Back to today’s briefing

Story perspectives

New AI Model Boosts Medulloblastoma Diagnosis Accuracy to 78%

11/7/2025

50 10

Summary

Story summary

Researchers developed MB-CNN to classify medulloblastoma subtypes from MRI scans with about 78% accuracy.
The model also predicts prognosis-related genetic alterations with high sensitivity: TP53 mutations 91%, MYC amplifications 84%, and chromosome 11 deletions 87%.
Led by Dr. Yanong Li of Capital Medical University, the study suggests MB-CNN could reduce biopsies and speed treatment decisions, though multicenter validation is needed.

1 of 1

Story summary

Researchers developed MB-CNN to classify medulloblastoma subtypes from MRI scans with about 78% accuracy.
The model also predicts prognosis-related genetic alterations with high sensitivity: TP53 mutations 91%, MYC amplifications 84%, and chromosome 11 deletions 87%.
Led by Dr. Yanong Li of Capital Medical University, the study suggests MB-CNN could reduce biopsies and speed treatment decisions, though multicenter validation is needed.