Background: There is no consensus on how to best treat patients (pts) with high-risk prostate cancer. Methods: The outcomes for 2,736 high-risk prostate cancer pts treated with radical prostatectomy (RP), external beam radiotherapy (RT), and I-125 brachytherapy (BT) at a single institution from 1996 to 2012 were reviewed. The majority of RT pts were treated prior to 2002 because of our preference for RP and BT over time. High-risk was defined per the NCCN criteria. The outcomes assessed were biochemical failure (bF), clinical failure (cF), and prostate cancer mortality (PCM). Results: The distribution by treatment was RP 54%, RT 27%, and BT 19%. The median follow up for all pts was 4.6 years (y) (range 0.1-19.5): 3.8 y (0.1-18.7) for RP, 7.7 y (0.1-19.4) for RT, and 4.1 y (0.1-16.8) for BT pts. No patient received RT+BT, and 44% received androgen deprivation therapy (ADT). On multivariable analysis (see table) RP pts were at higher risk for bF vs. RT; BT pts and RT pts were at higher risk for cF vs. RP; and RT pts were at higher risk for PCM vs. RP. All multivariable analyses were adjusted for clinical stage, biopsy Gleason score, pre-treatment PSA, and duration of ADT. Conclusions: RP is associated with worse bF but better cF and PCM. There is no difference between BT and RT for bF, cF, or PCM while BT and RP had similar PCM. These outcomes may be a result of selection bias or differences in follow up time among the three treatment arms so no demonstration of modality superiority is possible.