A Scanner Smartly: How Artificial Intelligence Is Making Healthcare Imaging Quicker, Kinder And More Efficient
The music festival Lollapalooza is held every year in downtown Chicago, but that’s nothing compared to the gathering that’s been described as “Lollapalooza for radiologists”: the annual meeting of the Radiological Society of North America, which kicked off Dec. 1 at Chicago’s McCormick Place.
OK, nobody actually compares RSNA to Lollapalooza. But it is the world’s largest gathering of radiologists, who come to hear from leaders in the field, take in exhibitors’ booths and talk over their profession’s latest trends and technologies. In recent years, that means talk has turned often to artificial intelligence, which is changing the face of medicine in general — enabling speedier test results, clearer diagnostic scans and greater amounts of one-on-one time between doctors and patients. At RSNA 2019, the 105th Scientific Assembly and Annual Meeting, participants will get to check out an expanded AI Showcase that highlights the technology’s implications for their field.
GE Healthcare, whose scientists and physicians will be deeply engaged at this year’s meeting, has done a lot of work at the intersection of healthcare imaging and artificial intelligence. Here’s a selection.
Dr. Keith Dreyer wears a lot of hats: A radiologist who teaches at Harvard Medical School, he holds a degree in mathematics and a doctorate in computer science, and serves as chief data science officer at Partners HealthCare. It’s in the latter capacity that Dreyer is involved in a collaboration between Partners and GE Healthcare, which in 2017 signed a 10-year agreement to “integrate artificial intelligence into every aspect of the patient journey.” Dreyer spoke to GE Reports at RSNA 2017, laying out the “huge opportunities” AI provides in medical diagnostics, particularly radiology. “I firmly believe that a radiologist plus an AI will beat a radiologist, and will also beat an AI working alone,” Dreyer says. “We have to figure out how to make them work together.”
What’s that look like in practice? Take stroke detection, Dreyer says: “Let’s say we do 200,000 MRI exams of the brain per year and 20,000 are stroke. We can annotate those 20,000, measure the brain lesions caused by the stroke and so on. Next we use the entire 200,000-image set to train the algorithm and use it to identify the type of stroke. When it’s finished, we come back with a test set to see how accurate it was and repeat the process.”
Helping a radiologist reading a CT scan, the AI can then be a tremendous resource: “The machine has already reviewed every orientation of the spine and run through every permutation — much faster than the physician can — and alerted the physician to any abnormalities it’s identified,” Digard said. “It could save the radiologist an enormous amount of time for the software to take care of those universal things. That is, in fact, the entire point of AI — to allow physicians to focus their attention on what is critical for their patients, not spend time on the tedious tasks.”
There needs to be a way to quickly triage the images, and that’s where AI comes in. GE Healthcare’s Critical Care Suite is a collection of algorithms embedded in a mobile X-ray device that can sort through hundreds of images in minutes and call attention to anything that looks suspicious. The team that developed Critical Care Suite trained the algorithm on chest X-rays culled from hospitals around the world; with the suite recently cleared for use by the FDA, doctors can begin working with the software in pilot tests, and it should start showing up more widely in hospitals in 2020.
Finally, radiologists are getting the kind of image clarity that’s long been available to … gamers? It’s true — and it’s part of the appeal of GE Healthcare’s most advanced radiology ultrasound system, the LOGIQ E10, which is powered by advanced algorithms and the same artificial intelligence technology behind advanced gaming. It can process 10 times more data and generate images faster than previous ultrasound systems. The ability to see an injury or problem spot clearly, for instance, gives patients a measure of security — they can look at their scanned images alongside doctors and find their concerns validated. And for doctors it’s a bonanza. “We don’t even bother marking the body with ink before saphenous vein surgery anymore,” says radiologist John Cronan, who uses the system at Lifespan, Rhode Island’s largest health system. He’s referring here to the practice of marking the location of troublesome varicose veins to guide the surgeon during the operation. “We just send the surgeon a photo.” In other words, it’s a snap.