Improving risk stratification in patients with prostate cancer managed by active surveillance: a nomogram predicting the risk of biopsy progression

Abstract

To develop a clinical tool that integrates different risk factors and provides individual predictions of the risk of biopsy progression in patients with prostate cancer managed by active surveillance. Our analysis included 205 patients on active surveillance, each of whom had had at least two surveillance biopsies. We used the Cox proportional hazard regression model to analyse the association between different risk factors and progression-free survival over successive biopsies. This multivariate model was then used to develop a nomogram. Discrimination and calibration of the nomogram were internally validated using 200 bootstrap resamplings. The median follow-up of patients free of progression was 4.6 years. A total of 58 (28%) patients experienced progression. Factors significantly associated with progression were: overall number of positive cores in the diagnostic and first surveillance biopsies, race and prostate-specific antigen density. The bootstrapping concordance index of the nomogram including these variables was 81%. The nomogram tended to underestimate the probability of progression but it identified fairly accurately the distinct groups of patients at low, intermediate and high risk of progression. In the development cohort, the nomogram was able to separate patients with respect to their risk of biopsy progression. Since accurate risk stratification is essential to optimize patient care, this tool, if external validation confirms its performance, may prove useful for both the counselling and management of patients with low-volume, Gleason 6 prostate cancer.