IBM's Watson Moving Closer to Patient Care
As one of the nation's foremost lung cancer experts, Mark Kris knew big data analysis could help doctors harness and make sense of the ever-growing deluge of cancer research. But after working with IBM's Watson system, Kris is now a true believer in the power of cognitive computing.
"It's going to be the ultimate in personalized medicine," said Kris, the chief of Thoracic Oncology at New York's Memorial Sloan-Kettering Cancer Center, "because it is going to be able to learn more facts about you than any one doctor or health-care system can, and can process them in a way that will ultimately be useful to your care."
Kris and a team of oncologists from Sloan-Kettering have been working with IBM's Watson programmers and engineers to train the system to understand cancer treatment. "Then, teaching the system how to read the medical records and medical reports, which is not a simple task," he said.
Over the last year, they have collected more than 2.5 million pages of cancer data, written text from medical journals, clinical trial research, and 1,500 case files.
But after repeated testing and training, Watson now consistently recommends the best options for evidence-based cancer treatment, Kris said. What has impressed him most is how quickly its recommendations have improved, with each test and the addition of new data.
"It totally flips the whole quality movement on its head," explained Kris. "Giving treatments based on it—they're all bench-marked against everything we know."
With estimates that new cancer data will double in four years, much of it in unstructured written form, Manoj Saxena, IBM's general manager for Watson, said cognitive computing is the only way doctors can keep up with the best treatment options.
"It's humanly impossible for the doctors to pull all of the medical records about pathology reports, family history, the most recent diseases and drugs, and to bring it together to the point of treatment," Saxena said. "That's where Watson is a decision-support system we believe will have breakthrough implications."
And that's why health-care analytic programs like Watson should see strong growth over the next five years, said Steven Van Kuiken, chief of McKinsey's Health and Information Technology division.
(Read More: Is IBM's Watson the Answer for Consumers?)
"The market is very immature, as opposed to the retail or banking big data market, where you have very big players. That's not true in health care yet," said Van Kuiken, co-author of the study, The Big-Data Revolution in US Health Care: Accelerating Value and Innovation.
One big driver of big data adoption will be push from payers—Medicare, insurers and patients themselves—for more evidenced-based, quality care when it comes to the costly treatment of cancer and other diseases.
"Just the sheer cost of health care on the system is creating more demand for big data," he said.
(Read More: Jolie's Decision Boosts BRCA Test-Maker Myriad)
Beta-testing of Watson with current cancer patients is on track to begin this summer at two cancer facilities, WestMed Group outside New York City, and the Maine Center for Cancer Medicine. If the trial goes as expected, the commercial Watson product for cancer providers could come to market later year.
For Sloan-Kettering's Mark Kris, being able to make high-quality cancer care more widely available for patients will mark an important milestone.
"The chance to develop one system that virtually every doctor in the world can tap into and tap into research advances expertise from the top folks at Sloan Kettering and all over the country," he explained, " It's just an amazing opportunity."
—By Bertha Coombs. Follow her on Twitter: