Genomic analysis of comprehensive next generation sequencing data to explore the criteria for MET amplification as an actionable biomarker in NSCLC.

Introduction: MET amplification (METamp) can be a de novo or acquired resistance driver; however, the definition of METamp that best captures patients who may respond to targeted therapy remains debated. We explored the genomic landscape of METamp NSCLC including degree of amplification, co-drivers, amplicon size, and outcomes to MET inhibitors.

Methods: Hybrid-capture NGS-based genomic profiling from 88,547 tissue and 12,428 liquid NSCLC samples were queried for METamp (copy number (CN) ≥ ploidy + 4, or amplification ratio (AmpRatio; [CN/sample ploidy] ≥ 3). A nationwide de-identified real-world (rw) clinico-genomic database (CGDB) of NGS results linked to deidentified, electronic health record-derived clinical data was used to assess treatment and outcomes.

Results: Among 10,760 evaluable patients in CGDB, 362 (3.4%) had a METamp. In targeted therapy-naïve cases, MET AmpRatio negatively correlated with non-METex14 co-drivers (median 4.1 vs 2.9, p < 0.0001). MET AmpRatio was not significantly correlated with tumor mutational burden (p = 0.79) but was inversely correlated with amplicon size (p < 0.001). Among paired METamp tissue samples, 8/30 had METamp detected in liquid; higher tumor fraction and AmpRatio were associated with liquid detection. Among 39 METamp patients receiving MET inhibitors, longer median real-world progression free survival was observed with MET AmpRatio ≥ 3 vs < 3 (4.9 vs. 1.7mos, HR 0.53 [95 %CI:0.21-1.3]).

Conclusions: MET AmpRatio positively correlated with focal amplification and absence of co-drivers and trended with increased benefit from MET inhibitors. Further studies evaluatingcombinatorial data including MET AmpRatio, amplicon size and presence of other potential drivers, as predictive biomarkers for therapies targeting MET amplification in NSCLC are warranted.