4655 Background: Many patients with biochemical relapse after definitive therapy for prostate cancer receive androgen deprivation therapy. PSA levels eventually will rise despite castrate levels of testosterone. Many of these patients have nonmetastatic disease and do not develop metastases for a median of 30 mos (Smith MR JCO, 2005). However, there is no standard therapy for patients with nonmetastatic CRPC. ENTHUSE M0 (Study 15) is a global phase III study, now closed to enrollment, comparing zibotentan 10 mg vs placebo in nonmetastatic CRPC patients. Coprimary endpoints are overall survival and radiographic progression or death. Patients are screened for metastases by bone and CT/MRI scans. An unexpectedly high number of patients failed screening, prompting this analysis. Methods: All screened patients were included. For each investigator, the number of patients who enrolled and who failed screening were recorded. Screen fail reasons were recorded. Screen failure rates were computed by clinical specialty (oncology or urology) and geographic region by dividing the number of screen failures by the number of enrolled patients in each category. Results: As of January 14, 2011, 2583 patients completed screening, and 1169 (45%) failed screening. Screen failure rates were not different by clinical specialty or by geographic region. The leading causes of screen failures by clinical specialty are listed in the table. Conclusions: These data suggest that detectable asymptomatic metastatic disease may exist in about 30% of presumed nonmetastatic CRPC. These findings stress the value of periodic staging studies in men with CRPC, as those with detectable metastatic CRPC may benefit from emerging therapies. Screen failure criteria, n (%) Oncology (631 patients enrolled) Urology (1,885 patients enrolled) Detection of metastatic disease 184 (29) 566 (30) Creatinine clearance of <50 mL/min 45 (7) 156 (8) Long QTc interval 24 (4) 82 (4) Absence of biochemical progression 27 (4) 76 (4) Testosterone level >70 ng/dL (noncastrate) 15 (2) 71 (4) Total screen failures, n 283* 873* * Some patients had >1 reason for screening failure.