AI model predicts breast cancer response using longitudinal MRI scans

Researchers have developed a new longitudinal framework that combines a frozen 3D foundation encoder, Pillar-0, with a Temporal Dynamics Network (TDN) to predict neoadjuvant chemotherapy response in breast cancer patients. This model analyzes serial Dynamic Contrast-Enhanced (DCE) MRI scans taken at four clinical timepoints, integrating this imaging data with clinical and treatment information. When evaluated on a cohort of 982 patients from the I-SPY2 and ACRIN-6698 trials, the framework achieved a test AUROC of 73.6% and a balanced accuracy of 69.1%, demonstrating that longitudinal 3D imaging provides valuable complementary information to clinical variables for predicting pathologic complete response (pCR). AI

IMPACT This model could improve treatment assessment for breast cancer patients by predicting response to chemotherapy from serial MRI scans.

RANK_REASON The item is a research paper detailing a new AI model for medical image analysis. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CV →

• ACRIN-6698
• arXiv
• Fidel Omar Tito Cruz
• Pillar-0
• Temporal Dynamics Network

• paper
• model release
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →