Deep Learning Radiomics Based on MRI for Differentiating Benign and Malignant Parapharyngeal Space Tumors

Area of Science:

• Oncology
• Radiology
• Artificial Intelligence

Background:

• Parapharyngeal space (PPS) tumors require accurate preoperative diagnosis to guide treatment.
• Distinguishing malignant from benign PPS tumors is clinically challenging.

Purpose of the Study:

• To develop and validate a diagnostic tool for parapharyngeal space tumors.
• To integrate deep learning and radiomics features for improved diagnostic accuracy.

Main Methods:

• Retrospective study of 217 patients with PPS tumors.
• Development of deep learning (DL), radiomics (Rad), and deep learning radiomics (DLR) models using neck MRI.
• Evaluation of model performance using AUC, specificity, and sensitivity; clinical utility assessed by DCA.

Main Results:

• The DLR model achieved the highest AUC (0.899 training, 0.821 test set).
• DLR model demonstrated superior performance compared to standalone DL and Rad models.
• Decision curve analysis confirmed the clinical utility of the DLR model.

Conclusions:

• The DLR model shows high predictive ability for diagnosing malignant and benign PPS tumors.
• This integrated model can significantly aid in preoperative diagnosis of PPS tumors.
• The DLR model serves as a valuable tool for clinical decision-making.