IBM's Watson Could Diagnose Cancer Better Than Doctors

Posted in Medical Computing by Qmed Staff on October 22, 2013
IBM's Watson
The third generation of IBM's Watson platform will be able to actively reason.

Several years ago, IBM’s Watson supercomputer gained fame after beating some of the world’s top Jeopardy! players. To accomplish that feat, researchers fed thousands of points of information into Watson’s database, allowing it to retrieve information presented through natural language. While winning Jeopardy! might be an exciting challenge for researchers, Watson’s next goal could revolutionize oncology. IBM is currently working on the third-generation of the Watson platform, which has the power to debate and reason, according to IBM CEO Ginni Rometty.

The latest version of Watson can absorb and analyze vast amounts of data, allowing it to make diagnoses that are more accurate than human doctors. If a Watson-style computer was deployed through a cloud interface, healthcare facilities may be able to improve diagnosis accuracy, reduce costs and minimize patient wait times.

In combination with the Memorial Sloan-Kettering Cancer Center and Wellpoint, a private healthcare company, researchers hope to see Watson available for rent to any clinic or healthcare facility that wants to get its opinion on an oncology diagnosis. On top of this, researchers state that the system will be able to suggest the most affordable way of paying for treatment.

Approximately two years ago, IBM researchers announced that Watson had the same level of knowledge as a second-year med-school student. Over the past year, Wellpoint, Sloan-Kettering and IBM have been teaching the system how to comprehend and analyze peer-reviewed medical data. While researchers have started off with just breast, prostate and lung cancers, IBM hopes to expand to other forms of cancer in the near future.

As of now, Watson has assimilated over 600,000 unique types of medical evidence. In addition, Watson’s database includes two million pages sourced from a variety of different medical journals. To improve the link between symptoms and a diagnosis, Watson also has the ability to search through 1.5 million patient records to learn from previous diagnoses. This amount of information is more than any human physician can learn in a lifetime.

According to a study by Sloan-Kettering, only one-fifth of knowledge used by physicians when diagnosing a patient is based on trial-based information. To stay on top of new medical knowledge as it is published, physicians would have to read for at least 160 hours every week. Since Watson can absorb this information faster than a human, it could potentially revolutionize the current model of healthcare.

According to Samuel Nessbaum of Wellpoint, Watson’s diagnostic accuracy rate for lung cancer is 90%. In comparison, the average diagnostic accuracy rate for lung cancer for human physicians is only 50%.

Wellpoint believes that Watson will be able to reduce waste. According to the company, approximately one-third of the money spent on U.S. healthcare is wasted every year. With Watson, utilization management can be approved, state researchers.

"What Watson is going to enable us to do is take that wisdom and put it in a way that people who don't have that much experience in any individual disease can have a wise counselor at their side at all times and use the intelligence and wisdom of the most experienced people to help guide decisions,” notes Dr. Larry Norton, a researcher at Sloan-Kettering.