AI model predicts brain metastasis risk in advanced breast cancer patients

A machine‑learning tool called CNSPredict was presented at the ASCO 2026 meeting in the United States. The model analyzes clinical information, tumor characteristics and genetic alterations to estimate the likelihood that women with metastatic breast cancer will develop brain metastases.

In validation studies the tool identified a high‑risk group in which roughly one‑quarter of patients developed brain metastases within two years, compared with about five percent in the low‑risk group. Risk factors associated with higher probabilities included triple‑negative disease, four or more metastatic sites, adrenal gland involvement and mutations in TP53, PTEN or ERBB2. Oncologist Miguel Zugman said the AI approach “could help identify which patients truly have higher risk of complications and allow for more individualized imaging, monitoring and treatment.” While promising, the technology remains under validation and current guidelines do not recommend routine brain imaging for asymptomatic patients.