Abstract
OBJECTIVE: To identify preoperative factors predicting Gleason score downgrading after radical prostatectomy (RP) in patients with biopsy Gleason score 3+4 prostate cancer and to determine if prediction of downgrading can identify potential candidates for active surveillance (AS). PATIENTS AND METHODS: We identified 1317 patients with biopsy Gleason score 3+4 prostate cancers who underwent RP at the Memorial Sloan-Kettering Cancer Center between 2005 and 2013. Several preoperative and biopsy characteristics were evaluated by forward selection regression, and selected predictors of downgrading were analysed by multivariable logistic regression. Decision curve analysis was used to evaluate the clinical utility of the multivariate model. RESULTS: Gleason score was downgraded after RP in 115 patients (9%). We developed a multivariable model using age, prostate-specific antigen density, percentage of positive cores with Gleason pattern 4 cancer out of all cores taken, and maximum percentage of cancer involvement within a positive core with Gleason pattern 4 cancer. The area under the curve for this model was 0.75 after 10-fold cross validation. However, decision curve analysis revealed that the model was not clinically helpful in identifying patients who will downgrade at RP for the purpose of reassigning them to AS. CONCLUSION: While patients with pathological Gleason score 3 + 3 with tertiary Gleason pattern ≤4 at RP in patients with biopsy Gleason score 3 + 4 prostate cancer may be potential candidates for AS, decision curve analysis showed limited utility of our model to identify such men. Future study is needed to identify new predictors to help identify potential candidates for AS among patients with biopsy confirmed Gleason score 3 + 4 prostate cancer.