Circulating Tumor DNA-Based Molecular Residual Disease in Early-Stage Non-Small-Cell Lung Cancer: Biology, Assays,...

Complete surgical resection remains the cornerstone of curative therapy for early-stage non-small-cell lung cancer (NSCLC), yet a substantial proportion of patients relapse despite a macroscopically complete operation and negative conventional staging. This residual risk reflects micrometastatic disease that is below the resolution of computed tomography and that current clinicopathologic risk factors only coarsely predict. Circulating tumor DNA (ctDNA) provides a direct, blood-based readout of this molecular residual disease (MRD), its detection after definitive treatment is among the strongest known predictors of recurrence, and ctDNA frequently rises months before radiographic progression. In this narrative review, based on a literature search closed on 12 September 2026, we synthesize the foundational and translational literature underpinning ctDNA-MRD in resected NSCLC. We outline the biology of ctDNA shedding and metastatic dissemination established by serial-sampling and phylogenetic studies; the prognostic performance of landmark and longitudinal MRD, tabulated quantitatively across the major cohorts, meta-analyses and trial-nested analyses; the perioperative kinetics that dictate when sampling is informative; the analytical design of tumor-informed versus tumor-agnostic assays and the determinants of their sensitivity; and the biological and technical sources of false-negative and false-positive results, of which clonal hematopoiesis and low-shedding histologies are the most consequential. We then position MRD within the contemporary perioperative treatment landscape-adjuvant targeted therapy and immunotherapy, and neoadjuvant and perioperative chemoimmunotherapy-and review the emerging, still-investigational paradigm of MRD-guided escalation and de-escalation, including the early termination of the MRD-selected MERMAID trials. We conclude with the analytical, evidentiary and implementation barriers-and the current positions of guideline bodies and regulators-that separate a powerful prognostic biomarker from a validated predictive tool, and we set out the trial-design requirements needed to close that gap.