Background: Detection of circulating tumor DNA (ctDNA) in blood is a powerful predictive biomarker for identifying molecular residual disease (MRD) post-surgery. Not much is known about the utility of identifying MRD using ctDNA in pts with resected CCA, despite its associated poor prognosis. The STAMP trial (NCT03079427), a randomized phase 2 study comparing adjuvant capecitabine (CAP) to gemcitabine plus cisplatin (GemCis) showed no difference in recurrence-free survival (RFS) and overall survival (OS) between the 2 arms. The goal of this analysis was to evaluate the feasibility of monitoring ctDNA to predict the risk of recurrence in pts with resected CCA. Methods: A total of 254 plasma samples were collected from a cohort of 89 CCA (hilar, N = 43; distal, N = 46) pts post-surgery at 3 separate time points; pre-ACT (baseline, N = 89), on-ACT C5D1 (after 5 cycles of treatment) (N = 88), and on-ACT C8D1 (after 8 cycles of treatment) (N = 77). Longitudinal ctDNA testing was performed using a personalized, tumor-informed ctDNA assay (Signatera, bespoke mPCR-NGS assay). ctDNA results were analyzed and evaluated for its correlation with RFS and OS. Results: At the MRD timepoint (baseline, pre-ACT), 24.7% (22/89) pts were ctDNA-positive, 90.9% (20/22) of whom recurred demonstrating a trend for shorter RFS (HR = 1.7, 95%Cl 0.98-2.8, p = 0.069) compared to ctDNA-negative pts. All pts who remained ctDNA-positive on-ACT at C5D1 (N = 17) and C8D1 (N = 15), recurred clinically with a significantly shorter RFS (HR = 8.1, 95%Cl 4.3-15, p < 0.0001 and HR = 5.0, 95%Cl 2.7-9.4, p < 0.0001, respectively). The primary cancer site (hilar vs distal) or type of ACT (CAP vs GemCis) had no impact on RFS irrespective of ct-DNA result. Pts who remained ctDNA-negative on ACT (N = 56) showed significantly better RFS compared to pts who remained positive (N = 10, HR = 6.6, 95%Cl 3.1-14, p < 0.001), or who turned positive (N = 11, HR = 5.2, 95%Cl 2.54-10.5, p < 0.001). In multivariate analysis including primary tumor site, sex, pathological grade and ACT regimen, positive ctDNA status at baseline (pre-ACT) remained significant for poor RFS (HR = 1.82, 95%CI 1.04-3.18, p = 0.035). Conclusions: Personalized monitoring of ctDNA both pre and on-ACT is feasible and predictive for recurrence in pts with resected CCA. To our best knowledge, this is one of the first reports presenting the clinical implication of ctDNA monitoring on ACT, in pts with CCA. Our findings suggest ctDNA monitoring may help optimize clinical decision-making in pts with resected CCA. Larger prospective studies would be needed to validate the findings of this study. Clinical trial information: NCT03079427.