A large real-world study of molecular testing for non-small cell lung cancer (NSCLC) has identified important limitations in detecting MET exon 14 skipping variants, demonstrating that both assay design and bioinformatics can affect diagnostic accuracy. The researchers also reported results from the first multinational external quality assessment (EQA) programs for MET exon 14 testing, which showed high performance for tissue testing but lower accuracy for liquid biopsy analysis. The findings were published in The Journal of Molecular Diagnostics.
MET exon 14 skipping alterations occur in approximately 3–4 percent of NSCLCs and are established predictive biomarkers for targeted therapy. Detecting these alterations is challenging because they can arise from a wide variety of mutations affecting exon 14 and its surrounding splice sites.
To evaluate current testing approaches, researchers analyzed 379 NSCLC samples collected between 2016 and 2024. They identified 171 distinct MET exon 14 and splice-site variants, illustrating the broad range of alterations laboratories may encounter in routine practice.
Among 114 variants assessed with both DNA- and RNA-based next-generation sequencing (NGS), RNA testing confirmed exon 14 skipping in 107. However, the comparison also revealed potential pitfalls of DNA-based testing. Two large deletions were missed by routine variant-calling software but were visible in the raw sequencing data, and one synonymous splice-site variant was initially filtered out despite causing exon 14 skipping. In contrast, seven variants identified by DNA sequencing did not produce exon 14 skipping on RNA analysis, indicating that not every sequence change near exon 14 has a functional effect.
RNA testing also had practical limitations. 57 variants could not be evaluated because insufficient tissue remained or RNA quality was inadequate. Of those, 18 had previously been reported as exon 14 skipping variants, while 39 could not be functionally confirmed.
The study also examined the first multinational QuIP EQA programs for MET exon 14 testing. Tissue-based testing showed strong performance, with 49 of 50 participating laboratories (98 percent) successfully completing the assessment using a variety of molecular methods.
Performance was lower for liquid biopsy testing. Pass rates improved from 38 percent in 2022 to 63 percent in 2024, but false-negative results remained common, particularly for variants present at low allele fractions or involving intronic deletions. Analysis of unsuccessful cases found that inadequate sequencing coverage and bioinformatics filtering frequently prevented variant detection, even when mutations were present in the raw sequencing data.
The findings highlight several considerations for diagnostic laboratories. The diversity of MET exon 14 alterations means that assay validation should include large deletions, synonymous splice-site variants, and intronic changes. The study also emphasizes that bioinformatics pipelines require validation alongside laboratory workflows and that combining DNA- and RNA-based testing, when sufficient material is available, may reduce the likelihood of clinically relevant variants being missed. Continued optimization of liquid biopsy assays and participation in external quality assessment programs may further improve testing performance.
Newsletters
Receive the latest pathologist news, personalities, education, and career development – weekly to your inbox.
