The protocol was designed to randomize men with intermediate-to high-risk prostate cancer to 76 Gy in 38 fractions at 2.0 Gy per fraction (Arm I, conventional fractionation intensity-modulated radiation therapy [CIMRT]) vs. 70.2 Gy in 26 fractions at 2.7 Gy per fraction (Arm II, hypofractionation intensity-modulated radiation therapy [HIMRT]). The hypofractionation arm was hypothesized to be equivalent to 84.4 Gy in 2 Gy fractions assuming an α/β ratio of 1.5. This design was formulated to test whether dose escalation via hypofractionation results in a significant improvement in FFBF, without increasing late complications. The principal hypothesis was that the 8-Gy escalation in biologic dose between Arms I and II would result in a 15% gain in FFBF from 70% to 85%.
The estimated 5-year FFBF rate for intermediate-risk patients treated with radiotherapy alone to 76 Gy at 2 Gy per fraction is 70% (12
). The 5-year FFBF rate for high-risk patients treated with 76 Gy plus 2 years of androgen deprivation is also estimated to be about 70% (20
Classification of risk, eligibility, and stratification
Men with Stage T1-3 adenocarcinoma of the prostate and Gleason score ≥5 were eligible if they had intermediate to high-risk features. Intermediate risk was defined as Gleason score 7, pre-treatment initial PSA (iPSA) >10–20 ng/mL, or ≥3 biopsy cores of Gleason score ≥5, as long as no high-risk features were present. High risk was defined as Gleason score 8–10, Gleason score 7 in ≥4 cores, cT3 disease, or an iPSA >20 ng/mL. Up to 4 months of androgen deprivation with a luteinizing hormone releasing hormone (LHRH) agonist or antiandrogen before randomization was permitted. For those with intermediate risk who were receiving androgen deprivation when enrolled, androgen deprivation was discontinued, whereas for those with high risk (also referred to as unfavorable) androgen deprivation was to be continued for 2 years.
Patients were ineligible if they had a prior history of androgen deprivation for >4 months before randomization, an iPSA of >80 ng/mL, prior pelvic radiotherapy, prior radical prostatectomy, or prior malignancy other than nonmetastatic skin cancers or early stage small lymphocytic lymphoma within the last 5 years. Patients were stratified by iPSA (≤10, >10, ≤20, or >20 ng/mL), Gleason score (5–7 or 8–10), and whether or not long-course androgen deprivation was planned (high or intermediate risk).
Patients were simulated in the supine position in an α-cradle (Smithers Medical Products, Inc., North Canton, OH) with a Plexiglas holder to immobilize the feet. An enema per rectum was given before simulation to empty the rectum as much as possible. The patients were asked to have a moderately full bladder.
Target and normal structure definition and margins
Computed tomography and magnetic resonance imaging (MRI)-based simulations were routinely performed unless there was a medical contraindication to performing an MRI (e.g., pacemaker). Images were acquired from the top of the ilium to below the ischial tuberosities. A slice thickness of 3 mm was used from the bottom of the sacroiliac joints to the ischial tuberosities and 1-cm cuts above and below these levels. The scans were loaded into a planning computer and fused based on bony anatomy using either chamfer matching or maximization of mutual information methods. The computed tomography images were used to outline the external contour and femoral heads. All other contours, including the bladder, rectum, prostate, seminal vesicles, and lymph nodes were outlined using the MRI images.
The structures outlined included the prostate, proximal seminal vesicles (at least the first approximately 9 mm for intermediate-risk patients and >9 mm for high-risk patients), distal seminal vesicles, rectum (entire contents) from the ischial tuberosities to the sigmoid flexure, bladder (entire contents), femurs down to the superior aspect of the lesser trochanter, and the external contour. The penile bulb and corporal bodies were outlined for reference; no dose constraints were placed on these structures. The pelvic lymph nodes were added as a target structure in men with high-risk features after the results from Radiation Therapy Oncology Group (RTOG) protocol 94-13 became available (21
). Our policy now is to outline and treat the external iliac, obturator, and proximal internal iliac lymph nodes () using the vessels as a guide. The lymph nodes are outlined up to the bifurcation of the common iliac vessels. As described by Price et al.
), six artificial ring-shaped structures were also defined to aid in reducing dose to the normal tissues.
Fig. 1 Illustration of the target and normal tissue volumes. Magnetic resonance and computed tomography images were obtained at 3-mm intervals and fused. Every other image slice (every 6 mm) is displayed. The structures outlined are displayed as follows: urinary (more ...)
The gross tumor volume (GTV) was the prostate, any extraprostatic spread identified on imaging, and the proximal seminal vesicles. The primary clinical target volume (CTV1) encompassed the GTV, although an additional approximately 6–9 mm was added inferiorly below the last slice on which the prostate apex was seen on MRI because of lack of capsule in this region. The contour was generous (an extra 1–2 mm) around obvious areas of bulky/extraprostatic tumor seen on MRI. The CTV1 for intermediate-risk patients included the prostate and proximal seminal vesicles (approximately 9 mm); these structures were usually outlined separately to facilitate daily B-mode acquisition and targeting ultrasound alignment (NOMOS, Cranberry Township, PA) and then grouped as the CTV1 for planning. In high-risk patients, the CTV1 included at least 50% of the seminal vesicles (all gross disease extending to the seminal vesicles received the full dose), in addition to the prostate and any extraprostatic extension. In the high-risk patients, the CTV2 comprised the distal portions of the seminal vesicles and the CTV3 comprised the periprostatic, periseminal vesicle, external iliac, obturator, and internal iliac lymph nodes (). The PTV1s were planned to receive a D95% of the prescription dose or higher. The PTV2s and PTV3s were planned to receive a D95% of ≥56 Gy in the CIMRT arm and ≥52–50 Gy in the HIMRT arm (initially it was 52 Gy, but was changed to 50 Gy after the protocol opened to better equate the subclinical dose for the two treatment arms biologically). The PTV2 and PTV3 doses were applied over the full number of fractions (38 for Arm I and 26 for Arm II).
The PTV1, PTV2, and PTV3 margins were consistent within each arm, but were different for the two treatment groups. For CIMRT, the desired PTVs were 8 mm in all dimensions except posteriorly (the prostate-rectal interface for PTV1), in which the margin was 5 mm. For HIMRT, the desired PTVs were 7 mm in all dimensions except posteriorly, in which the margin was 3 mm. The PTV margins were smaller for the HIMRT arm to reduce the potential increased complication risk from hypofractionation. This strategy was based in part on the Cleveland Clinic experience using a similar technique (18
) and the rationale that the 90% line in the HIMRT plans would fall in about the same position as the 100% line in the CIMRT arms. The effective PTV in the CIMRT plans (i.e., where the prescription line was located relative to the CTVs as monitored on a transverse slice by slice basis) was 8–13 mm in all dimensions, except posteriorly, in which the effective PTV was 3–8 mm. The effective PTV in the HIMRT plans was 5–10 mm around the CTVs in all dimensions, except posteriorly, in which the PTV was 2–6 mm.
IMRT plan evaluation and acceptance
Step-and-shoot IMRT was planned using the Corvus (NOMOS) treatment planning system. A series of dose–volume histograms were generated and analyzed to determine the adequacy of the plan. At least 95% of the PTV (D95%) was to receive the prescribed dose; a variation was noted if <95% to 90% of the PTV received the prescribed dose and a protocol violation was noted if <90% of the PTV received the prescribed dose. There were no variations for the PTV D95% (median, 100%). The maximum dose heterogeneity allowable in the PTV was 20%. There were 8 patients in the HIMRT (overall median heterogeneity = 17.2%) and 5 patients in the CIMRT (median overall = 15.9%) groups that had dose gradients above 20%. These were considered variations (dose gradient > 20–25%); no violations (dose gradient > 25%) were observed. Because the dose is prescribed to the minimum isodose line encompassing the PTV, the dose variability was seen in portions of the target volume receiving higher than the specified dose.
The normal tissue planning limits for the bladder and rectum were set based on prior studies (23
). The plan was deemed acceptable under the following conditions. Less than or equal to 17% and 35% of the rectum should receive ≥65 Gy (V65 Gy) and ≥40 Gy (V40 Gy), respectively, for the conventionally fractionated patients (Arm I, 76 Gy total dose). The bladder V65 Gy and V40 Gy was ≤25% and ≤50% in Arm I patients. The rationale for these cutpoints has been described previously (25
). The criteria for the bladder were relaxed because a meaningful dosimetric cutpoint has not been defined.
For Arm II, the rectal V50 Gy and V31 Gy were ≤17% and ≤35%. The bladder V50 Gy and V31 Gy were ≤25% and ≤50%. The derivation of the V50 Gy and V31 Gy criteria for the Arm II patients was based on very conservative extrapolations from the V65 Gy and V40 Gy parameters used for Arm I patients. The α/β ratio for late effects was assumed to be the same as that for prostate cancer tumor control (α/β ratio 1.5) and was probably too conservative, as described in the Discussion section.
If the volume of the rectum or bladder exceeded the dose limits described by <7.5%, this was classified as a variation. The inclusion of rectal volumes beyond these constraints was considered a protocol violation. The inclusion of bladder volumes beyond these constraints was considered a secondary protocol variation; it was not considered a protocol violation because a distinct bladder dose–volume histogram relationship has not been defined previously. The variations and violations in Arm I (standard fractionation) were as follows: rectal V65 Gy, one variation (0.3% above constraint) and no violations; rectal V40 Gy, two variations (1.4% and 2.5% above constraint); bladder V65 Gy, 5 variations and two secondary variations (median, 4.8% above constraint; range, 0.2–18.6%); bladder V40 Gy, four variations and three secondary variations (median, 7.2% above constraint; range, 2.9–26.3%). The variations and violations in Arm II (hypofractionation) were as follows: rectal V50 Gy, 17 variations and no violations (median, 2.3% above constraint; range, 0.1–5.4%); rectal V31 Gy 10 variations and no violations (median, 4.0% above constraint; range, 0.1–6.5%); bladder V50 Gy, 12 variations and 12 secondary variations (median, 7.6% above constraint; range, 0.2–35.3%); and bladder V31 Gy, 3 variations and 11 secondary variations (median, 20.9% range, 2.0–41.9% above constraint).
Endpoint and statistics
The primary endpoint of the study is FFBF using the American Society for Therapeutic Radiology and Oncology consensus guidelines (28
). In this communication, we describe the acute side effects of radiotherapy using modified RTOG and Late Effects Normal Tissue Task Force (LENT) criteria, modeled after that described by Hanlon et al.
) and later Storey et al.
The acute gastrointestinal (GI) side effect grading scale is as follows. Grade I: Increased frequency or change in quality of bowel habits needing ≤2 antidiarrheals per week; Rectal discomfort not requiring analgesics; Mild rectal bleeding not needing medication: Grade II; Diarrhea needing more than two antidiarrheals per week; Mucous discharge requiring one sanitary pad per day; Rectal pain needing analgesics or occasional narcotics; Rectal bleeding needing Anusol HC or other medication; Rectal bleeding or other GI symptoms requiring a treatment break of ≤1 week: Grade III: Diarrhea needing more than two antidiarrheals per day or parenteral support; Severe mucous discharge requiring more than one sanitary pad per day; Rectal pain requiring frequent narcotics (≥1/day) for more than a week; GI bleeding requiring one transfusion; Rectal bleeding or GI symptoms requiring a treatment break of >1 week: Grade IV; Acute or subacute obstruction; GI bleeding requiring more than one transfusion; Fistula or perforation; Abdominal pain or tenesmus requiring bowel diversion.
The acute genitourinary (GU) grading scale is: Grade I: Frequency or nocturia twice pretreatment habit or non-narcotic medication (e.g., alpha blocker) once per day over baseline; Dysuria not needing medication; Microscopic or infrequent gross hematuria not needing medication: Grade II: Frequency or nocturia less frequent than hourly; Dysuria or bladder spasm needing an anesthetic (Pyridium or occasional narcotics); Hematuria or GU symptoms requiring medication or a treatment break of ≤1 week. Infrequent gross hematuria needing medical intervention. Urinary obstruction requiring temporary catheterization (including Foley or self-catheterization) for ≤1 week: Grade III: Frequency or nocturia hourly or more; Dysuria, pain, or spasm needing narcotics >1 dose/day for >1 week; Hematuria or GU symptoms requiring a treatment break >1 week; Gross hematuria requiring one transfusion; Urinary obstruction from prostate inflammation or clots requiring catheterization (including Foley or self-catheterization or suprapubic) for >1 week: Grade IV: Hematuria needing more than one transfusion; Hospitalization for sepsis from obstruction, ulceration, or necrosis of the bladder.
Two-sample t tests were used to assess differences between dosimetric parameters according to treatment arms. Confirmatory analyses were performed using nonparametric Wilcoxon tests. Similar methodology was used to evaluate differences in International Prostate Symptom Scores according to treatment groups. Stepwise ordinal logistic regression modeling was used to determine independent predictors of changes in GU and GI toxicity, relative to pretreatment function assessed using the same grading scale. The variable for the change in acute toxicity was coded as follows: no change in toxicity acutely was coded as a 0; an increase in toxicity from Grade 0 to Grade 1 was coded as a 1; from Grade 1 to Grade 2 as a 2; from Grade 2 to Grade 3 as a 3; from Grade 0 to Grade 2 as a 3; from Grade 1 to Grade 3 as a 3; and Grade 0 to Grade 3 as a 4. Covariates included: maximum dose received by the bladder (continuous), maximum dose received by the rectum (continuous), rectal volume (continuous), bladder volume (continuous), rectal V65 Gy/V50 Gy (continuous), rectal V40 Gy/V31 Gy (continuous), bladder V65 Gy/V50 Gy (continuous), bladder V40 Gy/V31 Gy (continuous), PTV1 volume (continuous), PTV1 mean dose (continuous), PTV1 maximum dose (continuous), androgen deprivation therapy (no vs. yes), iPSA (continuous), T-stage (T1–T2 vs. T3), risk group (intermediate vs. high), and treatment group (Arm I vs. Arm II).