C

Toctomno

(C

Int. J. Radiation Oncology Biol. Phys., Vol. 59, No. 4, pp. 1047–1052, 2004Copyright © 2004 Elsevier Inc.

Printed in the USA. All rights reserved0360-3016/04/$–see front matter

doi:10.1016/j.ijrobp.2003.12.016

LINICAL INVESTIGATION Prostate

PERIRECTAL SEEDS AS A RISK FACTOR FOR PROSTATEBRACHYTHERAPY-RELATED RECTAL BLEEDING

AMY MUELLER, M.D.,* K ENT WALLNER, M.D.,*†‡ GREGORY MERRICK, M.D.,§ ERIC FORD, PH.D.,*†

STEVEN SUTLIEF, PH.D.,*† WILLIAM CAVANAGH, M.S.,* AND WAYNE BUTLER, PH.D.§

*Radiation Oncology, Veterans Affairs Puget Sound Health Care System, Seattle, WA;†Department of Radiation Oncology,University of Washington, Seattle, WA;‡Radiation Oncology, Group Health Cooperative, Seattle, WA;

and§Schiffler Cancer Center, Wheeling, WV

Purpose: To correlate rectal wall doses and perirectal seed numbers with late rectal bleeding after prostatebrachytherapy.Methods and Materials: We studied 148 patients randomized to implantation with I-125 vs. Pd-103 at the VAPuget Sound HCS from 1998 through 2001 and for whom postimplant dosimetry was available. Implants wereperformed by standard techniques, using a modified peripheral loading pattern. A postimplant computedtomography (CT) scan (3 mm slice thickness) was obtained 1–4 h after implantation. Rectal doses were expressedas the R100, R200, and R300, defined as the rectal volume (cc) that received more than 100%, 200%, or 300%of the prescription dose, respectively. The rectum was considered to be a solid structure defined by the outer wall,without attempting to differentiate the inner wall or contents. In addition to conventional dose parameters, eachpatient’s postimplant CT scan was reviewed for the number of seeds within 0, 0.1–2, and 2.1–4 mm of the outerrectal wall. The proximal edge of the seed was used for distance determinations from the outer rectal wall.Patients who reported Grade 1 or higher Radiation Therapy Oncology Group morbidity were contacted bytelephone to obtain more details regarding their rectal bleeding. Those who reported persistent bleeding lastingfor more than 1 month were categorized as Grade 2.Results: Patients had a wide range of rectal wall doses, with R100 values ranging from 0.0 to 10.4 cc (median,0.95 cc). Similarly, the number of perirectal seeds within 0.0 to 2.0 mm of the rectum varied widely, ranging from0 to 12 seeds (median: 1 seed). Seven patients (7 of 144 � 5%) developed persistent rectal bleeding, one of whomrequired a colostomy. Both rectal radiation doses and the number of perirectal seeds were higher in patients withpersistent rectal bleeding. The number of perirectal seeds <2.0 mm of the rectal wall was higher in patients withrectal bleeding (p � 0.037), but the number of seeds 2–4 mm from the wall were not related (p � 0.72). Inmultivariate regression analysis including prostatic D90 (the dose that covers 90% of the postimplant prostate),preimplant transrectal ultrasound volume, R300, and the number of seeds <2 mm from the rectal wall asindependent variables, only the R300 was statistically significantly associated with the likelihood of persistentrectal bleeding (p � 0.025).Conclusion: A limited number of errant perirectal sources in itself does not appear to place patients at increasedrisk of rectal bleeding, providing that the overall rectal wall doses are within acceptable values. © 2004 ElsevierInc.

Prostatic carcinoma, Brachytherapy, Rectum, Morbidity.

pidlytermting,ingof

, byood

0%

o afteri m-s fria-bB ns.R

d toh lat-e ctald1 ctal

ogy60 S.

356;

F003.

A

INTRODUCTION

ransperineal prostate brachytherapy has grown raver the past 10 years. Although data regarding long-ancer control and morbidity outcomes are accumulahere is still limited detailed technical information regardptimization of cancer control rates or minimizationorbidity. Probably the most vexing morbidity reportedearly all reliable investigators, has been a small likelihf late rectal bleeding.Radiation proctopathy with bleeding occurs in 2% to 1

Reprint requests to: Kent Wallner, M.D., Radiation Oncol#174), Veterans Affairs Puget Sound Health Care System, 16olumbian Way, Seattle, WA 98108-1597. Tel: (206) 768-5

1047

f patients, generally starting between 6–18 monthsmplantation(1–5). Endoscopy typically reveals a circucribed area of intense erythema, telangiectasias, andility of the anterior rectal wall overlying the prostate(6).lood loss is usually minimal, rarely requiring transfusioarely, proctitis progresses to prostatic-rectal fistula(7–9).The risk of rectal bleeding has been consistently linke

igher rectal wall doses, but the two are only loosely red—only a small proportion of patients with higher reoses actually develop clinically recognized bleeding(3,0). The loose correlation between dosimetry and re

ax: (206) 768-5331. E-mail: kent.wallner@med.va.govReceived Jun 10, 2003, and in revised form Dec 9, 2

ccepted for publication Dec 12, 2003.

mpttsdposfrlld

ICp

pewttsds

sCpssfrbcgoo

1048 I. J. Radiation Oncology ● Biology ● Physics Volume 59, Number 4, 2004

orbidity might be explained by a predisposition of someatients to rectal morbidity due to inherent radiation sensi-ivity or microvascular damage from other causes. Alterna-ively, it has occurred to the authors that a few misplacedources, even in the setting of acceptable overall rectal walloses, might be responsible for rectal bleeding in someatients (Fig. 1). Previous investigators have looked only atverall rectal wall doses, and have not addressed the pos-ibility that one or more misplaced sources could accountor rectal bleeding, even in the absence of generally highectal dose. If so, technique changes to minimize the like-ihood of perirectal seeds might decrease the likelihood ofate rectal bleeding. Accordingly, we correlated rectal walloses and perirectal seed numbers with late rectal bleeding.

METHODS AND MATERIALS

We studied 148 patients randomized to implantation with-125 vs. Pd-103 at the Veterans Affairs Puget Sound Healthare System from 1998 through 2001 and for whomostimplant dosimetry was available (11). Implants were

Fig. 1. Postimplant CT images showing seeds at 0, 1, 3,the surface were included in the analysis presented hereR300% � 0.12 cc.

erformed by standard techniques, using a modified periph-ral loading pattern (12–14). No external beam radiationas given. The preimplant prostate volumes ranged from 12

o 85 cc (median, 28 cc). I-125 mCi/seed ranged from 0.41o 0.9 mCi (median, 0.86 mCi/seed). The number of I-125eeds ranged from 30 to 96 (median, 50). All Pd-103 waselivered as 2.0 mCi/seed. The total number of Pd-103eeds ranged from 60 to 189 (median, 94).

A postimplant computed tomography (CT) scan (3 mmlice thickness) was obtained 1–4 h after implantation. TheT-derived postimplant target volume was determined asreviously described (15). The contoured images andources were entered into a Varian 7.0 treatment planningystem (Charlottesville, VA). A redundancy check was per-ormed on seed localization to prevent seed duplication. Theectum was contoured at least one CT image above andelow the prostate, and further if rectal tissue was stillontacted by the prescription isodose. Dose–volume histo-rams of the prostate and rectum were calculated using theuter prostatic and rectal margins identified on CT scan byne investigator (K.W.). Dosimetric parameters analyzed

mm from the rectal surface. Only seeds within 4 mm ofpatient had a R100% � 1.5 cc, R200% � 0.29 cc, and

and 5. This

ioDoRrtst

tswd

qcm

N

1049Perirectal seeds ● A. MUELLER et al.

ncluded the V100—the percent of the postimplant prostater rectal volume covered by the prescription dose, and the90—the dose that covers 90% of the postimplant prostater rectal volume. Rectal doses were expressed as the R100,200, and R300, defined as the rectal volume (cc) that

eceived more than 100%, 200%, or 300% of the prescrip-ion dose, respectively. The rectum was considered to be aolid structure defined by the outer wall, without attemptingo differentiate the inner wall or contents (16).

Table 1. Modified

Recta

Grade 0 Grade 1

ormal Bowel movements 3–5 times daily Bow

Slight rectal discharge or bleeding Exc

Fig. 2. R100 (Top) or seeds 0–2 cm from rectum (Bottnumber of seeds 0–2 mm from outer rectal wall.

In addition to conventional dose parameters, each pa-ient’s postimplant CT scan was reviewed for the number ofeeds within 0, 0.1–2, and 2.1–4 mm of the outer rectalall. The proximal edge of the seed was used for distanceeterminations from the outer rectal wall (see Fig. 1).Treatment-related morbidity was monitored by mailed

uestionnaires to all patients, using Radiation Therapy On-ology Group (RTOG) criteria at 1, 3, 6, 12, 18, and 24onths (Table 1). Patients who reported Grade 1 or higher

morbidity criteria

idity

Grade 2 Grade 3

vements �5/day Obstruction or bleedingrequiring surgery

mucus or intermittent bleeding

ith patients arranged in order of increasing RV100 or

RTOG

l morb

el mo

essive

om), w

RmllOtpRalm

v22s

brstrwb

owTwhppsT

o

TSAVDRRR#####

osptd

TVR#

1050 I. J. Radiation Oncology ● Biology ● Physics Volume 59, Number 4, 2004

TOG morbidity were contacted by telephone to obtainore details regarding their rectal bleeding. No patient was

ost to follow-up. Those who reported persistent bleeding,asting for more than 1 month, were categorized as Grade 2.ne of the 7 patients with bleeding required a colostomy—

he details of his case will be summarized in a futureublication. His rectal radiation doses (R100 � 0.96 cc,200 � 0.0 cc) were not excessive, but he had undergonenterior rectal wall biopsies for postimplant bleeding. Fol-ow-up times ranged from 18 to 56 months (median, 34onths).

RESULTS

Patients had a wide range of rectal wall doses, with R100alues ranging from 0.0 to 10.4 cc (median, 0.95 cc) (Fig.). Similarly, the number of perirectal seeds within 0.0 to.0 mm of the rectum varied widely, ranging from 0 to 12eeds (median, 1 seed).

Seven patients (7/144 � 5%) developed persistent rectalleeding, one of whom required a colostomy. None of theemaining 6 patients required blood transfusions or invasiveurgical procedures. Four patients with rectal bleeding werereated with Pd-103, and 3 were treated with I-125. Bothectal radiation doses and the number of perirectal seedsere higher in patients with persistent rectal bleeding (Ta-le 2).Rectal bleeding was statistically linked with all measures

f rectal wall dose, from 100% to 300% of prescription, butas more closely related to the higher dose values (Table 2).he incidence of bleeding increased with higher R100, butas relatively low even in patients with higher doses origher perirectal seed numbers (Fig. 3). The number oferirectal seeds �2.0 mm of the rectal wall was higher inatients with rectal bleeding (p � 0.037), but the number ofeeds 2–4 mm from the wall were not related (p � 0.72,able 2).The number of perirectal seeds and rectal wall doses were

nly loosely related (r � 0.52, Fig. 4). In multivariate

eated to prescription dose or higher (left) or the number

Table 3. Multivariate relationship between rectal bleeding andvarious parameters

p value

RUS 0.43100 (%) 0.29300 (cc) 0.025seeds �2 mm 0.23

Abbreviations as in Table 2.

Table 2. Relationship between rectal bleeding and variousparameters

No bleeding Bleeding p value

RUS 31 cc 41 cc 0.02eed # 77 77 0.98ge 65 64 0.54100 (%) 86% 92% 0.1690 96% 110% 0.13100 (cc) 1.4 3.2 0.0041200 (cc) 0.12 0.60 0.0002300 (cc) 0.026 0.18 �0.0001seeds 0.0 mm 0.67 1.6 0.07seeds 0.1–2 mm 1.3 3.3 0.001seeds 2.1–4 mm 2.1 1.9 0.72seeds �2 mm 2.0 4.9 0.037seeds �4 mm 4.1 6.7 0.075

Abbreviations: TRUS � transrectal ultrasound; V100 � percentf the postimplant prostate or rectal volume covered by the pre-cription dose; D90 � the dose that covers 90% of the postimplantrostate or rectal volume; R100, R200, R300 � the rectal volumehat received more than 100%, 200%, or 300% of the prescriptionose, respectively.*P values are based on two-tailed Student’s t test.

Fig. 3. Incidence of rectal bleeding vs. the volumes of rectum trof seeds 0–2 mm from the rectal wall (right).

rtsotT

ocit

srtcwdvs

ocmprtatrmrira

alptbpbawRh

Fmp

1051Perirectal seeds ● A. MUELLER et al.

egression analysis including prostatic D90, preimplantransrectal ultrasound volume, R300, and the number ofeeds �2 mm from the rectal wall as independent variables,nly the R300 was statistically significantly associated withhe likelihood of persistent rectal bleeding (p � 0.025,able 3).

DISCUSSION

Several investigators have documented the contributionf rectal radiation doses to postimplant bleeding. Althoughomparisons between series are hampered by different tim-ng of the postimplant CT scans and variations in the way

ig. 4. Correlation between R300 and number of seeds within 2.0m of the outer rectal wall (some points represent more than 1

atient).

hat rectal doses are described, nearly all investigators have t

1

1

hown a higher incidence of rectal bleeding with higherectal doses (1, 3). Despite the consistent relationship be-ween rectal wall doses and morbidity, it is clear that theorrelation is loose, in that only a small minority of patientsith high rectal wall doses actually develop clinically evi-ent bleeding. Why some patients with higher doses de-elop morbidity while others do not is purely a matter ofpeculation at this time.

We undertook the current detailed study of the geographyf perirectal seeds with the hypothesis that perirectal seedsould be an independent risk factor in themselves for rectalorbidity. In fact, we have found in univariate analysis that

erirectal seeds are another predisposing cause of higherectal doses and late rectal bleeding. However, it appearshat their contribution to rectal bleeding is due to theirssociation with higher rectal doses, and not necessarily toheir proximity to the rectal wall per se. Accordingly, weecommend that care be taken to keep sources 2 mm orore away from the outer rectal wall, to limit the overall

ectal dose. A limited number of errant perirectal sources intself does not appear to place patients at increased risk ofectal bleeding, providing that the overall rectal wall dosesre within acceptable values.

Caution is in order when dosimetry data summarized herere compared to other centers. Postimplant edema typicallyeads to lower Day 0 rectal dosimetric parameters, so thatractitioners using Day 30 dosimetry should consider thatheir R100 values associated with an increased risk of rectalleeding will likely be higher (17). Second, the choice ofostimplant imaging modality may alter the relationshipetween rectal doses and proctitis, an issue that has yet to beddressed. Finally, for our own quality assurance program,e have chosen to rely primarily on R100 values rather than200 or R300 (Figs. 2 and 3) because of our concern thatigher dose parameters are more sensitive to minor varia-

ions in how the rectum is contoured.

REFERENCES

1. Snyder KM, Stock RG, Hong SM, et al. Defining the risk ofdeveloping grade 2 proctitis following 125-I prostate brachy-therapy using a rectal dose–volume histogram analysis. Int JRadiat Oncol Biol Phys 2001;50:335–341.

2. Gelblum DY, Potters L. Rectal complications associated withtransperineal interstitial brachytherapy for prostate cancer. IntJ Radiat Oncol Biol Phys 2000;48:119–124.

3. Waterman FM, Dicker AP. Probability of late rectal morbidityin I-125 prostate brachytherapy. Int J Radiat Oncol Biol Phys2003;55:342–353.

4. Merrick GS, Butler WM, Dorsey AT, et al. Rectal dosimetricanalysis following prostate brachytherapy. Int J Radiat OncolBiol Phys 1999;43:1021–1027.

5. Hu K, Wallner K. Clinical course of rectal complicationsfollowing I-125 prostate brachytherapy. Int J Radiat OncolBiol Phys 1998;41:263–265.

6. Smith S, Wallner K, Han B, et al. Argon plasma coagulation

for rectal bleeding following prostate brachytherapy. Int JRadiat Oncol Biol Phys 2001;51:636–642.

7. Howard A, Wallner K, Han B, et al. Clinical course anddosimetry of rectal fistulas after prostate brachytherapy.J Brachyther Int 2001;17:37–42.

8. Theodorescu D, Gillenwater JY, Koutrouvelis PG. Prosta-tourethral-rectal fistula after prostate brachytherapy: Incidenceand risk factors. Cancer 2000;89:2085–2091.

9. Wallner K, Blasko J, Dattoli M. Morbidity. In: Dattoli M,editor. Prostate brachytherapy made complicated, 2nd ed. Se-attle: SmartMedicine Press, 2001: p. 15.1–15.21.

0. Han B, Wallner K. Dosimetric and radiographic correlates toprostate brachytherapy-related rectal complications. Int J Can-cer 2002;96:372–378.

1. Wallner K, Merrick G, True L, et al. I-125 versus Pd-103 for lowrisk prostate cancer: Morbidity outcomes from a prospectiverandomized multicenter trial. Ca J Sci American 2002;8:67–73.

1

1

1

1

1

1

1052 I. J. Radiation Oncology ● Biology ● Physics Volume 59, Number 4, 2004

2. Mueller A, Wallner K, Corriveau J, et al. A reappraisal oflocal anesthesia for prostate brachytherapy. Radiother Oncol2003;67:309–312.

3. Merrick GS, Butler WM. Modified uniform seed loading forprostate brachytherapy: Rationale, design, and evaluation.Techniques Urol 2000;6:78–84.

4. Han B, Wallner K, Aggarwal S, et al. Treatment margins forprostate brachytherapy. Semin Urol Oncol 2000;18:137–141.

5. Badiozamani KR, Wallner KE, Cavanagh W, et al. Compa-rability of CT-based and TRUS-based prostate volumes. Int JRadiat Oncol Biol Phys 1999;43:375–378.

6. Hilts M, Spandinger I, Keyes M. Comparison of methods forcalculating rectal dose after 125-I prostate brachytherapy im-plants. Int J Radiat Oncol Biol Phys 2002;53:775–785.

7. Waterman FM, Dicker AP. Effect of post-implant edema onthe rectal dose in prostate brachytherapy. Int J Radiat OncolBiol Phys 1999;45:571–576.