Objective To explore the clinical application value of multimodal radiomics in differentiating parotid pleomorphic adenoma(PA)from adenolymphoma(AL).Methods The clinical and imaging data of 68 cases of PA and 52 cases of AL were retrospectively analyzed.All patients underwent ultrasound examination,enhanced CT scan and enhanced MRI scan of the neck before the operation.All patients were randomly divided into training group(n=84)and validation group(n=36)according to the ratio of 7∶3.The 3D Slicer software was used to manually draw the lesion area of the preoperative images and perform radiomics feature extraction.The best feature subset was selected to establish the radiomics model,and the diagnostic efficacy of different models was evaluated by the receiver operating characteristic(ROC)curve.Results The study found that age,gender,and smoking history were effective in differentiating parotid PA from AL,and were used to construct a clinical diagnostic model.Eighteen features were selected through dimensionality reduction to establish the radiomics model.A multimodal combined diagnostic model was then constructed by integrating the radiomics model with gender,age,and smoking history.Compared to the clinical model and the radiomics model,the multimodal combined diagnostic model demonstrated the highest area under the curve(AUC)for distinguishing PA from AL in both the training group and the validation group.Conclusion The combination of radiomics based on neck ultrasound,CT,and MRI with clinical characteristics shows significant clinical application value in the preoperative differentiation of PA and AL.

Objective To explore the clinical application value of multimodal radiomics in differentiating parotid pleomorphic adenoma(PA)from adenolymphoma(AL).Methods The clinical and imaging data of 68 cases of PA and 52 cases of AL were retrospectively analyzed.All patients underwent ultrasound examination,enhanced CT scan and enhanced MRI scan of the neck before the operation.All patients were randomly divided into training group(n=84)and validation group(n=36)according to the ratio of 7∶3.The 3D Slicer software was used to manually draw the lesion area of the preoperative images and perform radiomics feature extraction.The best feature subset was selected to establish the radiomics model,and the diagnostic efficacy of different models was evaluated by the receiver operating characteristic(ROC)curve.Results The study found that age,gender,and smoking history were effective in differentiating parotid PA from AL,and were used to construct a clinical diagnostic model.Eighteen features were selected through dimensionality reduction to establish the radiomics model.A multimodal combined diagnostic model was then constructed by integrating the radiomics model with gender,age,and smoking history.Compared to the clinical model and the radiomics model,the multimodal combined diagnostic model demonstrated the highest area under the curve(AUC)for distinguishing PA from AL in both the training group and the validation group.Conclusion The combination of radiomics based on neck ultrasound,CT,and MRI with clinical characteristics shows significant clinical application value in the preoperative differentiation of PA and AL.