The international Intermediate Clinical Endpoints in Cancer of the Prostate (ICECaP) working group established metastasis-free survival (MFS) as a surrogate endpoint for overall survival (OS) for men with localized prostate cancer (PCa). ICECaP required individual patient data which limited their analyses to <25% of all eligible randomized controlled trials (RCTs). Since surrogate endpoints are both disease stage and treatment specific, the performance of MFS and other intermediate clinical endpoints (ICEs) across the landscape of localized PCa patients and treatments is unclear. Herein, we comprehensively assess all commonly reported ICEs in localized PCa. A systematic search utilizing Medline to identify RCTs in localized PCa was performed in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines. Hazard ratios for OS and the ICEs were collected. ICEs included local failure (LF), biochemical failure (BF), distant metastases (DM), BF-free survival (BFS; includes BF, DM, or death), progression-free survival (PFS; LF, DM, or death), and MFS (DM or death) with OS. Candidacy for surrogacy was evaluated using the second condition of the two-stage meta-analytic approach testing whether the treatment effect on the ICE and OS are correlated using R2, weighted by the inverse variance of the log of the ICE hazard ratio. Of 4,217 RCTs screened, 59 trials including 42,272 patients were eligible for analysis. Overall, LF (R2 0.036), BF (R2 0.176), BFS (R2 0.238), and DM (R2 0.364) had poor correlation to the treatment effect on OS. PFS (R2 0.655) had moderate correlation, and MFS (R2 0.781) had the strongest correlation. These trends held true in trials that included or were exclusively in high risk disease. Radiotherapy (RT) RCTs (n = 50, 36,872 patients) showed similar trends with low correlation of LF (R2 0.073), BF (R2 0.334), and BFS (R2 0.229). MFS had stronger correlation to OS (R2 0.777). In RCTs testing use/duration of hormone therapy (n = 26 trials, 24,071 patients), MFS had very strong correlation to OS (R2 0.957). By contrast, in trials evaluating RT dose escalation or fraction size, MFS had lower correlation (R2 0.691). MFS also had strong a correlation to OS (R2 0.947) in patients treated with radical prostatectomy which to date has not been demonstrated. This novel approach corroborates the findings of ICECaP in a substantially larger sampling of RCTs in localized PCa, and supports MFS as the most robust surrogate endpoint for OS across disease stage and treatment type. Importantly, LR and BF have almost no correlation to the treatment effect observed for OS, even in high risk trials, and these ICEs should not be used as surrogate endpoints for phase 3 randomized trials.