Artificial Intelligence in Cancer Care - the lofty promises by IBM Watson supercomputer – hype or hope? Where are we right now?
We often read overly optimistic news coverage of medical advances such as stem cells, personalized medicine, CRISPR technology and 3 D printer for organs. Furthermore digital technology is another healthcare arena where uncritical media reports may raise false hopes. One salient example is the artificial intelligence The internet is full of publications, touting the new dimension of machine learning devices which are about to revolutionize health care. And here IBM Watson is clearly heading the headlines.
So what is IBM Watson?
IBM Watson is the most renowned cognitive computing system that uses natural language processing software can understand written language and generate probabilistic algorithms. In the medical area Watson is able to read and understand millions of pages of medical articles and based on this continously extended knowledge generating multiple possibilities in making diagnostic or treatment decisions.
IBM Watson for Oncology
2011 IBM chose oncology (cancer care) as the first health care application, since cancer diagnoses and therapies are both due to the nature of cancer very schematized, thus facilitating access to machine learning. IBM Watson computer is told to be able to summarize a medical synopsis of each cancer patient, provided that the supercomputer gets all their documents (e.g. blood tests, pathology and imaging reports detailing the type, size, and location of the tumor, cancer genetic mutations) in a digital form. Additionally to the synopsis the supercomputer combs through huge amounts of uptodate medical literature to prioritize evidence-based indvidualized treatment options for the particular patient. What a good luck for the patient and a bad lack for the oncologists, who has just to check the plausiblity and who may not be able to compete with the continously updated superknowledge of IBM Watson. As a matter of course this supercomputer was originally meant to be a high quality decision support tool. Those readers who do hard to imagine such a procedure, examples for this artificial superdoctor supporting the doctor can be found in short Youtube videos like the following one:
e.g. https://www.youtube.com/watch?v=hbqDknMc_Bo
A bumpy start
Oncology was meant to be the gateway for other medical specialities. While IBM has offensively marketed Watson for cancer care, most people did not notice that IBM had not published any scientific papers in peer journals demonstrating how this technology was superior to good old human intelligence. Even worse, the below attached link reveals an investigation which found that six years after the launch of IBM Watson Oncology, the supercomputer cannot yet match the lofty expectations IBM created. It still seem to struggle with the basic step of learning about different forms of cancer. Surprisingly only more than 50 hospitals on all five continents have agreements with IBM or intermediary technology companies, to use IBM Watson. And adopting this supercomputer to all internal IT technologies was a huge integration project which was clearly underestimated. On the other hand several cancer centres, mainly in the USA, are working with these platforms trying to further develop the capacities of this cognitive computer. But also in the research area not everything is running smoothly. Just a few months ago the Anderson Cancer Center stopped ist collaboration with IBM.
So this sounds more like a hype?
We are entering a new era of oncology, shaped by medical advances in genomics and technologic advances in artificial intelligence. IBM Watson for Oncology however is still in its infancy and once you have read the below mentioned link, you will be sceptical whether IBM will be ever superior to experienced oncologists. At least it will take much more time before we may see on a large scale success stories. The potential of this technology is certainly big as it will further learn and improve with new ingested data and growing computer power. Triggererd by the offensive marketing and uncritical optimistic news coverage, the expectations have been far too high. What is neglected however is the fact, that IBM has gained in the meantime a wider access to big health organisations, where startups or even giant rivals like Apple, Facebook or Google do hard to catch up.
And (re-) insurance industry?
But is not just health care - even (re)insurance industry is currently checking applications of the IBM Watson technology. We are curious to learn more about their success stories?!
Chief Medical Officer PartnerRe
7 年We all cannot predict future, but I dare say that one should not compare AI's capability is comparable to play chess with the much more sophisticated and also empiric and empathic (!) approach to take care for cancer patients. Neither I do agree that human's brain does not further develop - just think of a scientist of the 19th century - he would feel pretty overstrained. It is given fact, that the collective IQ increases over time. I did not say that AI does not further develop - rather the contrary I am convinced, but the the speed takes much longer than anticipated, at least in the medical area. This is not an issue of Moore's law. But anyway, let us exchange in five years - by the way the colleagues told me already five years ago that Watson will penetrate oncology - so neither them as generators of Watson were able to predict short-term future.
Partner at ACENT
7 年Don′t underestimate the Impact of the exponential curve of Hardware developmen. And it is only a question of a short time, when Computers will be better at analyzing digital Pictures (e.g.in oncology) than man. Like in chess or go the Computer will win at the end, it is just a matter of a short time. Simply because mankind is not improving ist brain Hardware. My recommendationwould be not to deny IT Progress, but to see, how you can integrate it. Look as an example at free style chess, where the combination of man and machine can still be better, than each of them alone.
Healthcare Innovator | Chief Innovation Officer @ Promaxo | Co-Founder @ Santé Value | Bridging Clinical Innovation & Tech-Driven Business Growth
7 年Kudos to IBM Watson for enabling common man to understand the power of computing and AI. A good learning to avoid putting the cart before the horse. Marketing claims should be backed by sound scientific proofs which takes time. In this day of Tweets and instant gratification a possible conclusion of a hype doesn't surprise me.