Background: ¹⁷⁷Lu-PSMA-617 is approved in adults with PSMA+ metastatic castration-resistant prostate cancer (mCRPC). This analysis of VISION data sought to build predictive models for clinical outcomes after ¹⁷⁷Lu-PSMA-617. Methods: In VISION, adults with PSMA+ mCRPC received ¹⁷⁷Lu-PSMA-617 (7.4 GBq every 6 weeks, ≤6 cycles) plus protocol-permitted standard of care (SoC) or SoC alone. In this post hoc analysis, 29 baseline parameters were assessed for prognostic and predictive value on overall survival (OS), radiographic progression-free survival (rPFS) and PSA response (≥50% reduction; PSA50). Parameters associated with outcome regardless of treatment were prognostic, and those associated with outcome after treatment with ¹⁷⁷Lu-PSMA-617+SoC vs SoC were predictive. Cox proportional hazards (OS and rPFS) or logistic regression (PSA50) models were used in univariate analyses for parameter selection and to build statistical models. Here, univariate analyses are presented. Multiplicity was corrected by use of q values (α = 0.05). Results: Data from 831 adults were analyzed. In initial univariate analyses, 76%, 69% and 38% of parameters assessed were prognostic for OS, rPFS and PSA50, respectively; fewer were predictive of better outcomes after ¹⁷⁷Lu-PSMA-617+SoC vs SoC (subset shown). Conclusions: Baseline prognostic and predictive parameters were identified for OS, as well as for rPFS and PSA50, in adults with mCRPC in VISION receiving ¹⁷⁷Lu-PSMA-617+SoC or SoC alone. For clinical application, select parameters will be reassessed categorically, and multivariate predictive models will be developed using identified parameters and parameter shrinkage.

Parameter treated as *categorical or ^†continuous.

For prognostic parameters, symbols denote correlation of higher/positive values with worse outcome (‘–‘) or with better outcome (‘+’).

For predictive parameters, symbols denote correlation of higher/positive values with decreased (‘–‘) or increased (‘+’) effect of ¹⁷⁷Lu-PSMA-617 vs SoC.