Prediction of EGFR gene mutations in non-small cell lung cancer using clinic combined with CT imaging features
ZUO Lihua
CHANG Jiangli
ZHAO Qirui
ZHU Yonghua
JI Hongbo
HAN Dan
Abstract：Objective To explore the predictive role of clinical and CT imaging features in epidermal growth factor receptor(EGFR)gene mutations in non-small cell lung cancer(NSCLC).Methods A retrospective collection was performed on 412 patients with pathologically confirmed NSCLC who had clinical data,genetic testing results,and CT imaging data were con-ducted,including 292 cases in the training set and 120 cases in the validation set.Univariate analysis was performed to assess differences in clinical and imaging features between the mutation-positive and mutation-negative groups in the training set.Fea-tures with statistical significance in univariate analysis were included in multivariate analysis to identify independent predictive factors for EGFR mutations.A logistic regression model was established,and a nomogram was created to visualize the model.The area under the curve(AUC)was used to evaluate the model's effectiveness in predicting EGFR gene mutations,while calibra-tion curves and decision curves were used to assess the model's practicality.Results 1)Significant differences were found be-tween the EGFR mutation-positive and negative groups regarding gender,smoking history,pathological type,lesion type,halo sign,liquefactive necrosis,air bronchogram,vascular bundle sign,and pleural indentation sign(P<0.05);2)Multivariate analysis identified no smoking history,halo sign,liquefactive necrosis,air bronchogram,vascular bundle sign as independent predictive factors for EGFR mutations;and 3)The AUC value for predicting EGFR gene mutations in the training set ROC curve was 0.746,with sensitivity,specificity,and accuracy of 71.2%,66.2%,and 64.2%,respectively;the AUC value for the valida-tion set was 0.708,with sensitivity,specificity,and accuracy of 74.6%,62.3%,and 65.0%.Calibration curves indicated good consistency between the predicted model and observed results in both the training and validation sets.Conclusion The clinical-CT imaging model for predicting EGFR mutations in NSCLC patients holds certain value,which can serve as a non-invasive method for predicting EGFR mutations in NSCLC patients.
Keywords：Non-small cell lung cancerEpidermal growth factor receptorPredictionTomographyX-ray computed
Publication Date：2025-02-28
Online Publishing Date：2025-08-15(First online date of this platform, not the publication date of the document)
Pages：6( 33-38 )