3623 Background: Circulating tumor DNA (ctDNA) has emerged as a novel marker with high specificity and positive predictive value for identifying radiographically occult minimal residual disease (MRD) in colorectal cancer (CRC) patients (pts). A widely acknowledged unmet need is the development of therapeutic strategies to eradicate MRD and increase the cure rate of the disease. Ongoing trials are evaluating MRD as a biomarker for determining the intensity of post-operative adjuvant therapy. However, the feasibility and efficacy of therapy in pts who are MRD+ during surveillance (after all curative intent treatments including adjuvant chemotherapy) is unknown. Methods: In this phase II study, pts with stages II-IV CRC noted to be MRD+ during surveillance (ctDNA+ but no radiographic evidence of disease on imaging in last 30 days) received 6 months (m) of TAS-102. Pts were identified through the MD Anderson INTERCEPT program that integrates MRD testing with a tissue informed assay (Signatera) into standard of care for CRC pts. The primary endpoint was 6-m ctDNA clearance and secondary endpoints included 3-m ctDNA clearance, and disease-free survival (DFS). For comparison, a synthetic cohort (SC) was retrospectively created through the MD Anderson INTERCEPT Program of MRD+ pts not enrolled in MRD trials. Pts in the SC were matched by stage, BRAFmt status, MSI-H status, and treatment history. Results: Between 7/22/22 and 11/7/23, 15 MRD+ pts were enrolled and received TAS-102. Baseline characteristics included a median age of 62 (range: 37-73), resected Stage IV (80%), MSS (100%), BRAF wildtype (100%) and neoadjuvant therapy (87%). 30 pts were included in the SC and baseline characteristics included median age of 57 (range: 32-79), resected stage IV (80%), MSS (100%), BRAF wildtype (100%) and neoadjuvant therapy (90%). At data cutoff, 14 pts receiving TAS-102 had 3-m ctDNA levels. 7 pts (50%) had undetectable ctDNA levels at 3-m compared to 2 pts (6.7%) in the SC (p=0.0022). 12 had a decrease in their ctDNA level (86%) compared to 5 (17%) in the SC (p<0.0001). At 6-m, 82% of pts on trial had ctDNA levels at or below prior levels. 5 pts (45%) maintained undetectable ctDNA at 6-m compared to 2 pts (6.7%) in the SC (p=0.0096). Median DFS was 9.4-m with TAS-102 compared to 5.8-m in the SC. Conclusions: It is possible to identify MRD+ pts through screening during surveillance and enroll them onto therapeutic trials. Preliminary results suggest that treatment with TAS-102 is effective in inducing ctDNA clearance in the short term, and in doing so, may prolong DFS for pts receiving additional therapy compared to the SC. However, without evidence of persistent clearance this is unlikely to result in an increased cure fraction. Updated results, including survival data, will be presented. Clinical trial information: NCT05343013 .
