Contemporary evaluation of the National Comprehensive Cancer Network prostate cancer risk classification system.

OBJECTIVE To analyze the National Comprehensive Cancer Network prostate cancer guidelines pretreatment risk groups in a contemporary series of patients treated with radical prostatectomy. METHODS We analyzed our institutional radical prostatectomy database, including all patients with clinically localized disease treated from 2000 to 2010. Using the National Comprehensive Cancer Network guidelines, the patients were classified into low-, intermediate-, or high-risk groups. The pathologic outcomes were assessed, and the biochemical recurrence (BCR)-free survival rates were calculated and compared using the log-rank test and Cox proportional hazards analysis. RESULTS A total of 12 821 men met the inclusion criteria. The pathologic and 10-year BCR-free survival rates differed significantly by risk group (low risk, 92.1%; intermediate risk, 71.0%; and high risk, 38.8%; P < .01). Among the intermediate-risk men, the 10-year BCR-free survival was significantly greater for men assigned to the intermediate-risk group by clinical stage (88.8%) than for those deemed intermediate risk by the Gleason score (73.6%) or prostate-specific antigen (PSA) level (79.5%; P = .01). Likewise, in the high-risk men, a trend was seen toward improved 5-year BCR-free survival for patients with clinical stage T3a tumors (77.8%) compared with those considered high risk because of the Gleason score (53.7%) or PSA level (41.0%; P = .13). On multivariate analysis, clinical stage, Gleason score, and PSA level were all significantly associated with BCR. CONCLUSION We observed heterogeneous outcomes among patients within the National Comprehensive Cancer Network intermediate- and high-risk groups. The BCR-free survival rates were superior for men with an advanced clinical stage compared with those with an advanced Gleason score or elevated PSA level. This within-group heterogeneity must be considered when choosing the treatment modality and predicting an individual patient's prognosis.