Multifocality and prostate cancer detection by multiparametric magnetic resonance imaging: correlation with whole-mount histopathology.

Abstract

Multiparametric magnetic resonance imaging (mp-MRI) is increasingly used in prostate cancer (CaP). Understanding the limitations of tumor detection, particularly in multifocal disease, is important in its clinical application. To determine predictors of CaP detection by mp-MRI as confirmed by whole-mount histopathology. A retrospective study was performed of 122 consecutive men who underwent mp-MRI before radical prostatectomy at a single referral academic center. A genitourinary radiologist and pathologist collectively determined concordance. The odds of tumor detection were calculated for clinical, MRI, and histopathologic variables using a multivariate logistic regression model. The 122 patients had 283 unique histologically confirmed CaP tumor foci. Gleason score was 6 in 21 (17%), 7 in 88 (72%), and ≥8 in 13 (11%) patients. Of the 122 cases, 44 (36%) had solitary and 78 (64%) had multifocal tumors. Overall mp-MRI sensitivity for tumor detection was 47% (132/283), with increased sensitivity for larger (102/141 [72%] >1.0 cm), higher-grade (96/134 [72%] Gleason ≥7) tumors, and index tumors (98/122 [80%]). Index tumor status, size, and prostate weight were significant predictors of detection in a multivariate analysis, and multifocality did not adversely impact detection of index tumors. A prostatectomy population was necessary by design, which may limit the ability to generalize these results. Sensitivity for tumor detection increased with tumor size and grade. Index tumor status and tumor size were the strongest predictors of tumor detection, regardless of tumor focality. Some 80% of index tumors were detected, but nonindex tumor detection, even of high-grade lesions, was poor. These findings have important implications for focal therapy. We evaluated the ability of magnetic resonance imaging (MRI) to detect cancer in patients undergoing prostatectomy. We found that tumor size and grade were important predictors of tumor detection, and although cancer is often multifocal, MRI is often able to detect the worst focus of cancer.