Abstract Background: Retention of a normally functioning bladder would likely improve patient quality of life in the treatment of muscle-invasive bladder cancer (MIBC). Detection of residual cancer in bladder-sparing approaches relies on TURBT/bladder biopsy and clinical assessment. Historically, there is a 15-50% salvage cystectomy or metastasis rate with bladder sparing after chemotherapy alone or chemoradiotherapy. We hypothesized that clinical recurrence in bladder-sparing approaches was due to undetected subclinical disease at the time of (false) cT0 bladder assessment. We designed a next-generation sequencing (NGS) test to determine if residual mutations detected in the urine might be a surrogate for residual disease in MIBC patients undergoing RC. Methods: Patient samples were collected on IRB-approved protocols. DNA was isolated from whole urine (n=4 precystectomy samples), urine supernatant (n=17 prechemotherapy specimens, 15 precystectomy specimens, n=18 patients total), and germline (n=22) and sequenced using a targeted NGS panel of commonly mutated bladder cancer genes with sample indices and unique molecular barcodes. Twenty-one patients had neoadjuvant chemotherapy. TURBT tissue collected prior to neoadjuvant chemotherapy also underwent whole-exome sequencing (WES; n=17) as a benchmark. Single-nucleotide variants in urine (SNVU) and in TURBT tissue (SNVT) were compared. Persistence or clearance of SNVU was correlated to residual disease status at the time of radical cystectomy. Results: In benchmark samples from 15 patients, 46/76 SNVT identified by WES were also identified as SNVU in prechemotherapy urine samples. In precystectomy urine samples, absence of SNVU occurred in 5/6 patients achieving pT0 cystectomy specimens. In comparison, SNVU were detected in the precystectomy urine in 12/13 patients with residual disease. Absence/presence of SNVU was associated with residual disease status (p=0.0029, Fisher). In addition, TERT promoter hotspot mutations that are not detectable using WES-based approaches were detected in 9 of 22 patients in at least one urine specimen, and additional SNVU were detected that were not detected in tumor tissue. Conclusion: Presence/absence of SNVU strongly associates with residual disease status. Additional samples from urothelial cancer patients are currently being analyzed to determine the optimal urine compartment (cell-free vs. pellet vs. whole urine). Validation studies may encourage the use of urine biomarkers to improve residual disease detection in cystectomy avoidance algorithms. Citation Format: Phillip H. Abbosh, Uttam Satyal, David Liu, Ilsiya Ibragimova, Lauren Szeto, Benjamin Miron, James Ford, Daniel Parker, Aeen Asgar, Daniel M. Geynisman, R. Katherine Alpaugh, Paul Cairns, Costas D. Lallas, Edouard J. Trabulsi, Jean Hoffman-Censits, Eliezer M. Van Allen, Alexander Kutikov, Elizabeth R. Plimack. Use of deep sequencing of urinary DNA as a biomarker of residual disease status at the time of radical cystectomy abstract. In: Proceedings of the AACR Special Conference on Bladder Cancer: Transforming the Field; 2019 May 18-21; Denver, CO. Philadelphia (PA): AACR; Clin Cancer Res 2020;26(15_Suppl):Abstract nr B29.