Background: Current methodologies for monitoring treatment response are largely based on conventional scans and/or tumor biopsies, which may be limited in their ability to accurately assess disease burden at the molecular level. Circulating tumor DNA (ctDNA) detection in blood has emerged as a prognostic and predictive biomarker and has shown to better assess treatment response in patients receiving immune checkpoint inhibitors (ICI). We conducted a prospective, pilot study to investigate the concordance of serial ctDNA detection and dynamics with radiographic response in patients with advanced GU malignancies undergoing ICI-based treatment. Methods: Twenty patients with histologically confirmed advanced GU malignancies (renal, urothelial, and prostate) were enrolled in the prospective study. All eligible patients received ICI treatment for at least 12 weeks and were followed by serial collection of blood samples every 6-8 weeks until disease progression. Conventional scans were performed approximately every 12 weeks until disease progression. Overall response rate (ORR) by investigator was reported and associated with ctDNA detection. Results: ctDNA analysis was performed on 122 plasma samples obtained from 20 patients (N=15 renal cell carcinoma; N=4 urothelial carcinoma; N=1 prostate cancer). Prior therapies to ICI-based treatment included chemotherapy (10%), hormonal therapy (5%) and anti-VEGF (5%). After study enrollment, patients received anti-PD-1 (95%), anti-CTLA-4 (30%) or anti-PD-L1 (5%) with an ORR of 70% as best response. With a median follow-up of 19 months (range: 4-48), progressive disease was observed in 7 patients. Nineteen patients had longitudinal plasma samples available and ctDNA detection at any time point was 45% (9/20). The overall concordance between ctDNA dynamics and radiographic response at 12 weeks was observed in 89% (17/19) of patients. Of these 17 concordant patients, one patient showed transient ctDNA positivity followed by clearance at the last two timepoints on treatment. The two patients with discordant results included the ones with CNS-only metastasis (ctDNA negative). Of the 7 patients who progressed on ICI, ctDNA was detected in five (71%); the remaining two had CNS-only metastases. The last patient had a single time point available on treatment that showed ctDNA-positivity and passed away 7 weeks after molecular evidence of disease. Conclusions: In this study, serial collection of blood samples for ctDNA analysis to monitor response to ICI-based therapiesin patients with advanced GU tumors was feasible. There was a high concordance rate between radiological imaging and ctDNA data, especially in extra-CNS disease. Further studies are needed to validate ctDNA as a tool to aid disease monitoring in patients treated with ICI.