Developing an AI Model for Diagnosing Various Brain Diseases Using Multiple Types of Medical Data

What this Trial Is About

Researchers are developing a new, comprehensive AI vision-language foundation model to improve diagnosis and interpretation of various brain diseases. This observational study aims to use multiple types of data including patient demographics, medical imaging like MRI, CT, and PET scans, histopathological and genomic data when available, plus other lab tests to create a detailed screening and diagnostic tool for brain diseases. The study also seeks to establish a diagnostic model that includes uncertainty quantification and automated report generation for a wide range of brain conditions. Additionally, the model will be validated using MRI scans from large populations. This study does not involve any interventions or treatments. Instead, it collects and analyzes a variety of clinical and imaging data from patients diagnosed with brain diseases such as brain tumors and neurodegenerative disorders, as well as from individuals without brain disease who have complete brain MRI scans and clinical data. Data quality is important, so cases with incomplete or poor-quality MRI scans are excluded from analysis. Participants will contribute their existing clinical case data and brain imaging results to support model development and validation. The main outcome measure is the performance of the AI model in diagnosing brain diseases during the perioperative period. Since this is an observational study, there are no active treatments or procedures for participants. The study is sponsored by Xiangya Hospital of Central South University and aims to advance brain disease diagnosis through large-scale data analysis and AI technology.

A Vision-Language Foundation Model for Brain Disease Diagnosis From Multimodal Data