Multinational proficiency tests for EGFR exon 20 insertions reveal that the assay design matters

Michaela A. Ihle, Carina Heydt, Anne M. Schultheis, Robert Stöhr, Florian Haller, Sylvia Herold, Daniela Aust, Wolfgang Dietmaier, Matthias Evert, Markus Eszlinger, Anja Haak, Silke Laßmann, Daniela Vorholt, Frank Breitenbücher, Martin Werner, Anna Streubel, Thomas Mairinger, Maja Grassow-Narlik, Sabine Merkelbach‐BruseScientific 2024 · DOI

TL;DR

This study reveals that current proficiency tests for detecting EGFR exon 20 insertions (EGFR exon20ins) in non-small cell lung cancer (NSCLC) highlight a critical issue: assay design significantly impacts diagnostic accuracy. Next-generation sequencing (NGS) consistently outperforms commercially available quantitative PCR (qPCR) methods, particularly for the diverse and often rare EGFR exon20ins mutations, underscoring the need for broader-spectrum detection methods in clinical practice.

What Problem Does This Paper Address?

EGFR exon 20 insertions (EGFR exon20ins) are a rare but clinically significant group of mutations found in non-small cell lung cancer (NSCLC). These mutations confer resistance to many standard EGFR tyrosine kinase inhibitors (TKIs), but new targeted therapies are emerging that specifically address these alterations. To ensure patients receive the most effective treatment, accurate and reliable detection of EGFR exon20ins is crucial. However, the EGFR exon 20 region is known for its high mutational heterogeneity, meaning there are many different types of insertion mutations. This complexity poses a challenge for diagnostic laboratories, as existing testing methods may not be capable of detecting the full spectrum of these mutations, potentially leading to missed diagnoses and suboptimal patient care. This paper addresses the need to assess the performance of different diagnostic methods in detecting these challenging mutations across a multinational cohort of laboratories.

Methods

The study employed two multinational proficiency testing programs, one focused on German-speaking countries (DACH) and another broader international one, to evaluate the performance of clinical laboratories in detecting EGFR exon20ins. These tests utilized both formalin-fixed, paraffin-embedded (FFPE) tissue samples and liquid biopsy samples (plasma-derived cell-free DNA). Participating institutes were free to use their preferred diagnostic methods, which included Next-Generation Sequencing (NGS), mutation-/allele-specific quantitative PCR (qPCR), Sanger sequencing, and pyrosequencing. The performance of each institute was assessed based on their ability to correctly identify the presence or absence of specific EGFR exon20ins mutations in the provided samples.

Datasets

The study used reference samples for proficiency testing. These samples were derived from two sources:

  • Clinical samples: A retrospective analysis of a patient cohort from Cologne identified a spectrum of EGFR exon20ins mutations.
  • Artificial samples: These were specifically designed and manufactured by SensID GmbH to mimic clinical samples, ensuring a controlled and consistent source of mutations for testing. These artificial samples were used in both FFPE tissue and liquid biopsy formats, with varying allelic fractions to simulate real-world scenarios, including low-frequency mutations.

The study did not introduce a new benchmark dataset but rather used these carefully prepared samples to evaluate existing laboratory practices.

Key Insights and Ideas

  • Assay design is paramount: The most striking insight is that the design of the detection assay, rather than the specific kit or platform within a method, is the primary determinant of success in identifying EGFR exon20ins.
  • NGS reigns supreme: Next-generation sequencing (NGS) consistently demonstrated superior performance across both tissue and liquid biopsy samples, achieving high success rates. This is attributed to its ability to cover the entire EGFR exon 20 region and detect a broad spectrum of mutations, including rare ones.
  • qPCR's limitations: Commercially available mutation-/allele-specific qPCR assays, while often used, showed significantly lower success rates, particularly in liquid biopsies. This is because these assays are typically designed to detect only a limited number of common EGFR exon20ins mutations, leading to a high rate of false negatives for the broader spectrum of variants.
  • Heterogeneity is the enemy of narrow assays: The inherent heterogeneity of EGFR exon20ins mutations means that assays with a limited detection spectrum are inherently prone to missing clinically relevant variants.
  • Liquid biopsy is challenging: Detecting EGFR exon20ins in liquid biopsy samples proved more difficult, with lower overall success rates compared to tissue samples, especially for qPCR methods. This is likely due to lower allelic fractions and the complexities of cfDNA analysis.
  • Standardization is key: The study implicitly highlights the need for standardized nomenclature (like HGVS) and robust validation of diagnostic methods to ensure accurate and reproducible results.

Key Results

  • Tissue samples:
    • DACH proficiency test: 86.5% of participating institutes successfully detected EGFR exon20ins in tissue samples. Institutes using NGS had a 100% success rate, while only 28.5% of those using mutation-/allele-specific qPCR succeeded.
  • International proficiency test: 68.7% of institutes passed the tissue part. Again, NGS and Sanger sequencing showed high success rates (89.8% for NGS/Sanger combined), while only 11.1% of qPCR users succeeded.
  • Liquid biopsy samples:
    • DACH proficiency test: 66.7% of participants achieved successful results. 84% of NGS users passed, while none of the qPCR users did.
  • International proficiency test: 45.8% of participants were successful. 81.5% of NGS users passed, while none of the qPCR or pyrosequencing users succeeded.
  • False negatives: A significant number of false negatives were observed, particularly with qPCR, for samples carrying EGFR exon20ins. This was often due to the assay's inability to detect the specific mutation present.

Limitations and Future Work

  • Limitations:
    • The study acknowledges that while NGS is superior, some NGS users still failed, particularly with low allelic fraction samples. The reasons for these failures (e.g., bioinformatic analysis issues) are not fully elucidated for all cases.
  • The international proficiency test had a smaller sample size for liquid biopsies, potentially limiting the statistical power for certain analyses.
  • The study focused on specific EGFR exon20ins mutations and did not cover every possible variant.
  • Future Work:
    • The authors suggest that further validation of the entire workflow, from cfDNA extraction to bioinformatic analysis, is needed for liquid biopsy testing.
  • They emphasize the importance of scrutinizing the detection spectrum of any chosen assay, especially for qPCR, to ensure it covers the relevant mutations.
  • The findings strongly advocate for the broader adoption of NGS for EGFR exon20ins testing in clinical practice, aligning with recommendations from professional societies.
  • Further research could explore the specific reasons for NGS failures in low-allelic fraction samples and optimize bioinformatic pipelines.

Summary updated

Abstract

Insertion mutations in exon 20 of the epidermal growth factor receptor gene (EGFR exon20ins) are rare, heterogeneous alterations observed in non-small cell lung cancer (NSCLC). With a few exceptions, they are associated with primary resistance to established EGFR tyrosine kinase inhibitors (TKIs). As patients carrying EGFR exon20ins may be eligible for treatment with novel therapeutics-the bispecific antibody amivantamab, the TKI mobocertinib, or potential future innovations-they need to be identified reliably in clinical practice for which quality-based routine genetic testing is crucial. Spearheaded by the German Quality Assurance Initiative Pathology two international proficiency tests were run, assessing the performance of 104 participating institutes detecting EGFR exon20ins in tissue and/or plasma samples. EGFR exon20ins were most reliably identified using next-generation sequencing (NGS). Interestingly, success rates of institutes using commercially available mutation-/allele-specific quantitative (q)PCR were below 30% for tissue samples and 0% for plasma samples. Most of these mutation-/allele-specific (q)PCR assays are not designed to detect the whole spectrum of EGFR exon20ins mutations leading to false negative results. These data suggest that NGS is a suitable method to detect EGFR exon20ins in various types of patient samples and is superior to the detection spectrum of commercially available assays.
BibTeX
@inproceedings{ihle2024multinational,
  author = {Michaela A. Ihle and Carina Heydt and Anne M. Schultheis and Robert Stöhr and Florian Haller and Sylvia Herold and Daniela Aust and Wolfgang Dietmaier and Matthias Evert and Markus Eszlinger and Anja Haak and Silke Laßmann and Daniela Vorholt and Frank Breitenbücher and Martin Werner and Anna Streubel and Thomas Mairinger and Maja Grassow-Narlik and Sabine Merkelbach‐Bruse},
  title = {{Multinational proficiency tests for EGFR exon 20 insertions reveal that the assay design matters}},
  booktitle = {Scientific},
  year = {2024},
  doi = {10.1038/s41598-024-63821-2},
}
Figure