Perioperative Circulating Tumor DNA in Resectable Gastric and Gastroesophageal Junction Adenocarcinoma: Molecular Residual Disease, Recurrence Prediction, and a Trial-Ready Framework for Treatment Adaptation.

Recurrence remains a major cause of treatment failure after curative-intent therapy for resectable gastric and gastroesophageal junction adenocarcinoma. Clinicopathological staging, pathological response, serum tumor markers, and imaging incompletely identify patients with molecular residual disease (MRD). Circulating tumor DNA (ctDNA) has therefore emerged as a minimally invasive biomarker for perioperative risk stratification, response assessment, postoperative MRD detection, and recognition of molecular recurrence before radiographic relapse. In gastric and gastroesophageal junction cancer, postoperative ctDNA positivity has shown the most reproducible association with recurrence and survival, while serial studies indicate that early clearance during neoadjuvant therapy, persistence after preoperative therapy, and residual ctDNA after surgery or adjuvant chemotherapy represent distinct molecular states. However, several barriers remain: assay heterogeneity, variable sampling windows, clonal hematopoiesis, low shedding in diffuse-type and peritoneal-predominant disease, limited negative predictive value, and the absence of randomized evidence that ctDNA-guided intervention improves survival. Lessons from colorectal cancer are cautionary: randomized escalation and de-escalation trials have not uniformly translated strong prognostic validity into clinical utility. This review summarizes perioperative ctDNA evidence in resectable gastric and gastroesophageal junction adenocarcinoma, compares tumor-informed and tumor-agnostic platforms, examines peritoneal and molecular-subtype-specific limitations, and proposes a trial-ready framework for escalation, de-escalation, and molecular surveillance. At present, ctDNA is suitable for risk stratification and clinical-trial enrollment as an adjunct to, rather than a replacement for, conventional surveillance; it is not ready for routine treatment adaptation outside prospective studies.