Dive Brief:
- Medtronic announced launch of an artificial intelligence enhanced endoscopy system in European markets designed to detect pre-cancerous polyps.
- The medtech giant is partnering with Cosmo Pharmaceuticals, the manufacturer of the AI software and device, to distribute the system, called GI Genius.
- Medtronic expects to bring the system to the U.S. market in the second half of 2020, spokesman John Jordan told MedTech Dive in an email. U.S. clinical trials are planned for the device, he said.
Dive Insight:
Artificial intelligence in healthcare is luring investment because of its potential to improve and speed diagnosis, particularly in radiology and image analysis. The Cleveland Clinic this year placed AI in its annual list of Top 10 Medical Innovations and unveiled a new center to develop clinical uses in areas such as diagnostics, disease prediction and treatment planning.
FDA to date has approved more than 30 machine learning algorithms for use in medicine, ranging from GE Healthcare’s AI-augmented X-ray system to detect collapsed lungs to AliveCor’s KardiaAI software library to assess electrocardiogram rhythms. An analysis published last month in The Lancet Digital Health concluded AI can detect diseases from medical imaging with the same accuracy as healthcare professionals.
Medtronic previously demonstrated its interest in AI with its partnership with Viz.ai to accelerate the adoption of AI stroke detection and triage software.
In colonoscopy, AI could help prevent doctors from missing polyps and assist in making a precise diagnosis. Medtronic said the GI Genius will act as a second observer, noting that studies have shown that having another set of eyes can increase polyp detection rates. The device, which the company claims is the first AI system for colonoscopy, recognizes pre-cancerous lesions in real time and highlights them with a visual marker.
Medtronic’s Jordan said the system was trained and validated using a series of videos of 2,684 histologically confirmed polyps from 840 patients who underwent high-definition, white-light colonoscopy as part of a previous randomized controlled study. A total of 1.5 million images showing the polyps from different perspectives were taken from videos and manually annotated by expert endoscopists, he said.
"It’s important to note that the AI technology does not need to compare in real-time between images. The AI algorithms are trained and capable of identifying polyps without the need to access image databases during the product’s operation," Jordan said.
Alessandro Repici, head of gastroenterology at Humanitas Hospital in Milan, Italy, said in a Medtronic press release the GI Genius module "can be extremely precise in identifying lesions in the colonic mucosa that can be difficult to detect and may have been missed," thereby offering potential to improve diagnosis and outcomes for patients that may have colorectal cancer.
