Background: Although clinical features of multi-parametric magnetic resonance imaging (mpMRI) have been associated with biochemical recurrence in localized prostate cancer, such features are subject to inter-observer variability. Objective: To evaluate whether the volume of the dominant intraprostatic lesion (DIL), as provided by a deep learning segmentation algorithm, could provide prognostic information for patients treated with definitive radiation therapy (RT). Design, Setting, and Participants: Retrospective study of 438 patients with localized prostate cancer who underwent an endorectal coil, high B-value, 3-Tesla mpMRI and were treated with RT between 2010 and 2017. Intervention: RT. Outcome Measurements and Statistical Analysis: Biochemical recurrence and metastasis risk, assessed with a cause-specific Cox regression and time-dependent receiver operating characteristic analysis. Results and Limitations: The artificial intelligence (AI) model identified DILs with an area under the receiver operating characteristic curve (AUROC) of 0.827 at the patient level. For the 233 patients with available PI-RADS scores, with a median follow-up of 5.6 years, AI-defined DIL volume was significantly associated with biochemical failure (adjusted hazard ratio 1.54, 95% confidence interval 1.09-2.17, p=0.014) after adjustment for PI-RADS score. Among all 438 patients with a median follow-up of 6.9 years, the AUROC for predicting 7-year biochemical failure for AI volume (0.790) was similar to that for an expanded National Comprehensive Cancer Network (NCCN+) category (p=0.17). The AUROC for predicting 7-year metastasis for AI volume trended towards being higher compared to NCCN+ categories (0.854 vs 0.769, p=0.06). Conclusions: A deep learning algorithm could identify the DIL with good performance. AI-defined DIL volume may be able to provide prognostic information independent of the NCCN+ risk group or other radiologic factors for patients with localized prostate cancer treated with RT.