Erectile Dysfunction and Absorbed Dose to Penile Base Structures in a Randomized Trial Comparing Ultrahypofractionated and Conventionally Fractionated Radiation Therapy for Prostate Cancer.

PURPOSE:To study the relationships between absorbed dose to penile base structures and erectile dysfunction (ED) in patients treated with ultrahypofractionated (UHF) radiation therapy (RT) or conventionally fractionated (CF) RT for prostate cancer. METHODS AND MATERIALS:This dose-response study comprises 673 patients (57%) of the 1180 per-protocol patients included in the HYPO-RT-PC trial (median follow-up 5, years), where patients were randomized to CF (39 × 2.0 Gy, 8 weeks) or UHF (7 × 6.1 Gy, 2.5 weeks). No androgen deprivation therapy was allowed. Only patients with erectile function sufficient for intercourse at baseline and complete RT data were included in this study. Erectile function was assessed by physician at regular follow-ups. The main endpoint was severe ED (EDs). The penile bulb (PB) and crus were retrospectively delineated on the treatment planning computed tomography scans. Dose-volume descriptors were derived from EQD2 converted dose matrices (α/β = 3 Gy). Univariable and multivariable Cox proportional hazard regression and logistic regression were used to find predictors for EDS. RESULTS:No significant difference in EDs was found between CF and UHF. During the follow-up period, EDs occurred in 27% of the patients in both treatment groups. Average (median) PB mean dose, Dmean, was 24.5 (20.2) in CF and 18.7 (13.1) Gy3 in UHF. Age was the only significant predictor for EDs in Cox analyses. All dose-volume variables contributed significantly in univariable logistic regression at 2-year follow-up. Age and near maximum dose (D2%) were significant predictors for EDs in multivariable logistic regression analyses at both 1 and 2 years. CONCLUSIONS:The frequency of EDS was similar in the CF and UHF treatment groups. Age at radiation therapy was the strongest predictor for EDs, followed by dose to PB, and was most evident for younger patients. We propose D2 % <50 Gy3 and Dmean <20 Gy3 to the PB as the primary objectives to be applied in the treatment planning process.