MP53-02 IMPACT OF MULTIFOCALITY ON PROSTATE CANCER DETECTION BY MULTI-PARAMETRIC MRI

INTRODUCTION AND OBJECTIVES: Higher Gleason score (GS) is associated with increased level of suspicion on multi-parametric magnetic resonance imaging (mp-MRI) of the prostate, but some tumors still evade detection by MRI. The objective of this study was to characterize the performance of mp-MRI for detection of high GS tumors and identify risk factors for missed tumor detection. METHODS: A retrospective study was performed of 122 consecutive men who underwent mp-MRI prior to radical prostatectomy from 10/2010-2/2013. Clinical, MRI, and pathologic features were obtained. A genitourinary radiologist and pathologist collectively reviewed each case and matched the MRI to whole-mount pathology lesions. Index tumor was defined as the highest GS tumor; in cases of multifocal tumors of identical GS, the largest tumor was considered the index. Chi-square or Fisher’s exact analysis was performed for categorical and t-test for continuous variables. RESULTS: 180 regions of interest were identified by mp-MRI and 283 tumors were histologically confirmed among 122 men. Among the 57 tumors determined to be GS 4+3, 15/57 tumors (26%) were missed by mp-MRI in 11/50 men (22%), corresponding to a sensitivity of 74%. Index tumor status, tumor size, number of whole mount levels involved, and presence of extra-capsular extension were all associated with tumor detection (p<0.05) (table). 37/44 (84%) of index tumors were correctly identified by mp-MRI. Missed tumors were: less likely to be index tumors (47% vs 88% of MRI-pathology concordant tumors); likely to be smaller (1.24 cm 0.78 cm SD vs 2.41 2.06 SD), involving fewer whole-mount levels; and less likely to have extra-capsular extension (13% vs 55%). Of missed non-index tumors, 4/8 (50%) of the corresponding index tumors were detected by mp-MRI. Of the missed index tumors, no corresponding non-index tumors were detected; review of these slides demonstrated tumor infiltrating benign glands (3/7), small volume tumor within large overall prostate volume (3/7), or abundant foamy gland variant pathology (2/7) as possible reasons for missed tumor detection. CONCLUSIONS: High Gleason score tumors evaded mp-MRI detection in 26% of cases. Smaller, non-index, organ-confined tumors were more likely to be missed; there may be a histologic basis for nonvisualization of tumors by mp-MRI. Table Characteristics of tumors missed versus detected by mp-MRI.