RESUMO
PURPOSE: The five grade group system has been validated for men treated with radical prostatectomy. However, the prognostic value for men treated with radiation therapy is uncertain, with prior studies utilising old techniques and doses. We aimed to validate the International Society of Urological Pathology (ISUP) groupings for men treated with contemporary radiation therapy. METHODS: Men with localised prostate cancer treated with image-guided, dose-escalated (≥78 Gy) external beam radiation were identified across four institutions. Primary outcome was time to biochemical failure. Harrell's C index assessed performance of the ISUP system against other grading stratifications. RESULTS: 2205 men were included, withmedian follow-up of 5.6 years. Seven-year actuarial rates of biochemical failure for grade groups 1-5 were 9.3%, 10.4%, 13.2%, 12.4% and 23.4%. On multivariate analysis, hazard ratios for biochemical failure were1.19, 1.00, 1.10, 1.05 and 2.10 for grade groups 1-5, relative to 2. P values were only significant for grade group 5. Harrell's C index favoured an alternative three group model (comprising Gleason scores [6 and 3 + 4 = 7] vs [4 + 3 = 7 and 8] vs [9 and 10]) over ISUP grade groups. CONCLUSIONS: The ISUP grade groups were not validated in a contemporary cohort treated with dose-escalated, image-guided radiation therapy. Grade groups 1-4 were not statistically different from each other; however, grade group 5 had a significantly worse prognosis. We identified a new three group model that better predicted biochemical outcomes. Further work is requiredto validate optimal groupings for modern radiation therapy and investigate the contrasting prognostic capability of grade groups in surgical and radiation therapy patients.