AI in Medicine - Hollywood Sci-Fi comes to life
We’ve all seen movies in which robots become doctors. We’ve seem movies in which robots perform brain surgery like Ender’s Game. And who can forget the iconic transformation that Luke Skywalker goes through to become Darth Vader - all robotic surgery. And who can forget Prometheus and Micheal Fasbender’s role as android doctor David. In fact, in that movie, the robot pod for heath care conducts a whole delivery and delivers an alien. Yech.
But all these fantastical interpretations of the role of robotics are a little out of our reach. Automatization of health care is only in its beginning stages. So today, lets take a look at what is going on the real world.
One of the phrases that are being thrown around a lot these days is “Artificial Intelligence” or AI. As an oversimplification, AI is basically the ultimate reflection of the laziness of human beings. Imagine that we are teaching machines to think for us. Because we don’t want to. No, of course I’m joking. The purpose of AI, like any new technology, is to ease human life.
AI, in the world of Hollywood, has been turned into a villain. The idea of a salient being able to out think human beings is a scary thought but I suppose it is a part of evolution. So how has this AI been applied in the field of medicine?
Google’s AI has a 92.4% accuracy of detecting cancer from pathology slides in comparison to the 73.2% accuracy that actual pathologists have. The system uses pattern recognition to recognise visual patterns which are predictive of metastasis. In addition, most of errors that were made by the AI system were secondary to improper slide preparation and labelling and out of focus slides. So there is a lot of room for improvement.
IBM’s Watson is another powerful AI that is already in place in over 230 medical centres. And 55 of these medical centres use Watson as an active resource to detect metastasis. In fact, the amount of data fed to the Watson system was so enormous that there have been 11 system updates in just the past year. What did you expect, the beast can read 800 pages of journal entries a second!
Statistically, Watson has been shown to recommend the same treatment as doctors more than 90% of the time irrespective of the type of cancer. It has also been showed to be faster than most pathologists.
But AI isn’t only useful in looking for cancer in pictures. It has been used to do more than that. AI is now being used to identify the patterns of genetic expression that govern oncogenesis. The machine collects multidimensional data and de-noises it to a simpler level. And then this simple data is used to map areas of high genomic expression. Then retrospective analysis is done to match the data with the true neoplasms. This has allowed us to learn about numerous more newer genes associated with oncogenesis.
Obviously any amazing technological advancement cannot go without contribution from MIT. Now the smarties at MIT have taken a more innovative approach. Their AI machine is used to improve the quality of life in cancer patients who are on chemotherapy. The aim is to minimise tumor size and maximise quality of life.
By now you’re probably thinking what happened to all the other disease except cancer. They are all in the works. Now there is AI that can find diabetic retinopathy in half the time that an ophthalmologist can. This speeds up screening time. Many other illnesses and uses for AI in their management are being looked into. This is the next level.
While technological giants like Elon Musk warn us against the perils of such technology, I think that it is time that we embraced the changes that are about to come. If it means that we might have to face Skynet and the Terminator one day, them so be it.
Author: Narendran Sairam (Facebook)
Sources and citations