Abstract

Mesenchymal-epithelial transition exon 14 skipping mutations (METex14) define a targetable molecular subset of non-small cell lung cancer (NSCLC), but optimal circulating tumor DNA (ctDNA) biomarkers to guide MET tyrosine kinase inhibitor therapy remain undefined. Within the EMBRACE trial of ensartinib, we conducted a prospective biomarker analysis with plasma collected at baseline (V0) and 4 weeks after treatment initiation (V1). ctDNA was profiled using a hybrid-capture next-generation sequencing panel covering 229 cancer-related genes, and mutations were classified into three monitoring paradigms: MET-specific alterations, canonical alterations (tier I/II), and pan-alterations (any variant in any gene on the panel). Baseline ctDNA positivity rates were 48.3%, 48.3%, and 75.9% for MET-specific, canonical, and pan-alterations, respectively, and declined across all categories at V1. Conversion to ctDNA negativity in any paradigm was associated with a tendency toward higher objective response rate (ORR) and longer progression-free survival (PFS). Moreover, clearance of MET-specific alterations conferred the greatest clinical benefit, with an ORR of 75.0% versus 16.7% and a median PFS of 9.3 months versus 2.2 months (p = 0.005). Among the three paradigms, MET-specific monitoring provided the most favorable diagnostic performance to identify long-term responders, with a specificity of 90% and a positive predictive value of 80%. These data demonstrate that early MET-specific ctDNA clearance is a robust on-treatment biomarker for ensartinib benefit in METex14 NSCLC, while broader ctDNA profiling remains valuable for uncovering emerging resistance mechanisms.