Background: By predicting those most likely to relapse, cfDNA-based MRD testing has potential clinical utility for adjuvant therapy decision making. Ideally, MRD would be paired with validated treatment selection testing to enable precision adjuvant therapy. Herein we report the development and application of a combined personalized MRD and treatment selection test to patients with GU cancer. Methods: Patients with clinically localized solid tumors undergoing definitive therapy were enrolled on a prospective trial (NCT05082701) where precision MRD testing was performed with standard imaging-based disease recurrence monitoring. StrataMRD was performed on 2 tubes of peripheral blood through assessment of 1-12 personalized tracer mutations identified via tumor tissue profiling by Strata Select, which also provides validated genomic profiling, and angiogenesis inhibitor and immunotherapy treatment response scores (Angio TRS and IRS, respectively). Patients with GU cancers and valid StrataMRD results from at least one time point were eligible for this analysis. Tissue biomarkers were compared vs. 792 patients with advanced GU cancers tested by the same tissue platform. Results: A total of 49 patients with GU cancers who underwent definitive surgery and had at least one valid cfDNA based MRD test were eligible (*Table). The median age was 64 years, 92% of patients were male, and 49% of patients were stage I or II (vs. 51% stage III); patients underwent initial MRD testing at a median of 11 weeks after surgery and a median of 3 MRD tests have been performed per patient. Overall, 4/49 (8.2%) patients had an initial MRD+ test, and 8/49 (16.3%) had a MRD+ test at any time point (4 converted from MRD- to MRD+). Patient level MRD sensitivity and specificity in the evaluable cohort (with concurrent or subsequent imaging) was 100% (6 MRD+ / 6 recurrences) and 100% (24 MRD- / 24 not recurred), respectively. In this cohort, 34% and 52% of the renal cell carcinoma (RCC) patients were IRS- and Angio TRS-High, respectively, compared to 29% and 50%, respectively, of patients with advanced RCC undergoing similar tissue testing. Similarly, 15%, and 14% of the patients with bladder or other GU cancers were IRS-High, respectively, compared to 27% and 13% of those with advanced bladder or other GU cancers, respectively. Conclusions: Personalized cfDNA based MRD testing had high sensitivity and specificity vs. routine imaging for detecting disease recurrence in patients with GU cancers. Combined tumor testing with validated angiogenesis inhibitor and immunotherapy treatment selection biomarkers enables individualized adjuvant therapy decision making. Clinical trial information: NCT05082701.