Merge’s technology is in use in 7500 healthcare facilities in the US, according to information supplied by IBM . The firm processes billions of images such as x-rays, MRIs and CT scans with 30 billion images processed to-date, a number that is growing by the day. In fact, IBM researchers estimate that up to 90 percent of all medical data today is in the form of images.
Up until now, Watson has been focused primarily on understanding natural language processing, whether that’s spoken word or in writing, and it’s gotten quite good at that. More recently IBM has been experimenting with getting Watson to understand pictures and video, John Kelly, senior vice president for solutions and research at IBM told TechCrunch.
The company is buying Merge Healthcare specifically to showcase this new ability and to provide new research and understanding in the area of radiological diagnostics.
When IBM announced Watson Health last spring, it was hoping to usher in a new kind of data-driven healthcare, one in which a medical professional can work together with Watson to optimize (at least in theory) the healthcare experience.
Too Much Information
The amount of information a medical professional has to deal with today is mind boggling. A medical record can run into hundreds or even thousands of pages, making it nearly impossible for a human to read and comprehend the contents quickly.
When you factor in increasingly complex medical images, even a trained technician has trouble processing these and understanding how they might tie into the latest medical research and the patient’s medical history.
“What’s happening is that scanning and imaging machines have gotten incredibly good. They are generating extremely dense images and images in motion in some cases. Doctors are overwhelmed by the content of these images. Watson will help the doctor interpret and understand these complex images,” Kelly said.
Radiologists in particular are being challenged by this growing amount of information, says Dr. Elliot Siegel, a physician who has worked with IBM on Watson since its earliest days. Siegel is professor and vice chair at University of Maryland School of Medicine, Department of Diagnostic Radiology and chief of radiology for the Veterans Affairs Maryland Healthcare System.
He says the number of images radiologists have to deal with has gone up 100 fold in the last 20-25 years with these doctors sometimes looking at 100,000 images a day. They have to process all of this visual information while trying to cross correlate with patient historic data, lab data end even genomic data — and make an accurate diagnosis. He believes faced with that amount of information, it is becoming clear that computers can help.
“Being able to automatically detect diseases and make measurements and quantify disease is increasingly being done with assistance of computers,” Siegel said.
If you’re concerned about the notion of machines making medical decisions, you shouldn’t be Kelly said. It’s more about humans working together with machines to help understand and interpret this tidal wave of information.
Dr. Siegel agrees. “We are looking at how we can take computer technology, which is getting faster and smarter and applying that to making more rapid diagnoses.” As we move forward, he believes we need to build an understanding of how to use these tools to assist medical professionals and learn what humans are good at and what machines are good at.
The Merge Healthcare sale is not a done deal. It’s still subject to regulatory and shareholder approval, but if it passes all of these hurdles, it is expected to close later this year.