Computed tomography-based texture analysis for predicting adjuvant therapy response in postoperative patients with EGFR-mutant non-small cell lung cancer.

Background: Emerging research suggests that epidermal growth factor receptor (EGFR)-tyrosine kinase inhibitor (TKI) agents are not universally effective in patients with EGFR-mutant non-small cell lung cancer (NSCLC), with many developing varying degrees of acquired resistance. Studies have found that such resistance is significantly associated with certain imaging features. Therefore, this study aimed to examine the application value of combining texture analysis techniques with computed tomography (CT) images in predicting the efficacy of targeted adjuvant therapy in patients with EGFR-mutant NSCLC following surgery. Methods: The basic clinical data of patients with EGFR-mutant NSCLC who underwent surgery followed by targeted therapy with first-generation EGFR-TKIs at the First Affiliated Hospital of Hebei North University between January 2019 and September 2024 were retrospectively collected. Texture features of the tumor were extracted from chest CT images via 3D Slicer software, and after standardization, feature dimensionality reduction and selection were performed through correlation analysis and the least absolute shrinkage and selection operator (LASSO) algorithm. Univariate and multivariate logistic regression analyses were then conducted on clinical and texture features to identify independent prognostic factors. A clinical model, a radiomics texture model, and a joint model were developed with R software (The R Foundation for Statistical Computing). Model performance was assessed with the area under the curve (AUC), calibration curves, and decision curve analysis (DCA). A nomogram was constructed based on the combined model. Results: In this study, 150 texture features were extracted, and dimensionality reduction was conducted with the LASSO algorithm. At the optimal l-value of 0.0753, three candidate features were preliminarily selected. These three features were then subjected to univariate and multivariate logistic regression analyses, ultimately yielding one significant texture feature. Smoking was found to be an independent prognostic factor for patients with EGFR-mutant NSCLC (P<0.05). The AUC for predicting poor prognosis in patients with EGFR-mutant NSCLC was 0.756 for the clinical model, 0.771 for the texture-analysis model, and 0.90 for the combined model. The combined model demonstrated significantly better predictive performance than the individual models (P<0.05). DCA further confirmed the superior clinical utility of the combined model. A nomogram was constructed to provide an intuitive and quantitative tool for evaluating treatment efficacy in individual patients. Conclusions: CT texture-based analysis demonstrated favorable predictive performance in assessing the efficacy of postoperative adjuvant targeted therapy in patients with EGFR-mutant NSCLC. The proposed model offers an intuitive and reliable reference for individualized treatment planning.