Primary Endpoint Analysis of a Randomized Phase III Trial of Hypofractionated vs. Conventional Post-Prostatectomy Radiotherapy: NRG Oncology GU003

Purpose/Objective(s)To determine if hypofractionated post-operative prostate bed radiotherapy (HYPORT) does not increase patient-reported genitourinary (GU) or gastrointestinal (GI) toxicity over conventionally fractionated post-operative radiotherapy (COPORT).Materials/MethodsEligibility criteria were: 1) an undetectable PSA (< 0.1 ng/mL) with either margin negative pT3pN0/X or margin positive pT2pN0/X adenocarcinoma of the prostate or 2) a detectable PSA (≥ 0.1 ng/mL) and pT2/3pN0/X disease. HYPORT was 62.5 Gy to the prostate bed in 25 fractions of 2.5 Gy. COPORT was 66.6 Gy in 37 fractions of 1.8 Gy. Lymph node RT was not allowed. Androgen deprivation therapy (ADT) ≤ 6 months was allowed. Patients were stratified according to baseline Expanded Prostate cancer Index Composite (EPIC) score (four tiers based on GU and GI scores) and ADT use (yes vs. no) then randomized 1:1. The co-primary endpoints were based on change scores (24-month score minus baseline score) from the GU and GI domains of the EPIC. The hypothesis is that the mean change scores at 24 months are no worse for HYPORT than it is for COPORT. The non-inferiority margins were based on 0.5*standard error from NRG Oncology/RTOG 0415: -5 for GU and -6 for GI. Two hundred eighty-two patients provide ≥ 90% power with a one-sided alpha = 0.025 for each domain while inflating for non-compliance/loss to follow-up.ResultsBetween July 2017 and July 2018, 298 patients were screened and 296 were randomized: 144 to HYPORT and 152 to COPORT. Compliance with the EPIC was 100% at baseline, 83% at the end of RT, 77% at 6 months, 78% at 12 months, and 73% at 24 months. At the end of RT, the HYPORT and COPORT mean GU change scores were neither clinically significant nor significantly different and remained so at 6 and 12 months. The mean GI change scores for HYPORT and COPORT were both clinically significant and significantly different at the end of RT (HYPORT mean GI = -15.0 vs COPORT mean GI = -6.8 P ≤ 0.01). However, both the HYPORT and COPORT mean GI change scores clinically and statistically significant differences were resolved at 6 and 12 months. The 24-month mean GU and GI change scores for HYPORT and COPORT remained neither clinically nor statistically significant (HYPORT mean GU = -5.2 vs COPORT mean GU = -3.0, P = 0.81; HYPORT mean GI = -2.2 vs COPORT mean GI = -1.5, P = 0.12). With a median follow-up for censored patients of 2.1 years, there was no difference between HYPORT versus COPORT for biochemical failure defined as a PSA ≥ 0.4 ng/mL followed by a value higher than the first by any amount (2-yr actuarial, 12% vs 8%, P = 0.29) or local failure (2-yr actuarial, 0.7% vs 0.8%, P = 0.35).ConclusionHYPORT is non-inferior to COPORT in terms of late patient-reported GU or GI toxicity. More follow-up is needed to appropriately assess disease control endpoints. In some clinic scenarios, HYPORT may be considered an acceptable practice standard. To determine if hypofractionated post-operative prostate bed radiotherapy (HYPORT) does not increase patient-reported genitourinary (GU) or gastrointestinal (GI) toxicity over conventionally fractionated post-operative radiotherapy (COPORT). Eligibility criteria were: 1) an undetectable PSA (< 0.1 ng/mL) with either margin negative pT3pN0/X or margin positive pT2pN0/X adenocarcinoma of the prostate or 2) a detectable PSA (≥ 0.1 ng/mL) and pT2/3pN0/X disease. HYPORT was 62.5 Gy to the prostate bed in 25 fractions of 2.5 Gy. COPORT was 66.6 Gy in 37 fractions of 1.8 Gy. Lymph node RT was not allowed. Androgen deprivation therapy (ADT) ≤ 6 months was allowed. Patients were stratified according to baseline Expanded Prostate cancer Index Composite (EPIC) score (four tiers based on GU and GI scores) and ADT use (yes vs. no) then randomized 1:1. The co-primary endpoints were based on change scores (24-month score minus baseline score) from the GU and GI domains of the EPIC. The hypothesis is that the mean change scores at 24 months are no worse for HYPORT than it is for COPORT. The non-inferiority margins were based on 0.5*standard error from NRG Oncology/RTOG 0415: -5 for GU and -6 for GI. Two hundred eighty-two patients provide ≥ 90% power with a one-sided alpha = 0.025 for each domain while inflating for non-compliance/loss to follow-up. Between July 2017 and July 2018, 298 patients were screened and 296 were randomized: 144 to HYPORT and 152 to COPORT. Compliance with the EPIC was 100% at baseline, 83% at the end of RT, 77% at 6 months, 78% at 12 months, and 73% at 24 months. At the end of RT, the HYPORT and COPORT mean GU change scores were neither clinically significant nor significantly different and remained so at 6 and 12 months. The mean GI change scores for HYPORT and COPORT were both clinically significant and significantly different at the end of RT (HYPORT mean GI = -15.0 vs COPORT mean GI = -6.8 P ≤ 0.01). However, both the HYPORT and COPORT mean GI change scores clinically and statistically significant differences were resolved at 6 and 12 months. The 24-month mean GU and GI change scores for HYPORT and COPORT remained neither clinically nor statistically significant (HYPORT mean GU = -5.2 vs COPORT mean GU = -3.0, P = 0.81; HYPORT mean GI = -2.2 vs COPORT mean GI = -1.5, P = 0.12). With a median follow-up for censored patients of 2.1 years, there was no difference between HYPORT versus COPORT for biochemical failure defined as a PSA ≥ 0.4 ng/mL followed by a value higher than the first by any amount (2-yr actuarial, 12% vs 8%, P = 0.29) or local failure (2-yr actuarial, 0.7% vs 0.8%, P = 0.35). HYPORT is non-inferior to COPORT in terms of late patient-reported GU or GI toxicity. More follow-up is needed to appropriately assess disease control endpoints. In some clinic scenarios, HYPORT may be considered an acceptable practice standard.