The identification of <em>MET </em>exon 14 skipping mutations in patients with non–small cell lung cancer presents a complex diagnostic challenge that requires both DNA and RNA analysis, according to results of a University of Michigan pathology study.
David M. Manthei, MD, PhD
The identification ofMETexon 14 (METex14) skipping mutations in patients with nonsmall cell lung cancer (NSCLC) presents a complex diagnostic challenge that requires both DNA and RNA analysis, according to results of a University of Michigan pathology study.1
The question of how best to detect this mutation becomes more clinically relevant as therapies are designed for this population, David M. Manthei, MD, PhD, noted in presenting the findings at the Association for Molecular Pathology 2019 Annual Meeting and Expo.
Three tyrosine kinase inhibitors have received breakthrough therapy designations from the FDA for patients withMETex14 aberrations: 2 investigational agents capmatinib and tepotinib, as well as crizotinib (Xalkori), which is approved inALK- andROS1-positive NSCLC settings.2
Overall,METex14 skipping mutations have been found in 2% to 5% of lung adenocarcinomas, depending on the study, said Manthei, a fellow in the Department of Pathology at Michigan Medicine in Ann Arbor. These alterations have been described as somatic mutations that affect sites of RNA splicing, resulting in the skipping ofMETex14and activation of kinases.3
Investigators have found that these alterations can occur due to point mutations and deletions that affect splicing or through exon 14 deletion, Manthei said. Alterations can occur between exons 13 and 14 and between exons 14 and 15. When exon 14 is excluded, “the tyrosine kinase protein is more stable and can have prolonged kinase signaling, which can lead to the deleterious effect,” he said.
After analyzing testing approaches, Manthei and colleagues concluded that a combined DNA and RNA workflow detectsMETex14 mutations that would not be identified with DNA-only analysis. They also found that RNA-based methods yield information on variant splicing that might not surface otherwise.
DNA Versus RNA Sequencing
Manthei discussed the advantages and disadvantages of current methods of detectingMETex14 alterations: DNA testing offers a more robust substrate and mutations in exons that are often well defined but coverage depends on the design of the testing panel being used and less common variants may not be described in functional studies.
RNA testing provides information on the functional effect of mutations with coverage of many aberrations in a single assay, but the substrate is more changeable, the analytical process involves additional steps, and alternative splicing makes the results harder to interpret.
The experience with the number of fusions detected in mutation testing led investigators to question assay findings more specifically, Manthei said. The team conducted a retrospective analysis of tissue samples from 482 patients using the Oncomine Focus Assay (OFA),1a next-generation sequencing test designed for clinical research.4
When OFA was used to detectMETex14 DNA mutations, 7 cases (1.5%) with no alternative drivers were identified. In 61 cases (12.7%), no mutation inMETex14 was detected but 42 of these samples had alternative drivers includingKRAS,EGFRexon 19 mutations, and ALKfusions. Fusion reads ranged from 16 to 59 per sample in the mutation group and from 1 to 12 in those without a mutation.
“We saw a wide range of fusion reads and no other drivers that were present,” Manthei said. “However, there were an additional 61 cases that had at least some fusion reads that were present based on the assay. The question is what do you do with them? How do you interpret them?”
When there are alternative drivers, “it may not make sense to think of it as an underlyingMETfusion event there,” he added.
Although OFA can analyze both materials, its DNA coverage ofMETex14 detects mutations in the splice donor site (3’ end) but not the splice acceptor site (5’ end), investigators said in their conference poster. For RNA, the assay can detectMETexon 13 to 15 fusions, but interpretive thresholds have not been defined.1
To analyzeMETex14 mutations with RNA fusions, investigators used OFA and then Sanger sequencing. The Sanger method was used on the 5’ end ofMETex14 on residual DNA for cases in which RNA suggested a skipping mutation, defined as the percentage ofMETex14 fusion (%fusion) reads ≥0.03% (0.05% from lowest confirmed case).
Overall, investigators identified 11 cases (2.3%) withMETex14 mutations: 8 with 3’ end alterations by OFA sequencing and 3 with 5’ end aberrations by Sanger. Ten of the 11 cases had adequate RNA reads by OFA; those samples demonstrated %fusion reads ranging from 0.082 to 22.1 and %fusion reads perMETcopy of 0.106 to >25.
Additionally, 58 cases with RNA fusion reads by OFA analysis that did not have detectableMETex14 DNA mutations, which investigators said could indicate alternative splicing.
The combined approach offers a better picture of theMETex14 landscape, according to Manthei. “The OFA assay is picking up what it’s supposed to and we’re also able to now interrogate whether additional DNA mutations might be underlying these fusion readings,” he said.
Nevertheless, Manthei said, some fusion reads still have no clear answer. In such instances, “we might want to be [reflexively sending a borderline case] to something like Sanger sequencing to really inform whether a mutation is present or not,” he said.
Although the use of RNA-based methods improves the analysis, Manthei said, caution is needed in interpreting the results: RNA fusion reads can be detected in nonmutated samples, and determining a threshold for true positivity is challenging. The specificity of the threshold can be improved significantly by normalizing the analysis to total RNA reads andMETcopy number, he said.