Background: Data suggests that changes in ctDNA quantity correlate with response to therapy in patients with advanced solid malignancies. However, there is little consistency on how to calculate and interpret such changes. Here, we apply a clinically-validated molecular response algorithm to a RW cohort of patients with aCRC to further evaluate its ability to assess treatment outcomes. Methods: We queried the Guardant INFORM database, which comprises aggregated commercial payer health claims and de-identified records from patients with comprehensive ctDNA testing via Guardant360 (G360) from September 2018-March 2022. Patients with aCRC who received G360 within 15 weeks prior to new treatment initiation (any line of therapy) and a second test 3-15 weeks after treatment initiation were retrospectively evaluated using the G360 Response algorithm. Cox proportional hazards (CPH) were used for RW time to next treatment (TTNT), time to treatment discontinuation (TTD), and overall survival (rwOS) analyses. Patients categorized as molecular responders had a >50% decrease in mean variant allele fraction (VAF) ratio from pre-treatment to on-treatment. Gender, age, line of therapy (LOT) and comorbidities were included as covariates. Median TTNT, TTD, and rwOS were calculated by Kaplan Meier. Results: Of 185 aCRC patients with eligible MR results, 65% received chemotherapy +/- VEGF, 21% received regimens containing anti-EGFR monoclonal antibodies, and 14% received other therapies. 43% of aCRC patients were classified as molecular responders, 42% were non-responders, and 15% were not evaluable by the algorithm due to no/low ctDNA at one or both timepoints. 16% of patients cleared their ctDNA on treatment (i.e., ctDNA became undetectable). Molecular responders had significantly longer TTNT (median 10.1 months vs 6.1 months; HR p < 0.005), TTD (median 5.2 Months vs 3.9 months, HR p=0.041), and rwOS (not reached vs 17.8 months, HR p=0.017). Conclusions: Patients with aCRC classified as molecular responders, as calculated by this algorithm, had prolonged time on treatment and overall survival. Compared to tumor markers, ctDNA has a short half-life, which can allow for early response assessment, as shown by our study. These findings are relevant for clinical care, with future potential to allow for adaptive clinical trial design.