PurposeThe outcome of the present study is to determine variables available at the time of diagnosis able to predict disease reclassification in prostate cancer (PCa) patients on active surveillance (AS).Materials and methodsFrom January 2014 to December 2018, 114 consecutive low-risk PCa patients were enrolled in AS protocol according to inclusion criteria: PSA ≤ 10 ng/ml, Gleason score (GS) ≤ 6 or International Society of Urological Pathology (ISUP) Gleason grade group (GG) 1, maximum cancer core length (MCCI) < 50%, and ≤ 2 positive cores on biopsy. Patients were followed with confirmatory and yearly prostate biopsy, semi-annually with prostate-specific antigen (PSA), and digital rectal examination (DRE). Disease reclassification was defined as upgrading biopsy: GS ≥ 3 + 4 = 7 or ISUP GG ≥ 2, more than two positive cores, MCCI > 50%, or changes in serum PSA > 10 ng/ml. Uni- and multivariate Cox proportional hazards regression models, receiver performance curves (ROC), and Kaplan-Meier analysis were performed to characterize AS criteria and identify variables that predict disease reclassification. Finally, decision curve analysis (DCA) was performed to evaluate the net benefit of using PV in addition to standard variables to predict disease reclassification.ResultsPCa was diagnosed by systematic transrectal ultrasound-guided prostate biopsy (TRUS-Bx). The mean (range) follow-up was 32.7 (12-126) months. Disease reclassification occurred in 46 patients (40%). On univariate statistical analysis prostate specific antigen (PSA) (p = 0.05), prostate volume (PV) (p = 0.022), PSA density (PSAD) (p < 0.001) and number of positive cores (p = 0.021) were significant factors for disease reclassification. On the multivariate analysis, PSAD (p < 0.001) and PV (p = 0.003) were the only statistically significant independent variables to predict disease reclassification. A PSAD cut-off of 0.16 ng/ml² and a PV cut-off of 44 ml gave a maximal area under the curve, 0.69 and 0.63, respectively. Kaplan-Meier analysis showed that the median survival free from disease reclassification during AS was almost doubled in patients with PSAD < 0.16 ng/ml2 or PV > 44 ml. DCA showed a positive net benefit and clinical usefulness of the model, including PV, to predict disease reclassification between threshold probabilities of 20-50%.ConclusionsPV and PSAD significantly predicted failure from AS in our patients. Patients with a baseline PV of fewer than 44 ml would be more likely to have disease reclassification and unsuitable for acceptable AS protocols. Therefore, we believe that PV may help to select PCa patients for AS, especially in populations where the use of mpMRI is limited.