Fujifilm wins FDA nod for AI-powered endoscopy tech that could ...

Fujifilm announced today that the FDA granted 510(k) clearance for its CAD EYE AI-powered detection system for endoscopic imaging.

CAD EYE enables the real-time detection of colonic mucosal lesions like polyps and adenomas during colonoscopy procedures. It supports endoscopists in detecting and removing pre-cancerous lesions, regardless of size, shape and color. The system enters a market currently cornered by Medtronic and its GI Genius system.

GI Genius, a computer-aided polyp detection system powered by AI, first became available in the U.S. in 2021. Read more about Medtronic’s AI-related efforts for GI Genius here.

Fujifilm’s system features a compatible expansion unit (the Fujifilm EX-1) and endoscopy support software (EW10-EC02). CAD Eye builds upon the company’s Eluxeo endoscopic imaging system, with AI image processing for integration with the system’s processor and endoscope.

The company plans to make CAD EYE commercially available this spring after it completes a limited market evaluation.

“Every day we pride ourselves on delivering what we believe is the highest quality imaging and optics, arming endoscopists with the tools they need to combat this public health issue,” said Tai Fujita, GM, endoscopy, Fujifilm Healthcare Americas Corporation. “Today, we’re thrilled to take that a step further with the introduction of CAD EYE, which has the potential to dramatically improve the quality of colonoscopy.”

The company developed CAD EYE using deep learning technology in its Tokyo-based global AI technology center. Fujifilm validated the system using histologically confirmed polyps in clinical images.

CAD EYE works with both white light imaging and Linked Color Imaging (LCI). Fujifilm included an enhanced visualization mode to differentiate the red color spectrum, enhancing mucosal visualization.

According to a news release, the technology supports the detection of lesions that could prove easy to miss. This includes flat lesions, those at the corner of the endoscopic view and multiple lesions present in a single frame. When CAD EYE identifies a suspicious polyp, the physician automatically receives both visual and auditory alerts. Visual cues overlay, but don’t interfere with the clinical images within the existing workflow.

Fujifilm said studies demonstrated CAD EYE’s ability to lead to considerable advances in colorectal cancer detection and diagnosis. The system detects significantly more adenomas during screening and surveillance compared to conventional high-definition colonoscopy without AI assistance. Plus, that occurs without any increase in procedure time.

Additionally, Fujifilm saw 17% higher adenoma per colonoscopy when compared to high-definition conventional colonoscopy. The system detects colorectal neoplastic lesions at a level comparable to that of an expert and superior to that of a beginner, the company said.

“As a gastroenterologist dedicated to patient care and safety, I’m heartened by the FDA’s approval of Fujifilm’s AI CAD polyp detection algorithm. This breakthrough not only enables the early detection of precancerous lesions that can lead up to colorectal cancer but also significantly reduces the risk of missed lesions, enhancing our precision and improving patient outcomes,” said DR. Sravanthi Parasa of the Swedish Medical Center. “It marks a pivotal advancement in our ability to safeguard patient health, reinforcing our commitment to embracing technological innovations in gastrointestinal healthcare.”