- Email: [email protected]
Impact of pelvic nodal irradiation with intensity-modulated radiotherapy on treatment of prostate cancer
Int. J. Radiation Oncology Biol. Phys., Vol. 66, No. 2, pp. 583–592, 2006 Copyright © 2006 Elsevier Inc. Printed in the USA. All rights reserved 0360-3016/06/$–see front matter
doi:10.1016/j.ijrobp.2006.05.033
PHYSICS CONTRIBUTION
IMPACT OF PELVIC NODAL IRRADIATION WITH INTENSITY-MODULATED RADIOTHERAPY ON TREATMENT OF PROSTATE CANCER ROBERT A. PRICE JR., PH.D.,* JEAN-MICHEL HANNOUN-LEVI, M.D.,‡ ERIC HORWITZ, M.D.,* MARK BUYYOUNOUSKI, M.D., M.S.,* KAREN J. RUTH, M.S.,† C.-M. MA, PH.D.,† AND ALAN POLLACK, M.D., PH.D.* *Department of Radiation Oncology, and †Biostatistics Facility, Fox Chase Cancer Center, Philadelphia, PA; ‡Centre Antoine Lacassagne, Nice, France Purpose: The aim of this study was to evaluate the feasibility of treating the pelvic lymphatic regions during prostate intensity-modulated radiotherapy (IMRT) with respect to our routine acceptance criteria. Methods and Materials: A series of 10 previously treated prostate patients were randomly selected and the pelvic lymphatic regions delineated on the fused magnetic resonance/computed tomography data sets. A targeting progression was formed from the prostate and proximal seminal vesicles only to the inclusion of all pelvic lymphatic regions and presacral region resulting in 5 planning scenarios of increasing geometric difficulty. IMRT plans were generated for each stage for two accelerator manufacturers. Dose volume histogram data were analyzed with respect to dose to the planning target volumes, rectum, bladder, bowel, and normal tissue. Analysis was performed for the number of segments required, monitor units, “hot spots,” and treatment time. Results: Both rectal endpoints were met for all targets. Bladder endpoints were not met and the bowel endpoint was met in 40% of cases with the inclusion of the extended and presacral lymphatics. A significant difference was found in the number of segments and monitor units with targeting progression and between accelerators, with the smaller beamlets yielding poorer results. Treatment times between the 2 linacs did not exhibit a clinically significant difference when compared. Conclusions: Many issues should be considered with pelvic lymphatic irradiation during IMRT delivery for prostate cancer including dose per fraction, normal structure dose/volume limits, planning target volumes generation, localization, treatment time, and increased radiation leakage. We would suggest that, at a minimum, the endpoints used in this work be evaluated before beginning IMRT pelvic nodal irradiation. © 2006 Elsevier Inc. Intensity-modulated radiotherapy, Pelvic lymphatics, Whole pelvic radiotherapy.
reduction of dose to the organs-at-risk (OAR) within the pelvis, including bowel, when using IMRT vs. conventional threedimensional conformal radiotherapy (3D–CRT) delivery. Sanguinetti et al. (3) also evaluated the effects on target coverage and OARs with emphasis on the rectum when pelvic lymphatics are included in the treatment of prostate cancer with conventional methods and IMRT. Ashman et al. (4) have recently made comparisons between patients treated with WPRT using 3D–CRT and IMRT with respect to observed clinical morbidity and dosimetric parameters. In this work we evaluate the dosimetric impact of lymphatic inclusion on our routine IMRT prostate plan acceptance criteria as well as treatment delivery time and leakage radiation.
INTRODUCTION Over the past decade or so we have experienced a relative decrease in radiotherapy treatment volume for prostate cancer. Target volumes are smaller as whole pelvic radiotherapy (WPRT) is used less and margins are smaller with the common use of daily localization techniques. The use of intensity-modulated radiotherapy (IMRT) for treatment (Tx) of these relatively small targets, surrounded by dose limiting structures, has become relatively routine. However, a randomized trial by Roach et al. (1) suggests that the irradiation of the pelvic lymphatic regions may be a benefit for a certain subset of men with prostate cancer. Critical structure sparing is a primary concern when treating prostate cancer and the addition of the nodal regions places larger volumes of bladder and bowel in the treatment area that may result in unwanted complications. The use of IMRT for these treatments appears obvious but presents additional challenges to be considered. Nutting et al. (2) had previously shown a
Ten prostate cancer patients with nonmetastatic lymph node negative disease were randomly selected for this study. All men
Reprint requests to: Robert A. Price, Jr., Ph.D., Department of Radiation Oncology, Fox Chase Cancer Center, 333 Cottman Avenue, Philadelphia, PA 19111. Tel: (215) 728-2819; Fax: (215)
728-4789; E-mail: [email protected] Received Jan 17, 2006, and in revised form May 25, 2006. Accepted for publication May 26, 2006.
METHODS AND MATERIALS
583
584
I. J. Radiation Oncology
● Biology ● Physics
Volume 66, Number 2, 2006
Impact of IMRT on prostate cancer
underwent noncontrast computed tomography (CT) (PQ5000, Philips Medical Systems, Cleveland, OH) and magnetic resonance imaging (MRI) (0.23T open MRI scanner, Philips Medical Systems, Cleveland, OH) simulation with a full bladder and empty rectum. The target and normal tissue volumes were defined on the fused MR/CT data sets. The prostate, proximal seminal vesicles, distal seminal vesicles, periprostatic/peri seminal vesicle lymph nodes, external iliac lymph nodes, proximal obturator, proximal internal iliac nodes, presacral/perirectal nodes were all outlined separately. The proximal and distal seminal vesicles are defined separately because the proximal seminal vesicles receive the full dose, whereas the distal seminal vesicles are given the same dose as the lymph nodes. The rectum was defined as the rectal wall and its contents from the anal verge to the sigmoid flexure superiorly. The bladder volume was defined as the entire bladder and its contents. Bowel was conservatively delineated for each patient to include all space that could potentially be occupied by bowel. The potential bowel space is the area between the pelvic nodal regions, beginning at the sigmoid flexure inferiorly from just above the rectum and extending superiorly to 1 cut above the most superior lymph nodes outlined. A more detailed description of the target and normal tissue volume definitions has been described elsewhere (5, 6). The clinical target volume (CTV) for each group was determined by using a combination of sub volumes (i.e., CTV1, CTV2, etc.). The CTV for Group 1 included the prostate and proximal seminal vesicles (CTV1). For Group 2, the distal seminal vesicles (CTV2) were also included. Group 3 further adds the periprostatic and peri-seminal vesicle lymph nodes (CTV3). Group 4 (extended lymphatics) includes the external iliac, proximal obturator, and proximal internal iliac nodes (CTV4). Group 5 is the most comprehensive treatment and also includes the presacral/perirectal lymph nodes (CTV5). Groups 2 to 5 represent a progression of potential treatment volumes for high-risk patients, with Group 5 comprehensively including potential lymph node metastasis sites—the presacral/presciatic lymph nodes to S3 and the perirectal lymph nodes below that to the level of the seminal vesicles (7–9). This progression, illustrated in Figs. 1a to 1e, resulted in 5 different planning scenarios of increasing geometric difficulty. Volume expansion to define PTV3 and PTV4 is somewhat problematic because prostate motion is independent of the lymph nodes; prostate motion corrected using transabdominal ultrasound may result in a shift of the dose distribution away from the lymph nodes, assuming the isocenter remains aligned with the bony anatomy. Thus, PTV3 and PTV4 should use a larger margin than PTV1. However, this would lead to compromise in the treatment of the prostate in terms of achieving high target doses and sparing of the bladder and rectum. We are using the same margins for all PTVs (8 mm everywhere and 5 mm posteriorly), although we have found that sometimes 6 mm lateral margins for PTV4 are necessary to limit bladder and bowel dose. Since lateral interfraction prostate displacement is typically small, this seems reasonable. Because of the generous definition of CTV5 and its extremely complex geometry, no additional margin was used for this targeting group in this study.
● R. A. PRICE et al.
585
Intensity-modulated radiotherapy plans were generated for each targeting group (associated with each PTV) with the goal of meeting our routine clinical prostate plan acceptance criteria (5). These criteria include 95% of the planning target volume (PTV95) receiving at least the prescription dose. The rectal volume receiving ⱖ65 Gy (R65) should not be more than 17% and the volume receiving ⱖ40 Gy (R40) should not be more than 35%. The bladder volume receiving ⱖ65 Gy (B65) should not be more than 25%, and the volume receiving ⱖ40 Gy (B40) should not be more than 50%. The 50% isodose line should fall within the rectal contour and the 90% isodose line should not exceed half the diameter of the rectum on any CT slice. Additionally, the distance between the posterior edge of the CTV and the prescription isodose line should be between 3 and 8 mm on each axial slice. A 40-Gy bowel limit of 150 cc was based on experience gained from pelvic irradiation for rectal cancer (10) as well as dose levels discussed in the treatment of anal and gynecologic diseases (11–15). This is likely to be conservative because of the lower dose per fraction to this region and the absence of concurrent chemotherapy. Plans were generated for delivery on 2 accelerators. The first accelerator was a 10 MV Siemens Primus (Siemens Medical Systems, Concord, CA) using a minimum beamlet size of 10 ⫻ 10 mm2. The second accelerator was a 10MV Varian Ex 21 (Varian Medical Systems, Palo Alto, CA) utilizing a minimum beamlet size of 10 ⫻ 5 mm2 with the collimator rotated to 90 degrees (16). By using the leaf width (5 mm) as the short side of the beamlet and rotating the collimator, we are matching the resolution of a 5 ⫻ 5 mm2 beamlet in the direction perpendicular to the prostate-rectum interface. In addition, using a step size of 10 mm results in fewer MLC positions and a faster treatment delivery. The increased area of the elongated beamlet results in increased output and fewer segments and monitor units (MU). This MU reduction also results in a decrease in the amount of accelerator head leakage. All plans were generated using the Corvus inverse planning system, version 5.0 (NOMOS Corp., Cranberry, PA) by the same experienced planner (R.P.) using 5 intensity levels per beam. The step-and-shoot delivery method was used throughout. Each plan was started using 6 beam directions progressing to 9 as needed. In addition, regions for dose constraint were used for all plans (17). This technique is based on the idea that all regions that are not designated as target, or some normal or critical structure, are combined as 1 region during the IMRT planning process. This region (termed “tissue” for this planning system) is controlled by a single set of dose constraints. However, the volume of tissue is far greater than that of the structures of interest. Furthermore, tissue is given the lowest priority during the optimization process. The addition of the dose-conforming regions allows more partial volume input data and provides extra control over the dose distribution in the regions of interest with respect to plans run without the regions. By designating subsets of tissue as concentric regions around the target(s) and defining each region’s dose constraints, an increased measure of control over the dose gradient outside the target boundaries is realized resulting in increased dose conformity and decreased treatment time.
Fig. 1. (a– e) Illustration of anatomy, that with the associated planning target volumes (PTVs), make up the targeting groups (TG) used in this study. (a) (TG1) prostate and proximal seminal vesicles (SVs). (b) (TG2) TG1 and the addition of the distal seminal vesicles. (c) (TG3) TG2 and the addition of the periprostatic and peri-seminal vesicle lymph nodes. (d) (TG4) TG3 and the addition of the external iliac, proximal obturator and proximal internal iliac lymph nodes. (e) (TG5) TG4 and the addition of the presacral lymph nodes.
586
I. J. Radiation Oncology
● Biology ● Physics
Dose volume histogram (DVH) data were analyzed with respect to dose to the PTVs, rectum, bladder, bowel, and normal tissue. Additional analysis was performed for the number of segments required, MU, “hot spots” (i.e., unacceptable high-dose regions), and treatment time. Normal distributions of the endpoints were evaluated with the Shapiro-Wilk test of normality. For the 4 more complicated targeting groups, comparisons were made to Group 1 by paired t tests (normally distributed variables) or by Wilcoxon signed rank tests (non-normally distributed variables). The McNemar Exact Test (MET) was used to evaluate endpoints with respect to meeting or failing each criterion. This test is based on the number of patients that meet the criterion on 1 plan and fail on another and uses exact binomial probabilities because of the small sample size.
RESULTS All plans were normalized such that 95% of the prostate and proximal seminal vesicle PTV received 76 Gy. The remaining PTVs for each group received at least 56 Gy to 95% of their volumes. Table 1 contains the results for all endpoints studied resulting from plans generated for the Siemens Primus unit (10 ⫻ 10 mm2 minimum beamlet
Volume 66, Number 2, 2006
size). Included are the mean values for all 10 patients and all 5 targeting groups as well as the respective ranges and statistical significance (p-values). All endpoints except the bowel volume variables were normally distributed. Statistical significance was calculated for all groups with respect to Group 1. Since the conformity index is not defined for multiple co-joined targets with differing prescription doses and the fact that our target volume(s) changes with targeting progression, we have decided to evaluate the volume of normal tissue (NT) receiving some reference dose as a function of targeting progression. The reference dose used was our 56 Gy (NT56) nodal target prescription dose. Table 2 shows the results for all endpoints studied resulting from plans generated for the Varian Ex 21 unit (10 ⫻ 5 mm2 minimum beamlet size). Values for p and NT are defined as in Table 1. Table 3 contains the results from a head-to-head comparison for plans generated for the 2 accelerators. In this table, statistical significance is calculated between the results for the 2 linacs within each targeting group. Figure 2a illustrates the rectal results for all patients when
Table 1. Average results from plans generated with a minimum beamlet size of 10 ⫻ 10 mm2 10 ⫻ 10 mm2
Group 1
Group 2
Group 3
Group 4
Group 5
R65 (%) Range (%) p R40 (%) Range (%) p B65 (%) Range (%) p B40 (%) Range (%) p Bowel-65 (cc) Range (cc) p Bowel-40 (cc) Range (cc) p MU Range p Segments Range p Tx time (min) Range (min) p Max dose (%) Range (%) p NT56 (cc) Range (cc) p
11.9 7.0–16.6
11.8 7.4–15.8 0.791 28.8 19.8–36.3 0.071 21.6 7.3–39.4 0.018 39.8 17.6–75.4 0.126 0.7 0.0–2.4 0.250 9.2 0.0–36.3 0.016 1047 873–1286 0.170 55 43–79 0.788 9.1 7.1–13.0
11.6 7.6–16.2 0.640 29.0 19.7–36.0 0.039 23.8 10.8–50.3 0.017 47.8 21.4–84.9 ⬍0.005 1.3 0.0–7.3 0.063 14.0 0.0–68.8 ⬍0.005 1205 1053–1333 ⬍0.005 76 56–90 ⬍0.005 12.6 9.2–14.9
11.3 8.1–15.0 0.321 29.8 22.9–36.6 0.011 27.8 14.2–43.0 ⬍0.005 70.9 61.6–85.9 ⬍0.005 7.3 0.4–27.9 ⬍0.005 210.8 48.4–382.8 ⬍0.005 1521 1296–1743 ⬍0.005 117 89–136 ⬍0.005 19.3 14.7–22.4
10.8 6.4–14.7 0.082 35.4 21.0–43.7 ⬍0.005 30.2 20.1–49.2 ⬍0.005 72.0 59.1–81.0 ⬍0.005 9.1 0.4–62.9 ⬍0.005 209.3 57.7–358.5 ⬍0.005 1701 1533–1935 ⬍0.005 145 130–163 ⬍0.005 23.9 21.5–26.9
0.788 118.5 117.5–119.4 0.986 356 256–449 0.122
⬍0.005 118.7 116.7–120.4 0.645 409 284–509 ⬍0.005
⬍0.005 119.0 115.9–122.1 ⬍0.005 780 612–980 ⬍0.005
⬍0.005 120.2 117.8–121.4 ⬍0.005 920 679–1170 ⬍0.005
27.1 19.5–33.9 19.9 7.7–35.8 36.8 15.9–63.0 0.3 0.0–2.5 3.9 0.0–16.3 1013 855–1210 54 40–78 9.0 6.6–12.9 118.1 114.9–119.9 337 253–447
Abbreviations: Max ⫽ maximum; MU ⫽ monitor units; Tx ⫽ treatment. For reference, primary limits are as follows: R65 ⱕ17%, R40 ⱕ35%, B65 ⱕ25%, B40 ⱕ50%, Bowel40 ⱕ150 cc.
Impact of IMRT on prostate cancer
● R. A. PRICE et al.
587
Table 2. Average results from plans generated with a minimum beamlet size of 10 ⫻ 5 mm2 10 ⫻ 5 mm2
Group 1
Group 2
Group 3
Group 4
Group 5
R65 (%) Range (%) p R40 (%) Range (%) p B65 (%) Range (%) p B40 (%) Range (%) p Bowel-65 (cc) Range (cc) p Bowel-40 (cc) Range (cc) p MU Range p Segments Range p Tx time (min) Range (min) p Max dose (%) Range (%) p NT56 (cc) Range (cc) p
10.3 6.3–15.2
10.6 6.8–14.4 0.285 24.5 15.9–34.0 0.073 20.0 8.2–37.4 0.122 38.9 17.9–73.7 0.012 0.7 0.0–2.7 0.031 8.3 0.0–34.3 0.016 1516 1248–1836 0.986 203 141–291 0.338 10.4 8.2–13.7
10.3 6.1–14.4 0.841 25.1 15.6–34.8 0.031 23.1 10.4–45.7 ⬍0.005 47.8 21.7–86.1 ⬍0.005 0.9 0.0–3.2 0.063 13.6 0.0–63.0 0.016 1484 846–1870 0.645 262 174–351 ⬍0.005 12.2 9.4–15.3
10.4 6.3–14.0 0.588 26.1 17.0–33.4 ⬍0.005 26.9 15.3–40.5 ⬍0.005 67.7 58.8–81.8 ⬍0.005 4.4 0.0–20.1 0.008 199.1 36.2–340.2 ⬍0.005 2314 2138–2467 ⬍0.005 628 489–737 ⬍0.005 21.3 12.2–25.8
10.2 6.4–13.2 0.799 30.5 18.3–37.9 ⬍0.005 29.2 17.3–47.6 ⬍0.005 70.4 61.3–83.9 ⬍0.005 5.9 0.0–35.1 ⬍0.005 197.1 49.2–322.6 ⬍0.005 2252 2072–2340 ⬍0.005 725 640–792 ⬍0.005 25.4 22.4–27.7
0.267 116.9 114.3–118.9 0.886 331 224–421 0.021
0.007 117.0 114.4–119.5 0.982 381 266–485 ⬍0.005
⬍0.005 117.5 114.5–121.8 0.554 735 573–892 ⬍0.005
⬍0.005 119.9 117.2–126.3 0.019 859 631–1071 ⬍0.005
23.2 14.2–30.8 19.1 7.3–36.1 35.1 13.8–64.7 0.3 0.0–1.5 3.3 0.0–15.5 1515 1309–1781 189 137–324 9.8 7.9–11.5 117.0 114.7–119.5 316 216–417
Abbreviations as in Table 1. For reference, primary limits are as follows: R65 ⱕ17%, R40 ⱕ35%, B65 ⱕ25%, B40 ⱕ50%, Bowe140 ⱕ150 cc.
prescribing to targeting Group 4 (extended lymphatics). It can be seen that our clinical criteria are met in almost all cases. This is to be expected in that the lymphatic regions involved are not directly adjacent to the rectum. It is seen that the smaller leaf width MLC plans, particularly for the R40 endpoint, result in lower percentages of rectum being irradiated indicating increased dose conformity. The bladder endpoints for all patients, illustrated in Fig. 2b, are not met in the majority of cases. The geometry of targeting Group 4 with the extended lymphatics bounding the bladder laterally, and the associated PTVs, make it impractical to limit the bladder to our clinical criteria in all cases. The paired t-test evaluates statistical significance by comparing the mean values between groups for a given endpoint. We also evaluated endpoints on a case-by-case basis with respect to meeting or failing to meet our criteria using McNemar’s Exact Test (MET). This test indicates a highly significant trend in failing to meet the B40 criteria with the addition of the extended lymphatics (p ⫽ 0.008). Attempts to increase the bladder constraints further during optimization resulted in the unacceptable high-dose regions known
as hot spots. It should be noted that the values indicated for hot spots represent a single point in the dose matrix defined by voxels with sides approximately 0.9 mm in length. If we evaluate the dose distribution for clinically meaningful high-dose regions, at least 2 cm2 on an individual CT slice for example, the listed values decrease on average by approximately 5%. For example, an indicated hot spot of 118% would typically represent a clinically significant hot spot of 113%. These hot spots are always found within the prostate PTV. Given that there is some overlap between the prostate PTV and critical structures efforts are made to ensure that the high-dose regions do not fall in these overlap areas if possible. The affects on our rectal endpoints with the addition of the presacral region is illustrated in Fig. 2c. The high-dose limit is consistently met. The low-dose limit is not met in the majority of cases (MET p ⫽ 0.03 for the 10 ⫻ 10 mm2 beamlets; the smaller beamlets showed no significant trend). The addition of the presacral region, bounding a portion of the rectum laterally and posteriorly, results in a tube-like dose distribution. Maintaining the R40 limit was not always
588
I. J. Radiation Oncology
● Biology ● Physics
Volume 66, Number 2, 2006
Table 3. Results of head-to-head comparison of plans generated for the two different minimum beamlet sizes and accelerator manufacturers
R65 (10 ⫻ 10) % R65 (10 ⫻ 5) % p R40 (10 ⫻ 10) % R40 (10 ⫻ 5) % p B65 (10 ⫻ 10) % B65 (10 ⫻ 5) % p B40 (10 ⫻ 10) % B40 (10 ⫻ 5) % p Bowel-65 (10 ⫻ 10) cc Bowel-65 (10 ⫻ 5) cc p Bowel-40 (10 ⫻ 10) cc Bowel-40 (10 ⫻ 5) cc p MU (10 ⫻ 10) MU (10 ⫻ 5) p Segments (10 ⫻ 10) Segments (10 ⫻ 5) p Tx time (min) (10 ⫻ 10) Tx time (min) (10 ⫻ 5) p Max% (10 ⫻ 10) Max% (10 ⫻ 5) p NT56 (cc) (10 ⫻ 10) NT56 (cc) (10 ⫻ 5) p
Group 1
Group 2
Group 3
Group 4
Group 5
11.9 10.3 0.004 27.1 23.2 ⬍0.005 19.9 19.1 0.113 36.8 35.1 0.020 0.3 0.3 1.000 3.9 3.3 0.125 1013 1515 ⬍0.005 54 189 ⬍0.005 9.0 9.8 0.122 118.1 117.0 0.103 337 316 ⬍0.005
11.8 10.6 0.005 28.8 24.5 ⬍0.005 21.6 20.0 0.078 39.8 38.9 0.071 0.7 0.7 0.750 9.2 8.3 0.102 1047 1516 ⬍0.005 55 203 ⬍0.005 9.1 10.4 0.041 118.5 116.9 0.006 356 331 ⬍0.005
11.6 10.3 0.005 29.0 25.1 ⬍0.005 23.8 23.1 0.144 47.8 47.8 0.950 1.3 0.9 0.688 14.0 13.6 0.461 1205 1484 0.007 76 262 ⬍0.005 12.6 12.2 0.697 118.7 117.0 ⬍0.005 409 381 ⬍0.005
11.3 10.4 0.006 29.8 26.1 ⬍0.005 27.8 26.9 0.230 70.9 67.7 0.004 7.3 4.4 ⬍0.005 210.8 199.1 ⬍0.005 1521 2314 ⬍0.005 117 628 ⬍0.005 19.3 21.3 0.031 119.0 117.5 ⬍0.005 780 735 ⬍0.005
10.8 10.2 0.016 35.4 30.5 ⬍0.005 30.2 29.2 0.225 72.0 70.4 0.104 9.1 5.9 0.232 209.3 197.1 ⬍0.005 1701 2252 ⬍0.005 145 725 ⬍0.005 23.9 25.4 ⬍0.005 120.2 119.9 0.621 920 859 ⬍0.005
Abbreviations as in Table 1. For reference, primary limits are as follows: R65 ⱕ17%, R40 ⱕ35%, B65 ⱕ25%, B40 ⱕ50%, Bowel40 ⱕ150 cc.
possible. As expected the bladder endpoints are not met in the majority of cases as illustrated in Fig. 2d (MET for B40, p ⫽ 0.008 for both linacs). The plans that met the B65 endpoints are consistent with those for targeting Group 4 indicating the major contributing factor to increased bladder dose being the addition of the extended lymphatics. Figure 3 illustrates the results for all plans with respect to our bowel criteria. On average we were unable to meet the 40 Gy cut point with the addition of targeting Groups 4 and 5. Additionally these groups exhibit statistical significance with respect to failing this endpoint on a case-by-case basis (MET p ⫽ 0.016, both groups, both linacs). There is little change in the results between plans for these groups indicating the addition of the extended lymphatics as being the major contributor to increased dose to bowel. It is of note that the smaller MLC leaf widths resulted in a decreased volume of bowel being irradiated indicating increased target conformity. We have chosen to treat all targets, regardless of the intended dose, with a single IMRT plan. The prostate prescription of 76 Gy delivered in 38 fractions assures approx-
imately 2.0 Gy per fraction is being delivered. However, the prescription dose for the distal seminal vesicles and all nodal groups is 56 Gy in 38 fractions resulting in a gradient between approximately 1.5 Gy and 2.0 Gy per fraction. This prescription was derived with the aid of biologically equivalent dose calculations and is equivalent to 49 to 50 Gy delivered at 2.0 Gy per fraction. An alternative approach would be to use a series of “cone down” IMRT plans to maintain 2.0 Gy per fraction for all targets. We do not believe the dosimetric results will differ from those found in this study. DISCUSSION In this report, the dosimetric impact of pelvic lymphatic irradiation for prostate cancer was studied in terms of rectum, bladder, and bowel sparing for various nodal volumes. For all organs at risk, the nodal volume used was a determinant in fulfilling the specific DVH criteria. Although PTV coverage was assured through our normalization procedure, we found a significant increase in
Impact of IMRT on prostate cancer
● R. A. PRICE et al.
589
Fig. 2. (a) Rectal endpoint results for all patients with the inclusion of the extended lymphatics (TG4). The 40- and 65-Gy limits are delineated as well as the results for both multileaf collimator (MLC) sizes (v indicates 5 mm leaf widths and s indicates 10-mm leaf widths). (b) Bladder endpoint results for all patients with the inclusion of the extended lymphatics (TG4). The 40- and 65-Gy limits are delineated as well as the results for both MLC sizes (v indicates 5-mm leaf widths and s indicates 10-mm leaf widths). (c) Rectal endpoint results for all patients with the inclusion of the presacral lymphatics (TG5). The 40 and 65 Gy limits are delineated as well as the results for both MLC sizes (v indicates 5-mm leaf widths and s indicates 10-mm leaf widths). (d) Bladder endpoint results for all patients with the inclusion of the presacral lymphatics (TG5). The 40- and 65-Gy limits are delineated as well as the results for both MLC sizes (v indicates 5-mm leaf widths and s indicates 10-mm leaf widths).
dose to normal tissues and an increase in the maximum dose delivered with increasing nodal volume. These effects were decreased for plans generated for delivery with 5-mm leaf widths. While the difference in treatment delivery time between the 2 linacs studied was not believed to be clinically significant, the use of the smaller leaf MLCs (specifically the smaller minimum beamlet size) resulted in a significant increase in MU. This MU increase yields increased head leakage and should be evaluated with respect to vault shielding. Rectum It can be seen from Tables 1 and 2 that the R65 criteria can be met for all targeting groups; this is believed to be
essential, as this limit is set based on the probability of increased rectal complications (18). However, caution should be exercised with the introduction of the presacral lymphatics. At Fox Chase Cancer Center, the patient is simulated with the rectum empty, which allows for a maximum diameter of 3 cm on any individual CT slice. The length of the rectum is defined from the recto-sigmoid junction to the bottom of the ischial tuberosities including the wall and contents. Our experience is that the rectum is relatively empty after 2 weeks and planning under this condition is conservative. For example, if our DVH acceptance criteria are met with the rectum empty, by forcing a high (compressed) dose gradient across the rectal volume, we should be able to assure rectal sparing despite rectal
590
I. J. Radiation Oncology
● Biology ● Physics
Volume 66, Number 2, 2006
understood. The changing volume and shape of the bladder throughout the course of treatment will continue to complicate defining a dose/volume relationship with bladder complications. The bladder criteria were not met with the addition of the extended lymphatics or the presacral region (Tables 1 and 2). A statistically significant trend toward larger volumes receiving the dose limits is observed for these groups. Attempts to place stricter planning constraints on the bladder resulted in unacceptable hot spot occurrence. There was no significant difference found between plans generated for the 2 accelerators for the bladder 65 Gy cutpoint. No trend was found with targeting progression for B40 although a significant difference was found for Groups 1 and 4 with the smaller beamlet plans resulting in lower values. Fig. 3. Bowel-40 endpoint results for all patients and all targeting groups. The 150-cc limit is delineated as well as the results for both MLC sizes (minimum beamlet sizes are listed and correspond to linac vendor).
filling throughout treatment with the help of daily localization. On the other hand, rectal sparing may not be assured if the pre-sacral lymphatics are also treated. If the dose distribution is very conformal to the rectum, rectal filling during the course of treatment may result in unanticipated high dose being delivered to unintended volume(s) of rectum. Figures 4a to 4d illustrate an example of the relationship between the high-dose region and the rectum for target Groups 4 and 5. It is clear from these images that the addition of the presacral lymphatics increases the uncertainty in rectal sparing with respect to rectal filling during treatment. A strategy that may prove helpful is to add a margin to the posterior and lateral aspects of the rectum, forming a planning organ at risk volume (PRV) to account for this potential rectal filling and potentially keep the high-dose region away from the actual rectum. The R40 limit is somewhat arbitrary and is used to control the shape of the DVH. On average all plans met this criterion with a statistically significant trend toward larger volumes receiving 40 Gy with increased targeting from Groups 3 to 5. The head-to-head test between the 2 accelerators was made to evaluate the strengths and weaknesses of the different leaf widths and different minimum beamlet sizes. A significant difference was found for all targeting groups with the 10 ⫻ 5 mm2 beamlet plans resulting in lower values for both rectal end points. This may be attributed to the higher dose gradient attainable at the prostate-rectal interface with this beamlet configuration. Bladder A similar nodal volume dependency was observed for bladder sparing. Although B65 and B40 limits were set arbitrarily (5), as we have no data corresponding to increased complications at this time, they are useful for comparing bladder sparing between the various nodal volumes. The impact of treating larger bladder volumes is not well
Bowel On average the amount of bowel receiving 65 Gy was between 0.3 cc and 9.1 cc with no trends as a function of targeting progression. The 40 Gy criterion was not met in 60% of the cases with the addition of targeting Groups 4 and 5. Attempts to place increased planning constraints on the bowel led to increased bladder dose and unacceptable hot spots. Failure to limit the bowel receiving ⱖ40 Gy for these targeting groups is primarily attributed to the addition of the extended lymphatic regions and their associated PTV. Evaluation of the bowel end-points indicate that a significant difference is evident for Groups 4 and 5 with the smaller beamlet plans yielding lower values than Bowel-40. Although the smaller beamlets resulted in lower values in all cases, we should reiterate that this endpoint criterion was not met on average. Planning target volume The PTV needed for the uncertainty in position of the extended lymphatic regions should be similar to that of the prostate. The prostate and seminal vesicles move within the pelvis with respect to the bony structures while the extended lymphatic regions remain relatively fixed. We assign a PTV to the prostate and seminal vesicles in part because of this localization uncertainty and we actively localize daily with ultrasound. If our PTV for the extended lymphatic regions is inadequate and we move the prostate and seminal vesicles into the appropriate high-dose region using ultrasound, we may experience a geographic miss of the extended lymphatic regions because of their relative immobility. It is the medial aspect of the extended lymphatics PTV that overlaps with the bowel, resulting in increased dose. An alternative approach that may be successful would be to limit the PTV growth in the medial–lateral dimensions but maintain the current growth in the other directions. The medial–lateral dimension exhibits the least prostate motion (19). Subsequent daily localization shifts would need to be limited in the lateral dimension as well. The prostate and seminal vesicles could still be encompassed by their respective PTVs, and a limited amount of bowel lies in the anterior/ posterior and cranio/caudal directions with respect to the extended lymphatics.
Impact of IMRT on prostate cancer
● R. A. PRICE et al.
591
Fig. 4. (a, b) Axial and sagittal dose distributions (Patient Index #2) through the prostate region representing planning for TG4. The prescription isodose lines, 76 Gy and 56 Gy, as well as 38 Gy (50%) are shown. (c, d) Axial and sagittal dose distributions (Patient Index #2) through the prostate region representing planning for TG5, the addition of the presacral lymphatics. The prescription isodose lines, 76 Gy and 56 Gy, as well as 38 Gy (50%) are shown. Note the tube-like dose distrubution surrounding the rectum.
The maximum hot spot in the treatment plans did not vary significantly with target progression with the exception of the addition of the presacral region. Even though this difference was statistically significant, it is probably not clinically significant. Attempts to meet the set criteria for bladder and bowel tended to result in unacceptable hot spots with increase in targeting progression. There is a statistically significant increase in NT56 with targeting progression, beginning with targeting Group 3. This is to be expected, as our target volume is increasing. However, the amount of normal, nontarget tissue receiving 56 Gy approximately triples between Groups 1 and 5. This increase and the potential adverse implications should be evaluated with respect to the benefits gained from nodal irradiation. The head-to-head evaluation indicated a statistically significant decrease in maximum hot spot for targeting Groups 2, 3 and 4, although the clinical significance is arguable. There was a statistically significant decrease in NT56 for all targeting groups when using the smaller beamlets, hinting at increased conformity. However, it is again arguable if these decreases are clinically significant.
Monitor units A statistically significant trend in increasing MU with targeting progression is observed. Although this may be expected because of the relatively harsh dose limits of the normal structures of interest, the increasing target volume with targeting progression, and the complexity of the resultant intensity maps, this increase should be evaluated with respect to increased radiation leakage, vault shielding, and patient safety. This can be done by using variations of the modulation scaling factor (20). A corresponding statistically significant trend in increasing number of segments and subsequent treatment delivery time is associated with increasing targeting progression beginning with the introduction of the Group 3 lymphatic region. On average, treatment time more than doubles from delivery of Group 1 to Group 4. This increased treatment time has the obvious adverse effect on patient throughput but perhaps more importantly on patient comfort. Maintaining a full bladder for 20 min vs. 10 min may jeopardize patient immobilization and subsequent treatment accuracy. Although some dosimetric improvements are evident
592
I. J. Radiation Oncology
● Biology ● Physics
for the smaller beamlet plans a statistically significant increase in MU was found for all targeting groups when using this configuration. This is attributable in large part because of the decrease in output for the smaller field sizes. It is important to evaluate the effects on existing vault shielding this increase combined with IMRT patient load will have. A significant increase in the number of individual beam segments generated was also found with the smaller beamlet plans. This corresponded to an increase in treatment time in general with statistical significance evident for targeting Groups 2, 4, and 5. The increase in treatment time for Groups 4 and 5 are also affected by the need for an MLC carriage shift. Although statistically significant these increases may not be clinically significant.
Volume 66, Number 2, 2006
CONCLUSION The re-introduction of pelvic lymphatic irradiation during IMRT delivery for prostate cancer is increasing in popularity. However, a host of areas of concern are also introduced including dose per fraction, normal structure dose/volume limits, PTV generation, localization, treatment time, and increased radiation leakage and the effects on vault shielding. The feasiblility of pelvic lymphatic irradiation was performed by evaluating the clinical endpoints used at this treatment center. Although some of our acceptance criteria may be too strict (e.g., R65 and Bowel 40), others may be too loose (e.g., hot spots). More definitive answers with respect to endpoints will become evident with the completion of additional trials and research. In the interim we would suggest that, at a minimum, the endpoints used in this work be evaluated before beginning IMRT pelvic nodal irradiation.
REFERENCES 1. Roach M III, DeSilvio M, Lawton C, et al. Phase III trial comparing whole-pelvis versus prostate-only radiotherapy and neoadjuvant versus adjuvant combined androgen suppression: Radiotherapy Oncology Group 9413. JCO 2003;21:1904 – 1911. 2. Nutting CM, Convery DJ, Cosgrove VP, et al. Reduction of small and large bowel irradiation using an optimized intensitymodulated pelvic radiotherapy technique in patients with prostate cancer. Int J Radiat Oncol Biol Phys 2000;48:649 – 656. 3. Sanguinetti G, Cavey ML, Endres EJ, et al. Is IMRT needed to spare the rectum when pelvic lymph nodes are part of the initial treatment volume for prostate cancer? Int J Radiat Oncol Biol Phys 2006;64:151–160. 4. Ashman JB, Zelefsky MJ, Hunt MS, et al. Whole pelvic radiotherapy for prostate cancer using 3D conformal and intensity-modulated radiotherapy. Int J Radiat Oncol Biol Phys 2005;63:765–771. 5. Pollack A, Hanlon AL, Horwitz EM, et al. Dosimetry and preliminary acute toxicity in the first 100 men treated for prostate cancer on a randomized hypofractionation dose escalation trial. Int J Radiat Oncol Biol Phys 2006;64:518 –526. 6. Buyyounouski MK, Horwitz EM, Price RA Jr, et al. Prostate IMRT. In: Bortfield T, Rupert Schmidt-Ullrich R, de Neve W, Wazer DE, editors. IMRT handbook: concepts & clinical applications. Heidelberg: Springer-Verlag; 2006. p. 391– 410. 7. Shih HA, Harisinghani M, Zietman AL, et al. Mapping of nodal disease in locally advanced prostate cancer: Rethinking the clinical target volume for pelvic nodal irradiation based on vascular rather than bony anatomy. Int J Radiat Oncol Biol Phys 2005;63:1262–1269. 8. Ganswindt U, Paulsen F, Corvin S, et al. Intensity-modulated radiotherapy for high risk prostate cancer based on sentinel node SPECT imaging for target volume definition. BMC Cancer 2005;5:91. 9. Murray SK, Breau RH, Guha AK, Gupta R. Spread of prostate carcinoma to the perirectal lymph node basin: Analysis of 112 rectal resections over a 10-year span for primary rectal adenocarcinoma. Am J Surg Pathol 2004;28:1154 –1162. 10. Baglan KL, Frazier RC, Yan D, et al. The dose-volume relationship of acute small bowel toxicity from concurrent
11.
12.
13.
14. 15.
16.
17.
18.
19. 20.
5-FU-based chemotherapy and radiotherapy for rectal cancer. Int J Radiat Oncol Biol Phys 2002;52L:176 –183. Roeske JC, Lujan A, Rotmensch J, et al. Intensity-modulated whole pelvic irradiotherapy in patients with gynecologic malignancies. Intern J Radiat Oncol Biol Phys 2000;48:1613– 1621. Portelance L, Chao KSC, Grigsby PW, et al. Intensity-modulated radiotherapy (IMRT) reduces small bowel, rectum, and bladder doses in patients with cervical cancer receiving pelvic and para-aortic irradiation. Int J Radiat Oncol Biol Phys 2001;51:261–266. Adli M, Mayr NA, Kaiser HS, et al. Does prone positioning reduce small bowel dose in pelvic radiation with intensitymodulated radiotherapy for gynecological cancer? Int J Radiat Oncol Biol Phys 2003;57:230 –238. Mundt AJ, Lujan AE, Rotmensch J, et al. Intensity-modulated whole pelvic radiotherapy in women with gynecologic malignancies. Int J Radiat Oncol Biol Phys 2005;63:354 –361. Milano MT, Jani AB, Farrey KJ, et al. Intensity-modulated radiotherapy (IMRT) in the treatment of anal cancer: Toxicity and clinical outcome. Int J Radiat Oncol Biol Phys 2003;57: 230 –238. Price RA Jr, Paskalev K, McNeeley SW, Ma C-M. Elongated beamlets: A simple technique for segment and MU reduction for sMLC IMRT delivery on accelerators utilizing 5 mm leaf widths. Phys Med Biol 2005;50:N235–N242. Price RA Jr, Murphy S, McNeeley SW, et al. A method for increased dose conformity and segment reduction for sMLC delivered IMRT treatment of the prostate. Int J Radiat Oncol Biol Phys 2003;57:843– 852. Storey MR, Pollack A, Levy L, et al. Complications from radiotherapy dose escalation in prostate cancer: Preliminary results of a randomized trial. Int J Radiat Oncol Biol Phys 2000;48:635– 642. Chandra A, Dong L, Huang E, et al. Experience of ultrasoundbased daily prostate localization. Int J Radiat Oncol Biol Phys 2003;56:436 – 447. Price RA Jr, Ma CM, Chibani O. Shielding evaluation for IMRT implementation in an existing accelerator vault. J Appl Clin Med Phys 2003:231–238.
doi:10.1016/j.ijrobp.2006.05.033
PHYSICS CONTRIBUTION
IMPACT OF PELVIC NODAL IRRADIATION WITH INTENSITY-MODULATED RADIOTHERAPY ON TREATMENT OF PROSTATE CANCER ROBERT A. PRICE JR., PH.D.,* JEAN-MICHEL HANNOUN-LEVI, M.D.,‡ ERIC HORWITZ, M.D.,* MARK BUYYOUNOUSKI, M.D., M.S.,* KAREN J. RUTH, M.S.,† C.-M. MA, PH.D.,† AND ALAN POLLACK, M.D., PH.D.* *Department of Radiation Oncology, and †Biostatistics Facility, Fox Chase Cancer Center, Philadelphia, PA; ‡Centre Antoine Lacassagne, Nice, France Purpose: The aim of this study was to evaluate the feasibility of treating the pelvic lymphatic regions during prostate intensity-modulated radiotherapy (IMRT) with respect to our routine acceptance criteria. Methods and Materials: A series of 10 previously treated prostate patients were randomly selected and the pelvic lymphatic regions delineated on the fused magnetic resonance/computed tomography data sets. A targeting progression was formed from the prostate and proximal seminal vesicles only to the inclusion of all pelvic lymphatic regions and presacral region resulting in 5 planning scenarios of increasing geometric difficulty. IMRT plans were generated for each stage for two accelerator manufacturers. Dose volume histogram data were analyzed with respect to dose to the planning target volumes, rectum, bladder, bowel, and normal tissue. Analysis was performed for the number of segments required, monitor units, “hot spots,” and treatment time. Results: Both rectal endpoints were met for all targets. Bladder endpoints were not met and the bowel endpoint was met in 40% of cases with the inclusion of the extended and presacral lymphatics. A significant difference was found in the number of segments and monitor units with targeting progression and between accelerators, with the smaller beamlets yielding poorer results. Treatment times between the 2 linacs did not exhibit a clinically significant difference when compared. Conclusions: Many issues should be considered with pelvic lymphatic irradiation during IMRT delivery for prostate cancer including dose per fraction, normal structure dose/volume limits, planning target volumes generation, localization, treatment time, and increased radiation leakage. We would suggest that, at a minimum, the endpoints used in this work be evaluated before beginning IMRT pelvic nodal irradiation. © 2006 Elsevier Inc. Intensity-modulated radiotherapy, Pelvic lymphatics, Whole pelvic radiotherapy.
reduction of dose to the organs-at-risk (OAR) within the pelvis, including bowel, when using IMRT vs. conventional threedimensional conformal radiotherapy (3D–CRT) delivery. Sanguinetti et al. (3) also evaluated the effects on target coverage and OARs with emphasis on the rectum when pelvic lymphatics are included in the treatment of prostate cancer with conventional methods and IMRT. Ashman et al. (4) have recently made comparisons between patients treated with WPRT using 3D–CRT and IMRT with respect to observed clinical morbidity and dosimetric parameters. In this work we evaluate the dosimetric impact of lymphatic inclusion on our routine IMRT prostate plan acceptance criteria as well as treatment delivery time and leakage radiation.
INTRODUCTION Over the past decade or so we have experienced a relative decrease in radiotherapy treatment volume for prostate cancer. Target volumes are smaller as whole pelvic radiotherapy (WPRT) is used less and margins are smaller with the common use of daily localization techniques. The use of intensity-modulated radiotherapy (IMRT) for treatment (Tx) of these relatively small targets, surrounded by dose limiting structures, has become relatively routine. However, a randomized trial by Roach et al. (1) suggests that the irradiation of the pelvic lymphatic regions may be a benefit for a certain subset of men with prostate cancer. Critical structure sparing is a primary concern when treating prostate cancer and the addition of the nodal regions places larger volumes of bladder and bowel in the treatment area that may result in unwanted complications. The use of IMRT for these treatments appears obvious but presents additional challenges to be considered. Nutting et al. (2) had previously shown a
Ten prostate cancer patients with nonmetastatic lymph node negative disease were randomly selected for this study. All men
Reprint requests to: Robert A. Price, Jr., Ph.D., Department of Radiation Oncology, Fox Chase Cancer Center, 333 Cottman Avenue, Philadelphia, PA 19111. Tel: (215) 728-2819; Fax: (215)
728-4789; E-mail: [email protected] Received Jan 17, 2006, and in revised form May 25, 2006. Accepted for publication May 26, 2006.
METHODS AND MATERIALS
583
584
I. J. Radiation Oncology
● Biology ● Physics
Volume 66, Number 2, 2006
Impact of IMRT on prostate cancer
underwent noncontrast computed tomography (CT) (PQ5000, Philips Medical Systems, Cleveland, OH) and magnetic resonance imaging (MRI) (0.23T open MRI scanner, Philips Medical Systems, Cleveland, OH) simulation with a full bladder and empty rectum. The target and normal tissue volumes were defined on the fused MR/CT data sets. The prostate, proximal seminal vesicles, distal seminal vesicles, periprostatic/peri seminal vesicle lymph nodes, external iliac lymph nodes, proximal obturator, proximal internal iliac nodes, presacral/perirectal nodes were all outlined separately. The proximal and distal seminal vesicles are defined separately because the proximal seminal vesicles receive the full dose, whereas the distal seminal vesicles are given the same dose as the lymph nodes. The rectum was defined as the rectal wall and its contents from the anal verge to the sigmoid flexure superiorly. The bladder volume was defined as the entire bladder and its contents. Bowel was conservatively delineated for each patient to include all space that could potentially be occupied by bowel. The potential bowel space is the area between the pelvic nodal regions, beginning at the sigmoid flexure inferiorly from just above the rectum and extending superiorly to 1 cut above the most superior lymph nodes outlined. A more detailed description of the target and normal tissue volume definitions has been described elsewhere (5, 6). The clinical target volume (CTV) for each group was determined by using a combination of sub volumes (i.e., CTV1, CTV2, etc.). The CTV for Group 1 included the prostate and proximal seminal vesicles (CTV1). For Group 2, the distal seminal vesicles (CTV2) were also included. Group 3 further adds the periprostatic and peri-seminal vesicle lymph nodes (CTV3). Group 4 (extended lymphatics) includes the external iliac, proximal obturator, and proximal internal iliac nodes (CTV4). Group 5 is the most comprehensive treatment and also includes the presacral/perirectal lymph nodes (CTV5). Groups 2 to 5 represent a progression of potential treatment volumes for high-risk patients, with Group 5 comprehensively including potential lymph node metastasis sites—the presacral/presciatic lymph nodes to S3 and the perirectal lymph nodes below that to the level of the seminal vesicles (7–9). This progression, illustrated in Figs. 1a to 1e, resulted in 5 different planning scenarios of increasing geometric difficulty. Volume expansion to define PTV3 and PTV4 is somewhat problematic because prostate motion is independent of the lymph nodes; prostate motion corrected using transabdominal ultrasound may result in a shift of the dose distribution away from the lymph nodes, assuming the isocenter remains aligned with the bony anatomy. Thus, PTV3 and PTV4 should use a larger margin than PTV1. However, this would lead to compromise in the treatment of the prostate in terms of achieving high target doses and sparing of the bladder and rectum. We are using the same margins for all PTVs (8 mm everywhere and 5 mm posteriorly), although we have found that sometimes 6 mm lateral margins for PTV4 are necessary to limit bladder and bowel dose. Since lateral interfraction prostate displacement is typically small, this seems reasonable. Because of the generous definition of CTV5 and its extremely complex geometry, no additional margin was used for this targeting group in this study.
● R. A. PRICE et al.
585
Intensity-modulated radiotherapy plans were generated for each targeting group (associated with each PTV) with the goal of meeting our routine clinical prostate plan acceptance criteria (5). These criteria include 95% of the planning target volume (PTV95) receiving at least the prescription dose. The rectal volume receiving ⱖ65 Gy (R65) should not be more than 17% and the volume receiving ⱖ40 Gy (R40) should not be more than 35%. The bladder volume receiving ⱖ65 Gy (B65) should not be more than 25%, and the volume receiving ⱖ40 Gy (B40) should not be more than 50%. The 50% isodose line should fall within the rectal contour and the 90% isodose line should not exceed half the diameter of the rectum on any CT slice. Additionally, the distance between the posterior edge of the CTV and the prescription isodose line should be between 3 and 8 mm on each axial slice. A 40-Gy bowel limit of 150 cc was based on experience gained from pelvic irradiation for rectal cancer (10) as well as dose levels discussed in the treatment of anal and gynecologic diseases (11–15). This is likely to be conservative because of the lower dose per fraction to this region and the absence of concurrent chemotherapy. Plans were generated for delivery on 2 accelerators. The first accelerator was a 10 MV Siemens Primus (Siemens Medical Systems, Concord, CA) using a minimum beamlet size of 10 ⫻ 10 mm2. The second accelerator was a 10MV Varian Ex 21 (Varian Medical Systems, Palo Alto, CA) utilizing a minimum beamlet size of 10 ⫻ 5 mm2 with the collimator rotated to 90 degrees (16). By using the leaf width (5 mm) as the short side of the beamlet and rotating the collimator, we are matching the resolution of a 5 ⫻ 5 mm2 beamlet in the direction perpendicular to the prostate-rectum interface. In addition, using a step size of 10 mm results in fewer MLC positions and a faster treatment delivery. The increased area of the elongated beamlet results in increased output and fewer segments and monitor units (MU). This MU reduction also results in a decrease in the amount of accelerator head leakage. All plans were generated using the Corvus inverse planning system, version 5.0 (NOMOS Corp., Cranberry, PA) by the same experienced planner (R.P.) using 5 intensity levels per beam. The step-and-shoot delivery method was used throughout. Each plan was started using 6 beam directions progressing to 9 as needed. In addition, regions for dose constraint were used for all plans (17). This technique is based on the idea that all regions that are not designated as target, or some normal or critical structure, are combined as 1 region during the IMRT planning process. This region (termed “tissue” for this planning system) is controlled by a single set of dose constraints. However, the volume of tissue is far greater than that of the structures of interest. Furthermore, tissue is given the lowest priority during the optimization process. The addition of the dose-conforming regions allows more partial volume input data and provides extra control over the dose distribution in the regions of interest with respect to plans run without the regions. By designating subsets of tissue as concentric regions around the target(s) and defining each region’s dose constraints, an increased measure of control over the dose gradient outside the target boundaries is realized resulting in increased dose conformity and decreased treatment time.
Fig. 1. (a– e) Illustration of anatomy, that with the associated planning target volumes (PTVs), make up the targeting groups (TG) used in this study. (a) (TG1) prostate and proximal seminal vesicles (SVs). (b) (TG2) TG1 and the addition of the distal seminal vesicles. (c) (TG3) TG2 and the addition of the periprostatic and peri-seminal vesicle lymph nodes. (d) (TG4) TG3 and the addition of the external iliac, proximal obturator and proximal internal iliac lymph nodes. (e) (TG5) TG4 and the addition of the presacral lymph nodes.
586
I. J. Radiation Oncology
● Biology ● Physics
Dose volume histogram (DVH) data were analyzed with respect to dose to the PTVs, rectum, bladder, bowel, and normal tissue. Additional analysis was performed for the number of segments required, MU, “hot spots” (i.e., unacceptable high-dose regions), and treatment time. Normal distributions of the endpoints were evaluated with the Shapiro-Wilk test of normality. For the 4 more complicated targeting groups, comparisons were made to Group 1 by paired t tests (normally distributed variables) or by Wilcoxon signed rank tests (non-normally distributed variables). The McNemar Exact Test (MET) was used to evaluate endpoints with respect to meeting or failing each criterion. This test is based on the number of patients that meet the criterion on 1 plan and fail on another and uses exact binomial probabilities because of the small sample size.
RESULTS All plans were normalized such that 95% of the prostate and proximal seminal vesicle PTV received 76 Gy. The remaining PTVs for each group received at least 56 Gy to 95% of their volumes. Table 1 contains the results for all endpoints studied resulting from plans generated for the Siemens Primus unit (10 ⫻ 10 mm2 minimum beamlet
Volume 66, Number 2, 2006
size). Included are the mean values for all 10 patients and all 5 targeting groups as well as the respective ranges and statistical significance (p-values). All endpoints except the bowel volume variables were normally distributed. Statistical significance was calculated for all groups with respect to Group 1. Since the conformity index is not defined for multiple co-joined targets with differing prescription doses and the fact that our target volume(s) changes with targeting progression, we have decided to evaluate the volume of normal tissue (NT) receiving some reference dose as a function of targeting progression. The reference dose used was our 56 Gy (NT56) nodal target prescription dose. Table 2 shows the results for all endpoints studied resulting from plans generated for the Varian Ex 21 unit (10 ⫻ 5 mm2 minimum beamlet size). Values for p and NT are defined as in Table 1. Table 3 contains the results from a head-to-head comparison for plans generated for the 2 accelerators. In this table, statistical significance is calculated between the results for the 2 linacs within each targeting group. Figure 2a illustrates the rectal results for all patients when
Table 1. Average results from plans generated with a minimum beamlet size of 10 ⫻ 10 mm2 10 ⫻ 10 mm2
Group 1
Group 2
Group 3
Group 4
Group 5
R65 (%) Range (%) p R40 (%) Range (%) p B65 (%) Range (%) p B40 (%) Range (%) p Bowel-65 (cc) Range (cc) p Bowel-40 (cc) Range (cc) p MU Range p Segments Range p Tx time (min) Range (min) p Max dose (%) Range (%) p NT56 (cc) Range (cc) p
11.9 7.0–16.6
11.8 7.4–15.8 0.791 28.8 19.8–36.3 0.071 21.6 7.3–39.4 0.018 39.8 17.6–75.4 0.126 0.7 0.0–2.4 0.250 9.2 0.0–36.3 0.016 1047 873–1286 0.170 55 43–79 0.788 9.1 7.1–13.0
11.6 7.6–16.2 0.640 29.0 19.7–36.0 0.039 23.8 10.8–50.3 0.017 47.8 21.4–84.9 ⬍0.005 1.3 0.0–7.3 0.063 14.0 0.0–68.8 ⬍0.005 1205 1053–1333 ⬍0.005 76 56–90 ⬍0.005 12.6 9.2–14.9
11.3 8.1–15.0 0.321 29.8 22.9–36.6 0.011 27.8 14.2–43.0 ⬍0.005 70.9 61.6–85.9 ⬍0.005 7.3 0.4–27.9 ⬍0.005 210.8 48.4–382.8 ⬍0.005 1521 1296–1743 ⬍0.005 117 89–136 ⬍0.005 19.3 14.7–22.4
10.8 6.4–14.7 0.082 35.4 21.0–43.7 ⬍0.005 30.2 20.1–49.2 ⬍0.005 72.0 59.1–81.0 ⬍0.005 9.1 0.4–62.9 ⬍0.005 209.3 57.7–358.5 ⬍0.005 1701 1533–1935 ⬍0.005 145 130–163 ⬍0.005 23.9 21.5–26.9
0.788 118.5 117.5–119.4 0.986 356 256–449 0.122
⬍0.005 118.7 116.7–120.4 0.645 409 284–509 ⬍0.005
⬍0.005 119.0 115.9–122.1 ⬍0.005 780 612–980 ⬍0.005
⬍0.005 120.2 117.8–121.4 ⬍0.005 920 679–1170 ⬍0.005
27.1 19.5–33.9 19.9 7.7–35.8 36.8 15.9–63.0 0.3 0.0–2.5 3.9 0.0–16.3 1013 855–1210 54 40–78 9.0 6.6–12.9 118.1 114.9–119.9 337 253–447
Abbreviations: Max ⫽ maximum; MU ⫽ monitor units; Tx ⫽ treatment. For reference, primary limits are as follows: R65 ⱕ17%, R40 ⱕ35%, B65 ⱕ25%, B40 ⱕ50%, Bowel40 ⱕ150 cc.
Impact of IMRT on prostate cancer
● R. A. PRICE et al.
587
Table 2. Average results from plans generated with a minimum beamlet size of 10 ⫻ 5 mm2 10 ⫻ 5 mm2
Group 1
Group 2
Group 3
Group 4
Group 5
R65 (%) Range (%) p R40 (%) Range (%) p B65 (%) Range (%) p B40 (%) Range (%) p Bowel-65 (cc) Range (cc) p Bowel-40 (cc) Range (cc) p MU Range p Segments Range p Tx time (min) Range (min) p Max dose (%) Range (%) p NT56 (cc) Range (cc) p
10.3 6.3–15.2
10.6 6.8–14.4 0.285 24.5 15.9–34.0 0.073 20.0 8.2–37.4 0.122 38.9 17.9–73.7 0.012 0.7 0.0–2.7 0.031 8.3 0.0–34.3 0.016 1516 1248–1836 0.986 203 141–291 0.338 10.4 8.2–13.7
10.3 6.1–14.4 0.841 25.1 15.6–34.8 0.031 23.1 10.4–45.7 ⬍0.005 47.8 21.7–86.1 ⬍0.005 0.9 0.0–3.2 0.063 13.6 0.0–63.0 0.016 1484 846–1870 0.645 262 174–351 ⬍0.005 12.2 9.4–15.3
10.4 6.3–14.0 0.588 26.1 17.0–33.4 ⬍0.005 26.9 15.3–40.5 ⬍0.005 67.7 58.8–81.8 ⬍0.005 4.4 0.0–20.1 0.008 199.1 36.2–340.2 ⬍0.005 2314 2138–2467 ⬍0.005 628 489–737 ⬍0.005 21.3 12.2–25.8
10.2 6.4–13.2 0.799 30.5 18.3–37.9 ⬍0.005 29.2 17.3–47.6 ⬍0.005 70.4 61.3–83.9 ⬍0.005 5.9 0.0–35.1 ⬍0.005 197.1 49.2–322.6 ⬍0.005 2252 2072–2340 ⬍0.005 725 640–792 ⬍0.005 25.4 22.4–27.7
0.267 116.9 114.3–118.9 0.886 331 224–421 0.021
0.007 117.0 114.4–119.5 0.982 381 266–485 ⬍0.005
⬍0.005 117.5 114.5–121.8 0.554 735 573–892 ⬍0.005
⬍0.005 119.9 117.2–126.3 0.019 859 631–1071 ⬍0.005
23.2 14.2–30.8 19.1 7.3–36.1 35.1 13.8–64.7 0.3 0.0–1.5 3.3 0.0–15.5 1515 1309–1781 189 137–324 9.8 7.9–11.5 117.0 114.7–119.5 316 216–417
Abbreviations as in Table 1. For reference, primary limits are as follows: R65 ⱕ17%, R40 ⱕ35%, B65 ⱕ25%, B40 ⱕ50%, Bowe140 ⱕ150 cc.
prescribing to targeting Group 4 (extended lymphatics). It can be seen that our clinical criteria are met in almost all cases. This is to be expected in that the lymphatic regions involved are not directly adjacent to the rectum. It is seen that the smaller leaf width MLC plans, particularly for the R40 endpoint, result in lower percentages of rectum being irradiated indicating increased dose conformity. The bladder endpoints for all patients, illustrated in Fig. 2b, are not met in the majority of cases. The geometry of targeting Group 4 with the extended lymphatics bounding the bladder laterally, and the associated PTVs, make it impractical to limit the bladder to our clinical criteria in all cases. The paired t-test evaluates statistical significance by comparing the mean values between groups for a given endpoint. We also evaluated endpoints on a case-by-case basis with respect to meeting or failing to meet our criteria using McNemar’s Exact Test (MET). This test indicates a highly significant trend in failing to meet the B40 criteria with the addition of the extended lymphatics (p ⫽ 0.008). Attempts to increase the bladder constraints further during optimization resulted in the unacceptable high-dose regions known
as hot spots. It should be noted that the values indicated for hot spots represent a single point in the dose matrix defined by voxels with sides approximately 0.9 mm in length. If we evaluate the dose distribution for clinically meaningful high-dose regions, at least 2 cm2 on an individual CT slice for example, the listed values decrease on average by approximately 5%. For example, an indicated hot spot of 118% would typically represent a clinically significant hot spot of 113%. These hot spots are always found within the prostate PTV. Given that there is some overlap between the prostate PTV and critical structures efforts are made to ensure that the high-dose regions do not fall in these overlap areas if possible. The affects on our rectal endpoints with the addition of the presacral region is illustrated in Fig. 2c. The high-dose limit is consistently met. The low-dose limit is not met in the majority of cases (MET p ⫽ 0.03 for the 10 ⫻ 10 mm2 beamlets; the smaller beamlets showed no significant trend). The addition of the presacral region, bounding a portion of the rectum laterally and posteriorly, results in a tube-like dose distribution. Maintaining the R40 limit was not always
588
I. J. Radiation Oncology
● Biology ● Physics
Volume 66, Number 2, 2006
Table 3. Results of head-to-head comparison of plans generated for the two different minimum beamlet sizes and accelerator manufacturers
R65 (10 ⫻ 10) % R65 (10 ⫻ 5) % p R40 (10 ⫻ 10) % R40 (10 ⫻ 5) % p B65 (10 ⫻ 10) % B65 (10 ⫻ 5) % p B40 (10 ⫻ 10) % B40 (10 ⫻ 5) % p Bowel-65 (10 ⫻ 10) cc Bowel-65 (10 ⫻ 5) cc p Bowel-40 (10 ⫻ 10) cc Bowel-40 (10 ⫻ 5) cc p MU (10 ⫻ 10) MU (10 ⫻ 5) p Segments (10 ⫻ 10) Segments (10 ⫻ 5) p Tx time (min) (10 ⫻ 10) Tx time (min) (10 ⫻ 5) p Max% (10 ⫻ 10) Max% (10 ⫻ 5) p NT56 (cc) (10 ⫻ 10) NT56 (cc) (10 ⫻ 5) p
Group 1
Group 2
Group 3
Group 4
Group 5
11.9 10.3 0.004 27.1 23.2 ⬍0.005 19.9 19.1 0.113 36.8 35.1 0.020 0.3 0.3 1.000 3.9 3.3 0.125 1013 1515 ⬍0.005 54 189 ⬍0.005 9.0 9.8 0.122 118.1 117.0 0.103 337 316 ⬍0.005
11.8 10.6 0.005 28.8 24.5 ⬍0.005 21.6 20.0 0.078 39.8 38.9 0.071 0.7 0.7 0.750 9.2 8.3 0.102 1047 1516 ⬍0.005 55 203 ⬍0.005 9.1 10.4 0.041 118.5 116.9 0.006 356 331 ⬍0.005
11.6 10.3 0.005 29.0 25.1 ⬍0.005 23.8 23.1 0.144 47.8 47.8 0.950 1.3 0.9 0.688 14.0 13.6 0.461 1205 1484 0.007 76 262 ⬍0.005 12.6 12.2 0.697 118.7 117.0 ⬍0.005 409 381 ⬍0.005
11.3 10.4 0.006 29.8 26.1 ⬍0.005 27.8 26.9 0.230 70.9 67.7 0.004 7.3 4.4 ⬍0.005 210.8 199.1 ⬍0.005 1521 2314 ⬍0.005 117 628 ⬍0.005 19.3 21.3 0.031 119.0 117.5 ⬍0.005 780 735 ⬍0.005
10.8 10.2 0.016 35.4 30.5 ⬍0.005 30.2 29.2 0.225 72.0 70.4 0.104 9.1 5.9 0.232 209.3 197.1 ⬍0.005 1701 2252 ⬍0.005 145 725 ⬍0.005 23.9 25.4 ⬍0.005 120.2 119.9 0.621 920 859 ⬍0.005
Abbreviations as in Table 1. For reference, primary limits are as follows: R65 ⱕ17%, R40 ⱕ35%, B65 ⱕ25%, B40 ⱕ50%, Bowel40 ⱕ150 cc.
possible. As expected the bladder endpoints are not met in the majority of cases as illustrated in Fig. 2d (MET for B40, p ⫽ 0.008 for both linacs). The plans that met the B65 endpoints are consistent with those for targeting Group 4 indicating the major contributing factor to increased bladder dose being the addition of the extended lymphatics. Figure 3 illustrates the results for all plans with respect to our bowel criteria. On average we were unable to meet the 40 Gy cut point with the addition of targeting Groups 4 and 5. Additionally these groups exhibit statistical significance with respect to failing this endpoint on a case-by-case basis (MET p ⫽ 0.016, both groups, both linacs). There is little change in the results between plans for these groups indicating the addition of the extended lymphatics as being the major contributor to increased dose to bowel. It is of note that the smaller MLC leaf widths resulted in a decreased volume of bowel being irradiated indicating increased target conformity. We have chosen to treat all targets, regardless of the intended dose, with a single IMRT plan. The prostate prescription of 76 Gy delivered in 38 fractions assures approx-
imately 2.0 Gy per fraction is being delivered. However, the prescription dose for the distal seminal vesicles and all nodal groups is 56 Gy in 38 fractions resulting in a gradient between approximately 1.5 Gy and 2.0 Gy per fraction. This prescription was derived with the aid of biologically equivalent dose calculations and is equivalent to 49 to 50 Gy delivered at 2.0 Gy per fraction. An alternative approach would be to use a series of “cone down” IMRT plans to maintain 2.0 Gy per fraction for all targets. We do not believe the dosimetric results will differ from those found in this study. DISCUSSION In this report, the dosimetric impact of pelvic lymphatic irradiation for prostate cancer was studied in terms of rectum, bladder, and bowel sparing for various nodal volumes. For all organs at risk, the nodal volume used was a determinant in fulfilling the specific DVH criteria. Although PTV coverage was assured through our normalization procedure, we found a significant increase in
Impact of IMRT on prostate cancer
● R. A. PRICE et al.
589
Fig. 2. (a) Rectal endpoint results for all patients with the inclusion of the extended lymphatics (TG4). The 40- and 65-Gy limits are delineated as well as the results for both multileaf collimator (MLC) sizes (v indicates 5 mm leaf widths and s indicates 10-mm leaf widths). (b) Bladder endpoint results for all patients with the inclusion of the extended lymphatics (TG4). The 40- and 65-Gy limits are delineated as well as the results for both MLC sizes (v indicates 5-mm leaf widths and s indicates 10-mm leaf widths). (c) Rectal endpoint results for all patients with the inclusion of the presacral lymphatics (TG5). The 40 and 65 Gy limits are delineated as well as the results for both MLC sizes (v indicates 5-mm leaf widths and s indicates 10-mm leaf widths). (d) Bladder endpoint results for all patients with the inclusion of the presacral lymphatics (TG5). The 40- and 65-Gy limits are delineated as well as the results for both MLC sizes (v indicates 5-mm leaf widths and s indicates 10-mm leaf widths).
dose to normal tissues and an increase in the maximum dose delivered with increasing nodal volume. These effects were decreased for plans generated for delivery with 5-mm leaf widths. While the difference in treatment delivery time between the 2 linacs studied was not believed to be clinically significant, the use of the smaller leaf MLCs (specifically the smaller minimum beamlet size) resulted in a significant increase in MU. This MU increase yields increased head leakage and should be evaluated with respect to vault shielding. Rectum It can be seen from Tables 1 and 2 that the R65 criteria can be met for all targeting groups; this is believed to be
essential, as this limit is set based on the probability of increased rectal complications (18). However, caution should be exercised with the introduction of the presacral lymphatics. At Fox Chase Cancer Center, the patient is simulated with the rectum empty, which allows for a maximum diameter of 3 cm on any individual CT slice. The length of the rectum is defined from the recto-sigmoid junction to the bottom of the ischial tuberosities including the wall and contents. Our experience is that the rectum is relatively empty after 2 weeks and planning under this condition is conservative. For example, if our DVH acceptance criteria are met with the rectum empty, by forcing a high (compressed) dose gradient across the rectal volume, we should be able to assure rectal sparing despite rectal
590
I. J. Radiation Oncology
● Biology ● Physics
Volume 66, Number 2, 2006
understood. The changing volume and shape of the bladder throughout the course of treatment will continue to complicate defining a dose/volume relationship with bladder complications. The bladder criteria were not met with the addition of the extended lymphatics or the presacral region (Tables 1 and 2). A statistically significant trend toward larger volumes receiving the dose limits is observed for these groups. Attempts to place stricter planning constraints on the bladder resulted in unacceptable hot spot occurrence. There was no significant difference found between plans generated for the 2 accelerators for the bladder 65 Gy cutpoint. No trend was found with targeting progression for B40 although a significant difference was found for Groups 1 and 4 with the smaller beamlet plans resulting in lower values. Fig. 3. Bowel-40 endpoint results for all patients and all targeting groups. The 150-cc limit is delineated as well as the results for both MLC sizes (minimum beamlet sizes are listed and correspond to linac vendor).
filling throughout treatment with the help of daily localization. On the other hand, rectal sparing may not be assured if the pre-sacral lymphatics are also treated. If the dose distribution is very conformal to the rectum, rectal filling during the course of treatment may result in unanticipated high dose being delivered to unintended volume(s) of rectum. Figures 4a to 4d illustrate an example of the relationship between the high-dose region and the rectum for target Groups 4 and 5. It is clear from these images that the addition of the presacral lymphatics increases the uncertainty in rectal sparing with respect to rectal filling during treatment. A strategy that may prove helpful is to add a margin to the posterior and lateral aspects of the rectum, forming a planning organ at risk volume (PRV) to account for this potential rectal filling and potentially keep the high-dose region away from the actual rectum. The R40 limit is somewhat arbitrary and is used to control the shape of the DVH. On average all plans met this criterion with a statistically significant trend toward larger volumes receiving 40 Gy with increased targeting from Groups 3 to 5. The head-to-head test between the 2 accelerators was made to evaluate the strengths and weaknesses of the different leaf widths and different minimum beamlet sizes. A significant difference was found for all targeting groups with the 10 ⫻ 5 mm2 beamlet plans resulting in lower values for both rectal end points. This may be attributed to the higher dose gradient attainable at the prostate-rectal interface with this beamlet configuration. Bladder A similar nodal volume dependency was observed for bladder sparing. Although B65 and B40 limits were set arbitrarily (5), as we have no data corresponding to increased complications at this time, they are useful for comparing bladder sparing between the various nodal volumes. The impact of treating larger bladder volumes is not well
Bowel On average the amount of bowel receiving 65 Gy was between 0.3 cc and 9.1 cc with no trends as a function of targeting progression. The 40 Gy criterion was not met in 60% of the cases with the addition of targeting Groups 4 and 5. Attempts to place increased planning constraints on the bowel led to increased bladder dose and unacceptable hot spots. Failure to limit the bowel receiving ⱖ40 Gy for these targeting groups is primarily attributed to the addition of the extended lymphatic regions and their associated PTV. Evaluation of the bowel end-points indicate that a significant difference is evident for Groups 4 and 5 with the smaller beamlet plans yielding lower values than Bowel-40. Although the smaller beamlets resulted in lower values in all cases, we should reiterate that this endpoint criterion was not met on average. Planning target volume The PTV needed for the uncertainty in position of the extended lymphatic regions should be similar to that of the prostate. The prostate and seminal vesicles move within the pelvis with respect to the bony structures while the extended lymphatic regions remain relatively fixed. We assign a PTV to the prostate and seminal vesicles in part because of this localization uncertainty and we actively localize daily with ultrasound. If our PTV for the extended lymphatic regions is inadequate and we move the prostate and seminal vesicles into the appropriate high-dose region using ultrasound, we may experience a geographic miss of the extended lymphatic regions because of their relative immobility. It is the medial aspect of the extended lymphatics PTV that overlaps with the bowel, resulting in increased dose. An alternative approach that may be successful would be to limit the PTV growth in the medial–lateral dimensions but maintain the current growth in the other directions. The medial–lateral dimension exhibits the least prostate motion (19). Subsequent daily localization shifts would need to be limited in the lateral dimension as well. The prostate and seminal vesicles could still be encompassed by their respective PTVs, and a limited amount of bowel lies in the anterior/ posterior and cranio/caudal directions with respect to the extended lymphatics.
Impact of IMRT on prostate cancer
● R. A. PRICE et al.
591
Fig. 4. (a, b) Axial and sagittal dose distributions (Patient Index #2) through the prostate region representing planning for TG4. The prescription isodose lines, 76 Gy and 56 Gy, as well as 38 Gy (50%) are shown. (c, d) Axial and sagittal dose distributions (Patient Index #2) through the prostate region representing planning for TG5, the addition of the presacral lymphatics. The prescription isodose lines, 76 Gy and 56 Gy, as well as 38 Gy (50%) are shown. Note the tube-like dose distrubution surrounding the rectum.
The maximum hot spot in the treatment plans did not vary significantly with target progression with the exception of the addition of the presacral region. Even though this difference was statistically significant, it is probably not clinically significant. Attempts to meet the set criteria for bladder and bowel tended to result in unacceptable hot spots with increase in targeting progression. There is a statistically significant increase in NT56 with targeting progression, beginning with targeting Group 3. This is to be expected, as our target volume is increasing. However, the amount of normal, nontarget tissue receiving 56 Gy approximately triples between Groups 1 and 5. This increase and the potential adverse implications should be evaluated with respect to the benefits gained from nodal irradiation. The head-to-head evaluation indicated a statistically significant decrease in maximum hot spot for targeting Groups 2, 3 and 4, although the clinical significance is arguable. There was a statistically significant decrease in NT56 for all targeting groups when using the smaller beamlets, hinting at increased conformity. However, it is again arguable if these decreases are clinically significant.
Monitor units A statistically significant trend in increasing MU with targeting progression is observed. Although this may be expected because of the relatively harsh dose limits of the normal structures of interest, the increasing target volume with targeting progression, and the complexity of the resultant intensity maps, this increase should be evaluated with respect to increased radiation leakage, vault shielding, and patient safety. This can be done by using variations of the modulation scaling factor (20). A corresponding statistically significant trend in increasing number of segments and subsequent treatment delivery time is associated with increasing targeting progression beginning with the introduction of the Group 3 lymphatic region. On average, treatment time more than doubles from delivery of Group 1 to Group 4. This increased treatment time has the obvious adverse effect on patient throughput but perhaps more importantly on patient comfort. Maintaining a full bladder for 20 min vs. 10 min may jeopardize patient immobilization and subsequent treatment accuracy. Although some dosimetric improvements are evident
592
I. J. Radiation Oncology
● Biology ● Physics
for the smaller beamlet plans a statistically significant increase in MU was found for all targeting groups when using this configuration. This is attributable in large part because of the decrease in output for the smaller field sizes. It is important to evaluate the effects on existing vault shielding this increase combined with IMRT patient load will have. A significant increase in the number of individual beam segments generated was also found with the smaller beamlet plans. This corresponded to an increase in treatment time in general with statistical significance evident for targeting Groups 2, 4, and 5. The increase in treatment time for Groups 4 and 5 are also affected by the need for an MLC carriage shift. Although statistically significant these increases may not be clinically significant.
Volume 66, Number 2, 2006
CONCLUSION The re-introduction of pelvic lymphatic irradiation during IMRT delivery for prostate cancer is increasing in popularity. However, a host of areas of concern are also introduced including dose per fraction, normal structure dose/volume limits, PTV generation, localization, treatment time, and increased radiation leakage and the effects on vault shielding. The feasiblility of pelvic lymphatic irradiation was performed by evaluating the clinical endpoints used at this treatment center. Although some of our acceptance criteria may be too strict (e.g., R65 and Bowel 40), others may be too loose (e.g., hot spots). More definitive answers with respect to endpoints will become evident with the completion of additional trials and research. In the interim we would suggest that, at a minimum, the endpoints used in this work be evaluated before beginning IMRT pelvic nodal irradiation.
REFERENCES 1. Roach M III, DeSilvio M, Lawton C, et al. Phase III trial comparing whole-pelvis versus prostate-only radiotherapy and neoadjuvant versus adjuvant combined androgen suppression: Radiotherapy Oncology Group 9413. JCO 2003;21:1904 – 1911. 2. Nutting CM, Convery DJ, Cosgrove VP, et al. Reduction of small and large bowel irradiation using an optimized intensitymodulated pelvic radiotherapy technique in patients with prostate cancer. Int J Radiat Oncol Biol Phys 2000;48:649 – 656. 3. Sanguinetti G, Cavey ML, Endres EJ, et al. Is IMRT needed to spare the rectum when pelvic lymph nodes are part of the initial treatment volume for prostate cancer? Int J Radiat Oncol Biol Phys 2006;64:151–160. 4. Ashman JB, Zelefsky MJ, Hunt MS, et al. Whole pelvic radiotherapy for prostate cancer using 3D conformal and intensity-modulated radiotherapy. Int J Radiat Oncol Biol Phys 2005;63:765–771. 5. Pollack A, Hanlon AL, Horwitz EM, et al. Dosimetry and preliminary acute toxicity in the first 100 men treated for prostate cancer on a randomized hypofractionation dose escalation trial. Int J Radiat Oncol Biol Phys 2006;64:518 –526. 6. Buyyounouski MK, Horwitz EM, Price RA Jr, et al. Prostate IMRT. In: Bortfield T, Rupert Schmidt-Ullrich R, de Neve W, Wazer DE, editors. IMRT handbook: concepts & clinical applications. Heidelberg: Springer-Verlag; 2006. p. 391– 410. 7. Shih HA, Harisinghani M, Zietman AL, et al. Mapping of nodal disease in locally advanced prostate cancer: Rethinking the clinical target volume for pelvic nodal irradiation based on vascular rather than bony anatomy. Int J Radiat Oncol Biol Phys 2005;63:1262–1269. 8. Ganswindt U, Paulsen F, Corvin S, et al. Intensity-modulated radiotherapy for high risk prostate cancer based on sentinel node SPECT imaging for target volume definition. BMC Cancer 2005;5:91. 9. Murray SK, Breau RH, Guha AK, Gupta R. Spread of prostate carcinoma to the perirectal lymph node basin: Analysis of 112 rectal resections over a 10-year span for primary rectal adenocarcinoma. Am J Surg Pathol 2004;28:1154 –1162. 10. Baglan KL, Frazier RC, Yan D, et al. The dose-volume relationship of acute small bowel toxicity from concurrent
11.
12.
13.
14. 15.
16.
17.
18.
19. 20.
5-FU-based chemotherapy and radiotherapy for rectal cancer. Int J Radiat Oncol Biol Phys 2002;52L:176 –183. Roeske JC, Lujan A, Rotmensch J, et al. Intensity-modulated whole pelvic irradiotherapy in patients with gynecologic malignancies. Intern J Radiat Oncol Biol Phys 2000;48:1613– 1621. Portelance L, Chao KSC, Grigsby PW, et al. Intensity-modulated radiotherapy (IMRT) reduces small bowel, rectum, and bladder doses in patients with cervical cancer receiving pelvic and para-aortic irradiation. Int J Radiat Oncol Biol Phys 2001;51:261–266. Adli M, Mayr NA, Kaiser HS, et al. Does prone positioning reduce small bowel dose in pelvic radiation with intensitymodulated radiotherapy for gynecological cancer? Int J Radiat Oncol Biol Phys 2003;57:230 –238. Mundt AJ, Lujan AE, Rotmensch J, et al. Intensity-modulated whole pelvic radiotherapy in women with gynecologic malignancies. Int J Radiat Oncol Biol Phys 2005;63:354 –361. Milano MT, Jani AB, Farrey KJ, et al. Intensity-modulated radiotherapy (IMRT) in the treatment of anal cancer: Toxicity and clinical outcome. Int J Radiat Oncol Biol Phys 2003;57: 230 –238. Price RA Jr, Paskalev K, McNeeley SW, Ma C-M. Elongated beamlets: A simple technique for segment and MU reduction for sMLC IMRT delivery on accelerators utilizing 5 mm leaf widths. Phys Med Biol 2005;50:N235–N242. Price RA Jr, Murphy S, McNeeley SW, et al. A method for increased dose conformity and segment reduction for sMLC delivered IMRT treatment of the prostate. Int J Radiat Oncol Biol Phys 2003;57:843– 852. Storey MR, Pollack A, Levy L, et al. Complications from radiotherapy dose escalation in prostate cancer: Preliminary results of a randomized trial. Int J Radiat Oncol Biol Phys 2000;48:635– 642. Chandra A, Dong L, Huang E, et al. Experience of ultrasoundbased daily prostate localization. Int J Radiat Oncol Biol Phys 2003;56:436 – 447. Price RA Jr, Ma CM, Chibani O. Shielding evaluation for IMRT implementation in an existing accelerator vault. J Appl Clin Med Phys 2003:231–238.