Study of AI-Assisted Diagnosis for Common Malignant Brain Tumors Using MRI Images

What this Trial Is About

Researchers are building a large-scale MRI database focused on malignant brain tumors, including gliomas, brain metastases, and lymphomas. The study aims to develop an artificial intelligence AI system using deep learning to segment and classify multiple brain tumor subtypes. This approach seeks to improve non-invasive preoperative diagnosis accuracy and reduce the need for biopsies, addressing limitations in current methods due to small sample sizes and limited classification performance. The study collects retrospective data from two main centers, gathering 3,000 cases confirmed by histopathology with preoperative multimodal MRI scans, mainly CE-T1WI and T2-FLAIR images taken on 3.0T or 1.5T MRI machines. The AI system is designed to automatically segment complex tumor tissues and assist in diagnosing common malignant brain tumors. This deep learning-based diagnostic tool aims to enhance the accuracy of brain tumor classification and support clinical decision-making. Participants MRI images and pathology results are used to build the AI-assisted diagnostic system within 30 days. The study involves reviewing and analyzing imaging data retrospectively without active treatment. Researchers focus on constructing and validating the AI system to improve auxiliary diagnosis for brain malignancies. The total participation involves providing access to imaging and diagnostic information, with no additional procedures required from patients.

AI-assisted Diagnosis of Malignant Brain Tumors