Interlesional Heterogeneity of EGFR Mutations: A Systematic Review and Meta-analysis

BACKGROUND: Activating epidermal growth factor receptor (EGFR) mutations are key drivers in non-small cell lung cancer (NSCLC) and other solid tumours, predicting responses to tyrosine kinase inhibitors (TKIs). Tumour heterogeneity alongside sampling and technical factors may contribute to discordant EGFR status across biopsies, complicating treatment decisions. However, systematic evidence on prevalence and drivers of discordance remains limited. METHODS: This systematic review and meta-analysis followed Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines and was registered in PROSPERO (CRD42024615727). MEDLINE, Embase, and Scopus (2004-2024) were searched for studies reporting EGFR mutation discordance in adult solid tumours. Eligible studies compared primary and metastatic tumours (tissue-tissue), tissue and liquid biopsies (tissue-liquid), or different liquid biopsies. Data extraction and QUADAS-2 risk of bias assessment were performed independently. Discordance proportions were analysed on the logit scale with Haldane-Anscombe correction when needed. Random-effects meta-analysis was conducted using the Paule-Mandel estimator with Hartung-Knapp-Sidik-Jonkman confidence intervals. Subgroup analyses and meta-regression were used to estimate pooled discordance and explore potential predictors. RESULTS: A total of 154 studies (15,560 patients) predominantly involving NSCLC were included. The pooled discordance rate was 16.1% (95% confidence interval 14.2-18.2). Rates were similar for tissue-tissue (16.8%) and tissue-liquid (15.5%), but higher for liquid-liquid (34.0%). Plasma was the most-studied liquid source (16.5%), while cerebrospinal fluid showed the highest discordance (35.5%). Prior TKI exposure was associated with higher discordance (25.8%) compared with treatment-naive patients (14.6%; p = 0.003). Patients who later developed resistance also had higher baseline discordance (21.0% vs 15.0%; p = 0.042). Discordance varied by metastatic site, from 15.1% in lymph nodes to 17.9% in brain/central nervous system. Meta-regression identified TKI exposure, resistance, and mutation prevalence as predictors. CONCLUSIONS: EGFR mutation discordance is common and clinically relevant, particularly in NSCLC, but varies substantially by sampling strategy, biofluid, treatment context, and metastatic site. Given the high between-study heterogeneity and the predominance of NSCLC and Asian cohorts, pooled estimates should be interpreted as descriptive summaries rather than universally generalisable benchmarks. These findings support integrated and context-aware sampling strategies for EGFR-targeted therapy and resistance monitoring.