“Jeopardy” Computer Star Becomes a Cancer Hero!The Breast Cancer Site
IBM’s Watson supercomputer may be most famous among most of us for starring on an episode of “Jeopardy,” but for medical personnel, it’s a source of multiple breakthroughs. The supercomputer is helping oncologists in particular by using its amazing computational and analytical resources to find the exact right treatment for a specific cancer patient. Sequencing and analyzing a patient’s genome to spot mutations is a recognized way of tailoring treatment to the patient. However, conventional means of doing this involve a lot of sequencing time and a team of expert oncologists to examine and interpret the data. Watson’s ability to work with unstructured data, along with the immense amounts of computing power packed away within the supercomputer, mean that it can do the job in just three minutes per patient. Thanks to that speed, Watson’s capabilities may benefit up to 10,000 patients in the next two years, sourced from a range of VA hospitals. That’s around 30 times the number of patients who would get exams if not for Watson.
As well as sequencing genomes, Watson’s algorithms examine the patient’s medical records to find more clues about specific causes of and responses to the cancer. The computer then rapidly discards all of the irrelevant data and comes up with a set of best matches for treatment options based on current evidence-based medical approaches, allowing oncologists to focus solely on relevant information to pinpoint approaches and solutions.
This speeds up even the human portion of the treatment and allows patients to receive effective treatment more quickly, hopefully leading to better outcomes. Many cancers have a genetic basis, and these new techniques are likely to lead to breakthroughs in many areas. One of the areas where genetic links are well known, if not yet fully understood, is breast cancer. Find out how advanced computing power is already helping map the ways that cancer cells respond to treatment to help fight the disease.