Circulating tumor DNA in gastrointestinal cancers: promise, pitfalls, and the path forward

Circulating tumor DNA (ctDNA) has emerged as a powerful prognostic biomarker in gastrointestinal (GI) oncology, with the strongest evidence in colorectal cancer. Postoperative ctDNA positivity identifies patients at high risk of recurrence, and serial clearance patterns further refine risk. However, prognostic validity has not consistently translated into treatment-selection utility. In stage II colon cancer, DYNAMIC supported chemotherapy de-escalation without compromising recurrence-free survival, whereas DYNAMIC-III showed that escalation in ctDNA-positive stage III disease did not improve outcomes. More recent randomized data from CIRCULATE suggest that ctDNA-positive patients with proficient mismatch repair and microsatellite-stable stage II colon cancer may benefit from adjuvant chemotherapy, although the intention-to-treat primary endpoint was not statistically significant. COBRA further highlights the vulnerability of low-risk populations to assay-dependent interpretation and unvalidated clearance endpoints. Outside colorectal cancer, ctDNA is strongly prognostic in gastro-esophageal, pancreatic, and biliary tract cancers, but evidence supporting ctDNA-directed treatment remains limited. Assay heterogeneity, variable tumor shedding, postoperative sampling timing, false-positive and false-negative results, and uncertain benefit from treating molecular recurrence before radiographic disease remain major barriers. Prospective trials must show that biomarker-guided interventions improve patient-important outcomes before ctDNA can serve as a stand-alone treatment mandate across GI cancers.