Each week, results of new clinical trials and reports of researchers’ work are posted on the Internet. That information can be helpful to physicians treating cancer patients, but wading through it all and finding what’s applicable isn’t easy.
That’s why Dr. Ken Cowan, director of the Fred & Pamela Buffett Cancer Center, is excited about the cancer center’s collaboration with IBM and its Watson computer system.
“What Watson does,” Cowan said, “is scour the world literature every week for publications and for any new clinical trials that are open or in concept with drugs that target specific mutations in cancer cells.”
The computer system, he said, then assesses what’s important and not important.
IBM says Watson is able to analyze DNA information and rapidly identify treatments targeted to individual cancer patients. The company announced partnerships to test the Watson system with the Omaha cancer center, on the University of Nebraska Medical Center campus, and 13 others Tuesday.
Cowan expects the work to lead to other collaborations with IBM and others.
UNMC began using the system April 1 and will have access for no cost at least through the end of June, said Babu Guda, a UNMC associate professor and director of the bioinformatics and systems biology core facility. Information gleaned from the software test will be used only for research purposes, not for treating patients, Guda said.
IBM says Watson is a cognitive computing system that can process natural language and unstructured data and learn by experience, much the same way humans do. The system got a lot of publicity in 2011 when it won the game show “Jeopardy” over two former champions. An IBM video notes that the computer won after many prior “Jeopardy” questions had been fed into it and it learned how to answer them.
It’s learning about cancer and treating patients in the same way.
DNA analysis is only one area of cancer care where Watson’s computing power is being applied, Cowan said.
The system also is studying how tumor boards at cancer centers make treatment decisions about patients. Tumor boards, made up of doctors who are experts in different specialties, meet regularly to review and discuss the medical conditions and treatment options of patients.
Watson will “learn from everybody’s experience in large centers like ours and develop algorithms,” Cowan said. The information it generates “will help people who aren’t as experienced, who don’t have tumor boards to rely on.”
Information from UNMC’s cancer registry, which contains information on thousands of cancer patients across Nebraska, also could be analyzed by Watson to help it learn about what treatments are effective, Cowan said.
Guda said that during its 90 days of access to the system, UNMC researchers who are conducting a breast cancer study will be able to run at least 150 patients’ samples through Watson. In the project, DNA sequencing is done for normal breast tissue and the tumor breast tissue from the same patient.
“We can compare the mutations in the tumor DNA against the normal DNA as a reference. Then we can identify the specific mutations in the tumor. And then we feed that information into the IBM software package.”
Patients in the study consented only to research use of the data, he said, and the UNMC lab must be clinically certified to apply the data to a clinical setting.
All patient data is stripped of identifying information such as names and addresses before it’s uploaded into the IBM system, Guda said.
If the Watson testing goes well, he said, officials may decide to seek patients’ consent for clinical use and obtain the clinical certification. They then could pay IBM to run the data again, Guda said.
An IBM spokeswoman said cancer centers could extend access to Watson via a subscription-based system. She declined to say what the cost would be but noted that it would be less expensive than paying for supercomputers or servers.
In addition to the Buffett Cancer Center, other institutions that are part of the testing include Ann & Robert H. Lurie Children’s Hospital of Chicago; BC Cancer Agency, Vancouver, British Columbia; City of Hope, Los Angeles; Cleveland Clinic; Duke Cancer Institute, Durham, North Carolina; McDonnell Genome Institute at Washington University in St. Louis; New York Genome Center; Sanford Health, Fargo, North Dakota; University of Kansas Cancer Center; University of North Carolina Lineberger Cancer Center; University of Southern California Center for Applied Molecular Medicine; University of Washington Medical Center in Seattle; and Yale Cancer Center, New Haven, Connecticut.
Additional centers are expected to join later this year.