Comparison of in situ and extraction-based methods for the detection of MET amplifications in solid tumors

In EGFR-treatment naive NSCLC patients, high-level MET amplification is detected in approximately 2–3% and is considered as adverse prognostic factor. Currently, clinical trials with two different inhibitors, capmatinib and tepotinib, are under way both defining different inclusion criteria regarding MET amplification from proven amplification only to defining an exact MET copy number. Here, 45 patient samples, including 10 samples without MET amplification, 5 samples showing a low-level MET amplification, 10 samples with an intermediate-level MET amplification, 10 samples having a high-level MET amplification by a MET/CEN7 ratio ≥2.0 and 10 samples showing a high-level MET amplification with GCN ≥6, were evaluated by MET FISH, MET IHC, a ddPCR copy number assay, a NanoString nCounter copy number assay and an amplicon-based parallel sequencing. The MET IHC had the best concordance with MET FISH followed by the NanoString copy number assay, the ddPCR copy number assay and the custom amplicon-based parallel sequencing assays. The concordance was higher in the high-level amplified cohorts than in the low- and intermediate-level amplified cohorts. In summary, currently extraction-based methods cannot replace the MET FISH for the detection of low-level, intermediate-level and high-level MET amplifications, as the number of false negative results is very high. Only for the detection of high-level amplified samples with a gene copy number ≥6 extraction-based methods are a reliable alternative.