PSMA PET/CT-Derived Indicators and Outcomes After [Lu]Lu-PSMA-617: A Multicenter Retrospective Analysis from the U.S. Expanded-Access Program.

Pretherapeutic visual and quantitative indicators derived from prostate-specific membrane antigen (PSMA) PET/CT have been proposed as predictors of response to [Lu]Lu-PSMA-617 (Lu-PSMA) therapy in patients with metastatic castration-resistant prostate cancer. This study aimed to evaluate and compare the prognostic performance of these indicators in a cohort treated under the U.S. Expanded-Access Program. This retrospective analysis included 88 patients with metastatic castration-resistant prostate cancer from 3 U.S. institutions (University of California Los Angeles, University of California San Francisco, and Johns Hopkins) enrolled in the Expanded-Access Program who underwent baseline PSMA PET/CT before receiving Lu-PSMA. We assessed visual indicators-such as the visual PSMA PET tumor-to-salivary gland ratio, tumor heterogeneity, and intensity scores-and quantitative metrics including total tumor volume (TTV), total tumor SUV, total tumor SUV, total lesion uptake (TTV × total tumor SUV), total lesion quotient (TTV ÷ total tumor SUV), and the quantitative PSMA PET tumor-to-salivary gland score. Associations with clinical outcomes-a 50% or greater prostate-specific antigen decline (PSA50), prostate-specific antigen (PSA) progression-free survival (PFS), and overall survival (OS)-were analyzed using univariate and multivariate models. Predictive performance was evaluated via the concordance index from Cox proportional hazards regression. After a median follow-up of 36.1 mo (95% CI, 33.8-37.8 mo), the PSA50 rate was 43%, the median PSA PFS was 4.5 mo (95% CI, 3.7-7.2 mo), and the 2-y PSA PFS rate was 7% (95% CI, 3-16%). The median OS was 12.5 mo (95% CI, 10.4-17.1 mo), and the 2-y OS rate was 29% (95% CI, 20-40%). Among the evaluated metrics, total tumor SUV showed the highest predictive accuracy for PSA50 (area under the curve, 0.81; 95% CI, 0.73-0.90). In multivariate analyses adjusted for clinical factors, a higher total tumor SUV was independently associated with improved PSA PFS (hazard ratio, 0.58; 95% CI, 0.39-0.86; = 0.007) and OS (hazard ratio, 0.54; 95% CI, 0.34-0.86; = 0.009). Total tumor SUV also yielded higher concordance index values compared with models based on clinical variables alone (PSA PFS, 0.667 vs. 0.594; OS, 0.687 vs. 0.661, respectively). Baseline total tumor SUV on PSMA PET/CT provides independent prognostic information beyond clinical parameters and may serve as a useful biomarker for patient selection and treatment personalization with Lu-PSMA.