I. Sonni, Z. Mohammadigoldar, G. Furey, K. Kimura, P. C. Movva, A. Holzgreve, S. Bahri, T. Sevco, A. Chung, A. Sisk, T. Grogan, J. Czernin, K. Sung, J. Calais, S. E. Eggener, W. Brisbane, R. E. Reiter, S. Raman, A. B. Weiner
Background: The incremental value of PSMA PET/CT over multiparametric MRI (mpMRI) for selecting patients for prostate cancer focal therapy is undefined. Unlike lesion-detection studies, we evaluated whether PSMA PET/CT, mpMRI, and their combination correctly determine focal therapy eligibility when validated against whole-mount histopathology (WMHP). Methods: We retrospectively analyzed patients undergoing radical prostatectomy (RP) at UCLA (2017-2024) with preoperative PSMA PET/CT and mpMRI, biopsy grade group (GG) [≤]3, unilateral disease, and prostate-specific antigen <20 ng/mL. Two blinded readers per modality assessed extraprostatic extension (EPE), seminal vesicle invasion (SVI), bilateral disease, and multifocality; a third adjudicated discordances. Eligibility required absence of all four. On WMHP, ineligibility was EPE, SVI, multifocal or bilateral disease with a contralateral or secondary GG[≥]2 nodule, or any GG>3 focus at RP. Modalities were compared by McNemar tests. Results: Of 112 patients, 77 (69%) were ineligible on WMHP, including 19 (17%) with a GG>3 focus at RP. Sensitivity and specificity were 53% and 54% for PET and 70% and 49% for mpMRI. Combined imaging improved sensitivity to 83% (P<0.001 vs PET; P=0.002 vs MRI) but reduced specificity to 26%. Per 100 patients at this prevalence, this would correctly identify approximately 9 additional ineligible patients while falsely excluding approximately 7 true candidates. mpMRI was more sensitive than PET for EPE (77% vs 32%, P<0.001), whereas PET was more specific (90% vs 58%, P<0.001). Combined imaging improved detection of multifocal (44% vs 23%, P=0.001) and bilateral disease (45% vs 23%, P=0.002) versus MRI alone. Conclusions: Adding PSMA PET/CT to mpMRI significantly improves sensitivity for exclusionary features, reflecting complementary molecular and anatomic information. However, combined imaging still misses a substantial proportion of adverse pathology, so imaging alone should not determine eligibility.