| Objective: To explore the value of CT radiomics method in predicting lymph node metastasis of laryngeal carcinoma before operation.Methods: The imaging data of 304 patients with laryngeal carcinoma confirmed by surgical pathology examination were collected.243 cases and61 cases were randomly selected as training set and verification set,respectively.The volume of tumor was manually drawn on the venous phase CT images and radiomics features were extracted.In the training set,LASSO regression was used to screen the radiomics features,and logistic regression analysis was performed to construct the radiomics model,morphological model and their combined model.ROC curves were used to evaluate the discrimination efficiencies of different models,and Delong’s test was performed for the ROC curve of the model.Finally,the value of the model in the validation set was evaluated.Results: Five radiomics features were screened.In the training set,the AUC of the radiomics model was 0.85(95% CI: 0.77-0.93),the AUC of the morphological model was 0.77(95% CI: 0.69-0.82),the AUC of the combined model was 0.82(95% CI: 0.74-0.91);in the validation set,the AUC of the radiomics model was 0.83(95% CI: 0.69-0.97),the AUC of the morphological model was 0.72(95% CI: 0.65-0.79),and the AUC of the combined model was 0.73(95% CI: 0.56-0.91).Delong’s test showed that there were statistically significant differences between the radiomics model and the combined model as well as the radiomics model and the morphological model(P<0.05),but there was no significant difference between the combined model and the morphological model(P>0.05).Conclusion: The radiomics model can predict the lymph node metastasis of laryngeal carcinoma before operation,and its diagnostic efficiency is higher than those of the morphological model and combined model. |