Background: There is no standard imaging modality that specifically and accurately images prostate cancer metastases, hampering assessments of tumor distribution, prognostication, and response. J591 is a humanized antibody that targets the external domain of PSMA, which is copiously expressed in metastatic disease and upregulated in mCRPC. We have previously reported on the feasibility, and PK and biodistribution properties of Zr-89-J591 in 10 patients (Morris et al, GU ASCO 2013). We now report on the targeting/accuracy in 49 patients with mCRPC. Methods: Following standard CT/MRI, bone scintigraphy, and FDG PET imaging, 5 mCi of Zr-89-J591 was administered intravenously. Zr-89-J591 was imaged 6-8 days after injection. Positive scan findings were confirmed, where possible, with biopsies in the following preference: Zr-89-J591 and FDG positivity, then Zr-89-J591 and FDG mismatch, and lastly standard imaging and any PET mismatch. Results: A total of 535 unique bone lesions in 49 patients were identified using all imaging modalities. 339 (63%) of these were evident on bone scans, 302 (56%) on CT, 202 (38%) on FDG PET, and 490 (92%) on Zr-89-J591 PET. The concordance between bone scans and Zr-89-J591 was 303/490 (62%). 102 sites were seen only on Zr-89-J591 and not on any other imaging modality. 44 biopsies were performed from 33 patients; 21 of those biopsies were bone. The histopathologic correlation for bone is provided in the table below. All positive biopsies were J591 positive (16/16); sensitivity of J591 was =100%, and positive predictive value (16/18) was 89%. Conclusions: Imaging using J591 PET/CT detects bone disease at greater rates than any other imaging modality tested. Biopsy data of J591 positive lesions suggest that most J591 positive lesions represent biopsy confirmed prostate metastasis. We are now testing minibodies based on J591 that may target faster, offering a more favorable imaging schedule. Clinical trial information: NCT01543659.