Magnetic resonance imaging–based deep learning for predicting EGFR mutations in metastatic non-small cell lung cancer: A systematic review and meta-analysis.

Abstract

e14008 Background: Lung cancer is the most common cancer to metastasize to the brain. Identifying the EGFR mutation is critical for determining treatment strategies and prognosis. However, 17.26% of patients show EGFR mutation discordance between the primary tumor and CNS metastasis. On the other hand, it is not always feasible to obtain brain samples. Many studies have developed deep learning (DL) models to predict the existence of EGFR mutations using radiomics. In this paper, we aim to measure the accuracy of diagnostic tests in those studies. Methods: PubMed, Scopus, Cochrane Library and Web of Science databases were searched from inception to January 2025, to identify studies measuring the ability of DL models in differentiating between the types of EGFR mutation in patients with lung cancer brain metastases. The quality assessment of the included studies was performed using QUIPS tool. The diagnostic accuracy meta-analysis was conducted using Meta-DiSc software. Results: Nine studies were included in our systematic review and meta-analysis, encompassing 260 patients with EGFR mutant type and 264 patients with EGFR wild type. Our analysis revealed that DL approaches exhibit a high ability to differentiate between EGFR mutation types, with a sensitivity of 0.82 (95% CI: 0.73–0.88) and a specificity of 0.88 (95% CI: 0.78–0.94). Additionally, the positive likelihood ratio and negative likelihood ratio were 6.87 (95% CI: 3.5–13.3) and 0.20 (95% CI: 0.13–0.31), respectively. The diagnostic odds ratio was 33.4 (95% CI: 13.5–82.5), which is indicative of excellent diagnostic accuracy. Conclusions: DL models demonstrate exceptional diagnostic accuracy in detecting EGFR mutations in brain metastases from lung cancer. These findings highlight the potential of AI-driven approaches to enhance personalized treatment strategies and improve outcomes for patients with EGFR-mutant NSCLC brain metastases.