Can multiparametric MRI replace Roach equations in staging prostate cancer before external beam radiation therapy?

Purpose To investigate the agreement between Roach equations (RE) and multiparametric magnetic resonance imaging (mpMRI) in assessing the T-stage of prostate cancer (PCa). Materials and methods Seventy-three patients with biopsy-proven PCa and previous RE assessment prospectively underwent mpMRI on a 3.0T magnet before external beam radiation therapy (EBRT). Using Cohen’s kappa statistic, we assessed the agreement between RE and mpMRI in defining the T-stage (≥T3 vs.T ≤ 2) and risk category according to the National comprehensive cancer network criteria (≤intermediate vs. ≥high). We also calculated sensitivity and specificity for ≥T3 stage in an additional group of thirty-seven patients with post-prostatectomy histological examination (mpMRI validation group). Results The agreement between RE and mpMRI in assessing the T stage and risk category was moderate (k = 0.53 and 0.56, respectively). mpMRI changed the T stage and risk category in 21.9% (95%C.I. 13.4–33-4) and 20.5% (95%C.I. 12.3–31.9), respectively, prevalently downstaging PCa compared to RE. Sensitivity and specificity for ≥T3 stage in the mpMRI validation group were 81.8% (95%C.I. 65.1–91.9) and 88.5% (72.8–96.1). Conclusion RE and mpMRI show moderate agreement only in assessing the T-stage of PCa, translating into an mpMRI-induced change in risk assessment in about one fifth of patients. As supported by high sensitivity/specificity for ≥T3 stage in the validation group, the discrepancy we found is in favour of mpMRI as a tool to stage PCa before ERBT.