Multiparametric MRI is not sufficient for prostate cancer staging: A single institutional experience validated by a multi-institutional regional collaborative

Objective: Multiparametric magnetic resonance imaging (mpMRI) has been increasingly utilized in prostate cancer (CaP) diagnosis and staging. While Level 1 data supports MRI utility in CaP diagnosis, there is less data on staging utility. We sought to evaluate the real-world accuracy of mpMRI in staging localized CaP. Materials and Methods: Men who underwent radical prostatectomy (RP) for CaP in 2021 at our institution were identified. Sensitivity, specificity, positive predictive value and negative predictive value of mpMRI in predicting pT2N0 organ confined disease, extracapsular extension, seminal vesicle invasion, lymph node involvement, and bladder neck invasion were evaluated. Associations between MRI accuracy and AUA risk stratification (AUA RS), MRI institution (MRI-I), MRI strength (1.5 vs. 3T) (MRI-S), and MRI timing (MRI-T) were assessed. These analyses were repeated using Pennsylvania Urologic Regional Collaborative (PURC) data. Results: Institutional and community mpMRI CaP staging data demonstrated poor sensitivity (2.9%−49.2%% vs. 16.8%−24.4%), positive predictive value (40%−100% vs. 35.8%−68.2%), and negative predictive value (56.3%−94.3% vs. 68.4%−96.2%) in predicting surgical pathologic features – in contrast, specificity (89.1%−100% vs. 93.9%−98.6%) was adequate. mpMRI accuracy for extracapsular extension, seminal vesicle invasion, and lymph node involvement was significantly (p < 0.001) associated with AUA RS. There was no association between mpMRI accuracy and MRI-I, MRI-S, and MRI-T. Conclusion: Despite enthusiasm for its use, in a real-world setting, mpMRI appears to be a poor staging study for localized CaP and is unreliable as the sole means of staging patients prior to prostatectomy. mpMRI should be used cautiously as a staging tool for CaP, and should be interpreted considering individual patient risk strata.