Watson, the IBM computer that gained fame by beating human contestants on the US television game show Jeopardy! in February 2011, is to play a new role in helping oncologists diagnose and treat cancer. According to a joint statement released this week, Memorial Sloan-Kettering Cancer Center (MSKCC) and IBM scientists will be working together to create a decision support tool based on a computer that does a remarkable job of understanding tricky questions and finding the best answer.
The aim is to give doctors everywhere an “outcome and evidence-based decision support system” with improved and speedy access to comprehensive, up to date information about cancer data and practices so they can tailor diagnosis and treatments to the needs of individual patients.
David Ferrucci, an IBM researcher who has spent the last 15 years working on natural language problems and finding answers in the midst of unstructured information, said when Watson “beat” the Jeopardy! contestants, the idea is to “create a computer that can be more effective in understanding and interacting in natural language, but not necessarily the same way humans do it”.
Under the new initiative, scientists from both organizations will be combining the computational power of Watson, with its natural language processing ability, with MSKCC’s specialist clinical knowledge, huge repository of case histories, and molecular and genomic data.
Watson is an artificial intelligence system that runs on a cluster of Power 750 computers comprising ten racks of 90 servers, with a total of 2,880 processor cores running DeepQA software and storage. DeepQA is a software system developed by IBM that is capable of answering questions posed in natural language.
Watson can interpret queries posed in everday language, and uses statistical analysis, advanced analytics and its powerful array of processors to search millions of pages in seconds to deliver evidence-based statistically-ranked responses.
Oncologists from MSKCC will help to develop the computer to use a patient’s medical information to mine a vast array of continuously updated information. This information will include vetted treatment guidelines and published research, together with insights from experienced MSKCC clinicians.
Not only will the new tool provide doctors with individualized patient recommendations, it will also furnish them with a detailed account of the data and evidence behind it.
Oncology treatment is becoming more complicated, as more technology emerges, and we discover that cancer is not one disease but hundreds of sub-types, each with a different genetic fingerprint.
Add to this the significant discoveries in molecular biology and cancer biology, and new strategies for targeting individual molecular signatures in cancer tumors, and it is no wonder oncologists face a mind-boggling array of options to review when considering what is best for individual patients.
MSKCC President and CEO Craig B Thompson told the press their mission is to help doctors find and personalize cancer care for their patients, wherever they are:
“The combination of transformational technologies found in Watson with our cancer analytics and decision-making process has the potential to revolutionize the accessibility of information for the treatment of cancer in communities across the country and around the world.”
“We also expect tremendous new research opportunities to emerge from this collaboration,” he added.
Dr Martin Kohn, chief medical scientist for IBM said:
“Cancer care is profoundly complex with continuous clinical and scientific advancements to consider. This field of clinical information, given its importance on both a human and economic level, is exactly the type of grand challenge IBM Watson can help address.”
The first applications in lung, breast, prostate and other cancers are expected to be piloted in late 2012.
Written by Catharine Paddock PhD