569 Background: CPIs targeting PD-1/PD-L1 and CTLA-4 have transformed management of GU cancers. These agents are associated with a unique type of immune-mediated toxicities that may affect any organ, which may be severe or life-threatening. Development of irAEs may in part depend on type of CPI used, type of solid tumor, and individual patient factors including presence of subclinical inflammatory states. However, distinct biomarkers predictive of irAE development are poorly defined. This study aimed to examine possible associations between tumor mutation burden (TMB) and PD-L1 expression with occurrence of irAEs. Methods: A retrospective chart review was performed of all patients with urothelial carcinoma (UC) and renal cell carcinoma (RCC), who received therapy with CPIs from 1/2019 to 9/2022 at our institution and developed associated irAEs. Severe irAEs were defined as those necessitating initiation of immunosuppressive therapies or other supportive therapies. TMB and PDL-1 expression patterns were analyzed, if available. Given lack of standardized cutoffs for PD-L1 positivity across different GU cancer types, tumors were considered positive if a combined positive score of greater than 10 was present in UC or a total positive score of greater than 1% was present in RCC. Results: A total of 60 patients were included in this analysis - 33 with UC and 27 with RCC. Severe irAEs occurred in 19 patients with UC and 25 patients with RCC. Median TMB was 8.7 in patients with UC, while it was 2.5 in patients with RCC. PD-L1 positivity was seen in 39.3% (13/33) of UC patients and 40.7% (11/27) of RCC patients. In the UC cohort, 57.8% (11/19) had severe irAEs and TMB less than 10, while 36.8% (7/19) had severe irAEs and TMB greater than 10. Further, 47.3% (9/19) of UC patients had severe irAEs and positive PD-L1, while 42.1% (8/19) had severe irAEs and negative PD-L1. In the RCC cohort, 88% (22/25) had severe irAEs and TMB less than 10, while 8% (2/25) had severe irAEs and TMB greater than 10. Additionally, 40% (10/25) RCC patients had severe irAEs and positive PD-L1, while 56% (14/25) had severe irAEs and negative PD-L1. The incidence of TERT promoter mutations and TMB is noted in the table. Conclusions: Identification of predictive biomarkers of irAEs from CPIs is a major area of need in GU malignancies. Definitive conclusions from this dataset are limited by its small sample size. Interestingly, there appear to be differences in baseline expression of traditional immune biomarkers between UC and RCC, with lower prevalence of TMB greater than 10 and positive PD-L1 noted in the RCC cohort. As such, their utility when making treatment decisions in clinical practice may be of limited value. [Table: see text]