Background: Patients with localized high-risk prostate cancer have an increased risk of relapse following RP. We previously reported on Part 1 of a phase 2 trial testing neoadjuvant apalutamide, abiraterone acetate, prednisone, plus leuprolide (AAPL) or abiraterone, prednisone, and leuprolide (APL) for 6 months followed by RP. We demonstrated favorable pathologic responses (tumor ≤5 mm) in 20.3% of patients (n=24/118). Herein, we report the results of Part 2 testing adjuvant AAPL. Methods: At Part 2, patients were randomized 1:1 to AAPL for 12 months (Arm 2A) or observation (Arm 2B), stratified by neoadjuvant therapy and pathologic T stage (<ypT3 or ≥ypT3). The primary endpoint was biochemical progression-free survival (bPFS) at 3 years post-RP. Secondary endpoints included safety and time to testosterone recovery (>200 ng/dL). Results: Overall, 69% (n=82/118) of patients enrolled to Part 1 were randomized to Part 2, of whom 58.5% (n=48) had ≥ ypT3 disease at RP. Patient preference was the main reason for non-randomization in Part 2. Of patients not randomized to Part 2, 32.3% (n=10) had a pathologic complete response or minimum residual disease (tumor ≤5 mm) at RP. In the intent-to-treat analysis, 3-year bPFS was 81% for Arm 2A and 72% for Arm 2B; the difference was not statistically significant (HR 0.81 90% CI 0.43-1.49). In a post-hoc per-protocol analysis, the 3-year bPFS was 83% for AAPL treated patients and 69% for patients on observation (HR 0.50 90% CI 0.26-0.97). Of patients randomized to Part 2, 81% of patients had testosterone recovery in the AAPL treated group compared to 95% in the observation group with median to recovery of 8.7 months compared to 4.0 months, respectively. Of the 37 patients receiving AAPL on Arm 2A, 28 (75.7%) experienced any grade treatment-related adverse event (TRAE) and 4 (10.8%) had grade 3-4 TRAE. The most common any grade TRAE were fatigue (27%) and hypertension (16%). Conclusions: Neoadjuvant ADT in localized high-risk prostate cancer resulted in favorable pathologic responses and 3-year bPFS in a subset of patients. As 30% of patients declined an additional 12 months of adjuvant treatment, Part B of the study was underpowered to detect a difference between treatment arms. Subsequent studies will evaluate clinical and molecular predictors of bPFS. Clinical trial information: NCT02903368.