Dosimetry comparison of 192 Ir SourcesP. Papagiannis, A. Angelopoulos, E. Pantelis, L. Sakelliou, D. Baltas, P. Karaiskos, P. Sandilos, and L. Vlachos

Citation: Medical Physics 29, 2239 (2002); doi: 10.1118/1.1508378 View online: http://dx.doi.org/10.1118/1.1508378 View Table of Contents: http://scitation.aip.org/content/aapm/journal/medphys/29/10?ver=pdfcov Published by the American Association of Physicists in Medicine Articles you may be interested in Technical note: Monte-Carlo dosimetry of the HDR 12i and Plus 192 Ir sources Med. Phys. 28, 2586 (2001); 10.1118/1.1420398 Dosimetry close to an 192 Ir HDR source using N-vinylpyrrolidone based polymer gels and magnetic resonanceimaging Med. Phys. 28, 1416 (2001); 10.1118/1.1382603 Monte Carlo dosimetry of a new 192 Ir high dose rate brachytherapy source Med. Phys. 27, 2521 (2000); 10.1118/1.1315316 Spectra and air-kerma strength for encapsulated 192 Ir sources Med. Phys. 26, 2441 (1999); 10.1118/1.598763 Monte Carlo and TLD dosimetry of an 192 Ir high dose-rate brachytherapy source Med. Phys. 25, 1975 (1998); 10.1118/1.598371

Dosimetry comparison of 192Ir SourcesP. Papagiannis,a) A. Angelopoulos, E. Pantelis, and L. Sakellioub)

Nuclear and Particle Physics Section, Physics Department, University of Athens, Panepistimioupolis, Ilisia,157 71, Athens, Greece

D. BaltasDepartment of Medical Physics and Engineering, Strahlenklinik, Klinikum Offenbach, 63069 Offenbach,Germany and Institute of Communication and Computer Systems, National Technical University ofAthens, Zografou, 157 73 Athens, Greece

P. Karaiskos and P. SandilosMedical Physics Department, Hygeia Hospital, Kifisias Avenue and Erythrou Stavrou 4, Marousi,151 23 Athens, Greece

L. VlachosDepartment of Radiology, Medical School, University of Athens, Areteion Hospital, 76 Vas. Sofias Avenue,115 28, Athens, Greece

~Received 31 January 2002; accepted for publication 11 July 2002; published 12 September 2002!

192Ir sources besides being widely utilized in the field of conventional brachytherapy also find usein contemporary peripheral and coronal intravascular applications. In this study, the same MonteCarlo simulation code and input data were used to investigate differences between the dose ratedistributions of the most commonly used192Ir sources in the cm and mm distance range. Findingsare discussed in view of differences in source and encapsulation dimensions as well as structuraldetails. Results are presented in the AAPM TG-43 formalism, as generalized by AAPM TG-60, forfive 192Ir HDR source designs as well as an LDR seed and an LDR wire source. Dose rate constantsof the sources atr 051 cm andr 052 mm were found proportional to the corresponding geometryfactors along the transverse source bisectors and an equation of the formL r 0

(cGyh21 U21)51.123G(r 0,90°) provides results within clinical accuracy~less than 2%! for any 192Ir source.Radial dose functions do not depend significantly on source and encapsulation geometry and agreewithin 2% with that of a point192Ir source. Anisotropy is of importance for accurate dosimetry atthe cm distance range but it does not affect dose rate in the mm distance range significantly. At suchshort radial distances the source geometry factor defines the shape of isodose lines. Dose uniformityat given distances from the sources is strongly dependent on source dimensions as indicated by doserate profiles in polar and Cartesian coordinates. ©2002 American Association of Physicists inMedicine. @DOI: 10.1118/1.1508378#

Key words: dosimetry, Monte Carlo,192Ir, intravascular

tecnftuatm

ptine

i-elog

im-es

ving

e

the.n ac-

e--43ceser-sing

I. INTRODUCTION

192Ir sources are the most commonly used sources in institial brachytherapy and their dosimetry in the cm distanrange that is of interest to such applications is well knowHowever, currently utilized192Ir sources present a variety osource and encapsulation dimensions as well as strucdetails, and therefore it would be of interest to investigtheir dosimetric differences following the widespread dosietric formalism proposed by AAPM TG-43.1 Moreover,192Irsources have also presented encouraging results in thevention of restenosis after balloon angioplasty and stenof coronary and peripheral vessels. For coronary vessclinical trials utilize 192Ir radioactive seed ribbons~BestMedical! and the AngioRad™192Ir wire.2 For peripheral ves-sels, conventional high dose rate~HDR! remote after loadingdevices ~potentially including the microSelectron, VarSource and Buchler sources! are the modality of choice sincpatients can be safely transported to the radiation onco

2239 Med. Phys. 29 „10…, October 2002 0094-2405 Õ2002Õ29

r-e.

rale-

re-gls,

y

department with the treatment catheter in place.3 In theabove as well as intrabronchial applications, accurate dosetry in the mm distance range in the proximity of the sourcis required. Therefore, the AAPM TG-604 has recommendeda modified notation to generalize the TG-431 dosimetric for-malism and emphasize on this mm distance range by mothe reference dosimetric point ofr 051 cm to r 052 mm.

The aim of this work is to investigate differences in thdose rate distributions of the most commonly used192Irsources in the cm as well as the mm distance range, usingsame Monte Carlo~MC! simulation code and input dataResults are presented for source designs that present ative core length scale of one order of magnitude~from 1.3mm to 3 cm! and a variety of source and encapsulation gometries. The relative importance and behavior of the TGdefined dosimetric quantities in the transition to distanclose to the sources is investigated. Finally, overall diffences due to differences in source design are illustrated upolar and Cartesian coordinates dose rate distributions.

2239„10…Õ2239Õ8Õ$19.00 © 2002 Am. Assoc. Phys. Med.

io

trufus.

ansiathth

blrue

eto

egenotnth

gargho

erentofonso-lbe-itstom

idm-scmr,to

-

ton

arerere-

di-se

bleen-

en-of

alin

erylos

eence

ce

for

the

sing

onti

2240 Papagiannis et al. : Dosimetry comparison 2240

II. MATERIALS AND METHODS

A. Investigated sources

Calculations were performed around the old5,6 and new7

microSelectron, the old8,9 and new10 VariSource and theBuchler11 HDR brachytherapy source designs, the AngRad™ 192Ir wire12 and the Best Medical192Ir seed.13 Theactive core and encapsulation dimensions as well as stural details of the sources are presented in Table I andther details can be found in the corresponding reference

B. Monte Carlo simulations

Our own MC simulation code6,9,10,14–16was used to cal-culate the gamma dose rate distribution in water aroundinvestigated192Ir sources. In order to speed up calculatiothe water kerma approximation was utilized and results wbe presented for points at transverse distances greater thmm from the sources. At these points, that lie outsidecatheters utilized and are therefore of clinical importance,water kerma approximation safely applies for the192Irenergies.13 Recently however, we reported on a detectaeffect of the dose rate component owing to the beta spectemitted by192Ir, at radial distances in close proximity to thsources.17 Since air kerma strength is not applicable to bemitting sources no common frame exists for calibrationbeta and gamma emitters and the ratioDb /Dg normalizedper unit of192Ir activity was used to quantify the effect of thbeta dose rate component. Results from that study, suging that the magnitude of this effect depends stronglyonly on source and encapsulation geometry but also onencapsulation material, will be used in the following sectioto correct dose rate distributions at points close tosources.

Our code incorporates the exact structural details andometry of each source including its encapsulation. Primphotons are produced isotropically and uniformly throuthe active source core with a spectrum taken from Glasg

TABLE I. Structural details and geometries of the investigated192Ir HDRsources. All dimensions are in cm.

Source type

Active core EncapsulationOuter

diameterMaterial Length Diameter Material Thickness

microSelectron~old design!

Ir 0.35 0.0600 Stainlesssteel

0.0250 0.110

microSelectron~new design!

Ir 0.36 0.0650 Stainlesssteel

0.0125 0.090

VariSource~old design!

Ir 1.00 0.0340 Ti/Ni 0.0125 0.059

VariSource~new design!

Ir 0.50 0.0340 Ti/Ni 0.0125 0.059

Buchler Ir 0.13 0.1000 Stainlesssteel

0.0200 0.160a

Seed~Best Medical!

Pt/Ir 0.3 0.0100 Stainlesssteel

0.0200 0.050

AngioRad™ Ir 3 0.0127 Ti/Ni 0.0070 0.035b

aAn air gap of 0.01 cm exists between the active core and encapsulatibAn air gap of 0.004 15 cm exists between the active core and encapsula

Medical Physics, Vol. 29, No. 10, October 2002

-

c-r-

ll

lln 1ee

em

af

st-t

hese

e-y

w

and Dillman.18 Properly normalized probabilities for thmain processes of photoabsorption, coherent and incohescattering19–21 were used for the stochastic samplingtransport and interactions of primary and scattered photincluding characteristic photon emission following photelectric absorption.22 A photon history is terminated, with althe residual energy deposited on spot, if its energy fallslow a selected transport cutoff threshold of 10 keV or ifspatial coordinates lie outside the boundaries of the phanused.

The 192Ir sources were positioned at the center of a liquwater spherical phantom of 30 cm in diameter that is comonly used in similar studies.5,8,7 The phantom sphere wadivided into discrete concentric spherical shells of 0.025thickness up tor 50.5 cm and 0.1 cm thickness thereafteeach split into angular intervals of 1° with respect onlypolar angleu ~0, 180°! ~with the 180° angle referring to thedrive wire source side! since all dosimetric quantities involved are isotropic with respect to azimuthal anglew. Dosewas calculated from water kerma by weighting the phoenergy fluence at a specific (r ,u) surface with the corre-sponding mass-energy absorption coefficient.23 Thus, calcu-lated dosimetric data refer to exact radial distances andnot the result of averaging within a scoring volume whesteep dose gradients may occur. The scoring proceduresults in a three-dimensional histogram containing all invidual, (r ,u), point water-kerma values. Thus, polar doprofiles as well as radial dose profiles, for any radiusr and/orpolar angleu in the range discussed earlier, were availafor the calculation of the dosimetric parameters in the geralized TG-431 formalism recommended by TG-60:4

D~r ,u!5SKL r 0

G~r ,u!

G~r 0,90°!gr 0

~r !F~r ,u!, ~1!

wherer is the radial distance with respect to the source cter, r 0 corresponds to the reference dosimetric point(r 0 ,u0)5(1 cm,90°)(TG-43) or (r 0 ,u0)5(2 mm,90°)~TG-60! andu is the polar angle relative to the longitudinaxis of the source.SK is the source air kerma strengthunits of U ~1 U51 mGy m2 h2151 cGy cm2 h21! that wascalculated in separate simulations in free space for evinvestigated source as described in detail in Angelopouet al. and Karaiskoset al.10,16 L r 0

is the source dose ratconstant defined as the dose rate to water at the referpoint (r 0 ,u0) per unit source air kerma strength.gr 0

(r ) is theradial dose function normalized to unity at the referenpoint (r 0 ,u0) and F(r ,u) is the anisotropy function of thesource.G(r ,u) denotes the geometry factor that accountsthe distribution of radioactivity of densityr(rW8) at point rW8of a source active volume elementdV8 according to

G~rW !5EV

r~rW8!dV8

urW82rWu2 Y EV

r~rW8!dV8. ~2!

Although commercial treatment planning systems utilizepoint source approximation (1/r 2), the TG-43 report recom-mends the use of source geometry factors calculated uthe line source approximation:

.on.

tt

thtry

1

or

riesi

dtio.

g

f

ona

en

lerce

be-sug-

and

ob-

torcu-ein

ionrs

nly

ee-tant

d

the

e of

cat-eec-

in

ess-or,an

d

try

ofge-e,

us

2241 Papagiannis et al. : Dosimetry comparison 2241

G~r ,u!5b/Lr sinu, ~3!

whereb is the angle subtended by the active source corelengthL, with respect to calculation point (r ,u). While thisapproximation is safe in the cm distance range relativeconventional interstitial applications, it has been shownintroduce significant errors at radial distances close tosources (r ,L/2).14,24 Therefore the exact source geomefactor of each source, calculated according to Eq.~2! follow-ing the method described in Karaiskoset al.,14 will be usedin this study.

Results of this study refer to 109 primary photons for eachinvestigated source yielding statistical errors of less thanat 10°,u,170° and 3% atu,10° andu.170°, that are notsignificantly dependent on radial distance,r, since the scor-ing element increases liker 2 and scattering compensates fabsorption for the192Ir energies.10 Systematic uncertaintyoriginating by basic input data such as cross section libraas well as details of the energy spectrum used, do notnificantly affect dosimetry results for the192Ir energies.10

Moreover, overall propagated uncertainty cannot obscuresimetry comparisons of this study since the same simulacode and input data are used for all investigated sources

III. RESULTS AND DISCUSSION

A. Dose rate constant

The dose rate constant per unit source air kerma strenL r 0

, of the investigated sources as well as a point192Irsource was calculated at the reference distances or 0

51 cm ~TG-43! andr 052 mm ~TG-60! along the transversesource bisectors. Results are summarized in Table II alwith data from previous studies. It should be noted thatthough the dose rate due to the beta emitted spectrum~1electron versus 2.33 photons per decay of192Ir) is compa-rable to the gamma dose rate component at the refer

TABLE II. Dose rate constants,L r 0, of the investigated sources atr 0

51 cm and 2 mm in units of cGy h21 U21.

Source type

1 cm 2 mm

This workPreviousstudies This work

Previousstudies

microSelectron~old design!

1.11560.005

1.1155

1.116922.9760.12

22.975

microSelectron~new design!

1.10960.005

1.1087 22.7060.11

22.807

VariSource~old design!

1.04360.005

1.0448

1.043913.0360.07

12.978

VariSource~new design!

1.10160.005

1.10110 19.8460.10

19.8310

Buchler 1.11560.006

1.11511 27.8860.14

27.6211

Seed~Best Medical!

1.10460.005

1.10913 24.0460.12

23.7613

AngioRad™ 0.72760.004

0.71612 5.1660.03

5.1412

Point source 1.12160.002

27.9860.13

Medical Physics, Vol. 29, No. 10, October 2002

of

ooe

%

sg-

o-n

th,

gl-

ce

distance ofr 052 mm for a point source, it is of no detectabcontribution at this reference radial distance for real soudesigns.17

For r 051 cm, an excellent agreement is observedtween independent investigations for each source designgesting that MC dosimetry of192Ir sources is not sensitive indetails of the cross section libraries, the energy spectrumother input data used in the simulations.10 The greatest de-viation between results of this and previous studies isserved for the AngioRad™ source~1.5%!. This deviationcould be explained by the utilization of a conversion facrelating air kerma strength and source activity in the callations of Patelet al.,12 contradicting the specifications of thTG-43 formalism for calibration of brachytherapy sourcesair kerma strength units. Moreover, the utilized conversfactor of ~1/4.03! U per mCi apparent source activity refeto the Best Medical192Ir seed.12

For r 052 mm, dose rate constant data are scarce. OWang and Li13 have reported a value ofL2 mm

523.76 cGy h21 U21 for the Best Medical192Ir seed towhich the corresponding value of this study is in good agrment ~1%!. For the rest of the sources, dose rate consresults of this study are in close agreement~1%! compared toavailable dose rate results of previous studies at (r 0 ,u0)5(2 mm,90°) or extrapolation of previously publisheTG-43 dosimetric formalism data according to Eq.~1!, pre-sented in column 5 of Table II. For example, Patelet al.12

report a dose rate of 5.140 cGy h21 U21 at 2 mm along thetransverse bisector of the AngioRad™ source while fornew microSelectron, Daskalovet al.7 report L1 cm

51.108 cGy h21 U21 andg(2 mm)51.000 that extrapolatedusing the exact geometry factors yields a dose rate valu22.800 cGy h21 U21 at 2 mm.

Differences in the dose rate constant,L r 0, of the sources

per unit air kerma strength,SK , are, theoretically, due to thefiltration by the source core and encapsulation, in-water stering and spatial distribution of radioactivity within thsource.1 However, differences in the normalized energy sptrum emerging by the192Ir sources are not significant.8 Ad-ditionally, L r 0

presents only a minor energy dependence

the range close to the effective192Ir energy.25 Moreover, it iswell known that for the192Ir energies scattering compensatfor absorption.8,26,27This analysis implies that the spatial ditribution of radioactivity, addressed by the geometry factis the prevailing influence on the dose rate constant of192Ir source.

In Figs. 1~a! and 1~b!, L r 0results of this study presente

in Table II are plotted against the corresponding geomefactors G(1 cm,90°) andG(2 mm,90°), respectively. Datain Figs. 1~a! and 1~b! suggest that the dose rate constantsthe sources are indeed proportional to the correspondingometry factors within errors. A similar correlation of dosrate constant,L r 0

at r 051 cm, and geometry factor

G(1 cm,90°) was also observed for192Ir wire sources of ac-tive core length ranging from 0.5 to 12 cm in a previostudy.15 A linear fit of the form

l t

n

ei-ca

ford the

atny

edry

se.atthis

to

nifi-sedatioedtheblefor

t

2242 Papagiannis et al. : Dosimetry comparison 2242

L r 05a* G~r 0,90°! ~4!

was performed to the data of Figs. 1~a! and 1~b! and is alsopresented. The coefficients of the fit were found equaa5~1.12060.002! cGy cm2 h21 U21 for the r 051 cm dataand a5~1.11760.003! cGy cGy cm2 h21 U21 for the r 0

52 mm data. The two coefficients agree within errors acorrespond to the dose rate constant of an192Ir point source.A value of 1.12 cGy h21 U21 was also recommended by thTG-43 for192Ir sources.1 This recommended value is of clincally acceptable accuracy for sources of small length that

FIG. 1. ~a! Dose rate constant per unit source air kerma strength,L1 cm,plotted against the corresponding geometry factorG(1 cm,90°). ~b! Doserate constant per unit source air kerma strength,L2 mm, plotted against thecorresponding geometry factorG(2 mm,90°). A linear fit of the formL r 0

5a* G(r 0,90°) is also presented in both figures.

Medical Physics, Vol. 29, No. 10, October 2002

o

d

n

be approximated by a point source, but does not applyelongated source designs such as the old VariSource anAngioRad™. The equation proposed in this work

L r 0~cGy h21 U21!51.12~cGy cm2 h21 U21!

3G~r 0,90°!~cm22! ~5!

yields dose rate constant results with errors less than 2%both reference radial distances of 1 cm and 2 mm, for a192Ir source design.

B. Radial dose function

In Fig. 2, the radial dose functions,g(r ), of the investi-gated sources as well as a point192Ir source are plottedagainst radial distance,r. Presented results were calculatfollowing the TG-43 formalism, using the exact geometfactor values of each source.g(r ) results of Fig. 2 comparewell ~within 1%! with corresponding results of previoustudies5–10,12,13 except for the case of the Buchler sourcRegarding this HDR source, significant differences existradial distances close to the source between results ofstudy and the study of Ballesteret al.11 whereg(r ) valuespresent an abrupt increase. For example,g(2 mm)51.023 intheir study as opposed to 1.002 in this work. This is duethe fact that in their calculations Ballesteret al.11 used linesource approximated geometry factors that deviate sigcantly from the corresponding exact geometry factors uherein. This can be seen in the inset of Fig. 2 where the rof exact to line approximated geometry factors is plottagainst radial distance along the transverse bisector ofsources. While the line source approximation is acceptafor all other source designs it introduces significant errorsthe Buchler source~2.5% at 2 mm, rising up to 9% at 1 mm!.

FIG. 2. Radial dose function values,g(r ), plotted against radial distance,r,for the new~solid line! and old ~broken line! microSelectron designs, thenew~1! and old~3! VariSource designs, the Buchler source design~s!, theAngioRad™~h! and the seed source~L!. Corresponding data for a poin192Ir source are also presented as a dotted line.

epr

c

t-

atla

nc

e

me

seibeanapeisiresin

ort ot oe

a

con

tha

ifingr

ithsore

inuetica

its’

on,sesfactain

-

tive

theim-

fnge

e

ign

2243 Papagiannis et al. : Dosimetry comparison 2243

In Fig. 2 it can be seen thatg(r ) values of all sources arin close agreement except for the Buchler source thatsents a slight increase at radial distancesr ,5 mm possiblydue to hardening of the192Ir spectrum emerging from thesource due to its’ relatively increased core diameter.8 Scoringof the photon spectra at 2 mm along the transverse biseof the Buchler and the new VariSource~that presents thesmallest source core diameter! revealed a small but detecable difference~346.7 keV and 345.3 keV, respectively!.

Overall,g(r ) values of all sources are within 2% with thof a point192Ir source confirming that source and encapsution geometry does not significantly affect radial dose fution values of192Ir sources.7,10,15 However, source designstrongly affects the dose rate component owing to the bspectrum of192Ir that could increaseg(r ) at radial distancesclose to the sources. It has been shown17 that a significantincrease of almost 15% of the gamma dose rate at 1along the transverse bisectors of both VariSource designsists due to their relatively reduced encapsulation thickneswell as low encapsulation material density. This increasabout 5% for the new microSelectron source and negligfor the rest of the HDR source designs as well as the ssource due to their relatively increased core diameterencapsulation material density. Overall, in intravascularplications of large diameter peripheral vessels where ctered catheters are used, the beta dose rate componentof importance for HDR source designs. For a thin wsource of low encapsulation material density however, anificant beta dose rate contribution of about 40% was fouat 1 mm.17 This is the case for the AngioRad™ source fwhich Patelet al.12 have reported a dose rate enhancemen33% at 1 mm that could be of importance in the treatmensmall diameter coronary vessels where catheters of diamdown to 3.2 F are utilized.2

C. Anisotropy function

In Figs. 3~a! and 3~b! the anisotropy function,F(r ,u),polar angle profiles of four representative source designspresented at radial distances ofr 51 cm andr 52 mm. Thesetwo radial distance values where chosen because theystitute the reference dosimetric points of the TG-43 aTG-60 dosimetric formalisms~at u0590°). Moreover, theyare indicative of the cm and mm distance range fromsource centers that are of interest to interstitial and intravcular applications, respectively.

In Fig. 3~a!, it can be seen that all sources present signcant anisotropy, with anisotropy function values beistrongly dependent on source geometry. The two VariSoudesigns present anisotropy function values that agree w1% for polar angles 30°,u,150° and 5% at polar angleclose to their longitudinal axis due to the longer active cof the old VariSource design that leads to increasanisotropy.10 The new microSelectron design presentscreased anisotropy compared to the rest of the sources dits’ increased diameter. Finally, the anisotropy characterisof the Buchler source appear comparable to that of the VSource designs except foru,20°. In this polar angle region

Medical Physics, Vol. 29, No. 10, October 2002

e-

tor

--

ta

mx-

asisleedd-

n-not

g-d

ffter

re

n-d

es-

-

cein

ed-

tos

ri-

the Buchler source presents reduced anisotropy due tosignificantly shorter active core length.

Figure 3~b! reveals that anisotropy atr 52 mm is not sig-nificant at all points around all sources. This observaticontradicting the common belief that anisotropy increacontinuously as radial distance decreases, is due to thethat at such short radial distances from a source the mdose contributor to a specific (r ,u) point is the source segment closest to that point.15

D. Dose rate distributions in polar coordinates

The dose rate polar angle profiles of four representasource designs at the radial distances ofr 51 cm and r52 mm are presented in Figs. 4~a! and 4~b!, respectively.These dose rate profiles correspond to the product ofexact geometry factors of the sources and the TG-43 dosetric properties discussed above, according to Eq.~1!.

In Figs. 4~a! and 4~b! it can be observed that the form othe dose rate distributions in the cm and mm distance ra

FIG. 3. Anisotropy function,F(r ,u), polar angle profiles at radial distanc~a! r 51 cm and~b! r 52 mm for the new microSelectron design~solid line!,the new~1! and old~3! VariSource designs and the Buchler source des~s!.

heoriet

nsapexge-e-n

siinaaw-

testtedon

s is

cu-seer,vedthetry

n

o-oseallro-

pro-

thethe

cen,fer-l-

y ofpre-

io-toec-

orenewnted

lues

ethate to

se

d

eti-

ne

n

2244 Papagiannis et al. : Dosimetry comparison 2244

is totally different. At the radial distance ofr 51 cm @Fig.4~a!# it is points close to the transverse bisector of tsources that present higher dose rates than points close tlongitudinal axes of the sources, except for the old VaSource design that deviates significantly from all othsource dose rate distributions. This can be explained bygeometry factor profiles of the sources presented in the iof Fig. 4~a! where it can be seen that the point sourceproximation applies in this distance range for all sourcescept for the old VariSource design that presents the lonactive core (L51 cm).14,24 Therefore, the anisotropy function profiles of Fig. 3~a! are the decisive parameter that dtermines the dose rate polar angle profiles in the cm distarange.

At the radial distance ofr 52 mm of Fig. 4~b! it is pointsclose to the longitudinal axis of the sources that presentnificantly higher dose rates than points along the longitudaxes of the sources except for the Buchler source thatpears almost isotropic. This is due to the fact that as shoin the inset of Fig. 4~b!, the point source approximation ap

FIG. 4. Dose rate,dD/dt, polar angle profiles for the new microSelectrodesign~solid line!, the new~1! and old ~3! VariSource designs and thBuchler source design~s! at radial distance~a! r 51 cm and ~b! r52 mm. The geometry factor profiles of these sources are presented icorresponding insets.

Medical Physics, Vol. 29, No. 10, October 2002

the-rheet--st

ce

g-l

p-n

plies only for the Buchler source that presents the shoractive core and is unacceptable for all other investigasources. Geometry factor values depend significantlysource geometry, and given that the anisotropy functionclose to unity for all sources@see Fig. 3~b!#, the dose ratepolar angle profiles of Fig. 4~b! follow the correspondinggeometry factor polar angle profiles. This suggests that acracy in geometry factor calculation is of importance in dorate determination in the proximity of the sources. Howevit should be noted that accurate dosimetry may be achieby use of data that are self consistent with respect toapproximation used for the calculation of source geomefactors since according to Eq.~1! consistency in geometryfactor calculation would allow potential errors owing to aerroneous source approximation~line or point! to cancel out.

E. Dose rate distributions in Cartesian coordinates

Although the TG-43 dosimetric formalism uses polar cordinates, in intravascular applications where a uniform ddistribution is required at a given depth within a vessel wfor a lesion that may be several cm in length, dose rate pfiles in Cartesian coordinates are also of interest. Suchfiles are presented in Figs. 5~a! and 5~b! at distances,y, of 1cm and 2 mm away from the longitudinal source axes ofnew microSelectron, the old and new VariSource,Buchler and the AngioRad™ source designs.

In Fig. 5~a! it can be seen that away from the sourcenters (zÞ0) dose rate profiles of the new microSelectrothe new VariSource and the Buchler source present difences that follow their difference in anisotropy function vaues at the corresponding radial distances given the validitthe point source approximation for these sources in thesented radial distance range~from 1 cm to 1.8 cm!. The doserate profiles of the old VariSource design and the AngRad™ deviate significantly from all other profiles duetheir increased active core length of 1 cm and 3 cm, resptively.

In Fig. 5~b!, differences between the sources are mpronounced even between the new microSelectron, theVariSource and the Buchler sources since in the preseradial distance range~from 2 mm to 1.5 cm! dose rate valuesare governed by the corresponding geometry factor vathat are significantly dependent on source dimensions~seethe preceding section!. Only the Buchler dose rate profilresembles that of a point source while it can be seensources provide a uniform dose over a length comparabltheir active core length.

IV. CONCLUSIONS

We used our MC simulation code to compare the dorate distributions of the most commonly used192Ir HDRsource designs~the old and new microSelectron, the old annew VariSource and the Buchler!, the Best Medical192Ir seedand the AngioRad™192Ir wire source, as well as investigatthe relative importance of the AAPM TG-43 defined quanties in the cm and mm distance range from the sources.

the

cegon

ug

st

an

tble

aa

fa

esty atdentpro-es.

rch

ail:

ndc-up

red.

s-, P.y-asc.

C.cu-m-

c-

, L.n

.

rials.

K.-

y

ilos,y

o-f

. B.a-

.

-,’’

ofom-

s, P.

w

2245 Papagiannis et al. : Dosimetry comparison 2245

For the192Ir energies, dose rate constants of the sourdo not depend significantly on source and encapsulationometry or simulation input data and they were found proptional to the corresponding geometry factors along the traverse source bisectors. A simple equation was proposedthe accurate calculation of the dose rate constant of any192Irsource. Radial dose functions of the sources were also fonot to depend significantly on source and encapsulationometry and agree within 2% with that of a point192Ir source.

Geometry factor was found significant for accurate dorate determination especially at radial distances close tosource centers where anisotropy is almost negligiblecaution should be given in the utilization of192Ir HDRsources dosimetry data. These should be consistent inmethod of geometry factor calculation in order for possierrors owing to an erroneous source approximation~line orpoint! not to affect overall dosimetric accuracy.

In general, dose rate polar angle profiles are mainlyfected by the corresponding anisotropy function profileradial distances in the cm range from an192Ir source center.On the contrary, in the mm distance range the geometry

FIG. 5. Dose rate,dD/dt, profiles in Cartesian coordinates for the nemicroSelectron design~solid line!, the new~1!, and old ~3! VariSourcedesigns, the Buchler source design~s! and the AngioRad™~h! at distances~a! y51 cm and~b! y52 mm away from the longitudinal source axes.

Medical Physics, Vol. 29, No. 10, October 2002

se-r-s-for

nde-

ehed

he

f-t

c-

tor profile is the main influence in dosimetry of the sourcand consequently the shape as well as the dose uniformigiven radial distances from the sources is strongly depenon source dimensions. This was illustrated by dose ratefiles in polar and Cartesian coordinates around the sourc

ACKNOWLEDGMENT

This work was partly supported by the Special ReseaAccount of the University of Athens.

a!Electronic mail: ppapagi@cc.uoa.grb!Author to whom correspondence should be addressed. Electronic m

lsakell@cc.uoa.gr1R. Nath, L. L. Anderson, G. Luxton, K. A. Weaver, J. F. Williamson, aA. S. Meigooni, ‘‘Dosimetry of interstitial brachytherapy sources: Reommendations of the AAPM Radiation Therapy Committee Task Gro43,’’ Med. Phys.22, 209–234~1995!.

2A. E. Ajani, H-S. Kim, and R. Waksman, ‘‘Clinical trials of vasculabrachytherapy for in-stent restenosis: update,’’ Cardiovasc. Radiol. M2, 107–113~2001!.

3S. Nag, P. E. Cole, I. Crocker, S. K. Jani, K. V. Krishnasastry, V. Masullo, R. Nath, D. Nori, S. Parikh, P. Rubin, B. Speicer, P. S. TeirsteinTripurareni, R. Waksman, and J. F. Williamson, ‘‘The American brachtherapy society perspective on intravascular brachytherapy,’’ CardiovRadiol. Med.1, 8–19~1999!.

4R. Nath, H. Amols, C. Coffey, D. Duggan, S. Jani, Z. Li, M. Schell,Soares, J. Whiting, P. E. Cole, I. Crocker, and R. Schwartz, ‘‘Intravaslar brachytherapy physics: Report of the AAPM Radiation Therapy Comittee Task Group No. 60,’’ Med. Phys.26, 119–151~1999!.

5J. F. Williamson and Z. Li, ‘‘Monte Carlo dosimetry of the microSeletron pulsed and high dose rate192Ir sources,’’ Med. Phys.22, 809–819~1995!.

6P. Karaiskos, A. Angelopoulos, L. Sakelliou, P. Sandilos, C. AntypasVlachos, and E. Koutsouveli, ‘‘Monte Carlo and TLD dosimetry of a192Ir high dose-rate brachytherapy source,’’ Med. Phys.25, 1975–1984~1998!.

7G. M. Daskalov, E. Lo¨ffler, and J. F. Williamson, ‘‘Monte Carlo-aideddosimetry of a new high dose-rate brachytherapy source,’’ Med. Phys25,2200–2208~1998!.

8R. Wang and R. S. Sloboda, ‘‘Influence of source geometry and mateon the transverse axis dosimetry of192Ir brachytherapy sources,’’ PhysMed. Biol. 43, 37–48~1998!.

9P. Karaiskos, A. Angelopoulos, P. Baras, L. Sakelliou, P. Sandilos,Dardoufas, and L. Vlachos, ‘‘A Monte Carlo investigation of the dosimetric characteristics of the VariSource192Ir high dose rate brachytherapsource,’’ Med. Phys.26, 1498–1502~1999!.

10A. Angelopoulos, P. Baras, L. Sakelliou, P. Karaiskos, and P. Sand‘‘Monte Carlo dosimetry of a new192Ir high dose rate brachytherapsource,’’ Med. Phys.27, 2521–2527~2000!.

11F. Ballester, J. Pe´rez-Calatayud, V. Puchades, J. L. Lluch, M. A. SerranAndres, Y. Limami, F. Lliso, and E. Casal, ‘‘Monte Carlo dosimetry othe Buchler high dose rate192Ir source,’’ Phys. Med. Biol.46, N79–N90~2001!.

12N. S. Patel, S-T. Chiu-Tsao, P. Fan, H-S. Tsao, S. F. Liprie, and LHarrison, ‘‘The use of cylindrical coordinates for treatment planning prameters of an elongated192Ir source,’’ Int. J. Radiat. Oncol., Biol., Phys51, 1093–1102~2001!.

13R. Wang and X. A. Li, ‘‘A Monte Carlo calculation of dosimetric parameters of90Sr/90Y and 192Ir SS sources for intravascular brachytherapyMed. Phys.27, 2528–2535~2000!.

14P. Karaiskos, L. Sakeliou, P. Sandilos, and L. Vlachos, ‘‘Limitationsthe point and line source approximations for the determination of geetry factors around brachytherapy sources,’’ Med. Phys.27, 124–128~2000!.

15P. Karaiskos, P. Papagiannis, A. Angelopoulos, L. Sakelliou, D. BaltaSandilos, and L. Vlachos, ‘‘Dosimetry of192Ir wires for LDR interstitial

.

d

Nhy

ur

0

r,cta

on

s oom

a-

ef-for

st,’’

Mlcu-

ter-

is,at-

is,

u-

2246 Papagiannis et al. : Dosimetry comparison 2246

brachytherapy following the AAPM TG-43 dosimetric formalism,’’ MedPhys.28, 156–166~2001!.

16P. Karaiskos, P. Papagiannis, L. Sakelliou, A. Anagnostopoulos, anBaltas, ‘‘Monte Carlo dosimetry of the selectSeed125I interstitial brachy-therapy seed,’’ Med. Phys.28, 1753–1760~2001!.

17D. Baltas, P. Karaiskos, P. Papagiannis, L. Sakelliou, E. Loeffler, andZampoglou, ‘‘Beta versus gamma dosimetry close to Ir-192 bractherapy sources,’’ Med. Phys.28, 1875–1882~2001!.

18G. P. Glasgow and L. T. Dillman, ‘‘Specific-ray constant and exposrate constant for192Ir, ’’ Med. Phys.6, 49–52~1979!.

19J. H. Scofield, ‘‘Theoretical photoionization cross sections from 1 to 15keV,’’ Lawrence Livenmore Laboratory Report UCRL-51326, 1973.

20J. H. Hubbel, W. M. J. Veigele, E. A. Briggs, R. T. Brown, D. T. Cromeand R. J. Howerton, ‘‘Atomic form factors, incoherent scattering funtions, and photon scattering cross sections,’’ J. Phys. Chem. Ref. Da4,471–538~1975!.

21J. H. Hubbel and I. Overbo, ‘‘Relativistic atomic form factors and photcoherent scattering cross sections,’’ J. Phys. Chem. Ref. Data8, 69–105~1979!.

22E. F. Plechaty, D. E. Cullen, and R. J. Howerton, ‘‘Tables and graphphotointeraction cross sections from 0.1 keV to 100 MeV derived fr

Medical Physics, Vol. 29, No. 10, October 2002

D.

.-

e

0

-

f

the LLL Evaluated Nuclear Data Library,’’ Lawrence Livenmore Labortory UCRL-50400 6 Rev. 3, 1981.

23J. H. Hubbel and S. M. Seltzer, ‘‘Tables of x-ray mass attenuation coficients and mass energy-absorption coefficients 1 keV to 20 MeVelementsZ51 to 92 and 48 additional substances of dosimetric intereNISTIR 5632, 1995.

24M. J. Rivard, ‘‘Refinements to the geometry factor used in the AAPTask Group Report No. 43 necessary for brachytherapy dosimetry calations,’’ Med. Phys.26, 2445–2450~1999!.

25Z. Chen and R. Nath, ‘‘Dose rate constant and energy spectrum of institial brachytherapy sources,’’ Med. Phys.28, 86–96~2001!.

26A. Angelopoulos, A. Perris, K. Sakellariou, L. Sakelliou, K. Sarigiannand G. Zarris, ‘‘Accurate Monte Carlo calculations of the combinedtenuation and build-up factors, for energies~20–1500 keV! and distances~0–10 cm! relevant in brachytherapy,’’ Phys. Med. Biol.36, 763–778~1991!.

27L. Sakelliou, K. Sakellariou, K. Sarigiannis, A. Angelopoulos, A. Perrand G. Zarris, ‘‘Dose rate distributions around60Co, 137Cs, 198Au, 192Ir,241Am, 125I ~models 6702 and 6711! brachytherapy sources and the nclide 99Tcm,’’ Phys. Med. Biol.37, 1859–1872~1992!.