Traditional liquid biopsies for brain cancer are limited by the blood-brain barrier, but new AI-driven methods are improving detection. Researchers at Johns Hopkins University developed a technique that analyzes DNA fragments from tumors and immune cells in the blood, achieving about 75% detection in early studies. Another system, crossNN, examines epigenetic signatures and can classify tumors with over 99% accuracy using cerebrospinal fluid, potentially reducing the need for invasive biopsies.
AI is also enhancing imaging. A model from McGill University can detect how far metastatic brain cancer has spread using MRI scans, reaching around 85% accuracy. Combining MRI and PET with AI helps distinguish tumor growth from treatment effects, improving clinical decisions.
New surgical tools are increasing precision. FastGlioma helps surgeons identify hidden cancer cells during operations, while DeepGlioma can classify tumor subtypes in under 90 seconds, enabling faster, more accurate treatment planning.