Predicting isocitrate dehydrogenase status in glioma using hierarchical attention-based deep 3D multiple instance learning.

Authors: Xie Q, Sun Y, Liang Y, Shang Y, Wang H, Wang F, Wei R, Chen B, Zhang M, Niu C.

DOI: 10.3389/fonc.2025.1665690

Abstract Summary

AI model predicts glioma IDH mutation status from routine MRI scans, eliminating invasive biopsies. Tested on 840 patients, the HAB-MIL framework achieved 89-92% accuracy without requiring detailed image annotations. This non-invasive approach could streamline diagnosis and treatment planning for brain tumor patients while reducing physician workload.

Why Brain? 🧠

AI model predicts glioma IDH mutation status from MRI scans non-invasively with 89-92% accuracy, eliminating need for surgery and reducing physician workload in tumor diagnosis and treatment planning.

License: CC BY.

The image is AI-generated for illustrative purposes only. Courtesy of Midjourney.