Abstract:Parotid gland tumors are relatively common salivary gland disorders, with the majority being epithelial tumors and a minority originating from mesenchymal tissues (derived from mesodermal mesenchyme). Clinically, they primarily manifest as masses in the pre-auricular region, facial pain, numbness, and restricted mouth opening, with varying presentations depending on the type of tumor. The pathological types of parotid gland tumors are complex, and different types require different surgical approaches. Therefore, accurately diagnosing the type of parotid gland tumor is crucial for developing an appropriate surgical plan. Due to the abundance of soft tissue in parotid tumors, their complex histological composition, and significant variability in biological behavior, traditional imaging techniques often fall short in providing definitive qualitative assessments. With the advancement of medical digitization and technological improvements, modern intelligent healthcare has increasingly integrated big data and artificial intelligence, enabling the acquisition of deeper imaging data insights. Radiomics refers to the high-throughput extraction of large amounts of imaging information from modalities such as ultrasound, CT, and MRI, facilitating tumor segmentation, feature extraction, and model construction to assist physicians in making the most accurate diagnoses. Based on quantitative analysis of medical images, radiomics can quantify tumor heterogeneity and plays a significant role in the differential diagnosis of parotid gland tumors. This article reviews the latest advancements in the application of radiomics for the differential diagnosis of parotid gland tumors, providing valuable references for accurate clinical diagnosis.