ObjectiveTo investigate the value of dual-energy CT virtual plain scanning combined with radiomics in the preoperative diagnosis of lung squamous cell carcinoma and lung adenocarcinoma, and guide the clinical development of individualized treatment plan.

MethodsThe clinical data of lung squamous cell carcinoma and lung adenocarcinoma 162 patients confirmed by pathology were retrospectively analyzed.Fifty-four patients with lung squamous carcinoma and 108 patients with lung adenocarcinoma were randomly divided into the training group and test group according to the ratio of 3:2.The preoperative dual-energy CT(tube voltage 70 kVp and 120 kVp) plain scannig images of two groups were analyzed using the Darwin Intelligent Research Platform(http://122.115.38.132:7043/manage/project-manage) to sketch the region of interest (ROI) and establish the prediction model.The prediction model was verified by receiver operating characteristic (ROC) curve, and the value of imaging features in the prediction of lung squamous cell carcinoma and lung adenocarcinoma was evaluated.

ResultsIn the training group, the area under curve(AUC) value was 0.77, the accuracy rate was 71.13%, the specificity was 0.75, the sensitivity was 0.69, and the 95% confidence interval(CI) was 0.67-0.90.In the test group, the AUC value was 0.74, the accuracy rate was 64.62%, the specificity was 0.59, the sensitivity was 0.67, and the 95%CI was 0.53-0.93.

ConclusionsDual-energy CT virtual plain scannig combined with radiomics can be used as an auxiliary means to differentiate lung squamous carcinoma from lung adenocarcinoma, which has guiding significance in the clinical treatment of lung cancer patients.