by Conor Hale
SAN FRANCISCO—Just a week after backing its $60 million series B round, the Mayo Clinic has tapped healthcare data analytics firm nference to build out an artificial-intelligence-powered engine that will collate its years of clinical research and pathology work into something more searchable.
Dubbed the Clinical Data Analytics Platform, the project is Mayo’s first within an initiative to derive more insights from its large catalog of medical records—including millions of tissue slides and written physician notes that will be digitally annotated, linked and tagged. It will also aim to protect patient privacy by removing potentially identifying information.
“If someone asked how many patients have had lung cancer, who have been on pembrolizumab, and who have since recovered or gone on to have metastatic cancer—a question like that could take months for a researcher,” nference’s co-founder and chief scientific officer, Venky Soundararajan, said in an interview at the annual J.P. Morgan Healthcare Conference.
“They have to read the physician notes, they have to look at the radiology reports, they have to stare at the pathology screens,” Soundararajan said. “That triangulation process is extremely manual and intensive … a simple question like that would require you to put on a radiologist hat, a pathologist hat, talk to nurse practitioners and more.”
The project aims to digitize 25 million pathology slides within the next two to two-and-a-half years, which would provide the single largest source of labeled medical data in the world, he said.
That will feed into what Soundararajan describes as a universal knowledge graph, which will also work to incorporate handwritten notes from physicians and pathologists as well as medication histories and clinical records beyond the more structured data sets found with medical codes and genomics.
“It’s like if someone asked you what happened everywhere at JPM 2020, and you had to answer them with one spreadsheet—that would be really challenging, right?” he said. “There’s a reason why the text is important. Physicians have the same problem, and they work to describe what they’re seeing and detail every observation.”
The Clinical Data Analytics Platform will work to identify new targets for treatments as well as to develop new diagnostic methods of matching patients with different therapy regimens. Nference will also serve as the exclusive provider of analytic services for Mayo’s R&D work with the biopharmaceutical industry.
“Platform business models have been a force of disruption in many sectors, and the rapid digitalization of health care is affording us an unprecedented opportunity to solve complex medical problems and improve the lives of people on a global scale,” said John Halamka, president of Mayo Clinic’s platform efforts that launched last year to leverage technologies such as AI, connected devices and natural language processing through multiple partners.
Additionally, from a scientific standpoint, this knowledge will be able to help researchers better understand the concept of disease itself on a deeper level, including how different ailments are related to each other, Soundararajan said. “The lexicon of diseases, I think, will be fundamentally redefined.”