Introduction

The wide diffusion of multiparametric magnetic resonance imaging (MRI) has dramatically modified the scenario of prostate cancer (PCa) diagnosis. The detection rate of MRI-ultrasound (US) fusion biopsy increased as well as the need for an extended prostate biopsy sampling with saturation biopsy decreased. The aim of this study was to develop, internally validate and calibrate a nomogram to predict the probability of detecting a prostate cancer.

Methods

Prospectively collected data from 3 tertiary referral center series of 475 consecutive patients who underwent MRI-US fusion biopsy using the Koelis system were used to build the nomogram. A logistic regression model is created to identify predictors of PCa diagnosis with MRI-US fusion biopsy. Predictive accuracy was quantified using the concordance index (CI). Internal validation with 200 bootstrap resampling and calibration plot were performed.

Results

Mean age was 66.3 yrs (±7.98) and mean PSA levels were 9.8 ng/mL(±7.98). The overall PCa detection rate was 57.4%._x000D_ Age, PSA serum levels, PIRADS score at MRI report, number of targeted and number of systematic cores taken were included in the model (Figure 1).Predictive accuracy was 0.82. On internal validation the CI was 0.81 and predicted probability was well calibrated (Figure 2)._x000D_ Limitations include the lack of external validation and the absence of patients with races different by Caucasian in the development cohort._x000D_

Conclusions

This nomogram provides a high accuracy in predicting the probability of PCa diagnosis with MRI-US fusion biopsy. This is an easy to use clinical tool that physicians may use for patients counselling purposes.

Funding

none