International Journal of General Medicine (Aug 2022)
Prediction of VEGF and EGFR Expression in Peripheral Lung Cancer Based on the Radiomics Model of Spectral CT Enhanced Images
Abstract
Linhua Wu,1 Jian Li,1 Xiaowei Ruan,2 Jialiang Ren,3 Xuejun Ping,4 Bing Chen1 1Department of Radiology, General Hosipital of Ningxia Medical University, YinChuan, Ningxia Hui Autonomous Region, People’s Republic of China; 2Department of Radiology, People’s Hospital of Ningxia Hui Autonomous Region, Yinchuan, Ningxia Hui Autonomous Region, People’s Republic of China; 3Department of Pharmaceuticals Diagnosis, GE Healthcare, Beijing, People’s Republic of China; 4Department of Clinical Medical Faculty, Medical University of Ningxia, Yinchuan, Ningxia Hui Autonomous Region, People’s Republic of ChinaCorrespondence: Xuejun Ping, Department of Clinical Medical Faculty, Medical University of Ningxia, No. 1160 Shengli Street, Xingqing District, Yinchuan, Ningxia Hui Autonomous Region, People’s Republic of China, Tel +8613709516402, Email [email protected]: Energy spectrum CT is an effective method to evaluate the biological behavior of lung cancer. Radiomics is a non-invasive technology to obtain histological information related to lung cancer.Purpose: To investigate the value of the radiomics models on the bases of enhanced spectral CT images of peripheral lung cancer to predict the expression of the vascular endothelial growth factor (VEGF) and epidermal growth factor receptor (EGFR).Material and Methods: This study retrospectively analyzed 73 patients with peripheral lung cancer confirmed by postoperative pathology. All patients underwent dual-phase enhanced spectral CT scans before surgery. Regions of interest (ROI) were delineated in the arterial phase and venous phase. Key radiomics features were extracted and models were established to predict the expression of VEGF and EGFR, respectively. All models were established based on the expression levels of VEGF and EGFR in tissues detected by immunohistochemical staining as reference standards. Receiver operating characteristic (ROC) curve and calibration curve were used to evaluate the predictive performance of each model, and decision curve analysis (DCA) was used to evaluate the clinical utility of the models.Results: In predicting the expression level of VEGF, the combined (COMB) model composed of one spectral feature and two radiomics features achieved the best performance with area under ROC (AUC) 0.867 (95% CI: 0.767– 0.966), accuracy of 0.812, sensitivity of 0.879, and specificity of 0.667. According to the expression level of EGFR, three importance radiomics features were retained in the arterial and venous phases to establish the multiphase phase model which has the best performance with AUC of 0.950 (95% confidence interval: 0.89– 1.00), accuracy of 0.896, sensitivity of 0.868, and specificity of 1.Conclusion: The radiomics model of enhanced spectral CT images of peripheral lung cancer can predict the expression of EGFR and VEGF.Keywords: peripheral lung cancer, radiomics, energy spectrum CT, VEGF, EGFR
