intensity-modulated radiation therapy after hysterectomy: comparison with conventional treatment and...

C

Tpvnptd

RC7p

Int. J. Radiation Oncology Biol. Phys., Vol. 62, No. 4, pp. 1117–1124, 2005Copyright © 2005 Elsevier Inc.

Printed in the USA. All rights reserved0360-3016/05/$–see front matter

doi:10.1016/j.ijrobp.2004.12.029

LINICAL INVESTIGATION Cervix

INTENSITY-MODULATED RADIATION THERAPY AFTER HYSTERECTOMY:COMPARISON WITH CONVENTIONAL TREATMENT AND SENSITIVITY OF

THE NORMAL-TISSUE–SPARING EFFECT TO MARGIN SIZE

ANESA AHAMAD, M.B.B.S., F.R.C.R.,* WARREN D’SOUZA, PH.D.,† MOHAMMAD SALEHPOUR, PH.D.,†

REVATHY IYER, M.D.,‡ SUSAN L. TUCKER, PH.D.,§ ANUJA JHINGRAN, M.D.,*AND PATRICIA J. EIFEL, M.D.*

Departments of *Radiation Oncology, †Radiation Physics, ‡Diagnostic Radiology, and §Biostatistics and Applied Mathematics,The University of Texas M. D. Anderson Cancer Center, Houston, TX

Purpose: To determine the influence of target-volume expansion on the reduction in small-bowel dose achieved withuse of intensity-modulated radiation therapy (IMRT) vs. standard conformal treatment of the pelvis after hysterec-tomy, and to investigate the influence of patient body habitus on the normal-tissue sparing achieved with use of IMRT.Methods and Materials: A clinical target volume (CTV) was contoured on each of 10 planning computedtomography scans of patients who had been treated for cervical or endometrial cancer after a hysterectomy.Treatment planning was based on vaginal CTVs and regional nodal CTVs. To account for internal motion,margins were added to form an initial planning target volume (PTVA) as follows: 0.0 mm were added to theregional nodal CTV; 10 mm were added anteriorly to the vaginal CTV; and 5 mm were added to the vaginal CTVin all other directions. Two further PTVs (PTVB and PTVC) were produced by a 5-mm expansion of PTVA togive PTVB and a further 5-mm expansion to give PTVC. Treatment plans for all 3 PTVs were produced by useof 2 conformal fields (2FC), 4 conformal fields (4FC), or IMRT to deliver 45 Gy to more than 97% of the PTV.The primary goal of IMRT was to spare small bowel. The change in sparing that accompanied the increase inmargin size was assessed by comparison of dose–volume histograms that resulted from PTVA, PTVB, and PTVC.Measured patient dimensions were correlated with bowel sparing.Results: Significantly less small bowel was irradiated by IMRT than by 2FC (p < 0.0001) or 4FC (p < 0.0001)for doses greater than 25 Gy. Significantly less rectum was irradiated by IMRT than by 2FC (p < 0.0001) or 4FC(p < 0.0001). Significantly less bladder was irradiated by IMRT than by 2FC (p < 0.0001). However, themagnitude of the sparing achieved by use of IMRT decreased as margins increased. In particular, the volume ofsmall bowel spared by IMRT vs. 2FC or 4FC decreased as margin size increased (p � 0.0002 and p � 0.008 for2FC and 4FC, respectively). The amount of normal-tissue sparing achieved by use of IMRT vs. 4FC was inverselycorrelated with patient body mass index.Conclusion: Because the small-bowel sparing achieved with use of IMRT is markedly reduced by relatively smallexpansions of the target volume, accurate target delineation, highly reproducible patient immobilization, and aclear understanding of internal-organ motion are needed to achieve optimal advantage in the use of IMRT overconventional methods of posthysterectomy pelvic radiation therapy. © 2005 Elsevier Inc.

Female pelvis, Cervical cancer, IMRT, Radiation therapy.

isntascd

fO

G

INTRODUCTION

he standard radiation therapy techniques used to treat theelvis after hysterectomy for carcinoma of the uterine cer-ix or corpus involve either 2 or 4 fields. In these tech-iques, the cup-shaped tissue volume produced by theelvic floor and iliac lymph nodes can only be irradiated byreatment of the entire pelvic contents to the prescribedose. After hysterectomy, the pelvis often contains a signif-

Reprint requests to: Patricia J. Eifel, M.D., Department ofadiation Oncology, The University of Texas M. D. Andersonancer Center, Unit 97, 1515 Holcombe Blvd., Houston, TX7030. Tel: (713) 792-3400; Fax: (713) 563-2366; E-mail:[email protected]
Presented at the 44th Annual Meeting of the American Society A
1117

cant portion of the small bowel, which tends to fall into thepace left after removal of the uterus. A conformal tech-ique has been developed for intensity-modulated radiationherapy (IMRT) that can deliver a high dose of radiation ton irregular clinical target volume (CTV), with relativeparing of adjacent normal tissues (1). The modulated beaman produce dose distributions that have a relatively sharpose gradient between the target volume and adjacent un-

or Therapeutic Radiology and Oncology (ASTRO), New Orleans,ctober 2002.This work was supported in part by Cancer Center Support

rant CA16672 from the National Cancer Institute.Received Aug 2, 2004, and in revised form Nov 29, 2004.

ccepted for publication Dec 3, 2004.

iaadr

tRvaaofgbawteeiruaateI

pcc(wcmou

D

a

comrnhtaoctvptnvcs

fi

F(la

Ficand

PPP

t

1118 I. J. Radiation Oncology ● Biology ● Physics Volume 62, Number 4, 2005

nvolved critical structures. In the pelvis, IMRT can deliverspecified dose to the CTV (pelvic lymph nodes, vagina,

nd paravaginal tissues) and, simultaneously, a reducedose and dose per fraction to the intrapelvic small bowel,ectum, and bladder.

However, before routine clinical application of IMRT inhis setting, a number of uncertainties must be resolved.adiation oncologists have not traditionally outlined target-olume contours for adjuvant radiation therapy to the pelvisfter hysterectomy. Clinicians’ understanding of the tissuest risk for microscopic involvement and of the pelvic anat-my, particularly after surgical intervention, may be imper-ect. The optimal margins to be added to the CTV toenerate the planning target volume (PTV) are unknownecause little data are available on the motion of the CTVnd pelvic organs during the course of treatment in patientsho have undergone hysterectomy. To compensate for

hese uncertainties, clinicians may be tempted to generouslyxpand the CTV to ensure that a tumoricidal dose is deliv-red to involved tissues. However, as the margin size isncreased, the organ-sparing advantage of IMRT may beeduced. We, therefore, studied the amount of sparing withse of IMRT relative to conventional techniques as marginsround the CTV were increased. Because IMRT solutionsnd immobilization uncertainties may be influenced by pa-ients’ physical characteristics, we also evaluated the influ-nce of body habitus on normal-tissue sparing with use ofMRT.

METHODS

The planning CT scans from 10 patients who were treated withelvic radiation after hysterectomy for cervical or endometrialancer were used to generate IMRT and three-dimensional (3D)onformal plans for this study. Patients underwent CT simulationsPicker 5000, Marconi Medical Systems, Cleveland, OH). Theyere scanned in the supine position with their arms across the

hest. Three radiopaque pellet markers were placed as fiducialarkers at the anterior midline and at right and left lateral points

n the skin to define the isocenter. Patients were scanned from thepper abdomen to below the perineum.

efinition of CTVs and critical organsThe CTV included structures that would typically be considered

t risk for recurrence after hysterectomy in a patient with cervical

Table 1. Sizes of the margins added to the vaginal CTV andregional nodal CTV to generate PTVA, PTVB, and PTVC

Margin size (mm)

RegionalCTV

Anterior tovaginalCTV

Posterior, lateral,superior, and inferior to

vaginal CTV

TVA 0 10 5TVB 5 15 10TVC 10 20 15

Abbreviations: CTV � clinical target volume; PTV � planning
barget volume.
ancer. A central vaginal CTV and a regional nodal CTV wereutlined separately. The central vaginal CTV included the proxi-al vagina, paravaginal tissues, and paracervical tissues; its infe-

ior limit was just above the midsymphysis pubis. The regionalodal CTV included the common iliac, external and internal iliac,ypogastric, and presacral lymph nodes. These nodes were con-oured by delineating the perivascular fat and connective tissueround vessels of the common iliac, internal and external iliac, andbturator chain, as well as the presacral soft tissues and surgicallips. The superior limit of the regional nodal CTV was defined byhe common iliac nodes at the level of the midpoint of the L5ertebra. The internal iliac nodes were followed down into the trueelvis to include the hypogastric and obturator nodes inferiorly tohe level of the upper third of the obturator fossa. The presacralodes were included down to the level of the midpoint of the S3ertebra. Contouring was performed by a team of radiation on-ologists (P.J.E., A.J., and A.A.) and a diagnostic radiologist whopecialized in CT and MRI of the female pelvis (R.I.).

When the central vaginal and regional nodal CTVs were de-ned, bone and soft tissue were not included unless they were

ig. 1. Illustration of PTVA (red), PTVB (yellow), and PTVCgreen) as described in Table 1. Transverse sections shown at theevel of the common iliac nodes, internal and external iliac nodes,nd the vaginal vault respectively. PTV � planning tanget volume.

ig. 2. Discrepancy in reporting the volumes of normal tissuesrradiated: The arrow shows a part of the target volume that alsoontains normal bowel. In standard-plan calculations, pixels in thisrea would be counted only as part of the target and would not beot reported as bowel treated to 45 Gy (red bold line). Thisiscrepancy would result in an overestimation of volume of normal
owel spared in this high-dose region.

cecaapt

vItamcttalaaiPPpwi

C

oa(itbcwd

T

m9pwfiam

nocibwptnw

At

vmrnpt

FqetGb

B

R

B

1119Correlation between internal and external markers for abdominal tumors ● A. AHAMAD et al.

onsidered to be at risk for harboring microscopic disease. Forxample, the inferior part of the posterior bladder wall was in-luded where it abutted the upper vaginal vault and 1 to 2 cmbove because this region was considered to be part of the oper-tive bed at risk for microscopic disease involvement. However,elvic and vertebral bones were excluded from the CTV becausehey were considered to be at very low risk.

Margins of different magnitudes were added to the centralaginal CTV and regional nodal CTV to produce 3 different PTVs.n each case, greater margins were added anterior and posterior tohe vagina than to other portions of the CTV. The margins weredded in an effort to account for the as yet unknown internalotion of the vagina. Asymmetric expansion was performed be-

ause patients were simulated with a full bladder, which suggestedhat anterior deviation from the simulated position may be greaterhan posterior deviation. Studies are currently underway to moreccurately define the impact of internal-organ motion on the targetocation. PTVA consisted of the regional nodal CTV with nodditional margin plus the central vaginal CTV with 10 mm addednteriorly and 5 mm added posteriorly, laterally, superiorly, andnferiorly. PTVB and PTVC were created by uniformly expandingTVA by 5 mm and 10 mm, respectively (Table 1 and Fig. 1).TVs were generated on a contouring workstation and then ex-orted to the IMRT and the 3D planning computers. This processas necessary to ensure that the PTVs for the 3D plans were

dentical to those for the IMRT plans.

ig. 3. Intensity-modulated radiation therapy (IMRT) sparing. Thisuantity was used to correlate with body mass index. In thisxample, the volume of small bowel treated by conventional planso 30 Gy (V4FC) is shown in blue (A) and the volume treated to 30y by IMRT (VIMRT) is shown in yellow (B). The differenceetween these 2 volumes is shown in black (C).

Table 2. Irradiated normal-tissue volumes as stated by Corvus treIM

Organ Dose (Gy)

Median tissu

Standard plan (S

ladder 45 13.2 (6–48)30 77.9 (20–200)

ectum 45 16.5 (4–35)30 51.2 (19–121)

owel 45 35.3 (21–56)30 353.1 (262–477)

Abbreviation: IMRT � intensity-modulated radiation therapy.Note: Plans were generated by use of PTVB.
* The difference (H � S) is expressed as a percentage of S, the volum
ritical normal structuresThe superior and inferior extents of critical organs were outlined

n all CT slices in which portions of the PTV existed, as well ast an additional 2 cm superior and inferior to the limits of the PTVas recommended in ICRU 62) (2). Small bowel was outlined tonclude all intraperitoneal bowel and adjacent intraperitoneal softissue, with a superior limit 2 cm above the PTV. The differencesetween volumes of normal tissue treated by IMRT (VIMRT) vs. 2onformal fields (V2FC) and IMRT vs. 4 conformal fields (V4FC)ere derived by use of absolute tissue volumes determined byose–volume histograms (DVHs).

reatment planningIntensity-modulated radiation therapy (IMRT) and 3D confor-al treatment plans were prescribed to deliver 45 Gy to more than

7% of the PTV. IMRT plans were derived by use of inverselanning (Corvus, Nomos Inc., Sewickley, PA). Multiple iterationsere performed until an acceptable plan was achieved as con-rmed by inspection of the dose distribution in all 3 planes, DVHs,nd dose statistics. Intensity modulation was obtained by dynamicultileaf collimation.To select an optimum set of forward planning parameters (e.g.,

umber of beams and photon energy), preliminary plans werebtained by use of a variety of beam arrangements. In most cases,overage of the tumor and sparing of normal tissues progressivelymproved as the number of beams was increased from 4 to 8;eyond 8 fields, no significant improvement occurred. Plans alsoere generated by use of 8 fields with either 18-MV or 6-MVhotons. Six-megavolt IMRT plans tended to give better sparinghan 18-MV plans and were preferred because of their lowereutron production. For these reasons, 8-fields of 6-MV photonsere used for all of the IMRT plans in this study.

voiding the effect of overlapping target and normalissues on IMRT plans

If regions of the CTV overlap with normal structures, theolumes of normal tissues irradiated as reported by Corvus treat-ent planning software (Nomos) may not be accurate. Pixels in

egions of overlap are counted only as part of the CTV and not asormal tissue. Thus, the DVH from a standard optimized IMRTlan excludes the volume of normal tissues that also forms part ofhe CTV and underestimates the volume of normal tissues irradi-

t-planning software using standard plans (S) and hybridized (H)ns

mes (cm3) said to receivee (range)

Percentagedifference*

Hybridized plan (H) [(H � S)/S] � 100

61.6 (22–127) 367%122.4 (41–275) 57%

46.4 (24–88) 181%90.9 (40–180) 77%

128.4 (85–228) 263%466.2 (369–581) 32%

atmenRT pla

e voludos

)

e stated by the standard IMRT plan.

avCwhtbbm

thtFEwntaosth

T

c2TawBm

E

P

otIrtbti

P

tpo4

S

vawt

pt

ta(

o

S

t

Fm(wTu


ted. Because we added various margins to the CTV (with variousolumes of normal-tissue overlap) before exporting the data to theORVUS planning software, a simple comparison between plansould have yielded spurious results. Uncorrected plans wouldave significantly underestimated the volume of normal tissuereated to relatively high doses, particularly for plans that wereased on PTVs with large treatment margins. This error occursecause the normal tissues treated to the highest doses are at theargins of the CTV (Fig. 2).For this reason, all dose–volume data that pertains to normal

issues were calculated by use of composite DVHs derived fromybridized plans to accurately estimate the volumes of normalissue irradiated. These hybridized plans were produced as follows:or each case, a standard optimized IMRT plan was produced.ach optimized plan was convolved (without further optimization)ith the respective CT data that contained only the patient’sormal-tissue contours (without the CTV contours). Exclusion ofhe CTV from these dose–volume calculations, allowed accuratend consistent comparisons of the critical organ DVHs to bebtained. To confirm the suspected discrepancy that occurs withtandard-plan statistics from this planning system, we measuredhe difference between dose–volume statistics from standard vs.ybridized plans.

wo-field and four-field 3D plansThe ADAC (Pinnacle) system was used to produce 18-MV 3D

onformal plans by use of 2 fields (2FC) or 4 fields (4FC). TheFC plans utilized an anteroposterior and a posteroanterior field.he 4FC plans utilized anteroposterior, posteroanterior, and rightnd left lateral fields. The PTV imported for ADAC 3D planningas identical to that used for IMRT plans as given in Table 1.eams-eye-view blocks were created, with an additional 5-mmargin added to the block edge.

ffect of margin size on normal-tissue sparingFor each organ, the DVHs that resulted from IMRT plans with

TVA, PTVB, and PTVC were compared.The amount of volume sparing at a given dose achieved with use

f IMRT vs. 3D conformal technique was the difference betweenhe volume irradiated by the 3D plan and that irradiated by theMRT plan. The volume of normal tissue (bowel, bladder, andectum) irradiated by the IMRT plan (VIMRT) was subtracted fromhe volume irradiated by the 4FC plans (V4FC) for each doseetween 15 and 50 Gy at 5-Gy intervals (Fig. 3). A comparison ofhe magnitude of sparing was performed when margins werencreased by 5 mm from PTVA to PTVB to PTVC.

atient dimensions and small-bowel sparingThe body mass index (BMI), length of the CTV, and width and

hickness of the patient at mid-CTV were measured for eachatient. Each of these variables was correlated with the magnitudef small-bowel sparing with use of IMRT (VIMRT � V4FC) at the0-Gy level.

tatistical methodsThe differences in volumes of normal tissue irradiated by 2FC

s. IMRT and 4FC vs. IMRT were calculated at dose points 30, 40,nd 45 Gy. Data were analyzed by use of the sign test to determinehether the median volumes of bowel, bladder, or rectum treated

o various dose levels were different for IMRT vs. 2FC or 4FC (

lans. The sign test is a nonparametric test that makes no assump-ions about the underlying shapes of the distributions.

The effects of radiation treatment technique and margin size onhe doses and volumes of organs irradiated were compared bypplication of a 2-way repeated-measures analysis of varianceANOVA) that took unequal variances into account.

Patients’ physical characteristics were correlated with the amountf small-bowel sparing by use Spearman’s correlation.

RESULTS

tandard versus hybridized plansThe median volumes of small bowel, bladder, and rectum

reated to 30 and 45 Gy with standard vs. hybridized plans

ig. 4. Comparison of the effect on the normal tissue of intensity-odulated radiation therapy (IMRT) vs. the two conformal fields

2FC) and four conformal fields (4FC) conventional techniqueshen 45 Gy was given to initial planning target volume (PTVA).hese are composite dose volume histograms of the median vol-me (cc) of small bowel (top), rectum (middle), and bladder
bottom).

at3bstuapufePbb

Ir

ltostsIsI

iITnavI

maC

E

It4iceb

V

V

V

V

V

V


re given in Table 2. The standard-plan DVH underreportedhe volume of these organs that received 45 Gy by 181% to67% and the volume of these organs that received 30 Gyy 32% to 57%. The magnitude of underreporting wasignificantly correlated with the size of the margins used forhe PTV. When a random-effects regression model wassed, a highly significant effect of margin size on themount of underreported V45 was seen. Compared with thelans for PTVA, a change to PTVB led to an increase in thenderreporting of the V45 by 68.2 cc, 23.5 cc, and 23.6 ccor bowel, rectum, and bladder, respectively (p � 0.001 inach case). Compared with the plans for PTVA, a change toTV C led to an increase in the underreporting of the V45y 152.0 cc, 46.2 cc, or 43.7 cc for bowel, rectum, andladder, respectively (p � 0.001 in each case).

MRT versus conventional 3D conformaladiation therapy

Intensity-modulated radiation therapy plans deliveredess radiation to normal tissue than did the 2FC and 4FCechniques (Fig. 4). ANOVA performed across the rangef doses studied, demonstrated that a significantlymaller volume of small bowel was irradiated by IMRThan by 2FC (p � 0.0001) or 4FC (p � 0.0001). Aignificantly smaller volume of rectum was irradiated byMRT than by 2FC (p � 0.0001) or 4FC (p � 0.0001). Aignificantly smaller volume of bladder was irradiated byMRT than by 2FC (p � 0.0001) or vs. 4FC (p � 0.0001).

The median differences in small-bowel volume irradiated tondividual dose points 30, 40, and 45 Gy by 2FC and 4FC vs.MRT for margins A, B, and C are summarized in Table 3.hese differences represent the reductions in the volumes oformal tissue treated to each dose if IMRT is used rather thanstandard conformal plan. The median differences in the rectalolumes irradiated to 30, 40, and 45 Gy by 2FC and 4FC vs.

Table 3. Median reductions in volume of small bowel tror 4-field confor

Comparison groups

Median reduc

PTVA

V30: 2FC vs. IMRT 427 (185–767)p � 0.002

V30: 4FC vs. IMRT 219 (165–377)p � 0.002

V40: 2FC vs. IMRT 571 (310–903)p � 0.002

V40: 4FC vs. IMRT 275 (167–429)p � 0.002

V45: 2FC vs. IMRT 621 (343–941)p � 0.002

V45: 4FC vs. IMRT 284 (173–457)p � 0.002

Abbreviation: IMRT � intensity-modulated radiation tfields; 4FC � four conformal fields.

MRT for margins A, B, and C are given in Table 4. The

edian differences in the bladder volumes irradiated to 30, 40,nd 45 Gy by 2FC and 4FC vs. IMRT for margins A, B, and

are given in Table 5.

ffects of margin size and BMI on normal-tissue sparingThe magnitude of normal-tissue sparing with use of

MRT as opposed to standard conformal plans decreased ashe size of the margins added to the CTV increased. Tables 3,, and 5 show that the median difference in organ volumesrradiated to 30, 40, and 45 Gy decreased as the PTVhanged from PTVA to PTVB to PTVC. Fig. 5 shows theffect of changing margins on the small-bowel, rectal, andladder DVHs.When repeated-measures ANOVA were applied across

o dose X (VX) with IMRT vs. 2-field conformal (2FC)FC) techniques

nge) in volume (cc) treated to given dose andificance according to sign test

PTVB PTVC

332 (104–756) 291 (75–744)p � 0.002 p � 0.002

158 (113.5–243) 165 (94.7–307)p � 0.002 p � 0.002

499 (245.8–895) 476 (210–913)p � 0.002 p � 0.002

230 (154–393) 233 (168–310)p � 0.002 p � 0.002

579 (293–910) 511 (257–932)p � 0.002 p � 0.002

294 (185–432) 271 (199–393)p � 0.002 p � 0.002

; PTV � planning target volume; 2FC � two conformal

Table 4. Median reductions in volume of rectum treated to doseX (VX) with IMRT vs. 2-field conformal (2FC) or 4-field

conformal (4FC) techniques

Comparison groups

Median reduction (range) in volume (cc)treated to given dose and significance

according to sign test

PTVA PTVB PTVC

30: 2FC vs. IMRT 10 (2–42) 8 (0–22) 13 (5–20)p � 0.002 p � 0.022 p � 0.002

30: 4FC vs. IMRT 9.0 (2–41) 6 (0–18) 11 (4–17)p � 0.002 p � 0.022 p � 0.002

40: 2FC vs. IMRT 34 (20–94) 25 (9–65) 24 (10–57)p � 0.002 p � 0.002 p � 0.002

40: 4FC vs. IMRT 27 (12–87) 22 (7–64) 22 (9–56)p � 0.002 p � 0.002 p � 0.002

45: 2FC vs. IMRT 59 (26–127) 46 (21–109) 35 (15–88)p � 0.002 p � 0.002 p � 0.002

45: 4FC vs. IMRT 38 (7–107) 30 (5–98) 27 (6–79)p � 0.002 p � 0.002 p � 0.002

eated tmal (4

tion (rasign

herapy

Abbreviations as in Table 3.

aPIlst

ad

4sd0

stcpbwpmr

misPott

dp

cchra5hmd

Fmcrcv

V

V

V

V

V )

V


ll doses for all 10 patients, as margin size increased fromTVA to PTVB to PTVC, the volume of bowel spared byMRT vs. 4FC significantly decreased (p � 0.0079). Simi-arly, the volume of small bowel spared by IMRT vs. 2FCignificantly decreased as margin size increased from PTVAo PTVB to PTVC (p � 0.0002).

Table 6 shows the loss of bowel volume spared by IMRTs PTV increased from PTVA to PTVB to PTVC at specificoses.The amount of small-bowel sparing with use of IMRT vs.

FC was related to the body habitus of the 10 patients in thistudy. As BMI increased, the amount of sparing significantlyecreased (Spearman’s correlation coefficient � �0.64; p �.05).

DISCUSSION

Although a dose of 45 to 50 Gy improves relapse-freeurvival and reduces the risk of pelvic recurrences in pa-ients at high risk after hysterectomy for endometrial orervical cancer (3–5), clinicians frequently hesitate to useostoperative radiation therapy out of concern about possi-le late effects. The highly conformal treatments achievableith use of IMRT should, if accurately applied, make itossible to deliver standard doses more safely and ulti-ately to escalate the dose delivered to particularly high-

isk regions without causing unacceptable morbidity.Intensity-modulated radiation therapy is currently used inany centers to achieve more conformal treatment of this

rregular treatment volume (6–10). However, critical issuestill must be addressed: the influence of differing CTV andTV definitions; the effect of not reporting the true dose to theverlap regions of normal tissues; the influence of patient andarget-volume parameters on the relative benefit of IMRT; and

Table 5. Median reductions in volume of bladder treated to doseX (VX) with IMRT vs. 2-field conformal (2FC) or 4-field

conformal (4FC) techniques

Comparison groups

Median reduction (range) in volume (cc)treated to given dose and significance

according to sign test

PTVA PTVB PTVC

30: 2FC vs. IMRT 16 (6–101) 7 (1–59) 9 (4–74)p � 0.002 p � 0.002 p � 0.002

30: 4FC vs. IMRT 9 (4–79) 5 (0–44) 8 (4–74)p � 0.002 p � 0.022 p � 0.002

40: 2FC vs. IMRT 55 (9–173) 38 (3–122) 39 (10–126)p � 0.002 p � 0.002 p � 0.002

40: 4FC vs. IMRT 29 (6–93) 25 (3–74) 29 (10–83)p � 0.002 p � 0.002 p � 0.002

45: 2FC vs. IMRT 89 (12–222) 72 (5–213) 56 (123–142p � 0.002 p � 0.002 p � 0.002

45: 4FC vs. IMRT 50 (2–126) 50 (1–131) 46 (10–142)p � 0.002 p � 0.002 p � 0.002


he magnitude of internal-organ motion. Physicians must un- (

erstand these parameters to effectively deliver IMRT to ap-ropriately selected patients.

Doses of 60 to 72 Gy are routinely administered toontrol regional disease in patients with squamous cellarcinomas of the head and neck. For patients who have aigh risk of regional recurrence after neck dissection, a clearelationship has been demonstrated between radiation dosend disease-control rates (11). However, doses of more than0 to 55 Gy are rarely contemplated in patients who have aigh risk of recurrence after surgery for cervical or endo-etrial cancers, because the risks of serious bowel or blad-

er injury have been prohibitively high with use of standard

ig. 5. Comparison of the the effect on the normal tissue asargins are increased from PTVA to PTVB to PTVC. These are

omposite DVHs of the median volume (cc) of small bowel (top),ectum (middle), and bladder (bottom) as margin sizes are in-reased. This difference in volumes was significant by analysis ofariance: small bowel (p � 0.01), rectum (p � 0.01), and bladder
p � 0.01).

rahatdhatns

eoai

pucdaeasrggiaahTpw

tuaplmcl

mfotsogcmHhtdm

tbatbcs3raUecwiutdtcsitmptr

c

24


adiation techniques. Even when lower doses of 45 to 50 Gyre used, the risk of severe late small-bowel injury can beigh (5% to 15%) for patients who undergo radiation ther-py after hysterectomy with transperitoneal lymphadenec-omy (12, 13). The risk of injury is even greater if higheroses are used or if patients are very thin, smoke, or have aistory of pelvic infection (14). For these reasons, patientsre rarely treated with doses of more than 50 Gy, evenhough the rate of in-field recurrences continues to be sig-ificant, which suggests that lower doses are insufficient forome patients (3–5).

Recently, 2 groups have reported small clinical experi-nces with IMRT to the pelvis in patients who had uteriner cervical cancer (6–10). These authors demonstrated thebility to conform to complex pelvic target volumes, withmproved sparing of normal tissues.

The University of Chicago group has published severalapers that discuss their early clinical experience with these of IMRT to treat 10 women who had intact cervicalancer or had undergone hysterectomy for cervical or en-ometrial cancer (6–9). The authors defined the lymph nodereas of the CTV by adding a 2-cm margin around contrast-nhanced vessels. This method of target-volume definition,lthough easily described, inevitably includes normal tis-ues such as bone and small bowel that have a negligibleisk of harboring microscopic disease. Although this veryenerous definition of the CTV may reduce the chance of aeographical miss, it also reduces the degree of tissue spar-ng achievable with use of IMRT. To generate a PTV, theseuthors added 1 cm to the CTV in all directions, on thessumption of uniform motion of the target in all directions;owever, the validity of this assumption remains unknown.he authors noted that their margins could be reduced if theosition of the CTV throughout the course of treatmentere more accurately known.The second group explored the feasibility of IMRT to

reat the pelvis and para-aortic region in women with intactteri (10). They reported theoretical plans developed fromrchived CT data of 10 patients. The radiation dose wasrescribed to a target volume that contained the uterus andymph nodes only. No margin was added to allow for organotion or setup variability. The clinical use of such plans

ould result in underdosage of the target volume, particu-arly if internal-organ motion is significant.

Table 6. Median reduction (cc) in volume of small bowel sparconformal (4FC) plans

30 Gy, PTVA vs. PTVB 30 Gy, PTVA vs. PTVC

Medianreduction (range) p

Medianreduction (range)

FC 95.8 (11.5, 204.8) 0.002 147.9 (23.3, 312.6) 0.0FC 56.4 (11.4, 133.9) 0.002 66.3 (�131, 187.1) 0.1


One limitation noted in these studies was lack of data on p

otion of abdominal and pelvic organs during irradiationor cervical cancer. In our study, we found that the volumef the normal tissues spared by IMRT relative to conven-ional techniques was very sensitive to small increases in theize of the margin used to generate the PTV. The volumesf irradiated normal tissues increased markedly as the mar-ins for possible setup error and organ motion were in-reased. This observation emphasizes the importance ofinimizing unnecessary expansion of the target volume.owever, the steep fall-off in dose associated with theighly conformal distributions achieved with IMRT showshat we must not underestimate these uncertainties, becauseoing so would undoubtedly lead to undertreatment of tu-or-bearing tissues.Our data clearly demonstrate the importance of accurate

arget-volume definition and the benefit that can be gainedy improved immobilization and decreased uncertaintiesssociated with internal-organ motion. The intimate rela-ionship between the cervix, upper vagina, and posteriorladder wall can result in dramatic positional shifts, withhanges in bladder filling. Our anecdotal clinical experienceuggests that the position of the vaginal apex can change bycm or more as an empty bladder is filled. Variations in

ectal filling may also cause shifts in the target position,lthough these shifts are probably less dramatic. At Theniversity of Texas M. D. Anderson Cancer Center, when-

ver a boost to high-risk paravaginal tissues is needed, thelinical practice is to plan treatments by use of fused imagesith a full bladder and an empty bladder to produce an

nternal target volume (ITV). Alternatively, if this ITV isnacceptably large, we infuse a fixed volume of saline forhe treatment-planning CT and then catheterize patientsaily to fill the bladder with the same volume throughoutreatment. Although this method permits a more tightlyonforming treatment plan, with less critical organ expo-ure, the inconvenience of daily catheterization usually lim-ts this approach to relatively short boost treatments of lesshan 10 fractions. However, unless possible internal-organotion is well defined and taken into account in the planning

rocess, organ motion can lead to systematic errors throughouthe course of treatment as the CTV shifts out of the high-doseegion with its inherently steep dose-edge gradient.

Whereas uterine motion has been monitored in intactervical cancer patients by use of fluoroscopic electronic

given dose with IMRT vs. 2-field conformal (2FC) or 4-fieldhange in margin size

40 Gy, PTVA vs. PTVB 40 Gy, PTVA vs. PTVC



67.9 (�26.8, 105.3) 0.022 114.3 (�44.3, 219.8) 0.02231.3 (�1.9, 76.1) 0.022 42.5 (�26.1, 121.1) 0.109

ed thewith c

p

0209

ortal imaging and radiopaque markers (15), similar

swctt

irtm


tudies of vaginal motion have not been done in patientsho have undergone hysterectomy. We are currently

onducting a study to obtain sequential CT scans duringreatment to determine the consistency of patient posi-
ioning and internal-organ position. This study and sim- c
REFEREN

8. Mundt AJ, Roeske JC, Lujan AE. Intensity-modulated radia-

1

1

1

1

1

1

lar studies may allow us to refine IMRT technique byeducing some of the current uncertainties. In the mean-ime, IMRT should be applied with caution in the treat-ent of patients who have undergone hysterectomy for

ervical or endometrial cancer.

CES

1. Intensity Modulated Radiation Therapy Collaborative WorkingGroup. Intensity-modulated radiotherapy: Current status and is-sues of interest. Int J Radiat Oncol Biol Phys 2001;51:880–914.

2. International Commission on Radiation Units and Measure-ments. Report 62. Prescribing, Recording, and Reporting PhotonBeam Therapy (supplement to ICRU Report 50). Bethesda, MD,1999.

3. Creutzberg CL, van Putten WL, Koper PC, et al. Surgery andpostoperative radiotherapy versus surgery alone for patientswith stage-1 endometrial carcinoma: Multicentre randomisedtrial. PORTEC Study Group. Post Operative Radiation Ther-apy in Endometrial Carcinoma. Lancet 2000;355:1404–1411.

4. Roberts JA, Brunetto VI, Keys HM, et al. A phase III ran-domized study of surgery vs surgery plus adjunctive radiationtherapy in intermediate-risk endometrial adenocarcinoma (GOGNo. 99) (abstract). Gynecol Oncol 1998;68:135.

5. Sedlis A, Bundy BN, Rotman MZ, et al. A randomized trial ofpelvic radiation therapy versus no further therapy in selectedpatients with stage IB carcinoma of the cervix after radicalhysterectomy and pelvic lymphadenectomy: A Gynecol OncolGroup study. Gynecol Oncol 1999;73:177–183.

6. Brixey CJ, Roeske JC, Lujan AE, et al. Impact of intensity-modulated radiotherapy on acute hematologic toxicity inwomen with gynecologic malignancies. Int J Radiat OncolBiol Phys 2002;54:1388–1396.

7. Mundt AJ, Roeske JC, Lujan AE, et al. Initial clinical expe-rience with intensity-modulated whole-pelvis radiation ther-apy in women with gynecologic malignancies. Gynecol Oncol2001;82:456–463.

tion therapy in gynecologic malignancies. Med Dosim 2002;27:131–136.

9. Roeske JC, Lujan A, Rotmensch J, et al. Intensity-modulatedwhole pelvic radiation therapy in patients with gynecologic ma-lignancies. Int J Radiat Oncol Biol Phys 2000;48:1613–1621.

0. Portelance L, Chao KS, Grigsby PW, et al. Intensity-modu-lated radiation therapy (IMRT) reduces small bowel, rectum,and bladder doses in patients with cervical cancer receivingpelvic and para-aortic irradiation. Int J Radiat Oncol Biol Phys2001;51:261–266.

1. Peters LJ, Goepfert H, Ang KK, et al. Evaluation of the dosefor postoperative radiation therapy of head and neck cancer:First report of a prospective randomized trial. Int J RadiatOncol Biol Phys 1993;26:3–11.

2. Eifel PJ, Levenback C, Wharton JT, et al. Time course andincidence of late complications in patients treated with radia-tion therapy for FIGO stage IB carcinoma of the uterinecervix. Int J Radiat Oncol Biol Phys 1995;32:1289–1300.

3. Corn BW, Lanciano RM, Greven KM, et al. Impact of im-proved irradiation technique, age and lymph node sampling onthe severe complication rate of surgically staged endometrialcancer patients: A multivariate analysis. J Clin Oncol 1994;12:510–515.

4. Eifel PJ, Jhingran A, Bodurka DC, et al. Correlation of smok-ing history and other patient characteristics with major com-plications of pelvic radiation therapy for cervical cancer.J Clin Oncol 2002;20:3651–3657.

5. Kaatee RS, Olofsen MJ, Verstraate MB, et al. Detection oforgan movement in cervix cancer patients using a fluoroscopicelectronic portal imaging device and radiopaque markers. Int
J Radiat Oncol Biol Phys 2002;54:576–583.

intensity-modulated radiation therapy after hysterectomy: comparison with conventional treatment and...

Documents