Benefits and harms of the new prostate cancer grade grouping on the prediction of long‐term oncological outcomes in patients after radical prostatectomy

Abstract

To investigate whether the new prostate cancer grade groups model provides significant predictive value and better patient stratification on tumor progression after radical prostatectomy compared with the former Gleason grading models. Methods: Men treated at a tertiary center by radical prostatectomy between 2005 and 2017 were analyzed. The outcomes of interest were clinical progression-free and cancerspecific survival. Multivariate Cox regression analysis, C-index and decision curve analysis were carried out using three-tier (Gleason score 6, 7 and 8-10), four-tier (Gleason score 6, 7, 8 and 9-10) and new grade groups model. Results: In total, 1759 men were included in the analysis. At a median of 87 months (interquartile range 51-134 months) of follow up, clinical progression was detected in 78 (4.4%) and cancer-related death in 42 (2.4%) patients. The hazard ratio of clinical progression-free was 2.3, 5.7, 5.2 and 29.5; the hazard ratio of cancer-specific survival was 1.7, 3.2, 4.8 and 11.8 in the grade groups 2-5, relative to grade group 1, respectively. The grade groups model had higher C-index in comparison with four-and three-tier grading models for clinical progression-free survival 0.88 versus 0.85 versus 0.83 and for cancer-specific survival 0.82 versus 0.80 versus 0.80, respectively. In the decision curve analysis, the grade groups model shows marginally better net benefit on clinical progression-free and cancer-specific survival.Conclusions: The new model shows better performance in comparison with former Gleason grading models on the prediction of long-term oncological outcomes.