Application of MRI-Based Artificial Intelligence Models for Preoperative Molecular Subtyping and Prognostic Assessment of Midline Gliomas: A Multicenter Prospective Clinical Study

What this Trial Is About

Researchers are evaluating an MRI-based artificial intelligence (AI) model designed to predict key molecular markers H3K27M, IDH, and 1p/19q in patients with midline gliomas. This prospective multicenter study aims to validate the model's accuracy in preoperative molecular subtyping and its usefulness in assessing prognosis and supporting clinical decision-making across several neurosurgical centers. The study seeks to provide evidence for non-invasive, precise diagnosis and personalized management of these brain tumors. The AI model integrates tumor-containing slices from preoperative MRI scans with clinical variables such as patient age and sex. Using this information, the model predicts the molecular subtype of the glioma before surgery or biopsy. The study will assess the model's diagnostic performance by comparing its predictions to standard pathological results using measures like sensitivity, specificity, accuracy, F1 score, and AUC. The study will also evaluate the model's ability to predict patient survival outcomes and treatment response, and explore its potential role in guiding preoperative treatment planning. Participants will be patients with diffuse gliomas who have undergone preoperative MRI and have complete clinical data. Researchers will compare AI model predictions with postoperative or post-biopsy pathology results. The study includes monitoring diagnostic accuracy as the primary outcome. It is conducted prospectively in multiple centers, with ongoing collection of imaging and clinical data. The overall involvement duration varies with patient treatment and follow-up. The findings aim to enhance personalized diagnosis and prognosis in midline glioma care.

Multicenter Prospective Study on MRI AI Model for Midline Glioma Subtyping and Prognosis: