How an AI Scientist Turned Her Breast Cancer Diagnosis Into a Tool to Save Lives

November 4, 2019

When artificial intelligence researcher Regina Barzilay was first diagnosed with breast cancer in 2014, she says she was struck by immediate questions: “Am I going to survive? What’s going to happen to my son?”

But soon, the Massachusetts Institute of Technology scientist began asking a broader one: Why couldn’t her cancer have been diagnosed earlier?

Barzilay’s quest to find an answer would lead to a remarkable result: the development of an AI-based system for early detection of breast cancer, with the ability to predict whether a patient is likely to develop the disease in the next five years. A technology that had not yet penetrated the hospital setting now has the potential to save many thousands of lives each year.

Barzilay’s story unfolds as part of a new FRONTLINE documentary, In the Age of AI, that explores both the promise and the perils of artificial intelligence.

The promise includes the fact that, supplied with data from tens of thousands of mammograms and with the ability to understand diagnostic notes, recent advances in deep learning “can take hundreds of thousands of images where the outcome is known and learn… the very unique patterns that correlate highly with future occurrence of the disease,” Barzilay tells FRONTLINE in the above excerpt from the film.

It’s a demonstration of how artificial intelligence can outstrip the limitations of the “human capacity to recognize formalized patterns,” the MIT professor says of the software. It could result in cancers like hers being diagnosed far earlier — before invasive interventions like chemotherapy are necessary.

“When we find it early, we cure it, and we cure it without having the ravages to the body when we diagnose it late,” says Dr. Connie Lehman, head of the Breast Imaging Center at Massachusetts General Hospital, which partnered with Barzilay on developing the new predictive tool.

The above excerpt from In the Age of AI traces Barzilay and Lehman’s efforts and alliance in detail — from Barzilay’s early surprise that standard machine learning techniques weren’t already being applied in hospital settings, to the process of gaining access to thousands of mammograms and records from MGH’s breast imaging program, to what Barzilay and Lehman found when using deep learning to help analyze them.

“We are using technology not only to be better at assessing the breast density, but to get more to the point of what we’re trying to predict: ‘Does this woman have a cancer now and will she develop a cancer in five years?’” Lehman explains.

“It’s going to change the face of breast cancer,” she says of the AI technology.

To learn more, watch In The Age of AI starting November 5. Produced, written and directed by David Fanning and Neil Docherty, the documentary is a comprehensive survey of how artificial intelligence is transforming our world — both for good and for ill. It shows how, in cases like Barzilay’s, AI algorithms are ushering in a new age of great problem-solving potential. But the film also examines AI’s dark side — from fears about the future of work; the growing rivalry between the U.S. and China; and to privacy violations and threats to democracy, at home and abroad.

In the Age of AI premieres Tuesday, November 5 at 9 p.m. EST/8 p.m. CST. Tune in or stream on PBS, at, or on the PBS Video App.

Patrice Taddonio

Patrice Taddonio, Digital Writer & Audience Development Strategist, FRONTLINE



In order to foster a civil and literate discussion that respects all participants, FRONTLINE has the following guidelines for commentary. By submitting comments here, you are consenting to these rules:

Readers' comments that include profanity, obscenity, personal attacks, harassment, or are defamatory, sexist, racist, violate a third party's right to privacy, or are otherwise inappropriate, will be removed. Entries that are unsigned or are "signed" by someone other than the actual author will be removed. We reserve the right to not post comments that are more than 400 words. We will take steps to block users who repeatedly violate our commenting rules, terms of use, or privacy policies. You are fully responsible for your comments.

blog comments powered by Disqus

More Stories

Cheat Codes: Students Search For Shortcuts as Virtual Schooling Expands
Cheating has always been an issue in schools, but there is little getting in the way for students today. Shared answers have become even more accessible as districts have adopted or expanded their use of popular online learning programs.
October 23, 2020
As Purdue Pharma Agrees to Settle with the DOJ, Revisit Its Role in the Opioid Crisis
The proposed $8.3 billion settlement between Purdue Pharma, maker of OxyContin, and the federal government is the latest in a battle over who is responsible for the nation’s opioid crisis, as covered by FRONTLINE in "Chasing Heroin" and "Opioids, Inc."
October 21, 2020
With Election 2020 Underway, a Key Provision of the Voting Rights Act Languishes
Against the backdrop of a pandemic and a divisive presidential election, legislation to restore key provisions of the Voting Rights Act, following the landmark 2013 Supreme Court 2013 decision Shelby v. Holder, remains locked in Congress.
October 21, 2020
We Investigated 'Whose Vote Counts.' Our Findings Unfold Tonight.
A note from our executive producer about the new documentary 'Whose Vote Counts,' premiering Oct. 20.
October 20, 2020