Agreement Between Different Methodologies for Non-Invasive p.T790M and EGFR Sensitizing Mutation Testing

Abstract

Background. Tyrosine kinase inhibitors (TKIs) are the current standard of care for patients with advanced EGFR-mutant non-small cell lung cancer (NSCLC). However, most patients progressed within 1 to 2 years. The EGFR p.T790M mutation is the most common resistance mechanism to first and second generation EGFR TKIs. The identification of p.T790M mutation is of considerable clinical relevance as osimertinib has demonstrated clinical efficacy in this setting. Guidelines recommend testing for the p.T790M mutation in blood at relapse to TKIs, and re-biopsy only in case of a negative result. Several blood based methodologies for detection of EGFR mutations have been developed in the recent years. However, the number of comparison studies between platforms is very limited.

Method. This is a multicenter, cross-sectional study (ClinicalTrials.gov Identifier: NCT03363139) performed by the Spanish Lung Cancer Group. Samples from 75 consecutive EGFR mutant NSCLC patients were collected at disease progression to first line TKI treatment. The presence of EGFR mutations in the cfDNA was evaluated in 39 samples by 7 methodologies, namely: Cobas® EGFR Mutation Test v2 (Roche Diagnostics), Therascreen EGFR Plasma RGQ PCR Kit (Qiagen), QuantStudio® 3D Digital PCR System (Thermofisher), a 5′-nuclease real-time PCR (TaqMan®) assay in presence of PNA, OncoBEAM EGFR (Sysmex Inostics), NGS with two different gene panels: Oncomine® (Thermofisher) and Lung Cancer Panel (Qiagen). The agreement between methodologies was assessed using the kappa coefficient (K) and its corresponding 95% confidence intervals (95% CI). For quantitative variables the concordance correlation coefficient (ccc) was used.

Result. Complete results are available for 39 patients. Overall, the agreement between all methodologies for the detection of p.T790M mutation as well as the original EGFR sensitizing mutation was good (K=0.669; 95CI: 0.504-0.835 and K=0.750 95CI: 0.599-0.899 respectively). Remarkably, the agreement between FDA-approved methodologies for p.T790M detection was almost perfect (K=0.926; 95CI: 0.712-1) and good for the EGFR sensitizing mutations (K=0.657; 95CI: 0.417-0.902). Similarly, the agreement between NGS-based methodologies for the detection of p.T790M and the EGFR activating mutations was very high (K=0.843; 95CI: 0.567-1 and K=0.872 95CI: 0.595-1 respectively). Moreover, concordance between both technologies for p.T790M and EGFR sensitizing mutation mutant allele frequency was excellent (ccc=0.956; 95CI: 0.906-1 and ccc=0.980 95CI: 0.950-1 respectively). The proportion of samples that were positive for p.T790M detection varied from 28% (PCR based technologies) to 37% depending on the methodology.

Conclusion. NGS and PCR-based methodologies show a good to excellent agreement for the detection of EGFR mutations, including the p.T790M. Our results support the use of liquid biopsies for non-invasive testing of clinically relevant mutations (Data from the whole cohort will be presented at the meeting).