From self-driving cars to bionic arms, the dreams of artificial intelligence (AI) once imagined in science fiction novels are now reality.
But can AI tackle the complexities of diseases like cancer?
According to the International Agency for Research on Cancer (IARC), by 2030, the global burden of cancer is expected to grow to an estimated 21.7 million new cancer cases and 13 million cancer deaths because of the growth and aging of the population.
“To treat cancer more efficiently, we need to see patterns that we didn’t see before. I think that now we have digital tools that can enable us to do that,” says Ketil Widerberg, general manager, Oslo Cancer Cluster, an oncology research and industry cluster in Norway dedicated to accelerating the development of new cancer diagnostics and medicines.
As individual and population-health data grow – doubling about every 24 months – the ability to make sense of it and take action is limited for health care practitioners.
To keep up, doctors would have to read 29 hours each workday to understand the latest research and advances. And, as researchers uncover new findings about the origins and mutations of cancer, treatment approaches can vary down to the individual.
Now, we’re turning to the most successful player of Jeopardy!, a popular American television game show, to help reverse the course of the deadliest disease of all time.
Paging Dr. Watson
While IBM Watson is most famous for cruising to an easy victory on the popular U.S. quiz show in 2011, the artificial intelligence platform has not been idle since then.
IBM launched Watson Health in 2015, which, in part, aims to tackle some of the toughest challenges of cancer diagnosis and treatment.
Memorial Sloan Kettering, a famous cancer hospital in the United States, and IBM have been training Watson for Oncology in reading and understanding the oncology domain. Watson for Oncology is a specialized computing system within Watson Health designed to help the entire physician community when they consider cancer treatment options.
Watson for Oncology draws from an impressive corpus of information, including curated literature and rationales, as well as more than 290 medical journals, more than 200 textbooks and 12 million pages of text. Watson ranks identified treatment options and provides links to supporting evidence for each option to help oncologists as they consider treatment options for their patient.
No human being can keep up with the enormous amount of published literature – but Watson can by reading millions of pages in just seconds, according to Peter Mortensen, Watson Health executive, the Nordic Region.
The homework seems to have paid off.
In one study at the University of North Carolina School of Medicine, Watson matched 99 percent of the treatment recommendations made by oncologists; and in 30 percent of the cases, Watson identified additional options that the doctors missed.