Medical science uses some of the most advanced technologies known to man in order to understand, prevent and treat hundreds of thousands of ailments, illnesses and diseases. Remote Presence Robots, augmented reality and 3D bioprinting all have the potential to improve and personalize medical care in a myriad of ways and are some of the most futuristic technologies you can get your hands on.
Nevertheless, a not-so-new technology by IBM may beat them all to the crown.
Watson is an artificial intelligence system which first came to public attention back in 2011 when it famously beat two champion Jeopardy contestants and won the game. In order Watson to achieve this feat, it was trained on over 200 million pages of general knowledge for years prior to the show.
Medical doctors are now looking to harness Watson’s ‘brain power’ to diagnose extremely rare diseases; pretty much like House if he were a computer rather than an ethically dubious doctor.
A team from The University Hospital of Marburg’s Centre for Undiagnosed and Rare Diseases (or ZusE in German) have joined forces with IBM to give Watson a medical degree (in theory if not in practice. Or is that the other way around?)
Jürgen Schäfer and Tobias Müller of ZusE, along with IBM engineers, taught Watson about 500 past-patient files. It was then asked to use this information to provide a diagnoses for each patient, ranked by likelihood.
Incredibly, the top diagnoses for every single case file matched those provided by doctors when the case was ‘live’.
Unfortunately, the number of ‘lupus’ diagnoses has not been confirmed.
Due to the success of this experiment, Watson will now be used by ZusE to diagnose every new patient who attends the centre. What’s more, it may even be able to identify diseases which are completely new to medicine – if Watson cannot reach a diagnosis, it’s likely a diagnosis doesn’t exist.
So how does it work? It’s actually extremely simple. Patients fill in an exhaustive questionnaire detailing not only their current lifestyle, but also many details from their childhood. It may seem excessive but one patient was diagnosed precisely using this information; as explained by Muller, he had contracted bilharzia from tropical snails he owned as a child.
Once the questionnaire has been fed back to Watson, it then trawls through medical literature to identify the likely causes of a patient’s symptoms. This process, which can take doctors weeks, months or even years to complete, is accomplished by Watson in a matter of seconds.
Although using Watson may seem a little impersonal, the speed of diagnosis, and more important the accuracy of the diagnosis, cannot be ignored. If Watson’s success continues, one day we may balk at the idea of a diagnosis from a real person.
My only wish now is that Watson was called Wilson.
Original Source: New Scientist