Background: There is no standard imaging (SI) modality that specifically and accurately images prostate cancer (PC) metastases, hampering prognostication and response assessment. J591 is a humanized antibody that targets the external domain of PSMA. We have previously reported on the feasibility, PK and biodistribution of ⁸⁹Zr-J591 in 10 patients (pts) (Pandit-Taskar et al, Eur J Nuc Med Img 2014). We now report on the targeting/accuracy in 50 pts with mCRPC. Methods: Following standard CT/MRI, bone scan (BS), and FDG PET imaging, 5 mCi of ⁸⁹Zr-J591 was administered IV. ⁸⁹Zr-J591 was imaged 6-8 days after injection. Positive (pos) scan findings were confirmed, where possible, with biopsies (bxs) in the following preference: concordant ⁸⁹Zr-J591 and FDG pos, ⁸⁹Zr-J591 and FDG mismatch, and a mismatch between SI and any PET. Results: Imaging: A total of 703 lesions in 50 pts were identified using all imaging modalities. Bone:538 total bone lesions were detected. 491(91%) lesions were present on J591 of which 99 were only evident by J591. BS identified 339 (63%), CT 301 (56%), and FDG 207 (38%). Soft Tissue: 165 total soft tissue lesions were detected. 90 (55%) were seen on J591 of which 17 were only evident by J591. CT identified 124 (75%) and FDG 88 (53%). Pathology:46 bx’s were evaluable (21 bone, 25 soft tissue) in 34 pts. Of the unique J591 lesions bx’d, 5/7 were pos for PC. Bone: We bx’d 19 J591 pos lesions and 2 J591 neg sites. Overall, path concordance with J591 was: 89% true pos, 100% true neg, 11% false pos, and 0% false neg. Softtissue: We bx’d 16 J591 pos lesions and 9 J591 neg sites. Of these, we found 88% true pos, 11% true neg, 13% false pos, and 89% false neg. Conclusions: J591 PET identifies additional disease in bone not seen using other imaging modalities. These lesions are highly likely to correspond to disease by bx. The data also continues to affirm that J591 appears to be superior at identifying bony disease than soft tissue lesions. Clinical trial information: NCT01543659