Int. J. Radiation Oncology Biol. Phys., Vol. 80, No. 1, pp. 85–90, 2011Copyright � 2011 Elsevier Inc.

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

jrobp.2010.01.016

doi:10.1016/j.i

CLINICAL INVESTIGATION Prostate

DOSIMETRIC IMPACT OF INTERFRACTION CATHETER MOVEMENT INHIGH-DOSE RATE PROSTATE BRACHYTHERAPY

WILLIAM FOSTER, M.D., J. ADAM M. CUNHA, PH.D., I.-CHOW HSU, M.D., VIVAN WEINBERG, PH.D,DEVAN KRISHNAMURTHY, PH.D., AND JEAN POULIOT, PH.D

Department of Radiation Oncology, University of California San Francisco, San Francisco, California

ReprinQuebec, 1(418) 691ucsf.edu

This wof the A

Purpose: To evaluate the impact of interfraction catheter movement on dosimetry in prostate high-dose-rate(HDR) brachytherapy.Methods and Materials: Fifteen patients were treated with fractionated HDR brachytherapy. Implants were per-formed on day 1 under transrectal ultrasound guidance. A computed tomography (CT) scan was performed. In-verse planning simulated annealing was used for treatment planning. The first fraction was delivered on day 1. Acone beam CT (CBCT) was performed on day 2 before the second fraction was given. A fusion of the CBCTand CTwas performed using intraprostatic gold markers as landmarks. Initial prostate and urethra contours were trans-ferred to the CBCT images. Bladder and rectum contours were drawn, and catheters were digitized on the CBCT.The planned treatment was applied to the CBCT dataset, and dosimetry was analyzed and compared to the initialdose distribution. This process was repeated after a reoptimization was performed, using the same constraints usedon day 1.Results: Mean interfraction catheter displacement was 5.1 mm. When we used the initial plan on day 2, the meanprostate V100 (volume receiving 100 Gy or more) decreased from 93.8% to 76.2% (p < 0.01). Rectal V75 went from0.75 cm3 to 1.49 cm3 (p < 0.01). A reoptimization resulted in a mean prostate V100 of 88.1%, closer to the initialplan (p = 0.05). Mean rectal V75 was also improved with a value of 0.59 cm3. There was no significant change inbladder and urethra dose on day 2.Conclusions: A mean interfraction catheter displacement of 5.1 mm results in a significant decrease in prostateV100 and an increase in rectum dose. A reoptimization before the second treatment improves dose distribu-tion. � 2011 Elsevier Inc.

HDR brachytherapy, Prostate, Interfraction catheter displacement, Dosimetry.

INTRODUCTION

The American Cancer Society estimates that there were

186,320 new cases of prostate cancer reported and 28,860

deaths from prostate cancer in the United States in 2008,

which makes it the most frequent cancer and the second lead-

ing cause of death from cancer in men in this country. Radi-

ation therapy is an option for almost every patient with

localized disease and is commonly used as a definitive ther-

apy (1). Previous studies have shown that dose escalation to

the prostate is beneficial (2, 3). However, this improvement

in disease control was achieved at the cost of increased

toxicity because of an increased dose to normal tissues (2).

The optimal treatment would therefore maximize the dose

to the tumor cells, improving local control of the disease,

and limit the dose delivered to normal tissues, thereby limit-

ing treatment-related toxicities.

t requests to: William Foster, M.D., CHUQ Hotel-Dieu de1 Cote du Palais, Quebec, QC, Canada G1R 2J6. Tel:-5264; Fax: (418) 691-5268; E-mail: fosterw@radonc.

ork was presented as a poster at the 2009 Annual Meetingmerican Brachytherapy Society. Toronto, ON, Canada,

85

Efforts have thus been made to improve conformality

of the treatment. Different techniques can now be used, in-

cluding three-dimensional conformal radiation therapy,

intensity-modulated radiation therapy, stereotactic body radi-

ation therapy using a CyberKnife, proton beam radiation

therapy, and brachytherapy. One of the great advantages of

brachytherapy over other methods of radiation delivery is

the fact that the treatment applicator is inside the target vol-

ume and follows its movements, decreasing the possibility

of a geographical miss. Another advantage of brachytherapy

is the rapid dose falloff outside of the treated volume due to

the inverse square law (the dose delivered to a point a distance

r away from a source is proportional to 1/r2). The dose to the

surrounding tissues is therefore minimal with brachytherapy

compared with other methods of delivery, with a potential

benefit for reduced toxicity (4,5).

April 2009, Poster #76, Brachytherapy 8(2009):167.Conflict of interest: This work was supported in part by Nucletron

through a research contract.Received Sept 28, 2009, and in revised form Jan 8, 2010.

Accepted for publication Jan 12, 2010.

86 I. J. Radiation Oncology d Biology d Physics Volume 80, Number 1, 2011

High-dose rate (HDR) brachytherapy is used both as a radi-

ation boost to the prostate after external beam radiation ther-

apy and as monotherapy to treat prostate cancer. Most of the

fractionation schemes used in clinical practice result in the

delivery of two or more fractions spread over 2 days for a sin-

gle implant. The proper dose delivery of fractionated HDR

prostate brachytherapy relies on the stability of the implant

and on reproducible catheter positions for each fraction.

This is particularly critical given the current trend to reduce

the number of fractions and increase the dose per fraction.

A study previously performed by our group on patients

treated with HDR brachytherapy for prostate cancer showed

a mean craniocaudal catheter displacement of 4 mm between

the first and second fractions (6). Similar studies report

displacements ranging from 3 to 20 mm (7–11). However,

most of these studies did not look directly at the dosimetric

impact of these catheter displacements. While it might

be intuitively obvious that a decrease of the dose delivered to

the prostate on day 2 is to be expected with a movement of

the catheters, we still believe that it is of interest to quantify

this change in dose distribution. The aim of our current

study, therefore, was to evaluate the impact of interfraction

catheter movement on the dosimetry of subsequent fractions.

We also present here a technique to correct for these changes

while limiting the additional workload on the clinical team.

METHODS AND MATERIALS

Fifteen consecutive patients treated with HDR for prostate cancer

either as a part of their initial treatment, or as a salvage therapy, were

included in the study. All patients had intraprostatic gold markers

placed weeks prior to the implant. On day 1, patients were placed

in the lithotomy position under general anesthesia in the operating

room. A Foley catheter was placed in the urethra. Sixteen flexible

catheters were then inserted transperineally into the prostate under

transrectal ultrasound guidance. Two molds of dental putty were

used to secure the catheters to each other and were sutured to the per-

ineum. A cystoscopy was performed to make sure there were no

catheters in the bladder. After recovery from this surgical procedure,

patients were brought to the Radiation Oncology Department where

a planning computed tomography (CT) scan was performed. The

prostate, urethra, bladder, and rectum were contoured by a single

physician (IH). Inverse planning simulated annealing (12) was

used for treatment planning (OncentraBrachy; Nucletron Inc., Vee-

nendaal, The Netherlands) to deliver fractions ranging from 600 cGy

(salvage) to 950 cGy (prostate boost) to the prostate, depending on

the fractionation scheme used. The first fraction was delivered on

day 1. On day 2, the patient was brought to the brachytherapy suite,

shielded for HDR treatment, and equipped with a cone beam CT

(CBCT) (Simulix CBCT simulator; Nucletron Inc., Veenendaal,

The Netherlands). A CBCT was performed prior to delivery of the

second fraction by using the day 1 dose plan.

To obtain the data used for this study, images of the CBCT from

day 2 and the planning CT scan from day 1 were coregistered with

a rigid body registration using the intraprostatic gold markers as land-

marks. If fewer than three gold markers were present, another struc-

ture was used as a third landmark, usually intraprostatic calcifications.

This study includes dosimetric indices derived from five datasets.

The first dataset is the control, and it uses only the CT images. It uses

the original contoured volumes and the original day 1 dose plan. For

the second dataset, the initial prostate and urethra contours were

transferred to the CBCT images after image coregistration of the

CT and CBCT scans. The bladder and rectum contours were drawn,

and the catheters were digitized on the CBCT. The initial planned

treatment (same dwell positions and times) was applied to the

CBCT dataset using the updated catheter positions. This was done

by manually entering the dwell times used on day 1 on the CBCT

plan. The third dataset was obtained by using inverse planning sim-

ulated annealing to create another plan. This plan used the prostate

and urethra volumes from the CT scan, the catheters and rectum and

bladder volumes from the CBCT scan, and the same optimization

objective parameters used on day 1. The fourth dataset was obtained

following the same procedure as the second dataset; however, this

time, the physician recontoured the prostate and urethra volumes

on CBCT. These new volumes were used for an optimization to cre-

ate the fifth dataset. In summary, in addition to the initial plan from

CT used as a control, four plans were thus created for each patient

with the CBCT images by using old volumes with old planning

(OVOP), old volumes with new planning (OVNP), new volumes

with old planning (NVOP), and new volumes with new planning

(NVNP). For each plan, the prostate V100 (volume receiving 100

Gy or more) and V150, the urethra V120, bladder V75, and rectum

V75 were calculated and compared to the parameters of the initial

plan. Finally, the volumes of the prostate contours obtained on

days 1 (CT) and 2 (CBCT) were compared for each patient.

Statistical analysesAnalysis of variance for repeated measures was used to evaluate

the differences in dosing to the prostate, rectum, bladder, and urethra

with changes in planning or volume contouring, or both. The

Newman-Keuls post hoc test was used to compare each of the

four new dose distributions generated for day 2 with the initial

plan from day 1 (first dataset). A paired t statistic was used to com-

pare the initial and new prostate contoured volumes for each patient.

In addition, Fisher’s exact test was used to determine whether there

was any difference in the frequency of achieving threshold volumes

for specific doses for different structures between the initial plan and

each of the four alternatives.

RESULTS

Soft tissue contrast on CBCT was slightly lower than that

on CT, which made the contouring more difficult on day 2.

However, the catheters were easily identified on CBCT im-

ages. The gold markers were also clearly defined, which al-

lowed for an adequate image coregistration (Fig. 1). All

patients had 16 catheters implanted; the mean craniocaudal

catheter displacement in our study was 5.1 mm (Table 1).

The mean prostate V100 on the original plan on day 1 was

93.8%. The prostate V100 decreased when the initial plan was

applied to the CBCT images acquired on day 2, with mean de-

creases of 17.1% using old volumes (OVOP) and 11% using

new volumes (NVOP) (Fig. 2). Compared with the initial

plan, there was a significant difference in the mean V100, es-

pecially without a change in plan (analysis of variance posthoc tests for initial vs. OVOP, p = 0.0001; and NVOP, p =

0.0002). The median loss of prostate V100 was 19.8%

when the old volumes were used with the old planning on

day 2. Therefore, 7/15 patients showed a decrease in prostate

V100 of more than 20%. In fact, only 3/15 patients had

Fig. 1. (Left panel) CT planning on day 1, (middle panel) CBCT image on day 2, and (right panel) coregistration of the twoimaging studies. The catheters can be identified in black on both CT and CBCT images. On the coregistered images, one ofthe image datasets is inverted and the catheters appear white.

Dosimetric impact of interfraction catheter movement d W. FOSTER et al. 87

a decrease of less than 10% in their prostate V100. The use of

reoptimization with the original volumes (OVNP) consider-

ably reduced the loss in V100, although it was still lower

than that achieved at the time of the initial treatment (p <

0.05). Reoptimization therefore could only partly compensate

for catheter migration, since it was unable to deliver the dose

to areas no longer covered by catheters. The mean V100

values on day 2 were less than 90% for the OVOP, OVNP,

and NVOP plans. Only when both the volumes and the plan

were changed (NVNP) was there no significant difference

with the initial plan with a V100 of 92.4% (p = 0.56).

There were no differences in the mean prostate V150

values among the four plans for day 2 compared with that

of the initial plan. The frequency of achieving a V150 of

<40% was not different from the initial plan, occurring for

93% of the sample for the OVOP, OVNP, and NVNP plans

and 73% for the NVOP (data not shown).

Use of the initial plan on day 2 resulted in an increased rec-

tum V75. Significant mean increases above the initial plan of

0.74 and 0.80 cm3 were observed for OVOP and NVOP, re-

spectively (p < 0.01 for both). Using the old volumes and old

Table 1. Interfraction catheter displacement*

Patient Mean displacement (mm) Largest displacement (mm)

1 1.9 62 2.3 63 2.4 94 3.2 125 3.9 96 4.5 67 4.7 98 5.1 99 5.1 12

10 5.3 911 5.4 912 6 913 7.3 1214 9.2 1815 10.1 15Mean 5.1

* Data show mean and maximal interfraction catheter displace-ment for each patient. Each patient was implanted with 16 catheters.The mean overall displacement of catheters was 5.1 mm. Fivepatients had a maximal catheter displacement higher than 10 mm.

planning, 13/15 patients had an increase in their rectum V75.

In 12 patients, the resulting rectum V75 was higher than 1

cm3. There was no significant difference for both (OVNP

and NVNP) reoptimized plans. Similar to the initial plan,

the frequency of achieving a V75 of <1 cc was common

for both OVNP and NVNP (Fig. 3). There were no differ-

ences in mean urethra V120 cc values among the four plans

for day 2 compared with the initial plan (post hoc probabili-

ties are all nonsignificant) (Fig. 4). Even though 9/15 patients

had an increase in their urethra V120 with the use of the old

volumes and old planning, none of them reached a value of

more than 1 cm3. There was no difference in the mean blad-

der V75 cc volumes among the plans compared with the ini-

tial day 1 plan. All four versions of revised plans indicated

a decrease in the mean V75 cc, which was lowest when

both the contours and the dosing were revised, but not signif-

icantly different. In addition, all four options for day 2, in-

cluding OVOP, resulted in a high proportion of patients

with a bladder V75 <1 cc (most were 0) (Fig. 5).

Fig. 2. Prostate V100 (%) according to the planning/volumes used.The initial plan is the plan obtained on day 1, using the CT scan todelineate the prostate. The four other plans were based on the day 2CBCT images and used either the initial volumes from the CT scan(OV) or new volumes from the CBCT (NV) combined with eitherthe initial treatment plan used on day 1 (OP) or a new plan obtainedafter reoptimization (NP). Mean values are shown in red. Dashedline represent the clinical threshold of 90% used in our center.

Fig. 3. Rectum V75 (cc) according to the planning/volumes used.The initial plan is the plan obtained on day 1 using the CT scan to de-lineate the rectum. The four other plans were based on the day 2CBCT images and used either the initial volumes from the CT scan(OV) or new volumes from the CBCT (NV) combined with eitherthe initial treatment plan used on day 1 (OP) or a new plan obtainedafter a reoptimization (NP). Mean values are shown in red. Thedashed line represent the clinical threshold of 1 cc used in our center.

Fig. 5. Bladder V75 (cc) according to the planning/volumes used.The initial plan is the plan obtained on day 1 using the CT scan.The four other plans were based on the day 2 CBCT images andused either the initial volumes from the CT scan (OV) or new vol-umes from the CBCT (NV) combined with either the initial treat-ment plan used on day 1 (OP) or a new plan obtained aftera reoptimization (NP). Mean values are shown in red. The dashedline represent the clinical threshold of 1 cc used in our center.

88 I. J. Radiation Oncology d Biology d Physics Volume 80, Number 1, 2011

We compared the original prostate contours to those drawn

on the CBCT by the same physician. There was a significant

decrease in prostate volume with a mean decrease of 3.89

cm3 (p = 0.04). This represented an average decrease of

11% of prostate volume. Only two patients displayed an

increase in volumes of 17% and 42% (Fig. 6, increasing vol-

umes are shown above the dashed line).

DISCUSSION

Previous studies have shown interfraction catheter dis-

placement ranging from 3 to 20 mm in patients undergoing

Fig. 4. Urethra V120 (cc) according to the planning/volumes used.The initial plan is the plan obtained on day 1 using the CT scan. Thefour other plans were based on the day 2 CBCT images and used ei-ther the initial volumes from the CT scan (OV) or new volumes fromthe CBCT (NV) combined with either the initial treatment plan usedon day 1 (OP) or a new plan obtained after a reoptimization (NP).Mean values are shown in red. The dashed line represent the clinicalthreshold of 1 cc used in our center.

treatment with HDR brachytherapy for prostate cancer

(6–11). The current study showed similar results, with

a mean catheter displacement of 5.1 mm in the caudal

direction. Using coregistration of day 1 with day 2 images,

we were able to apply the day 1 planning to day 2 imaging

and therefore evaluate the impact of these movements on

the dose delivered on day 2.

Prostate coverage was poorer on day 2, as shown by the de-

creasing prostate V100 observed when the day 1 plan was ap-

plied to day 2 imaging. That observation was made when

both the day 1 and the day 2 volumes were used. The decrease

in coverage usually affected the base of the prostate because

of the distal migration of the catheters. The mean and median

Fig. 6. Volume of the prostate contours on CT on day 1 (initial con-tour) and on CBCT on day 2 (new contour). The dashed line repre-sents a 100% correlation between the 2 volumes. The lower qualityof the CBCT image and intraobserver variability explain the ob-served differences in prostate volumes.

Dosimetric impact of interfraction catheter movement d W. FOSTER et al. 89

losses of prostate V100 were 17.1% and 19.8% when the old

volumes were used with the old planning. Twelve of 15 pa-

tients had a loss of more than 10%, and 7/15 patients had

a loss of more than 20% in their prostate V100. The largest

observed decrease was 29.9%. Furthermore, only 2/15 pa-

tients had a V100 over the threshold value of 90% on day

2 when the old volumes and old planning were used.

When the day 1 plan was used, the dosimetry seemed bet-

ter when the new volumes were used, as opposed to the old

volumes transferred onto the CBCT images on day 2 (de-

creases of 11% vs. 17%, respectively). The difference in

dose parameters lies within the differences in volume delin-

eation. The volumes contoured on day 2 were significantly

smaller than the initial volumes (p = 0.04). This was surpris-

ing because we expected a possible increase in prostate size

due to acute trauma and swelling of the gland. Possible expla-

nations for the observed decrease in prostate volume are the

intraobserver variation in contour delineation and the poorer

quality of images on CBCT than on CT. The smaller volumes

on day 2 might lead to an underestimation of the loss of pros-

tate coverage. This explains why the decrease in prostate

V100 seems reduced with the use of the new contours as op-

posed to the original contours transferred onto CBCT images.

What is then the real loss of prostate coverage between frac-

tions 1 and 2? Since the CT images are clearer, the original

volumes probably are more reliable than the new volumes.

However, there are uncertainties on the initial volumes as

well. The intraobserver variability in prostate delineation

was previously described in the literature and results in 5%

variations in prostate volume (13). In our study, it was about

11%. The use of different imaging studies (CT and CBCT)

probably explains the higher value we observed. Taking

these facts into consideration, the real loss in mean prostate

V100 probably lies within the 11 to 17% interval, falling be-

low the standard threshold of 90% used in our center and pre-

viously described in an RTOG phase II study (14).

The bladder dose was lower on day 2, although it was not

statistically significant. This reflects the fact that catheters mi-

grate distally between fractions and are therefore pulled away

from the bladder. Only 1/15 patients reached a bladder V75

over 1 cm3 when the old volumes and old planning on day

2 were used. The urethra dose did not show significant vari-

ation. The only significant difference in dose delivered to or-

gans at risk was observed in the rectum dose. The use of the

initial plan on day 2 resulted in a rectal V75 approaching 1.5

cm3. Twelve of 15 patients had a rectum V75 of more than 1

cm3 on day 2, when the initial planning was applied on day 2.

The initial planning is done with the rectum position on a spe-

cific point in time. However, the rectum is a moving organ,

with random filling over time. The observed increased dose

is probably due to these changes in the shape of the organ.

It raises the question whether the planned rectum dose is re-

ally the one delivered at the time of initial treatment since

a movement of the rectal wall between planning and treat-

ment is difficult to assess.

The catheter displacements between fractions 1 and 2 re-

sulted in a significantly decreased dose to the prostate. In a ma-

jority of patients (13/15), the prostate V100 fell below the

usual threshold of 90%, with a decrease from 94 to 65% in

one patient. These results were obtained with an average cath-

eter displacement of 5.1 mm, a value that is consistent with

results previously published in the literature. It also increased

the dose to the rectum, with little effect on urethra and bladder

doses. Simnor et al. (11) recently published similar results.

The D90% delivered to the planning tumor volume was re-

duced by 28% and 32% on fractions 2 and 3, respectively,

while the dose delivered to 2 cm3 of the rectum was increased

by 0.69 and 0.76 Gy, respectively. These changes were ob-

served with mean interfraction catheter displacements of

7.9 and 3.9 mm on fractions 2 and 3, respectively.

Previous publications have shown a correlation between

the quality of treatment planning and clinical results in

HDR brachytherapy (15, 16). Therefore, do our results

imply that we should change the way HDR brachytherapy

is done in an effort to improve dose delivery on day 2? An

argument against this could be made that the results

published so far using HDR brachytherapy for prostate

cancer were accomplished using the initial treatment plan

on day 2 (14, 17–19). Therefore, these excellent outcomes

were achieved even though a suboptimal dose was probably

delivered on day 2. Also, our study has some limitations,

and it is possible that our results overestimate the loss of

prostate coverage on day 2. First, there is intraobserver

variability in contour delineation, which might have an

impact on dosimetric parameters. To limit that effect, we

used a coregistration of the CT and CBCT images to allow

us to transfer the day 1 volumes onto day 2 imaging.

Second, the image coregistration was done using landmarks

inside the prostate. However, the use of 3-mm slices on

both CBCT and CT images, the artifacts caused by the

presence of the gold markers, and the lesser image quality

on CBCT scans might have resulted in an imperfect image

coregistration. Because of the rapid dose falloff inherent in

brachytherapy, a small displacement in the prostate contour

because of the image coregistration could have had an

impact on the observed dosimetry. Nonetheless, the

interfraction migration of catheters has a clear impact on

dosimetry, and efforts should be made to limit these

impacts in the hope of a possible improvement in clinical

outcomes.

What then is the optimal strategy to make sure an optimal

dose is delivered on day 2? The first solution could be to cre-

ate a new plan on day 2, based on day 2 anatomy. The use of

reoptimization resulted in a better prostate coverage than the

use of the initial plan on day 2, while allowing for sparing of

normal tissues. An image coregistration between day 1 and

day 2 images can be used to transfer the urethra and prostate

volumes to avoid a change in the treated volume due to intra-

observer variability. The catheters can then be digitized as

they appear on day 2 and the bladder and rectum can be con-

toured. Reoptimization can be performed based on day 2

anatomy, all of this in a timely fashion. Another solution is

to push the catheters back in on day 2 and try to recreate

the initial implant anatomy. With this approach, one would

90 I. J. Radiation Oncology d Biology d Physics Volume 80, Number 1, 2011

need to ensure the prostate is not pushed in along with the

catheters, so another scan should be done after making the

catheter adjustment. In doing so, it is probably preferable to

run a new optimization to take into consideration the induced

displacement of the catheters. Finally, a third solution would

be to limit the number of fractions. Morton et al. (20) recently

presented their initial results with a cohort of patients treated

with a single fraction of 15 Gy of HDR as a boost in prostate

cancer (20). This method has the advantage of avoiding any

interfraction catheter migration. A longer follow-up is

needed to fully assess the safety and efficacy of that ap-

proach, but the initial results appear promising.

CONCLUSIONS

Our results show that interfraction catheter migration has

a deleterious impact on dosimetry in HDR prostate brachyther-

apy. The average catheter migration in our study was 5.1 mm,

similar to that previously reported in the literature. The average

prostate V100 went from 93.8% to 76.6% from day 1 to day 2

when the initial prostate volumes were used, a loss of 17.1% in

prostate coverage. On the other hand, the rectum dose was sig-

nificantly higher on day 2 when the initial plan was used,

though this did not induce significant changes in the urethra

and bladder doses. To improve dose delivery, the use of new

planning on day 2 is feasible, improving prostate coverage

and decreasing the dose to normal tissues, mainly the rectum.

Another option would be to use a single fraction of HDR bra-

chytherapy, therefore avoiding any interfraction catheter

movement. Preliminary results using single-fraction HDR

have been presented and seem promising (20), but longer

follow-up is needed before it can be suggested to be standard

practice. Until then, efforts should be made to limit the effects

of interfraction catheter movement on dose delivery.

REFERENCES

1. National Comprehensive Cancer Network. Clinical practiceguidelines in oncology: Prostate cancer. Available at: http://www.nccn.org/professionals/physician_gls/PDF/prostate.pdf.

2. Pollack A, Zagars GK, Rosen I, et al. Prostate cancer radiationdose response: Results of the M.D. Anderson phase III random-ized trial. Int J Rad Oncol Biol Phys 2002;53:097–1105.

3. Valicenti R, Lu J, Grignon D, et al. Survival advantage fromhigher-dose radiation therapy for clinically localized prostatecancer treated on the Radiation Therapy Oncology Group trials.J Clin Oncol 2000;18:2740–2746.

4. Fang F, Wang Y, Chiang P, et al. Comparison of the outcomeand morbidity for localized or locally advanced prostate cancertreated by high-dose-rate brachytherapy plus external beam ra-diotherapy (EBRT) versus EBRT alone. Jpn J Clin Oncol 2008;38:474–479.

5. Guix B, Bartrina JM, Tello JI, et al. Dose escalation with high-dose 3D-conformal radiotherapy (HD-3D-CRT) or low-dose3D-conformal radiotherapy plus HDR brachytherapy (LD-3D-CRT + HDR-B) for intermediate- or high-risk prostate cancer:Early results of a prospective comparative trial. J Clin Oncol2009;27:15.

6. Charra-Brunaud C, Hsu IC, Pouliot J, et al. Analysis of interac-tion between number of implant catheters and dose-volume-histograms in prostate high-dose-rate brachytherapy usinga computer model. Int J Radiat Oncol Biol Phys 2003;56:586–591.

7. Hoskin PJ, Bownes PJ, Ostler P, et al. High dose rate afterload-ing brachytherapy for prostate cancer: Catheter and glandmovement between fractions. Radiother Oncol 2003;68:285–288.

8. Mullokandov E, Gejerman G. Analysis of serial CT Scans toassess template and catheter movement in prostate HDR brachy-therapy. Int J Radiat Oncol Biol Phys 2004;58:1063–1071.

9. Damore SJ, Nisar Syed AM, Puthawala AA, et al. Needledisplacement during HDR brachytherapy in the treatment ofprostate cancer. Int J Radiat Oncol Biol Phys 2000;46:1205–1211.

10. Martinez AA, Pataki I, Edmunson G, et al. Phase II prospectivestudy of the use of conformal High-Dose Rate brachytherapy asmonotherapy for the treatment of favorable stage prostate can-cer: A feasibility report. Int J Radiat Oncol Biol Phys 2001;49:61–69.

11. Simnor T, Li S, Hoskin PJ, et al. Justification for inter-fractioncorrection of catheter movement in fractionated high dose-rate

brachytherapy treatment of prostate cancer. Radiother Oncol2009;93:253–258.

12. Lessard E, Pouliot J. Inverse planning anatomy-based dose op-timization for HDR-brachytherapy of the prostate using fastsimulated annealing algorithm and dedicated objective func-tion. Med Phys 2001;28:773–779.

13. Fiorino C, Reni M, Bolognesi A, et al. Intra- and inter-observervariability in contouring prostate and seminal vesicles: Implica-tions for conformal treatment planning. Radiother Oncol 1998;47:285–292.

14. Hsu I, Bae K, Sandler H, et al. Phase II trial of combined highdose rate brachytherapy and external beam radiotherapy foradenocarcinoma of the prostate: Preliminary results ofRTOG 0321. Abstract no. 1045 Proceedings of the 50th An-nual meeting of the American Society for Therapeutic Radiol-ogy and Oncology. Boston, MA 2008. p. S133.

15. Pinkawa M, Fischedick K, Eble MJ, et al. Dose-volumeimpact in high-dose-rate Iridium-192 brachytherapy asa boost to external beam radiotherapy for localized pros-tate cancer: A phase II study. Radiother Oncol 2006;78:41–46.

16. Pellizzon A, Salvajoli J, Novaes P, et al. The relationship be-tween the biochemical control outcomes and the quality of plan-ning of high-dose rate brachytherapy as a boost to external beamradiotherapy for locally and locally advanced prostate cancerusing the RTOG-ASTRO Phoenix definition. Int J Med Sci2008;5:113–120.

17. Bachand F, Martin AG, Vigneault E, et al. An eight-year expe-rience of HDR brachytherapy boost for localized prostate can-cer: Biopsy and PSA outcome. Int J Rad Oncol Biol Phys2009;73:679–684.

18. Mahmoudieh A, Tremblay C, Vigneault E, et al. Anatomy-based inverse planning dose optimization in HDR prostateimplant: A toxicity study. Radiother Oncol 2005;75:318–324.

19. Hsu IC, Cabrera AR, Shinohara K, et al. Combined modal-ity treatment with high-dose-rate brachytherapy boost forlocally advanced prostate cancer. Brachytherapy 2005;4:202–206.

20. Morton G, Loblaw A, Sankreacha R, et al. High-dose-rate pros-tate brachytherapy and supplemental external beam radiother-apy: A comparison of single fraction 15Gy high-dose-rate andhypofractionated external beam to a conventional fractionatedregimen. Brachytherapy 2009;8:110.