Research Team Using Artificial Intelligence to Reduce False Positive Diagnoses of Breast Cancer

A team of researchers at the Computer Science and Artificial Intelligence Laboratory at the Massachusetts Institute of Technology are developing artificial intelligence that they believe will greatly reduce the number of false positives in breast cancer detection when women undergo mammograms. The research team is made up of scientists from MIT, Harvard Medical School, and Massachusetts General Hospital.

False positives result from lesions that appear suspicious on mammograms and these lesions often have abnormal cells when tested by biopsy. As a result, patients undergo surgery to remove the lesions. But in 90 percent of the cases the lesions are benign. The goal of the research is to develop a program that will be able to determine whether the lesions are benign before surgery is considered. If successful, the artificial intelligence program could reduce the number of women who undergo unnecessary surgery.

In early tests, the artificial intelligence program correctly identified 97 percent of the breast cancers as malignant and reduced the number of benign surgeries by 30 percent.

Regina Barzilay, the Delta electronics Professor of Electrical and Computer Science at MIT and a breast cancer survivor, explains that “because diagnostic tools are so inexact, there is an understandable tendency for doctors to over-screen for breast cancer. When there’s this much uncertainty in data, machine learning is exactly the tool that we need to improve detection and prevent over-treatment.”

Constance Lehman, professor at Harvard Medical School and chief of the breast imaging division at Massachusetts General Hospital in Boston, adds that “in the past we might have recommended that all high-risk lesions be surgically excised. But now, if the model determines that the lesion has a very low chance of being cancerous in a specific patient, we can have a more informed discussion with our patient about her options. It may be reasonable for some patients to have their lesions followed with imaging rather than surgically excised.”

A paper about the research effort, “High-Risk Breast Lesions: A Machine Learning Model to Predict Pathologic Upgrade and Reduce Unnecessary Surgical Excision,” was published on the website of the journal Radiology. It may be accessed here.