The value of 18F-2-fluoro-2-deoxyglucose PET/computed tomography radiomics in epidermal growth factor receptor mutation subtypes prediction and progression-free survival in advanced lung adenocarcinoma patients with first-line epidermal growth factor receptor tyrosine kinase inhibitors therapy

Objective Epidermal growth factor receptor tyrosine kinase inhibitors (EGFR-TKIs) improve survival of EGFR-mutated lung adenocarcinoma (LUAD); however, outcomes vary with genetic subtypes and tumor heterogeneity in late-stage. We aimed to construct pretreatment 18 F-2-fluoro-2-deoxyglucose PET/computed tomography ( 18 F-FDG PET/CT) radiomics models for EGFR-subtype prediction and prognosis in first-line TKIs-treated patients. Methods We retrospectively analyzed 131 EGFR-mutated advanced LUAD patients from 2017 to 2024: 72 exon 19 deletion (19Del) and 59 exon 21 L858R (21L858R) mutations. After feature selection, support vector machine models: PET, CT, PET-CT, and clinical PET-CT combined models were built. Performance was evaluated by areas under the receiver operating characteristic curve (AUC), calibration curves, and decision curve analysis (DCA). Model-derived radscore was used to explore progression-free survival (PFS) in first-line EGFR-TKIs-treated patients. Multivariate Cox regression was conducted to identify independent factors. Results The clinical PET/CT combined model achieved AUCs of 0.854 [95% confidence interval (CI): 0.776–0.932] and 0.785 (95% CI: 0.639–0.932) in training and test sets. The calibration curves showed good agreement, and the DCA confirmed clinical utility. Among 125 successfully followed patients, 21L858R mutation patients showed poorer median PFS ( P = 0.008) compared to 19Del mutation. High radscore [hazard ratio (HR): 0.57, 95% CI: 0.34–0.94, P = 0.029], third-generation TKI therapy (HR: 0.45, 95% CI: 0.27–0.73, P = 0.001), and high maximum standardized uptake value (HR: 1.67, 95% CI: 1.03–2.69, P = 0.036) were independent factors of PFS. Conclusion Integrating 18 F-FDG PET/CT radiomics with clinical data precisely identifies EGFR mutation subtypes and guides initial TKI monotherapy in advanced LUAD.