiveness of a low-energy electronic brachytherapy source,”Phys. Med. Biol. 53, 7125–7135 (2008).

Medical Physics, Vol. 40, No. 3, March 2013