525 Background: Circulating tumor cells (CTCs) are promising biomarkers in metastatic urothelial cancer (UC). Unfortunately, efforts in localized disease have been unsuccessful, in part due to limitations of existing technologies that rely on counting cells and epithelial-marker expression. Here, we applied a novel selection-free digital pathology platform in a localized UC cohort. To date, this platform has associated CTC morphology with differential therapeutic response in metastatic UC and castrate-resistant prostate cancer. If feasible in localized UC, we may potentially identify best candidates for adjuvant therapy or bladder sparing, as well as enable sensitive monitoring for recurrence. Methods: N=16 consecutive UC pts included 8 (50%) metastatic controls and 8 (50%) localized (3 (37%) at TURBT and 5 (63%) at cystectomy). Peripheral blood was processed with the Epic CTC platform (pan-CK/CD45/PD-L1/DAPI staining). Approximately 3 million cells per slide were imaged. Unsupervised clustering categorized CTCs into 5 subtypes based on 11 morphologic features (nuclear solidity, speckling, nucleoli and entropy; cytokeratin speckling and ratio; and cytoplasmic/nuclear circularity, area, and convex area ratio). Results: 119 CTCs were detected from 11/16 (69%) pts (5/8 (63%) localized (2 NMIBC, 6 MIBC) and 6/8 (75%) metastatic). All MIBC pts had cystectomy (4/6 (67%) received NAC). 2/8 (25%) metastatic pts had stable disease, 3/8 (38%) were progressing, and 3/8 (38%) had newly detected M1. Median (range) CTC count/mL was similar for localized and metastatic pts (0.4 (0-58.6), 0.75 (0-1.9)). CTCs were detected in a pt with CIS, but not in a pt with TaHG disease. 1/16 (6.3%) pts had a single PD-L1+ CTC. CTCs were successfully assigned into 5 subtypes with predominant features of large, small, or linear cells, high cytoplasmic circularity, and prominent nucleoli. Conclusions: Digital pathology and subtype assignment of CTCs is feasible in localized UC. Ongoing efforts at our center include application of this technology in localized patients receiving investigational checkpoint inhibitor therapy to potentially predict best responders or conversely those at the highest risk for recurrence.