Impact of race on selecting appropriate patients for active surveillance with seemingly low-risk prostate cancer.

OBJECTIVE To evaluate active surveillance (AS) criteria on their ability to predict favorable pathology at prostatectomy within a low-risk African American (AA) cohort. METHODS The sensitivity, specificity, positive predictive value, receiver operator curve, and area under the curve (AUC) were compared for 6 published AS criteria (National Comprehensive Cancer Network; Memorial Sloan-Kettering Cancer Center; Prostate Cancer Research International: Active Surveillance Study; Johns Hopkinse-Epstein; University of California at San Francisco; and University of Miami) to predict organ-confined Gleason score 6 disease at prostatectomy in AAs and white Americans (WAs) with low-risk cancer. We also compared clinical parameters for AAs with favorable prostatectomy pathology with those for AAs with unfavorable pathology, and then used preoperative variables associated with unfavorable pathology as an additional exclusion criteria for AS. RESULTS Of 468 patients with low-risk disease, 308 of 402 (76.6%) WAs and 55 of 66 (83.3%) AAs were eligible for AS by one or more criteria (P =.23). For WAs, Prostate Cancer Research International: Active Surveillance Study criteria had the highest rate of favorable pathology (81.7%) and the best performance (AUC = 0.70) in determining appropriate candidates for AS. However, all 6 AS criteria performed poorly for AA patients, with all AUCs <= 0.52. When comparing AAs with favorable pathology with AAs with unfavorable pathology, only family history of prostate cancer was statistically significant (11 of 25 [47.8%] vs 8 of 41 [22.2%]; P = .04). When adjusting AS criteria in AAs to exclude those with a positive family history, the AUC increased most for the University of California at San Francisco (from 0.52 to 0.6) and Memorial Sloan-Kettering Cancer Center criteria (from 0.50 to 0.58). CONCLUSION Current criteria underperform in appropriately selecting AAs for AS. AAs considering AS should be counseled about their increased risk for occult adverse pathology, particularly if a family history of prostate cancer is present. (C) 2015 Elsevier Inc.