IBM says 14 U.S. cancer treatment centers will take part in a partnership to receive personalized treatment plans selected by its Watson supercomputer.
The cognitive computing capabilities of Watson will be focused on the genetic fingerprints of a patient's particular cancer to select the potentially best course of treatment, the company says.
Matching particular therapies to the DNA signatures of cancer tumors has yielded improved treatment outcomes for some patients, but the process of identifying the best drugs to target cancer-causing mutations can take weeks, researchers say.
However, Watson has the computing power to complete such tasks in minutes, and also has access to a database of scientific studies and clinical trials on specific cancers and best potential therapies, IBM says.
The amount of information available has grown so huge it has exceeded the ability of humans to survey it all, experts say.
"The solution is going to be Watson or something like it," says oncologist Norman Sharpless of the University of North Carolina's Lineberger Cancer Center, one of the centers entering into the IBM parnership. "Humans alone can't do it."
"Determining the right drug combination for an advanced cancer patient is alarmingly difficult, requiring a complex analysis of different sources of Big Data that integrates rapidly emerging clinical trial information with personalized gene sequencing," he says.
Watson can help reduce the time it takes to analyze the best treatment options for cancer patients by analyzing their genetic makeup, correlating that to research on the particular form of cancer, and suggesting the best targeted treatment, IBM says.
The computing power of Watson was first applied to cancer treatment in 2013, and last year IBM moved to "genomic medicine" in a partnership with the New York Genome Center to help identify genetic-based treatment options for brain cancer.
The cloud-based Watson supercomputer will be used in a number of cancer treatment centers by the end of this year. The program uses IBM Research advancements in analytics and existing Watson collaborations to help develop a genome data analysis solution.
The partnering centers will pay a subscription fee for the service, says Steve Harvey, vice president of IBM Watson Health, although he declined to disclose the exact costs.
Oncologists at the participating centers will upload the DNA fingerprint of a patient's tumor, showing which genes are mutated and possibly driving the malignancy, then Watson will examine thousands of mutations to try to identify which is driving the tumor and should therefore be the target of drug treatment.
"Watson will look for actionable targets," Harvey said, and them match to approved or experimental cancer drugs.