INTRODUCTION: We aimed to test the impact of International Society of Urological Pathology (ISUP) grade group (GG) on cancer-specific mortality (CSM) in organ-confined (pT2) prostate cancer (PCa) at radical prostatectomy (RP). METHODS: RP organ-confined PCa patients were identified (Surveillance, Epidemiology, and End Results [SEER] 2004-2015). Cancer-specific survival (CSS) rates were tested in Kaplan-Meier plots and multivariable Cox regression (MCR) models according to GG: 1-3 vs. 4 vs. 5. Sensitivity analyses addressed GG4 and GG5 patients with available primary and secondary Gleason score (GS). RESULTS: Overall, 61 172 patients with RP organ-confined PCa were identified. Of these, 57 715 (94.4%), 2036 (3.3%) and 1421 (2.3%) harbored GG1-3, 4, and 5, respectively. In Kaplan-Meier analyses, seven-year' CSS estimates were 99.6 vs. 98.2 vs. 93.8% for GG1-3 vs. 4 vs. 5, respectively (p<
0.002). In MCR models, GG4 (hazard ratio [HR] 2.72, p<
0.002) and 5 (HR 9.95, p<
0.002) independently predicted higher CSM, relative to GG1-3. Furthermore, GG5 also independently predicted higher CSM (HR 3.72, p<
0.002) vs. GG4. In sensitivity analyses, 1.2, 1.6, and 2.4 CSM events per 1000 person-years of followup were respectively recorded for GS 4+4, 3+5, and 5+3 patients. Conversely, 4.8 vs. 5.3 CSM events per 1000 person-years of followup were respectively recorded for GS 4+5 vs. 5+4/5+5 patients. CONCLUSIONS: In organ-confined PCa, at RP, a small proportion of patients harbor GG4-5. These patients exhibit higher CSM than their GG1-3 counterparts. Moreover, detectable mortality rate differences indicate a dose-response effect according to primary and secondary GS. This phenomenon applies in both GG4 and GG5, as well as between GG4 and GG5.