Liquid Biopsy in Gastrointestinal Cancers: Circulating Tumor DNA for Molecular Residual Disease Assessment and Early Treatment Monitoring

TL;DR

This paper reviews how blood-based tumor DNA testing can detect leftover cancer and track treatment response in gastrointestinal cancers, with strongest evidence in colorectal cancer.

Contribution

The study provides a comprehensive clinical synthesis of ctDNA's role in molecular residual disease detection and treatment monitoring across multiple gastrointestinal cancer types.

Findings

Postoperative ctDNA is a strong predictor of recurrence risk in colorectal cancer, guiding adjuvant treatment decisions.

ctDNA monitoring in metastatic colorectal cancer enables early response tracking and resistance detection.

ctDNA shows promise in gastroesophageal cancers but has limited sensitivity in pancreatic and biliary cancers due to low DNA shedding.

Abstract

Circulating tumor DNA (ctDNA) is a blood-based marker that may help doctors detect residual cancer after treatment and identify treatment failure earlier than standard imaging in gastrointestinal cancers. The strongest evidence currently comes from colorectal cancer, where postoperative ctDNA is a powerful predictor of recurrence risk and can guide adjuvant treatment decisions in selected settings. In gastroesophageal, pancreatic, biliary, and liver cancers, ctDNA is promising but less consistent because tumors often shed less DNA into the blood and test performance varies. Positive ctDNA results are usually clinically meaningful, while negative results must be interpreted cautiously. Broader routine use will require standardized testing, careful control of false positives, and trials showing improved patient outcomes.

Background: Liquid biopsy using circulating tumor DNA (ctDNA) is rapidly reshaping gastrointestinal (GI) oncology. The highest-impact applications are molecular residual disease (mRD) detection after curative-intent therapy and early recognition of progression or resistance during systemic treatment. Methods: We performed a structured, clinically oriented narrative synthesis by using explicit search, eligibility, evidence prioritization, and clinical interpretation rules, integrating landmark prospective cohorts, randomized ctDNA-guided strategy trials where available, meta-analyses, key methodological research (e.g., pre-analytics, assay design, and clonal hematopoiesis (CH)/clonal hematopoiesis of indeterminate potential (CHIP)), and selected trial registries. Results: In resected colorectal cancer (CRC), postoperative ctDNA positivity is among the strongest known biomarkers of recurrence risk; large prospective studies demonstrate clear separation of disease-free survival (DFS)/overall survival (OS) between mRD+ and mRD− patients. In stage II colon cancer, randomized data (DYNAMIC) show that a ctDNA-guided strategy reduces adjuvant chemotherapy exposure without compromising long-term outcomes. In metastatic CRC, ctDNA supports early response monitoring and resistance tracking; ctDNA-selected anti-EGFR rechallenge provides a model of biomarker-driven actionability (CHRONOS). In gastroesophageal cancers, longitudinal ctDNA dynamics correlate with relapse risk and treatment efficacy, and in esophageal squamous cell carcinoma, ctDNA after neoadjuvant chemoradiotherapy informs residual disease risk and adjuvant stratification. In pancreatic ductal adenocarcinoma and hepatobiliary malignancies, sensitivity is constrained by low shedding and background cell-free DNA (cfDNA), yet ctDNA positivity remains clinically meaningful, and emerging data in resected extrahepatic cholangiocarcinoma (STAMP-linked analyses) show that ctDNA dynamics during adjuvant therapy predict recurrence. Conclusions: ctDNA is a clinically validated biomarker for mRD in CRC, whereas in other GI cancers, it remains a promising but methodologically heterogeneous tool whose clinical utility is tumor- and context-dependent. The next phase requires interventional trials demonstrating outcome improvement, harmonized sampling and reporting standards, and rigorous control of confounders (notably CH/CHIP).

Figures