Watson Is Moving Beyond 'Jeopardy!' To Improve Cancer Treatment
Two weeks ago, IBM announced that access to Watson will now be available commercially to hospitals and insurers. They've been working in partnership with Memorial Sloan-Kettering and insurer WellPoint to offer two applications. One runs through thousands of medical records to help doctors diagnose and treat lung cancer, and another helps make insurance decisions.
Pricing terms haven't been disclosed, but hospitals will be able to rent or buy Watson capabilities either from their own server, or through a cloud based service. The WellPoint insurance tool will be used internally, and offered through a web portal to other providers, and it will split the revenue with IBM.
We spoke to Chris Coburn, the Executive Director of Cleveland Clinic Innovations, which is the corporate venturing arm of the Cleveland Clinic, and a member of the Watson advisory board about the potential uses for the supercomputer in health care.
According to Coburn, IBM thought about Watson for these applications "from the beginning." That's partially because David Ferrucci, the lead researcher on the project, was a biologist who once contemplated medical school.
Naturally, people thought the major applications would be in finance given the sheer amount of data and the complexity of hospital billing. But Coburn thinks that the first "big bang" result from these calculations will be in optimizing cancer care treatment paths.
That means for someone with cancer, Watson will have access to all of the information out there, and can use this to offer doctors a series of diagnoses and treatment options.
"The concept is while Sloan Kettering, and MD Andersen, and the Cleveland Clinic are fabulous institutions for cancer care, most cancer care is rendered in a community hospital or by a small practice doctor, and that's where the variability comes in," Coburn told Business Insider. "If you can create a meaningful care path where 80 percent of patients in a category are getting essentially the same standard of care, that's really a powerful breakthrough."
This is particularly essential for cancers, which mutate and change constantly as the disease progresses.
Watson has so much potential for this because it always has the maximum amount of current information, Coburn said. A new paper could have come out just that week, and the computer will factor the study's findings into its recommendations.
"[America] is spending an immense amount of money every year at the National Cancer Institute, or Leukemia Lymphoma Society — take your pick — the state of the art is in constant flux for cancer," Coburn said.
It's difficult for even cancer specialists to keep up with this deluge of data and process it all in meaningful ways that help patients get the best treatment.
The process isn't easy for computers, either. The technological barrier to digitally determining cancer treatment paths has been huge. There are so many decisions that go into determining the optimal treatment path for a cancer patient: Their history, the effectiveness any given drug in their ethnic group, the progression of different cancers, and the individual mutations that make a person's cancer cells unique.
Analyzing this information requires an enormous amount of computing power. The supercomputer also needs to be able to "read" studies and combine information from a huge variety of sources.
Because Watson can look at as many as 200 million pieces of information, all at once, and interpret the language used into data, it is the optimal tool to discover the best treatment options for each individual patient, and could help push smaller hospitals towards the same standard of care that we see in the best cancer hospitals.
Here's how the process would go:
Harnessing Watson could raise the standard of care at smaller cancer centers, and reduce misdiagnoses that put people's lives in danger.