“For at least 30 years now, it has been humanly impossible for doctors to keep up with all the material necessary to practice the art of medicine at the highest possible level’, says Herbert Chase, Professor of Clinical Medicine and Biomedical Informatics at the University of Columbia. And he’s right; the amount of medical literature is doubling every five years, and it is estimated that by 2020 it will be doubling every three months!”

Substantiation

In the meantime, much evidence has been accumulated [10]:

• In a study carried out by the Manipal Comprehensive Cancer Center in Bangalore, India, Watson for Oncology achieved 96% concordance for cases of lung cancer, 81% for colonic cancer and 93% for rectal cancer when compared with oncologists’ recommendations.
• Another study at the Bumrungrad International Hospital in Bangkok, Thailand, showed 83% concordance for several different types of cancer. The results have been reported at international conferences such as the ASCO[11].
• A technical feasibility study conducted by the Highlands Oncology Group, and in collaboration with Novartis with Watson for Clinical Trials Matching, showed that the time needed to screen patients for inclusion in a study could be reduced by 78%; an enormous time gain
  that benefits patients directly.

Watson

Named after the founder of IBM, the Watson computer system was developed at the start of this millennium in order to address the general challenge of dealing with Big Data: to rapidly extract insights, knowledge and wisdom from an immense mountain of different types of data. But how do you do that? How do you design and build a system like that? And what kind of software would you need?

A team of around thirty researchers set out to find the answer. At the same time, IBM made use of its knowledge in the field of
Natural Language Processing¹. With the interdisciplinary character of the study in mind, IBM thought up a new name for the technology: Cognitive Computing

The result was a collection of software modules which, combined with enormous amounts of data, can analyse and interpret data. Not only can they do this with structured data arranged neatly and labelled in databases (just 20% of the available data), but also with unstructured data such as videos, email attachments, images and sound. Based on a question or hypothesis, the system searches through all the available data and tests and calculates which data is most suitable as an answer [1].

However, once it had been developed, the system was not yet finished. It had to be tested in practice. The American television quiz Jeopardy! was picked. Before Watson could take part successfully, however, it had to be trained, taught how to analyse each question to precisely determine what is being asked, unravel the available content to extract accurate answers, and rapidly calculate the trustworthiness of the answer on the basis of useful and rejected information. This process took three years. The rest is history: Watson won the Jeopardy! quiz hands down in February 2011, relegating the very best players in the world to the second ranks [2].

Healthcare

Seven years later, Watson is supporting more and more healthcare professionals all over the world. An initial application is to find the most suitable cancer treatments [3,4]. IBM Watson for Oncology is used at tumour board meetings, a well-known type of multidisciplinary meeting in the world of healthcare during which a wide range of patient cases are discussed. Watson suggests alternatives or verifies conclusions. A second example is supporting researchers in discovering new pharmaceuticals [5]. In the USA, Quest Diagnostics provides an online service which, with the help of Watson for Genomics, links cancer cell mutations with treatments and the latest scientific insights. A third example is Watson for Clinical Trial Matching, which informs doctors of any clinical studies in which patients could be included [6-8]. The wishes and preferences of the patient are also taken into account, making the care provided more personal.

Healthcare professionals in over 50 hospitals in 13 countries across five continents are now working with Watson. Not that Watson makes the decisions; Watson is not an oracle or robot, but a highly-sophisticated device. It is precisely the combination of, and interaction between, man and machine that give doctors the confidence to feel supported by an intelligent assistant. Watson is, in fact, a continuation of the physician’s own rational thought process, and can make complex decision-making demonstrably more precise and patient oriented.

Watson has amassed the collective knowledge of countless health professionals all over the world through sources including reference books and publications. Watson would not have been able to take part in Jeopardy! without immense prior effort, and the same thing applied to healthcare. Concordance studies are set up everywhere where Watson is implemented in healthcare, studies that ultimately result in, or are required to result in, peer-reviewed scientific publications [9]. Systems must at minimum meet the clinical gold standard.

The Future

Today, less than three years after IBM’s Watson Health was founded, almost 40,000 patients have benefited from Watson healthcare interactions with regard to cancer, chronic diseases and even their personal well-being [12,13]. More and more ex-healthcare professionals are turning to Watson Health because they believe that the Watson technology can make a difference in their field of expertise. They are working on the clinical evidence programme on a daily basis, and help their former colleagues with the implementation of Cognitive Computing. Watson Health uses this in their mission statement, which is to improve lives and give hope by implementing innovation in order to deal with the most pressing healthcare challenges in the world with the aid of data and cognitive insights.

1. Also known as computational linguistics. Natural Language Processing is a specialization at the interface of linguistics and artificial intelligence, providing computational models of various kinds of linguistic phenomena. Computational linguistics is not limited to a specific domain within  linguistics (such as syntax, semantics, or sociolinguistics). Computational linguistics is a highly interdisciplinary field, with involvement of linguists, computer scientists, experts in artificial intelligence, logicians and cognitive scientists..

References

1. Noordhoff A. Hello Watson. Legadex Magazine 8 (2017 [Internet]. 2017 Jan 1 [cited 2017 Oct 17];28–31. Available from: https://issuu.com/legadex/docs/leg17751_digital_magazine_jan2017_0
2. IBM’s Watson Supercomputer Destroys Humans in Jeopardy | Engadget – YouTube [Internet]. [cited 2022 Jan 11]. Available from: https://www.youtube.com/watch?v=WFR3lOm_xhE
3. IBM-supercomputer Watson goed in plannen kankerbehandelingen | RTL Nieuws [Internet]. [cited 2022 Jan 11]. Available from: https://www.rtlnieuws.nl/tech/artikel/3931096/ibm-supercomputer-watson-goed-plannen-kankerbehandelingen
4. Watson Health: setting the record straight – Watson Health Perspectives [Internet]. [cited 2022 Jan 11]. Available from: https://www.ibm.com/blogs/watson-health/setting-the-record-straight/
5. Life Sciences Technology | Watson Health | IBM [Internet]. [cited 2022 Jan 11]. Available from: https://www.ibm.com/watson-health/solutions/life-sciences-technology
6. Empowering Heroes, Transforming Health. 2017;
7. Mayo Clinic finds IBM Watson increases enrollment of clinical trials | MobiHealthNews [Internet]. [cited 2022 Jan 11]. Available from: https://www.mobihealthnews.com/content/mayo-clinic-finds-ibm-watson-increases-enrollment-clinical-trials
8. IBM – Announcements [Internet]. [cited 2022 Jan 11]. Available from: https://newsroom.ibm.com/2018-03-08-Mayo-Clinics-clinical-trial-matching-project-sees-higher-enrollment-in-breast-cancer-trials-through-use-of-artificial-intelligence
9. 1. Watson for Oncology? | Kanker Strategie [Internet]. [cited 2022 Jan 11]. Available from: https://www.e-cancer.be/nl/step/watson-oncology
10. Ross C, Swetlitz I, Hogan A, Stat /. I IBM’s Watson supercomputer recommended “unsafe and incorrect” cancer treatments, internal documents show. 2017;
11. Meeting Library | 2017 ASCO Annual Meeting [Internet]. [cited 2022 Jan 11]. Available from: https://meetinglibrary.asco.org/browse-meetings/2017 ASCO Annual Meeting
12. IBM Watson Health | AI Healthcare Solutions | IBM [Internet]. [cited 2022 Jan 11]. Available from: https://www.ibm.com/watson-health
13. 1. International Business Machines’ (IBM) Management Presents at Morgan Stanley Healthcare Conference (Transcript) | Seeking Alpha [Internet]. [cited 2022 Jan 11]. Available from: https://seekingalpha.com/article/4105848-international-business-machines-ibm-management-presents-morgan-stanley-healthcare-conference