Identification of pathologically favorable disease in intermediate-risk prostate cancer patients: Implications for...

Intermediate-risk prostate cancer (PCa) represents a heterogeneous disease, where a non-negligible proportion of patients harbor favorable pathologic characteristics and are potentially eligible for active surveillance (AS). We aimed at developing a model for the identification of pathologically favorable PCa at radical prostatectomy (RP) among intermediate-risk patients.

Overall, 3,821 intermediate-risk patients treated with RP at two centers between 2005 and 2013 were identified. Pathologically favorable PCa was defined as low-grade organ-confined disease. Age, biopsy Gleason, PSA density (PSAD), and the percentage of positive cores were included in multivariable logistic regression analyses predicting favorable PCa and formed the basis for a logistic regression-based risk calculator. The internally validated discrimination and calibration of the risk calculator were quantified using 200 bootstrap resamples. Decision curve analysis (DCA) provided an estimate of the net benefit obtained using this model versus treating no one and treating everyone.

Overall, 10.0% of all intermediate risk patients had favorable disease. In multivariable analyses, patients with biopsy Gleason score ≤6 had higher probability of favorable disease compared to those with higher-grade disease (P < 0.001). Similarly, age, PSAD, and percentage of positive cores were associated with the probability of favorable disease (all P ≤ 0.01). The risk calculator achieved a validated accuracy of 82.5%. The DCA showed that our prediction model is better than both treating no one and treating everyone.

One out of ten intermediate-risk patients harbors favorable disease at RP. Our novel, pre-operative, validated risk calculator may help clinicians identifying patients who could be considered for conservative therapy approaches such as AS.