New AI tool shows promising accuracy when IDing survivors among breast cancer patients

Editor's Note

A new artificial intelligence (AI) tool has shown to accurately identify invasive breast cancer patients who are likely to be long-term survivors and could benefit from reduced chemotherapy, according to a new study. The findings were published in Nature Medicine on November 27. 

Researchers from Northwestern University collaborated with the American Cancer Society (ACS) for this study, which created a unique dataset of breast cancer patients representing patients from over 423 US counties. The researchers trained the AI with hundreds of thousands of human-generated annotations of cells and tissue structures within digital images of patient tissues supported by an international volunteer network of medical students and pathologists across several continents. 

This data was submitted over several years, until the AI model could reliably interpret images of breast cancer tissue. The AI system analyzes 26 different properties of a patient's breast tissue, generating an overall score and individual scores related to the cancer, immune, and stromal cells to help guide individual treatment plans. 

This study is the first to use AI to evaluate both cancerous and non-cancerous elements of invasive breast cancer and could serve as an important tool to prevent unnecessary and unpleasant chemotherapy treatments.