To determine the prognostic significance of neuroendocrine differentiation (NED) in Gleason score 8-10 prostate cancer treated with primary radiotherapy (RT).Chromogranin A (CgA) staining was performed and overseen by a single pathologist on core biopsies from 176 patients from the William Beaumont prostate cancer database. A total of 143 had evaluable biopsy material. Staining was quantified as 0%,1%, 1-10%, or10% of tumor cells. Patients received external beam RT alone or together with high-dose-rate brachytherapy. Cox regression and Kaplan-Meier estimates determined if the presence/frequency of neuroendocrine cells correlated with clinical endpoints.Median follow-up was 5.5 years. Forty patients (28%) had at least focal positive CgA staining (1% n = 21, 1-10% n = 11,10% n = 8). No significant differences existed between patients with or without staining in terms of age, pretreatment prostate-specific antigen, tumor stage, hormone therapy administration, % biopsy core involvement, mean Gleason score, or RT dose/modality. CgA staining concentration independently predicted for biochemical and clinical failure, distant metastases (DM), and cause-specific survival (CSS). For patients with1% vs.1% staining, 10-year DM rates were 13.4% vs. 55.3%, respectively (p = 0.001), and CSS was 91.7% vs. 58.9% (p0.001). As a continuous variable, increasing CgA staining concentration predicted for inferior rates of DM, CSS, biochemical control, and any clinical failure. No differences in outcomes were appreciated for patients with 0% vs.1% NED.For Gleason score 8-10 prostate cancer,1% NED is associated with inferior clinical outcomes for patients treated with radiotherapy. This relates most directly to an increase in distant disease failure.