Introduction: Genomic variants that lead to MET proto-oncogenem receptor tyrosine kinase (MET) exon 14 skipping represent a potential targetable molecular abnormality in NSCLC. Consequently, reliable molecular diagnostic approaches that detect these variants are vital for patient care.Methods: We screened tumor samples from patients with NSCLC for MET exon 14 skipping by using two distinct approaches: a DNA-based next-generation sequencing assay that uses an amplicon-mediated target enrichment and an RNA-based next-generation sequencing assay that uses anchored multiplex polymerase chain reaction for target enrichment.Results: The DNA-based approach detected MET exon 14 skipping variants in 11 of 856 NSCLC samples (1.3%). The RNA-based approach detected MET exon 14 skipping in 17 of 404 samples (4.2%), which was a statistically significant increase compared with the DNA-based assay. Among 286 samples tested by both assays, RNA-based testing detected 10 positives, six of which were not detected by the DNA-based assay. Examination of primer binding sites in the DNA-based assay in comparison with published MET exon 14 skipping variants revealed genomic deletion involving primer binding sequences as the likely cause of false negatives. Two samples positive via the DNA-based approach were uninformative via the RNA-based approach due to poor-quality RNA.Conclusions: By circumventing an inherent limitation of DNA-based amplicon-mediated testing, RNA-based analysis detected a higher proportion of MET exon 14 skipping cases. However, RNA-based analysis was highly reliant on RNA quality, which can be suboptimal in some clinical samples.