Regina Barzilay, PhD, a recently appointed Associate Member of the Ragon Institute and a MIT School of Engineering Distinguished Professor for AI and Health, is one of the principals of the new Artificial Intelligence/Machine Learning/human immunology collaborative initiative of the Ragon Institute and the Abdul Latif Jameel Clinic for Machine Learning in Health and the Schwarzman College of Computing of MIT. A winner of numerous scientific awards, including the MacArthur Fellowship “Genius Grant” in 2017, Barzilay’s groundbreaking discoveries include applying artificial intelligence and machine learning (“AI/ML”) to mammograms, predicting with impressive accuracy the incidence of breast cancer in patients up to five years before the emergence of clinical symptoms or detection by standard methods.

With philanthropic support from Mark and Lisa Schwartz, the Ragon Institute is well-positioned to expand upon AI/ML efforts already underway at the Ragon Institute, such as those from Bryan Bryson, PhD (Hie, B., et al. Science 2021). Barzilay has already begun working with Ragon Institute faculty Galit Alter, Gaurav Gaiha and Mathias Lichterfeld to develop algorithms to better understand and control the interaction of human immunology and diseases at the cellular level.

Barzilay spent years researching natural-language processing, an important AI/ML building block. Then, in 2014, she was diagnosed with breast cancer, prompting her to pivot her research toward a more urgent purpose — that of saving lives. Collaborating with Connie Lehman, MD, PhD, a breast radiologist at MGH, Barzilay and her team developed an AI/ML algorithm — called Mirai — to analyze hundreds of thousands of mammograms. The results accurately predicted future risk of cancer at more than twice the rate offered by traditional methods, a difference that could affect millions of patients. If confirmed in upcoming clinical trials, Mirai would signal a seismic shift in the diagnosis and treatment of breast cancer. Screenings, now age-based, could instead target each individual patient’s specific risk and would provide much earlier prognostic, not just diagnostic, information. These refinements would likely save countless lives and serve as an important building block in the development of many more AI/ML applications in clinical medicine.

Barzilay believes that AI/ML will have an equally groundbreaking impact on the field of human immunology, with the Ragon/MIT initiative as a catalyst. “AI/ML allows us to think differently about disease and its mechanisms, and using AI/ML at the cellular level allows us to understand human immunology and disease from a different perspective in ways that can be game-changing,” she says. She believes that the initiative is also filling an important gap because “the field of immunology and the field of AI/ML are both very advanced, but historically, little work has been done at the intersection.” One of Barzilay’s current projects is a collaboration with Galit Alter, PhD, a Ragon Institute principal investigator, to predict which antibodies might be the most effective at neutralizing new variants of viruses such as SARs-CoV-2, which causes COVID. These efforts have already helped expedite such research.

Appreciating that most of her computer science colleagues “are not fortunate enough to have a Ragon Institute for their work,” Barzilay anticipates that Ragon scientists will provide the crucial link of “translating the language of immunology into language AI/ML scientists can use to contribute to the field.” She fully envisions that the end result will not only train a new generation of scientists who will be “bilingual” in the language of both immunology and AI/ML, and thereby “able to push both fields further forward,” but will also enable computer scientists elsewhere to access and use such information. Barzilay is convinced that such accessibility will further empower this impending revolution in the field of human immunology — and, she is not shy to suggest, even in all of basic biology — being ignited by the Ragon Institute and MIT’s AI/ML/human immunology collaborative initiative.

Learn more about Barzilay’s work in the Washington Post and from her appearance on Good Morning America.