Inf. J. Radmion Oncology Bml. Phys. Vol. Pnnted in the U.S A. All rights reserved

21, pp

253-265 Copyright

0360-3016/91 $3 00 + .oO 0 1991 Pergamon Press plc

??Original Contribution

THREE-DIMENSIONAL TREATMENT PLANNING FOR POSTOPERATIVE TREATMENT OF RECTAL CARCINOMA BRENDA SHANK, M.D.,

PH.D. ,l THOMAS LOSASSO, PH.D. ,l LINDA BREWSTER, M.S.,’

CHANDRA BURMAN, PH.D.,’ ROBERT E. DRZYMALA, PH.D., LAWRENCE J. SOLIN, M.D.

ELIZABETH CHENG, B.S.

3 JANICE MANOLIS,

,* JAMES C. H. CHU, PH.D.

R.T.(T.),3

,* J. E. TEPPER, M.D.4

,’

MILJENKO V. PILEPICH, M.D.

AND MARCIA M.

,3

URIE, PH.D.~

‘Memorial Sloan-Kettering Cancer Center, New York, NY 10021; *University of Pennsylvania School of Medicine and the Fox Chase Cancer Center, Philadelphia, PA 19111; 3Mallinckrodt Institute of Radiology, Washington University School of Medicine, St. Louis, MO 63 110; and 4Massachusetts General Hospital, Department of Radiation Medicine, Boston, MA 02114 The role of three-dimensional (3-D) treatment planning for postoperative radiation therapy was evaluated for rectal carcinoma as part of an NC1 contract awarded to four institutions. It was found that the most important contribution of 3-D planning for this site was the ability to plan and localize target and normal tissues at all levels of the treatment volume, rather than using the traditional method of planning with only a single central transverse slice and simulation films. There was also a slight additional improvement when there were no constraints on the types of plans (i.e., when noncoplanar beams were used). Inhomogeneity considerations were not important at this site under the conditions of planning, i.e., with energies >4 MV and multiple fields. Higher beam energies (15-25 MV) were preferred by a small margin over lower energies (down to 4 MV). The beam’s eye view and dose-volume histograms were found quite useful as planning tools, hut it was clear that work should continue on better 3-D displays and improved means of translating such plans to the treatment area. Three-dimensional treatment planning, Dose-volume histograms, Beam’s eye view, Rectal carcinoma.

Rectal carcinoma was chosen as a site for three-dimensional (3-D) treatment planning because of its importance clinically (high incidence and low survival), and its challenge to the treatment planner (large volume to be treated to a relatively high dose adjacent to sensitive structures, especially small bowel). Colorectal carcinoma represents the leading site of cancer today with a relative incidence of 15%, of which 30% are cancers of the rectum (4). Although surgery is the mainstay of treatment for rectal carcinoma, the overall five-year survival is only 65%, even with the most aggressive surgical approach (5). This has prompted a large number of adjuvant radiation therapy studies, either prior to surgery (preoperative), after surgery (postoperative), or a combination of both (“sandwich” radiotherapy). Regardless of technique, doses required are high (8-11, 13, 14, 2 l), in the range of 4000-6000 cGy , even to control microscopic disease (7). This creates a treatment planning problem in that radiosensitive structures (small bowel,

bladder, femoral heads) are quite close to the planned target volume and, frequently, the target volume is difficult to delineate, either pre- or postoperatively. Furthermore, there are large inhomogeneities in pelvic fields, varying from bone in the pelvis to air within the bowel which is inconstant in location. Bone densities may be taken into consideration easily when computerized tomography (CT) data are used for planning; however, due to the mobility and distensibility of the small bowel, as well as the variable amounts of air present from day to day, relative locations and amounts of air and soft tissue in the pelvic area will vary from treatment to treatment. The influence of these inhomogeneities, as well as the role of beam quality in minimizing inhomogeneity effects and sparing of normal tissues, needs to be assessed. The depth-dose characteristics of high energy photon beams are usually thought to be desirable in treatment of rectal cancer since these tumors are deep-seated within the pelvis, a region that is often the thickest part of the body; however, this has not been subjected to a rigorous analysis. Prior to this contract, CT had been used to aid in treatment planning in relatively unsophisticated ways: for example,

Supported in part by NC1 Contracts NO1 CM-47316, NO1 CM-47695, NO 1 CM-47696, and NO1 CM-47697.

Reprint requests to: Brenda Shank, Mount Sinai Medical Center, One Gustave Levy Place, New York, NY 10029.

INTRODUCTION Rationale for choice of this site

253

I. J. Radiation Oncology 0 Biology 0 Physics

254

May 15, 1991, Volume 21, Number 1

Table 1. Postoperative rectum patient characteristics Patient Age (Y) Sex Body habitus Stage (Astler-Coller) Rectal exam Colonoscopy Barium enema

IVP

CXR Liver/spleen scan Type of resection

RECTUM- 1 52 6 Thin (21 x 38)* C2 Not palpable “Tumor at 9 cm” “Irreg. infilt. mass, 4 cm in length, on right lat. aspect of rectum. ” Left kidney larger than right. Clips in pelvis from prior surgery. Normal “Unremarkable” Abdominoperineal resection

RECTUM-2 56 0 Obese (24 X 42)* B2 (2nd 1”) Not palpable Not available “Diverticulosis. Intact anastomosis (distensible) from previous primary.” Normal

-

(4 Low anterior resection

* AP diameter x lateral width (cm).

normal tissue and target volumes within transverse planes with greater accuracy than previously available with orthogonal films (2). The advantages of pelvic CT are gradually being appreciated, but only the advent of interactive 3-D graphic displays has allowed the full potential of CT for treatment planning to be realized. Only with 3-D planning can one explore the possibility of utilizing noncoplanar beams to achieve high doses to the target areas with the avoidance of sensitive normal tissues. Postoperative treatment was chosen for this study because of its clinical relevance as judged by the prevalence of this type of treatment at the time of choosing this site (9, 21) and its use in a large cooperative group trial (8).

to localize

Protocol description Patient eligibility.

Patients with “curatively resected” adenocarcinoma of the rectum with Astler-Coller stage B2, Cl, or C2 (1) were eligible for this study. Stage B2: tumor extending through the bowel wall with negative pelvic lymph nodes. Stage Cl: tumor not extending through the bowel wall, but positive pelvic lymph nodes. Stage C2: tumor extending through the bowel wall with positive pelvic lymph nodes. “Curative resection” was defined as resection of all gross disease by an abdominoperineal resection (APR), low anterior resection (LAR), etc., with or without a pelvic lymph node dissection. Details of the pathology, staging, and surgical procedure were to be recorded on the patient accrual worksheet for this site. Patient evaluation. Required preoperative evaluation consisted of a physical examination, including a digital

Fig. 1. Representative CT images through area of BTV-2 (contours shown) for (a) thin male patient (RECTUM-l), and (b) obese female patient (RECTUM-2).

rectal examination, and a barium enema. Postoperatively, a CT exam for treatment planning was required, with the patient in the prone position. This CT exam encompassed the site of the primary and the regional lymph nodes with wide superior and inferior margins, in order to allow the use of noncoplanar beams from the caudad and/or cephalad directions. The inferior border was to be 10 cm inferior to the perineal skin, and the superior border was to be 10 cm superior to the IA-L5 vertebral junction. Slice spacing was to be a maximum of 1 cm throughout; slice thickness was not specified. Other laboratory tests, or radiographic or endoscopic examinations, such as an intravenous pyelo-

3-D planning for postop rectal carcinoma 0 B.

SHANK et al.

255

Fig. 2. Surface dose display of the bowel for RECTUM-2 patient. View is from left posterior inferior direction. “Hot” (red) region is apparent where bowel is adjacent to BTV-2.

gram or magnetic resonance imaging, were optional. Results of an IVP or MR imaging, if done, were to be reported on the patient accrual worksheet, along with the findings of the rectal exam, barium enema, and CTs. Radiation therapy specifications. The biological target volumes were defined as follows for this site: BTV-1: Pelvic nodes (para-rectal, presacral, internal iliac, common iliac); inguinal nodes (only for low-lying lesions, i.e., 55 cm from the anal verge). These were all considered to be nodes with potential microscopic disease. BTV-2: Primary tumor site (“tumor bed”) as defined by preoperative clinical exam, preoperative barium enema, preoperative CT scan, and/or operative note. Margins (for presumed microscopic disease) to extend 2.5 cm superiorly and inferiorly, and 2 cm radially. The prescribed tumor dose (PTD) to the mobile target volume- 1 (MTV-1) (17) for the nodal areas was specified as 46 Gy and to MTV-2, 56 Gy. All doses were to be delivered at conventional fractionation of 1.8-2.0 Gy/ fraction, 5 fractions/week. Normal tissues considered as at risk for injury were small bowel, bladder, femoral heads, skin, colon (if a low anterior resection had been done), and kidneys.

listed in Table 1. These particular patients were chosen to test planning in patients who were different in their tumor characteristics, body structure, and surgical procedure. Patient RECTUM-l (Fig. la) is a thin male with primary rectal cancer, Stage C2, who had an abdominoperineal resection with a permanent colostomy, whereas patient RECTUM-2 (Fig. lb) is an obese female with a Stage B2 rectal cancer, her second primary (the first being in the sigmoid colon), who had a low anterior resection for each primary.

METHODS AND MATERIALS

Biological target volumes BTV-1 description. Both patients had tumors that were >5 cm from the anal verge, so inguinal nodes were not to be included in the treatment plans, only pelvic nodes as defined above. B7’V-2 description. In patient RECTUM-l, BTV-2 was determined with the aid of 1) the colonoscopy report, 2) the preoperative barium enema, and 3) the postoperative CT scan (without contrast, no rectum postsurgery), showing surgical clips (Fig. la). In patient RECTUM-2, BTV-2 was determined solely by the postoperative CT scan [with contrast, 50 cc of dilute (1:25) Gastrografin per rectum], which showed the surgical clips. In both instances, the CT scan specifications adhered to the protocol requirements as stipulated above, and slice thickness was 1 cm.

Patient characteristics Both patients had histologically proven adenocarcinoma of the rectum. Details of the patients’ characteristics are

Traditional plan For this site, traditional plans were a pelvic “brick” for MTV- 1, with AP, PA, and two unwedged lateral fields, to

I. J. Radiation Oncology 0 Biology 0 Physics

256

May 15, 1991, Volume 21, Number 1

Table 2. Plans submitted for postoperative

Inst.

3-D std.

A B C D

X X X X

A B C D

X X X X

Convent.

X

Tradit .

X X

Inbomog .

Uncert.

Patient #l X* X* X* X* Patient #2 X*

X

rectum Beam qual.

3-D unconst.

X*

X X

X*

X

X X X

X

X

X*

X

X* X*

X

X X X

X X*

* Not formally evaluated at the evaluating institution.

a dose of 46 Gy, prescribed to the isocenter. The inclusion of the perineal scar is still controversial (18); for this contract, the perineal scar was not included in the MTV-2 fields. Beam energy was 10 MV photons. Following this, a cone down to MTV-2 was to be done; a suggested arrangement was two lateral wedged fields and a posterior field, for an additional 10 Gy. The superior border of the pelvic fields was the L4-5 junction. The posterior border of the lateral fields was through the center of the sacrum, and the anterior border was just posterior to the pubic symphysis, with care taken to note surgical clips and include superior nodal areas. Typical examples of whole pelvic and cone-down fields, and isodose contours generated without CT for a patient with an APR, were distributed to the participating institutions. From the above type of plan, with blocking delineated on either orthogonal simulation films or digital reconstructions, 3-D dose distributions with pixel-by-pixel inhomogeneity corrections were generated. Conventional plan The conventional plan for these two patients, i.e., how

they were actually planned and treated, was expected to be similar to the traditional plan, a four-field pelvic “brick” plan to 46 Gy with “boost” fields to 56 Gy total. This plan was included in order to ascertain how closely the traditional plan reflected the “reality” of actual treatment. 3-D standard plan The 3-D standard plan for these patients again was to utilize the same 4-field arrangement as used for the traditional plan, but 3-D graphics were to be used for planning, so that all target and normal tissue volumes would be defined and considered at all levels, and blocking defined accordingly. 3-D unconstrained plan For the 3-D unconstrained plan, an attempt at some degree of optimization was to be done and there were to be

no constraints on beam arrangements, as described in the prologue to these site-specific papers (17). Other plans

For this site, plans for inhomogeneity, beam quality, and uncertainty due to motion, etc., were also done (Table 2). Plan evaluations

For both individual plan evaluations and for blinded pairwise plan comparisons, the scoring system designed for this contract was used (12), which considers both tumor coverage and normal tissue complication potential. RESULTS AND DISCUSSION It is clear that the choice of postoperative treatment for rectal carcinoma remains an important adjuvant treatment option with many recent reports appearing in the literature of both nonrandomized (15, 20) and randomized (3, 6) studies. Treatment in these studies has been planned by traditional methods of simulation. The results presented herein will hopefully stimulate future clinical studies utilizing 3-D planning. Plans: examples and comparisons

Plans submitted for the postoperative rectum site are listed by coded institution in Table 2. Each category of plan will be discussed separately. Traditional plans. There were three traditional plans submitted: two plans for RECTUM-1 and one plan for RECTUM-2 (Table 3). Institution B adhered to the proposed traditional plan (see Methods and Materials), with field shaping designed on the simulation films for both patients. Institution D, however, demonstrated a different treatment philosophy in their traditional plan. Fields for RECTUM- 1 at Institution D were planned on digitally reconstructed radiographs from the CT images, and the field arrange-

3-D planning for postop rectal carcinoma 0 B. Table 3. Beam arrangements

SHANKer al.

251

for rectal traditional plans Beam arrangements

Patient

Institution

RECTUM- 1

B

4-field brick to 46 Gy; 10 MV; block design on simulation films for MTV-1 and MTV-2; weighting: A*-0.98, P-1.00, LL-0.77, RL-0.79

3-field boost to 10 Gy; 10 MV; weighting: wedged LPO- 1.23, wedged RPO- 1.27

D

4-field brick to 46 Gy with 15” lateral wedges; 10 MV; block design on digitally reconstructed radiographs for MTV-1 and MTV-2; weighting: equal

2 lateral boost fields to 4 Gy only with 15” wedges;

4-field brick to 46 Gy; 10 MV; block design on simulation films for MTV-1 and MTV-2; weighting: A-1.01, P-1.00, LL-0.82, RL-0.77

3-field boost to 10 Gy; 18 MV; weighting: wedged LPG-0.40, wedged RPG-0.70

RECTUM-2

B

MTV- 1

* A = anterior field, P = posterior, oblique.

MTV-2

weighting:

equal

LL = left lateral, RL = right lateral, LPO = left posterior oblique,

ment was similar to the proposed plan, but the dose to MTV-2 was limited to 50 Gy, due to concern regarding small bowel tolerance. Also the most superior lymph nodes were not included in the plan to the proposed dose for the same reason. Evaluations of traditional plans on an individual basis (single plan evaluations) reflected these two different institutional philosophies. The tumor control score (12) was lower for the Institution D plan as a result of the lower dose to MTV-2 and to the superior lymph nodes of MTV-1. If one accepts the premise that these nodes are relatively less important, i.e., less likely to contain tumor, the score would be higher, but no differential weighting of lymphatic areas was applied in this contract. 3-D standard plans. Each institution submitted a 3-D standard plan for each patient (Table 4). Although most plans were variations on the theme of a four-field brick to 46 Gy with a two- or three-field boost to follow, there were a few deviations from this expected class solution. Institution C used parallel opposed anterior and posterior fields for their initial fields to MTV-1 as well as to MTV-2 on both patients. In the obese patient, RECTUM-2, Institution A used three fields initially to MTV-1 (no anterior field), rather than four. All energies chosen were >lO MV (lo-18 MV). Single plan evaluations demonstrated that the volume of normal tissues other than small bowel (e.g., bladder or femoral heads) which received a high enough dose to be of concern was extremely small in these plans, as in the other plans at this site, and was considered to result in a very low complication probability, or that the doses to any portion of these organs (other than small bowel) were below the threshold for complication. In evaluating small bowel at the institution responsible for this site, one useful 3-D display was the surjiice dose display (Fig. 2), which could be rotated interactively in 3 dimensions on the screen, show-

P-1.00,

P-1.00,

RPO = right posterior

ing quickly “hot” regions on the surface envelope of bowel. Blinded plan comparisons (Table 5) from the same institution on the same patient of 3-D standard and traditional plans demonstrated that, in all cases, the 3-D standard plan was preferred over the traditional plan, primarily because of better tumor coverage, especially of MTV-1. This agrees with the expected result, since the consideration of the target volume information from all CT levels should intuitively be better than planning treatment on the basis of only one level. Conventional plans. Both patients were actually treated in a fashion similar to that suggested for the traditional plan, which demonstrates the relevance of the Contract’s choice of treatment protocols to actual practice. Both patients had a four-field plan to the pelvis planned with the aid of orthogonal simulation films and the use of a CT scan to aid in field shaping on these films. These equally weighted fields (to MTV-1) were treated to 48 Gy to the isocenter in 2 Gy fractions in RECTUM-l and to 46.8 Gy in 1.8 Gy fractions in RECTUM-2. In both patients, 10 MV photons were used for these fields. “Boosts” (to MTV-2 in essence) in both patients were with three fields, a posterior field and 60” wedged right and left posterior oblique fields. The total dose was 10 Gy delivered in 2 Gy fractions, using 10 MV photons in RECTUM-l and 18 MV in RECTUM-2 (the obese patient). Other issues. At this site, consideration of inhomogeneities was not found to be of major importance, at least when multiple fields were used as generally practiced (22). This was ascertained from the six plans submitted dealing with the issue of inhomogeneity (Table 2). Five plans were done regarding the issue of uncertainty, i.e., that arising from patient motion, or from errors arising during treatment setup, such as a 5” wedge rotation (collimator rotation) when 45” wedged pairs were used (19).

258

I. .I. Radiation Oncology 0 Biology 0 Physics

May 15, 1991, Volume 21, Number 1

Fig. 3. Beam’s eye views for RECTUM-2 3-D unconstrained plan: (a) left anterior oblique field, (b) right pc lstelrior infer ior oblique field (note complexity of gantry and couch angles), and (c) anterior “boost” field for M[TV‘-2. strut Stlu*es shown on these views include: shaped fields (green), BTV-1 (blue), BTV-2 (fuchsia), outer contour (re:d), bladl der (green), femoral heads (yellow), bowel (gold).

3-D planning for postop rectal carcinoma ??B.

259

SHANK et al.

Fig. 3. (con@

From this small rotation, the dose-volume histograms were indistinguishable. From the analysis of the effect of motion at this site (MTVs vs. BTVs), it was clear that the volumes of the MTVs were considerably larger than the BTVs (ratios of 2.5 for MTV-l/BTV-1 and 1.4 for MTV-2/

BTV-2 for the RECTUM-l patient). This demonstrates that, for this site, when motion is explicitly considered, one must treat a considerably larger target volume (with concomitant normal tissue toxicity implications). When standard plans for the RECTUM-l patient were analyzed, the

Table 4. Beam arrangements for rectal 3-D standard plans Beam arrangements* Patient RECTUM-l

A B C D

RECTUM-2

MTV- 1

Institution

A B C D

4-field brick to 46 Gy; 15 MV; weighting: P-0.28, LL-0.23, RL-0.23 4-field brick to 46 Gy; 10 MV; weighting: P-l .oo, LL-0.77, RL-0.77 2 parallel opposed equally weighted P and to 46.8 Gy; 18 MV 4-field brick to 46 Gy; 10 MV; weighting: wedges: 15” RL + LL

MTV-2 A-0.26, A-0.93, A fields equal;

3 fields to 46 Gy; 15 MV; weighting: equal to P, LL, and RL; wedges: 30” RL + LL 4-field brick to 46 Gy; 10 MV; weighting: A-0.95, P-1.00, LL-0.79, RL-0.74 2 parallel opposed equally weighted P and A fields to 46.8 Gy; 18 MV 4-field brick to 46 Gy; 10 MV; weighting: equal

* Field codes as in Table 3.

3-field boost to 10 Gy; 15 MV; weighting: P-0.51, LL-0.25, RL-0.24; wedges: 45” RL + LL 3-field boost to 10 Gy; 10 MV; weighting: P-1.00, LPO-0.43, RPG-0.43; wedges: 60” RPO + LPO 2 parallel opposed equally weighted P and A fields to 9 Gy; 18 MV 2 lateral fields to 10 Gy; 10 MV; weighting: equal; wedges: 15” RL + LL 3-field boost to 10 Gy; 15 MV; weighting: P-0.38, LL-0.31, RL-0.31; wedges: 30” RL + LL 3-field boost to 10 Gy; 10 MV; weighting: P-l .OO, LPO-0.60, RPO-0.74; wedges: 60” RPO + LPO 2 parallel opposed equally weighted P and A fields to 9 Gy; 18 MV 2 lateral fields to 10 Gy; 10 MV; weighting: equal

I. J. Radiation Oncology 0 Biology 0 Physics

May 15, 1991, Volume 21, Number 1

(b)

(4

(4 Fig. 4. The three “boost” fields (to MTV-2) for the 3-D unconstrained plan on RECTUM-l from Institution B (see Table 7): (a) transverse plane with posterior (P) field and the 2 shaped field outlines projected on this plane from the right and left posterior oblique fields (RPIO and LPIO), (b) reconstructed mid-coronal plane with the P, RPIO, and LPI0 field projections, (c) reconstructed mid-sag&al plane with the P field, and RPIO and LPI0 projections, and dotted white line indicating (d) the arbitrary oblique plane determined by the central axes of the RPIO and LPI0 beams shown. The shaped P field projection in this plane is also shown in the center of the intersection of the 2 oblique beams.

volumes receiving

Three-dimensional treatment planning for postoperative treatment of rectal carcinoma.

The role of three-dimensional (3-D) treatment planning for postoperative radiation therapy was evaluated for rectal carcinoma as part of an NCI contra...
15MB Sizes 0 Downloads 0 Views