ABSTRACT
Accurate prostate cancer (PCa) patient diagnosis and risk assessment are key to ensure the best outcome. Currently, low- and favorable intermediate-risk PCa patients may be offered AS due to the indolent nature of the disease. Nonetheless, deciding between active surveillance and curative-intent treatment remains an intricate task, as a subset of these patients may eventually progress, enduring poorer prognosis. Herein, we sought to construct risk calculators based on cancer biomarkers, enabling more accurate discrimination among patients which may benefit from active interventions.Ki67 immunoscore, GSTP1 and KLF8 promoter methylation levels (me) were assessed in PCa tissues. Study endpoints included overall and biochemical recurrence-free (BCR) survival. Combination with relevant clinicopathological parameters allowed for construction of graphical calculating tools (nomograms).Higher Ki67 index correlated with worse BCR-free survival, whereas higher KLF8me levels were associated with improved overall survival, especially in patients with lower-grade tumors. GSTP1me levels had no prognostic value. Among prognostic models tested, a BCR-risk calculator - ProstARK (including Ki67 and clinicopathologic parameters) - disclosed 79.17% specificity, 66.67% sensitivity, 55% positive predictive value, 86% negative predictive value, and 75.76% accuracy. Similar results were found using an independent PCa biopsy cohort, validating its prognostication ability.Combining clinicopathologic features and Ki67 index into a risk calculator enables easy and accurate implementation of a novel PCa prognostication tool. This nomogram may be useful for a more accurate selection of patients for active surveillance protocols. Nonetheless, validation in a larger, multicentric, set of diagnostic PCa biopsies is mandatory for further confirmation of these results.