Partners HealthCare looks to develop—and commercialize—artificial intelligence

Boston-based Partners HealthCare is jumping headlong into the artificial intelligence sector, inking a partnership with GE Healthcare to create new AI platforms that can be used well beyond the integrated health system.

The 10-year deal announced on Wednesday looks to unite the clinical capabilities and the massive data sets within the Partners HealthCare system—which includes Brigham and Women’s Hospital and Massachusetts General Hospital—with GE’s computing power to put AI in the hands of clinicians.

Neither GE Healthcare nor Partners would disclose the dollar amount the organizations were investing in the new partnership, but Partners HealthCare CEO David Torchiana, M.D., said both organizations were “putting in a significant amount of funding and workforce talent.”

A spokesperson for GE Healthcare also declined to divulge the exact dollar amount of the partnership, but said it is “one of the largest commitments GE Healthcare has ever made in the digital space.”

In November, the University of California, San Francisco, announced a partnership with GE Healthcare that uses machine learning to improve image-based diagnoses.

RELATED: UCSF puts machine learning in clinicians' hands

Similarly, Partners HealthCare is using the partnership to focus on developing applications to improve diagnostic imaging initially, although the system ultimately wants to “implement AI into every aspect of a patient journey—from admittance to discharge,” according to the announcement. Researchers have pointed to image-based specialties as an area where AI could have a significant impact.

RELATED: AI could prompt merge of radiology, pathology into one specialty

Stroke diagnosis, which often relies on the intuition of physicians to sniff out abnormalities, could see particular benefits from AI advancement, according to Keith Dreyer, D.O., Ph.D., chief data science officer at Mass General Hospital and Brigham and Women’s Hospital. Early detection of a potential stroke—something that might take a physician years to learn—could be processed in minutes using a neural network.

“Now the challenge is to make that part of your emergency room or part of your CT or MRI scan,” he said during a press briefing on Wednesday.

Getting AI to the front lines will be a critical part of this partnership, added Mark Michalski, M.D., executive director of the MGH & BWH Center for Clinical Data Science. He noted that researchers have watched novel models develop behind the scenes, but very few solutions have made it into the clinical environment.  

RELATED: Plenty of buzz for AI in healthcare, but are any systems actually using it?

“We’re starting to see deep learning in medicine translate now, but we’re just at the beginning of that discussion,” he said.

Commercializing those platforms will be a key part of this partnership, officials say. Although the algorithms will be open-sourced, GE and Partners plan to build AI platforms that can be distributed to healthcare organizations throughout the country. Ultimately, both organizations want to create a suite of apps that can be used on the front lines to provide clinical decision support for various specialties. 

"It has to be built to a commercial grade solution and it has to have distribution mechanism to get this into the front lines and into care delivery," Torchiana said. "And only by doing that will the substance of this collaboration have an impact on patient care."