5079 Background: 177 Lu-PSMA-617 (Lu-PSMA) contributes to prolong progression-free survival (PFS) and overall survival (OS) in metastatic castration-resistant prostate cancer (mCRPC) patients who progressed after chemotherapy. Pretherapeutic prostate-specific membrane antigen (PSMA) PET/CT information can be used to predict Lu-PSMA therapy response patterns and outcomes, with various quantitative and visual methods proposed. We aimed to test various proposed PSMA PET/CT-derived outcome predictors in the U.S. expanded-access program (EAP) cohort. Methods: Patients enrolled in the EAP (NCT04825652) for Lu-PSMA at 3 institutions with available pretherapeutic PSMA PET/CT and outcomes were included in this analysis. Quantitative analysis was performed for all tumor lesions on PSMA PET/CT with semi-automatically contouring. Total tumor volume (TV), total tumor SUVmean, total tumor SUVmax, total lesion uptake (TLU = TV * SUVmean), total lesion quotient (TLQ = TV / SUVmean), and quantitative PSMA PET tumor–to–salivary gland ratio (qPSG: high, ≥ 1.5; intermediate, 0.5–1.5; low, ≤ 0.5) were calculated for each patient. For visual analysis, visual PSG (vPSG: high, most of the lesions showed higher uptake than the parotid glands; intermediate, neither low nor high; low, most of the lesions showed lower uptake than the parotid glands) and heterogeneity and intensity of tumors (HIT: 1, SUVmax < 15; 2, 15–79 with heterogeneous intensity; 3, 15–79 with homogeneous intensity; 4, ≥ 80) scores were used for assessment. Outcomes included a prostate-specific antigen (PSA) PFS, and OS. We evaluated the predictive performance of each model using Cox proportional hazards regression analysis and assessed their performance with the concordance index (c-index). Results: In total, 88 patients who received Lu-PSMA within the EAP between May 2021 and March 2022 were eligible and included in this analysis. For the PSA PFS, the total tumor SUVmean achieved the highest c-index of 0.678 (HR 0.91 [95% CI, 0.85–0.97], p = 0.004), followed by the total tumor SUVmax with a c-index of 0.640 (HR 0.99 [95% CI, 0.99–1.00], p = 0.034). For OS, the TLQ achieved the highest c-index of 0.658 (HR 1.01 [95% CI, 1.00–1.01], p < 0.001), followed by the total tumor SUVmean with a c-index of 0.634 (HR 0.89 [95% CI, 0.83–0.96], p = 0.004). The HIT score showed the third highest c-index of 0.632; however, when using score 1 as the reference, the HR did not exhibit a sequential trend across ordinal categories as anticipated. Conclusions: Quantitative analysis outperformed visual analysis in predicting the outcome of mCRPC with Lu-PSMA therapy in the EAP cohort. Total tumor SUVmean was identified as the most robust predictor for PSA PFS, while TLQ showed promise as a predictor for OS. Incorporating these predictors into clinical decision-making for pre-Lu-PSMA therapy could aid in patient selection and treatment planning. Clinical trial information: NCT04825652 .