Taipei Medical University joins hands with IBM Watson to combat cancer
2017/12/04
Do you remember when the IBM supercomputer Watson won the American game show “Jeopardy!” in 2011?
With an eye toward expanding Watson’s potential, IBM created custom software allowing the supercomputer to help facilitate medical treatment.
In 2015, after acquiring medical data management and imaging companies, IBM for Health was conceived. The firm then introduced IBM Watson for Oncology in hospitals to assist physicians by analyzing vast amounts of medical data before suggesting professional treatment plans.
In use at 55 hospitals worldwide
In June, Taipei Medical University officially announced the formal integration of Watson into systems at affiliated facilities including Taipei Cancer Center and oncology clinics at Shuang-Ho Hospital and Taipei Municipal Wanfang Hospital. TMU said it was the first hospital system in Taiwan to incorporate the AI program.
TMU utilizes Watson by inputting vast amounts of medical data into one centralized cloud computing system. First, when a physician sees a cancer patient, he or she will log into Watson to view the patient’s medical records before also inputting new diagnoses or notes. Less than a minute later, Watson will provide a list of recommended treatment plans personalized to the patient’s needs. Watson will also highlight incomplete brackets, prompting the physician to input any missing information.
Chiou Jeng-fong, vice superintendent of Taipei Cancer Center, explained that in the case of breast cancer, Watson will ask oncologists for the size of the tumor, whether it is hormone receptor positive or negative, and inquire about its molecular markers. “The questions are all different for each patient; in other words, the system will always provide a personalized treatment plan,” he said.
Meticulous, accurate and scientific, Watson almost seems like a professionally trained oncologist. How does the system work?
According to Eric Wang, head of IBM Taiwan’s public, communication and distribution, Watson utilizes a learning algorithm and draws from a massive corpus of information, including 300 medical journals, 200 textbooks and almost 15 million pages of text from research papers.
In 2011, Watson went on trial at Memorial Sloan Kettering Cancer Center, a leading cancer treatment and research institution in New York City. Physicians evaluated the program and reviewed the practicality and rationality behind Watson’s proposed treatment plans.
According to statistics, Watson has been over 90 percent concordant with tumor board recommendations for breast cancer. In addition, the database is updated every three months. With input from the MSKCC medical team, Watson continues to improve.
Currently, Watson can be used to assist oncologists on breast, lung, colon, rectal, gastric, ovarian, cervical and prostate cancers. It is expected that by the end of this year, the system will be able to aid clinicians in treating four additional types of cancer. To date, the program is being used by more than 50 hospitals around the world.
After integrating the Watson system, TMU has mostly used the technology to treat relapse patients. Even for stage 3 cancer patients, Watson is still able to provide recommendations. “We wanted to challenge the system and use it for more difficult cases,” Chiou said.
For instance, the TCC vice superintendent once had a breast cancer patient who had undergone treatment for over 10 years. He had put the patient on all the latest medications, including one that cost up to NT$10,000 per dose.
In the beginning, the drug was effective, but after six months, the patient developed a resistance to the medication. Chiou then decided to enter this patient’s case into the system, which recommended a treatment that proved successful. “We had overlooked a very traditional drug. The medication, once given to the patient, proved to be very effective,” he said, noting that as an added bonus, the drug was very inexpensive.
Helping physicians make more informed decisions
In addition to oncology clinics, TMU also uses Watson’s recommended treatment plans as references during medical conferences when studying cancer cases. The system provides relevant medical information to physicians, such as crucial data on drug effectiveness, side effects, five-year survival rates and so on.
Chen Chun-You, who is in charge of Watson at the Taipei Cancer Center, said that the system can help physicians stay at the top of their game. Usually, oncologists need to take into account 30 signs and symptoms to make an accurate diagnosis, Chen said. However, sometimes they fall short an only register around 25. The system requires that doctors input 30 signs and symptoms before making a diagnosis, he explained.
Physicians need to constantly spend vast amounts of time absorbing new medical information, according to IBM Taiwan’s Wang. Watson has the capability to draw from a large database, saving oncologists a lot of time, he said. The system can benefit doctors by decreasing the time they need to spend researching so they can concentrate on patient care.
Addressing the concern many have on whether AI will make doctors obsolete, TCC’s Chiou said that he always informs his patients when he is enlisting the help of Watson. “AI technology can never replace doctors,” he said. Only trained physicians are able to make sense of the medical information provided by Watson and choose the most suitable treatment plan from the ones the system recommends, Chiou stressed. He said he expects Watson’s applications will expand rapidly to more hospitals nationwide. “The biggest winners are definitely the patients.”
Source: Taiwan Today (http://taiwantoday.tw/news.php?unit=11&post=125887)
