Background: MK-5684 (previously ODM-208) is a first-in-class, oral, non-steroidal, selective inhibitor of CYP11A1, the first and rate-limiting enzyme of steroid biosynthesis. MK-5684 suppresses the production of all steroid hormones and precursors that may activate the androgen receptor (AR) signaling pathway. Initial phase 1 (ASCO-GU 2022) and phase 2 results in AR-LBD mutation-positive (ESMO 2022) mCRPC patients were previously presented. Now we report phase 2 results for both AR-LBD mutation-positive and negative patients. Methods: MK-5684 5mg BID (with dexamethasone and fludrocortisone) was evaluated in an open-label expansion cohort in patients with progressing mCRPC who had previously received ≥1 line of 2^nd generation AR pathway inhibitor and ≥1 line of taxane-based chemotherapy. An initial expansion cohort enrolled 45 patients with an activating AR-LBD mutation identified in cell-free DNA (Guardant360 assay, 74-gene panel), a subsequent extension cohort included mainly AR-LBD mutation-negative patients to attain comparable groups of about 60 patients each, with and without AR-LBD mutations. Study objectives were safety and preliminary efficacy assessed by PSA and RECIST response and standard safety measures. MK-5684 treatment was continued until subsequent disease progression. The study was conducted at 18 sites in France, Finland, UK and USA. Data are based on a 17 July 2023 data cut-off. Results: A total of 66 AR-LBD mutation-positive and 68 AR-LBD mutation-negative patients (median age 68.3 years) were enrolled and received MK-5684 treatment. 53% and 33.8% of patients had previously received both abiraterone and enzalutamide, and 63.6% and 55.9% patients had received cabazitaxel in AR-LBD mutation positive and negative groups respectively. MK-5684 profoundly suppressed androgen synthesis resulting in PSA₅₀ responses in 55.6% and 16.7% of patients and PSA₃₀ responses in 69.8% and 30.0% of patients, with and without AR-LBD mutations respectively. At the time of abstract data cut-off, objective responses by RECIST had occurred in 8 patients, all with AR-LBD mutations (ORR 20.5% for AR-LBD positive). MK-5684 was well-tolerated: the most common treatment-related adverse events were related to adrenal suppression with the rate of hospitalization for adrenal insufficiency being much lower than in phase 1 when typically higher MK-5684 doses were administered (3.0% vs. 33% respectively). Data with a minimum follow up of approximately 4 months after last patient enrolment will be presented. Conclusions: Administration of MK-5684 to heavily pre-treated mCRPC patients showed promising antitumor activity. PSA₅₀ responses were most frequent among patients harboring activating AR-LBD mutations. Clinical trial information: NCT03436485.