Prognostic Significance of the Risk of Non-localized Disease on PSMA/PET: Comparative Performance of a Novel, PSMA/PET-Derived Risk Stratification Tool for High-Risk Prostate Cancer in a Large, Multi-Institutional Cohort

Abstract

<h3>Purpose/Objective(s)</h3> Occult non-localized disease (N1 or M1) at the time of definitive treatment of high-risk prostate cancer (HR-PCa) underlies most treatment failures. We recently developed a nomogram tuned to the identification of non-localized disease on prostate specific membrane antigen PET/CT (PSMA PET/CT) in patients with cN0M0 HR-PCa by conventional imaging, which incorporates iPSA, percent positive core on biopsy, Gleason Grade and cT stage. We systematically compared the prognostic performance of the PSMA nomogram against other validated risk stratification tools for HR-PCa. <h3>Materials/Methods</h3> We identified 5275 men from a multi-institutional database treated with radical prostatectomy (55%), external beam radiation therapy (EBRT, 31%), or EBRT and brachytherapy (14%). The prognostic performance of the PSMA nomogram was compared against the STAR-CAP score, the Cancer of the Prostate Risk Assessment (CAPRA) score, Cambridge Prognostic Groups (CPG) risk group, and Memorial Sloan Kettering Cancer Center (MSKCC) nomogram of 5-year recurrence probability. Accuracy was determined for 5- and 10-year distant metastasis (DM) and prostate cancer specific mortality (PCSM) using the concordance index (C-index). <h3>Results</h3> Median follow-up was 50.3 months (interquartile range 28.1-84.5 months). Overall, 16.1% developed DM, and 4.6% died from PCa. C-indices for the PSMA nomogram were 0.67 and 0.67 for 5- and 10-year DM, and 0.69 and 0.73 for 5- and 10-year PCSM, respectively. The PSMA nomogram outperformed CAPRA, CPG, and MSKCC for all endpoints, outperformed STAR-CAP for 10-year DM, and performed similarly to STAR-CAP for all other endpoints (see Table). The comparative performance between the tools was generally maintained at different follow-up times and in subgroup analyses stratified by primary treatment modalities. <h3>Conclusion</h3> A nomogram tuned to the detection of non-localized disease on PSMA/PET in HR-PCa patients is significantly prognostic of important clinical endpoints, with performance comparable to STAR-CAP and superior to other tools. These data strongly suggest that upstaging based on PSMA PET/CT provides clinically relevant information and warrant prospective validation.