Background: Antibodies (Abs) or small molecules can target PSMA with different biodistribution. Certain sites of PSMA expression (e.g. salivary/lacrimal glands, kidneys, small bowel) are not accessible to Abs. Given radiosensitivity of PC and potency of alpha emitters plus the ability to minimize targeting off tumor PSMA+ sites with J591, we performed a 1st in human study of ²²⁵Ac-J591. Methods: Men with progressive mCRPC following at least 1 potent AR-pathway inhibitor (ARPI) and chemo (or unfit/refuse) without limit of # prior therapies (Ra-223 and ¹⁷⁷Lu-PSMA allowed) with ECOG PS 0-2 and adequate organ function were eligible. Baseline ⁶⁸Ga-PSMA11 PET was performed, but not used for eligibility. Initially 1-subject cohorts treated until transition to 3+3 at dose level 5 (predicted by dosimetry to have moderate toxicity) with a single infusion of ²²⁵Ac-J591 (13.3 KBq/kg with planned escalation up to 93.3 KBq/kg). Dose-limiting toxicity (DLT) defined as attributable grade (Gr) 4 heme toxicity or Gr 3/4 non-heme tox. Imaging, genomic, patient-reported outcomes (PRO), and immune correlates embedded. Results: 22 men treated on 7 dose levels; median age 72.5 (range 58-89), PSA 147 (5-7168); 82% with >2 prior ARPI, 64% chemo, 23% Ra-223, 55% ¹⁷⁷Lu-PSMA. 1 (5%) CALGB (Halabi) good prognostic risk, 10 (45%) intermed, 11 (50%) poor risk. 1 of 6 in cohort 6 (80 KBq/kg) had DLT (Gr 4 anemia and platelets); no others had attributable Gr > 2 non-heme or Gr > 3 heme AE (including 0 of 6 at the highest dose level 93.3 KBq/Kg). Gr 1/2 AE’s: 17 (77%) fatigue, 11 (50%) pain flare, 10 (45%) anemia (+1 Gr 3), 10 (45%) platelets, 6 (27%) nausea, 6 (27%) xerostomia (5 of 6 with prior ¹⁷⁷Lu-PSMA), 5 (23%) neutropenia, 4 (18%) AST elevation. Despite prior treatment including ¹⁷⁷Lu-PSMA and no selection for PSMA expression, 14 (64%) with any PSA decline, 9 (41%) with > 50% PSA decline. 15 (68%) had initial PSA rise followed by decline from peak (delayed effect). 2 with response > 1 year despite prior ¹⁷⁷Lu-PSMA. Of 15 with paired baseline and 12-wk CTC counts, 8 declined, 4 remained undetectable, 3 increased. In subset with complete PRO data (baseline to 12 wks), pain was improved or absent by BPI-SF in 63% and by FACT-P in 89%. Social and emotional well-being domains of FACT-P improved or stabilized in majority; physical well-being improved in most responders. Conclusions: Alpha-emitter ²²⁵Ac targeting PSMA via J591 Ab is tolerable with early evidence of clinical activity in a pre-treated population with favorable PRO’s. Enrollment to expansion cohort being completed. Clinical trial information: NCT03276572.