Current and Emerging Insights on AI in Breast Imaging: An Interview with Mark Traill, MD, Part 3

For 40-year-old women identified as having a high lifetime risk for developing breast cancer, Mark Traill, M.D. said regular breast MRI screening for 30 years is “a very big ask” with poor compliance.

However, in an interview at the recent RSNA conference, Dr. Traill noted the emergence of image-based risk assessment artificial intelligence (AI) algorithms. He said the capability of these algorithms to ascertain short-term interval breast cancer risk offers promise in facilitating improved adherence to breast cancer screening protocols for women at high risk for breast cancer.

“You're not going to get an image with a circle around an abnormality. You're going to get a flat score that is an indicator of your short-term breast cancer risk. So rather than a lifetime risk, this is a short-term interval risk. It's much more actionable for the patient, much easier to manage for the patient, and that can be a tremendous advantage in getting the patient to come back in,” explained Dr. Traill, a breast radiologist affiliated with the University of Michigan Health West in Wyoming, Mich.

In his experience employing the algorithms in a research setting, Dr. Traill said the image-based risk assessment AI algorithms have demonstrated an ability to detect the “very earliest changes of breast cancer, often before the radiologist can recognize these changes.” He noted the algorithms are particularly effective at identifying distortion or very early calcification.

(Editor’s note: For additional content from the RSNA conference, click here.)

Dr. Traill added that current research comparing image-based risk assessment AI algorithms to clinical models for assessing breast cancer risk has shown that combining the models offer better results than either system used individually.

“Add genetics to that mix, and you're probably going to have a very powerful risk assessment that will give the patient some very specific information that may reveal a cancer in real time, or at least put (patients) on high alert that they need to come in when they're supposed to come in for a six month (follow-up),” maintained Dr. Traill,an assistant clinical professor at the Michigan State University College of Osteopathic Medicine.

(Editor’s note: For related content, see “Current and Emerging Insights on AI in Breast Imaging: An Interview with Mark Traill, MD, Part 1,” “Current and Emerging Insights on AI in Breast Imaging: An Interview with Mark Traill, MD, Part 2” and “Could a Mammography Worklist in Order of Increasing Breast Density Bolster Interpretation and Efficiency?”)

For more insights from Dr. Traill, watch the video below.