Background: Initial evidence suggests activity of pembro + enza in patients (pts) resistant to enza. We present results from the multicohort phase II study KEYNOTE-199 (NCT02787005) in chemotherapy-naive pts with mCRPC treated with pembro + enza after progression with enza and who had RECIST-measurable (C4) or bone-predominant (C5) disease. Methods: Pts who did or did previously receive abiraterone and for whom enza treatment failed after clinically meaningful response received pembro 200 mg Q3W, with continuation of enza for up to 2 y or until progression, toxicity, or withdrawal. End point was ORR per RECIST v1.1 (C4) by blinded independent central review (primary); DOR (C4), time to PSA progression, rPFS, OS, and safety. Results: A total of126 pts (C4, 81; C5, 45) were treated. Median (range) PSA was 31 ng/mL (0.4-1667) in C4 and 19 ng/mL (1-1750) in C5.Median (range) time from enrollment to data cut off was 15 mo (7-21) in C4 and 19 mo (7-21) in C5. In C4, ORR (95% CI) was 12% (6-22; 2 CRs, 8 PRs) and median (range) DOR was 6 mo (3+ to 13); 60% of pts had DOR ≥6 mo. DCR (CR + PR + SD) was 51% in C4 and C5. Median (95% CI) time to PSA progression was 4 mo (4-4) in C4 and 4 mo (4-4) in C5. Median (95% CI) rPFS was 4 mo (3-6) for C4 and 4 mo (3-6) for C5; 12-mo rPFS rate was 17% in C4 and 23% in C5. Median (95% CI) OS was NR (16-NR) in C4 and 19 (14-NR) mo in C5; 12-mo OS rate was 70% in C4 and 75% in C5. Shorter median OS was more associated with prior enza treatment <6 mo than with prior enza treatment ≥6 mo. Liver metastasis was associated with shorter median OS however, median OS in visceral disease subgroups appeared longer than expected. Any-grade/grade ≥3 treatment-related AEs occurred in 75%/26% of pts in C4 and 69%/24% in C5. Two pts in C5 died of immune-related AEs (Miller Fisher syndrome and myasthenia gravis). Any-grade/grade 3/4 rash (regardless of relatedness) was higher than that in prior reports for individual agents (33%/6%). Conclusions: Pembro + enza after enza resistance had manageable safety and showed antitumor activity for RECIST-measurable and bone-predominant mCRPC. This combination is being evaluated in an ongoing phase III combination trial. Clinical trial information: NCT02787005.