IBM’s Artificial Intelligence Watson Shows Ambition in Cancer Treatment

IBM’s History — Who Is Watson?

IBM was founded in New York in 1911 by Thomas J. Watson to use and develop the patents of punched-card machine inventor Herman Hollerith. Operating in fields ranging from hardware, software, servers, and supercomputers to nanotechnology, IBM takes its name from International Business Machines . IBM has produced hundreds of inventions used frequently in everyday life, including ATMs, floppy disks, personal computers, and barcode technology. With operations in approximately 170 countries and more than 370 thousand employees, it is one of the world’s largest technology companies. Its supercomputer Deep Blue , known for competing against Kasparov at chess, brought a new dimension to human–machine interaction. In 2011, we met IBM’s new supercomputer Watson, which answered questions on the United States quiz show Jeopardy and defeated its human opponents by a wide margin.

Watson is named after IBM co-founder Thomas J. Watson. Developed under David Ferruci’s leadership, it stood out for answering questions. Through its natural language processing capability and, beyond conventional information systems, cognitive computing , enables Watson to understand the roughly 80% of the internet made up of unstructured data and also train itself. To explain with an example, structured data are details such as a person’s height, weight, and eye color recorded in a defined format. These can readily be processed with algorithms and software to draw conclusions. Watson takes a different approach and imitates the human brain. Give Watson the word “cancer”: it connects this with all its data and presents cancer types, symptoms, and individualized treatment plans, showing references and probability percentages. Its predictive capability also comes to the fore. People often struggle to know what to ask, and Watson can use prediction to guide them.

IBM Watson Enters the Cancer Field

After winning the quiz show, Watson could have been used in many other fields, but was assigned to healthcare, specifically cancer treatment, at Memorial Sloan Kettering Cancer Center (MSKCC) in New York. Trained for oncology for approximately 5 years, IBM Watson significantly increased the speed of treatment decisions. Training combined blood tests, pathology and imaging reports describing tumor size, location and type, and genetic mutation status with extensive medical literature searches to create evidence-based treatment plans.

An advanced example of artificial intelligence, Watson reads and understands millions of pages, including doctors’ questions and decisions, and offers treatment plans based on probability calculations. Initially trained in lung, breast, and colorectal cancer, its recommendations for breast cancer from stage 1 to stage 3 were over 90% accurate. It also took half as long as experienced MSKCC oncologists to make lung cancer recommendations.

In a study of 638 patients at Manipal Cancer Center in India, physicians took approximately 20 minutes to review patient information and decide, whereas Watson took only 40 seconds.

A scientist can read approximately 200 to 300 research papers per year to keep up with developments, whereas Watson has no such limit. Its drug discovery platform was trained with 25 million Medline abstracts, over 1 million articles, and 4 million patents, and continues to be trained with 10 thousand new scientific papers and 100 new clinical trials each month.

Artificial intelligence strengthened by cognitive computing is helping physicians make decisions today, rather than in some distant future, and this training took Watson only 5 years. This is similar to training a young doctor, and Watson will become an indispensable tool for physicians in the future.

Further Reading: Watson in Every Area of Cancer

Cancer training that previously took Watson 6 to 9 months now takes 30 days. The software is being tested in approximately 30 cancer centers and dozens of research institutes in China and the United States.

Although Watson remains in testing at most United States cancer centers, a cloud-based system for breast, lung, colorectal, stomach, cervical, and ovarian cancers was recently established at Jupiter Medical Center in Florida. It is also used in some hospitals in Thailand, India, Korea, and Slovakia and is being adapted for certain hospital systems in China.

IBM Watson has other applications supporting oncology. Last January, it partnered with Illumina, which produces DNA sequencing equipment, aiming to provide genetic analysis results to the same standard as scientists in less time. Watson has also entered immunotherapy: last December, Pfizer worked with IBM Watson Health to help researchers identify new drug targets, potential combination treatments, and patients for clinical trials.

Watson was first used to offer patients different treatment plans. Currently it offers oncology plans only for certain cancers, but in the near future it will also be able to offer surgery, radiotherapy, and hormone therapy plans, showing what patients may expect in the short and long term.