RESUMEN

[177Lu]Lu-PSMA-617 was approved by the U.S. Food and Drug Administration for patients with prostate-specific membrane antigen (PSMA)-positive metastatic castration-resistant prostate cancer (mCRPC). Since the time of regulatory approval, however, real-world data have been lacking. This study investigated the efficacy, safety, and outcome predictors of [177Lu]Lu-PSMA-617 at a major U.S. academic center. Methods: Patients with mCRPC who received [177Lu]Lu-PSMA-617 at the Johns Hopkins Hospital outside clinical trials were screened for inclusion. Patients who underwent [177Lu]Lu-PSMA-617 and had available outcome data were included in this study. Outcome data included prostate-specific antigen (PSA) response (≥50% decline), PSA progression-free survival (PFS), and overall survival (OS). Toxicity data were evaluated according to the Common Terminology Criteria for Adverse Events version 5.03. The study tested the association of baseline circulating tumor DNA mutational status in homologous recombination repair, PI3K alteration pathway, and aggressive-variant prostate cancer-associated genes with treatment outcome. Baseline PSMA PET/CT images were analyzed using SelectPSMA, an artificial intelligence algorithm, to predict treatment outcome. Associations with the observed treatment outcome were evaluated. Results: All 76 patients with PSMA-positive mCRPC who received [177Lu]Lu-PSMA-617 met the inclusion criteria. A PSA response was achieved in 30 of 74 (41%) patients. The median PSA PFS was 4.1 mo (95% CI, 2.0-6.2 mo), and the median OS was 13.7 mo (95% CI, 11.3-16.1 mo). Anemia of grade 3 or greater, thrombocytopenia, and neutropenia were observed in 9 (12%), 3 (4%), and 1 (1%), respectively, of 76 patients. Transient xerostomia was observed in 23 (28%) patients. The presence of aggressive-variant prostate cancer-associated genes was associated with a shorter PSA PFS (median, 1.3 vs. 6.3 mo; P = 0.040). No other associations were observed between circulating tumor DNA mutational status and treatment outcomes. Eighteen of 71 (25%) patients classified by SelectPSMA as nonresponders had significantly lower rates of PSA response than patients classified as likely responders (6% vs. 51%; P < 0.001), a shorter PSA PFS (median, 1.3 vs. 6.3 mo; P < 0.001), and a shorter OS (median, 6.3 vs. 14.5 mo; P = 0.046). Conclusion: [177Lu]Lu-PSMA-617 offered in a real-world setting after regulatory approval in the United States demonstrated antitumor activity and a favorable toxicity profile. Artificial-intelligence-based analysis of baseline PSMA PET/CT images may improve patient selection. Validation of these findings on larger cohorts is warranted.