Abstract 737: Variation in prognostic value of Gleason score for lethal prostate cancer using time-varying area under the curve

Background: Gleason score is one of the strongest prognostic factors for prostate cancer survival in patients with primary prostate cancer. However, its predictive accuracy is less clear during extended follow-up periods. Methods: We included men diagnosed with prostate cancer in the Physicians’ Health Study (PHS) and Health Professionals Follow-up Study (HPFS) between 1982 and 2009. We included individuals with available radical prostatectomy tissue specimens and undertook a central standardized histopathologic re-review of H&E slides including Gleason score according to the 2014 International Society of Urological Pathology Consensus. Comparing pathologic Gleason score, we estimated hazard ratios (HR) and 95% confidence intervals (CIs) for lethal prostate cancer-free survival from the time of diagnosis using the Cox proportional hazards regression through 2018. Lethal prostate cancer was defined as development of distant metastases or prostate cancer-specific death. We assessed the time-dependent predictive accuracy of Gleason score over 10, 15, 20, and 25-years post-diagnosis with time-varying area under the curve (tAUC) using the Uno estimator. Results: Among the 1,404 prostate cancer patients, 113 developed lethal prostate cancer during a follow-up of up to 30 years (median 17.1 years). There were 11% men with Gleason score of 9-10 (grade group 5), 7.1% with 8 (grade group 4), 23% with 4+3 (grade group 3), 36% with 3+4 (grade group 2), and 23% with 6 (grade group 1). Compared to Gleason score 6, the HRs for lethal disease were 5.6 (95% CI: 1.3 to 24) for score 3+4, 13 (95% CI: 3.1 to 55) for 4+3, 17 (95% CI: 3.8 to 75) for 8, and 27 (95% CI: 6.5 to 117) for 9-10, adjusting for pathological tumor stage, age at diagnosis, and prostate-specific antigen (PSA) level at diagnosis. The overall AUC over the entire follow-up was 0.71 (95% CI: 0.63 to 0.79) for Gleason score alone and 0.76 (95% CI: 0.68 to 0.83) with the addition of stage, PSA, and age. Discrimination for lethal prostate cancer based on Gleason score alone varied over follow-up, peaking at a tAUC of 0.81 (95% CI: 0.77 to 0.85) at 10 years, and slowly declined to 0.78 (95% CI: 0.74 to 0.83) at 15 years, 0.78 (95% CI: 0.74 to 0.81) at 20 years, and 0.74 (95% CI: 0.69 to 0.80) at 25 years post-diagnosis. The tAUC had a similar trend with the addition of stage, PSA, and age, with higher point estimates at each time point. Conclusions: The prognostic value of Gleason score for lethal prostate cancer