If there has been a consistent message to come out of all the discussion about artificial intelligence (AI) in healthcare over the past few years it’s that in order to realize their vast potential, AI technologies must be able to fit seamlessly into existing clinical workflows.

A new collaboration between Partners HealthCare and Nuance Communications, Inc., an industry leader in the field of radiology and reporting, is designed to do just that.

Nuance will be working with the team at the Center for Clinical Data Science (CCDS) to develop new AI applications that integrate with Nuance’s suite of radiology diagnostic and reporting programs.

For AI algorithms to be useful to radiologists and patients, it is critical that they are both clinically validated and accessible through the existing reporting workflow, says Keith Dreyer, DO, PhD, Chief Data Science Officer of Partners HealthCare.

“Through this collaboration with Nuance, radiologists will be able to access and leverage validated AI and deep learning algorithms without leaving their preferred workflow,” Dreyer says. “Partnerships like this are a game changer given the time savings, increases in productivity and improved quality outcomes enabled for radiology staff.”

“The goal is to help clinicians better interpret and report on the radiological cases that they are responsible for and then eventually transition those images, reports and knowledge back into our system to support research and discovery,” says Trung Do, Vice President of Business Development for Partners HealthCare Innovation.

“Nuance owns a significant portion of the market share, but, more importantly, they are a vendor within our system,” Do adds. “For us, it makes a lot of sense to collaborate because it can fundamentally improve our healthcare delivery.”

Strength in Numbers

The CCDS, which was formed in 2016 by Massachusetts General Hospital and Brigham and Women’s Hospital, is leveraging the vast amounts of data in the PHS system to develop new AI and machine learning models that address critical needs in healthcare.

A key part of the center’s mission is to ensure that new AI technologies fit into existing systems and processes.

“Accessibility and efficiency are major drivers of the work taking place at the CCDS,” says Mark Michalski, MD, Executive Director of the CCDS. “We spend a lot of energy thinking about how to develop and efficiently deliver tools into the clinical workflow, starting upstream with designing smarter imaging modalities to downstream automation of reporting workflows.”

This is the third collaboration the CCDS has announced since its launch. The first was with NVIDIA in 2016 and focused on developing the compute power needed to process the large amounts of data that inform AI technologies. The next year, the CCDS launched a collaboration with GE Healthcare that’s focused on building out AI platforms and applications in diagnostics, imaging and MRI and CT scans. The collaboration with Nuance was announced in spring of 2018.

“An important aspect of this group of companies is that we share a common vision to enable AI in all parts of the healthcare system, mainly in diagnostics to start and then to move it to other domains and fields like cardiology, cancer and neurology,” Do says.

A Marketplace for AI Technology

Nuance will make the new AI technologies that are developed through the agreement available through the Nuance AI Marketplace for Diagnostic Imaging—an open platform where technology partners can deploy algorithms into Nuance’s cloud-based PowerShare network and users can select them through the company’s PowerScribe radiology reporting platform.

“They have built up a marketplace to distribute these technologies, so the nice thing is that not only does it solve problems for us at Partners but we have the ability and the channels to distribute this commercially as well as collaborate with other entities who have interest in this space,” says Do.

“They are a great distribution partner for the wonderful things that might come out of our collaboration.”