Concordance of <i>EGFR</i> and <i>KRAS</i> mutation status between primary non–small cell lung cancer and corresponding metastases: A systematic review and meta-analysis.

Abstract

e20563 Background: Accurate molecular profiling is critical for NSCLC treatment selection. Metastatic tissue is frequently used for genotyping when primary samples are unavailable. However, reported concordance of driver mutations between primary tumors and metastases varies. We conducted a meta-analysis to quantify concordance rates for EGFR and KRAS alterations. Methods: We searched PubMed, Embase, and Scopus till January 2026 for studies reporting paired primary metastatic testing for EGFR and/or KRAS in NSCLC. Random effects meta analyses were performed using generalized linear mixed models. Subgroup analyses evaluated metastatic site and detection method. Risk of bias was assessed using adapted QUADAS-2; certainty using GRADE. Results: Thirty studies comprising 1,668 paired samples were included. Seventeen studies (1,015 pairs) evaluated EGFR concordance, yielding 86.5% (95% CI, 78.1–91.9; I² = 60.3%). Concordance varied by metastatic site: 78.4% in brain to 91.2% in lymph nodes, though differences were not statistically significant. Eleven studies (369 pairs) assessed KRAS concordance at 85.4% (95% CI, 76.3–91.4; I² = 59.2%). EGFR protein expression and gene amplification showed lower concordance (76.4% and 73.0%). Certainty was moderate for EGFR and low for KRAS. Conclusions: EGFR and KRAS mutation status demonstrates high but imperfect concordance between primary NSCLC and metastases. Clinically meaningful discordance persists, supporting cautious interpretation and consideration of repeat testing when feasible. GRADE summary. Outcome Pairs Estimate (95% CI) Certainty EGFR mutation 1,015 86.5% (78.1–91.9%) Moderate KRAS mutation 369 85.4% (76.3–91.4%) Low EGFR IHC 195 76.4% (66.3–84.1%) Low EGFR FISH 89 73.0% (62.9–81.2%) Low