Introduction: Magnetic resonance imaging (MRI) of the prostate has been incorporated as a tool to confirm eligibility for active surveillance (AS). The impact of radiologist expertise on MRI diagnostic performance has been widely demonstrated. However, the increased number of centers performing MRI has increased the variability of both quality and performance. In this study, we assessed the diagnostic accuracy of MRI performed at a single high volume center vs external centers for patients under AS. Methods: We identified 535 men on AS at our Institution with baseline MRI and at least one surveillance biopsy (sBx). MRIs were stratified according to centers performing the scan (high volume single center performing > 500 MRIs per year vs external centers) and to the detection of visible lesions (negative: PI-RADS 1-2, positive: PI-RADS 3-5). Reclassification was defined as increase in Gleason score at sBx, or higher disease volume ( >33% of involved cores, or >50% of a single core involved). Reclassification-free survival (RFS) was estimated with the Kaplan-Meier method. A Cox model tested the prognostic impact of the center performing MRI according to lesion visibility, after adjusting for covariates. Results: Median age was 65 years (IQR 59-70), 52% (n=280) and 48% (n=255) of patients underwent baseline MRI at our Institution (internal) vs external Institutions. A lower proportion of negative scans was observed among the external MRI scans (18% vs 35%, p<0.001). After a median follow-up of 43 months, 47% (n=253) of men were reclassified. The 5-year RFS was 76% vs 40% for internal vs external negative MRI scans (p < 0.001). A negative external MRI was associated with a 1.98-fold increased risk of reclassification compared to a negative internal MRI (95% CI 1.07-3.65, p=0.03). In contrast, the 5-year RFS was 37% vs 35% for internal vs external positive MRI scans, respectively (p=0.2). Conclusions: We demonstrated that the negative predictive value of MRI scans performed at high volume centers is significantly higher compared to MRIs performed externally. Our findings indirectly support the need for centralization of MRIs in men on AS, as well as the invariable need of follow-up biopsies for those with negative MRIs performed externally. SOURCE OF Funding: None.
