A fascinating article was just published by Ariana Cha in the Washington Post describing how IBM’s Watson is being used at 14 major cancer centers to help interpret genomic data.
The program is still in the early stages, which means that Watson is still learning, but healthcare has become IBM’s major focus with the supercomputer. The company recently committed $1 billion to further advance the program based on its early success.
By referencing vast amounts of published literature and patient data Watson is able to provide a range of treatment recommendations along with a confidence level for each.
Aiding in leukemia treatment decisions at MD Anderson is an application that is profiled in Cha’s article, and MD Anderson is impressed enough with the results that it will soon begin training Watson for lung cancer and diabetes.
IBM is positioning Watson as an adjunct to physician-led care, but one must wonder at the prescience of an MD Anderson physician who is quoted as saying “They keep telling me it will not replace me…but I am pretty sure it will replace me.”
One thing that is not addressed in the article is whether or not Watson has the ability to utilize the combined patient phenotype and genotype data from all of the cancer centers at which it is installed.
If so then IBM has the potential to become an incredibly powerful player in healthcare, by connecting independent data silos and providing highly-informed analyses all in one platform.