On prostate biopsy, multiparametric magnetic resonance imaging (mpMRI) and the Prostate Health Index (PHI) have allowed prediction of clinically significant prostate cancer (csPCa).
To predict the likelihood of csPCa, we created a nomogram based on a multivariate model that included PHI and mpMRI. We assessed 315 males who were scheduled for prostate biopsies.
We used the Prostate Imaging Reporting and Data System version 2 (PI-RADS V2) to assess mpMRI and optimize PHI testing prior to biopsy. Univariate analysis showed that csPCa may be identified by PHI with a cut-off value of 77.77, PHID with 2.36, and PI-RADS with 3 as the best threshold. Multivariable logistic models for predicting csPCa were developed using PI-RADS, free PSA (fPSA), PHI, and prostate volume. A multivariate model that included PI-RADS, fPSA, PHI, and prostate volume had the best accuracy (AUC: 0.882). Decision curve analysis (DCA), which was carried out to verify the nomogram’s clinical applicability, showed an ideal advantage (13.35% higher than the model that include PI-RADS only).
In conclusion, the nomogram based on PHI and mpMRI is a valuable tool for predicting csPCa while avoiding unnecessary biopsy as much as possible.