SKOUT® system

The SKOUT® system is a computer-aided detection tool utilizing an artificial intelligence algorithm to analyze real-time colonoscopy video, aiming to improve the identification of potential colorectal polyps. Its methodology involves superimposing visual cues on the endoscopic display to assist gastroenterologists, supported by updated hardware for better performance and usability, and software designed for efficient image and error handling.

The SKOUT® system is a software-based medical device designed to assist gastroenterologists in the real-time detection of potential colorectal polyps during colonoscopy examinations for adult patients undergoing screening or surveillance. It functions as a computer-aided detection (CADe) tool, analyzing high-definition endoscopic video from compatible Olympus and FUJI endoscopes using an artificial intelligence-based algorithm. This algorithm processes the video feed and, upon identifying a suspected polyp, superimposes a blue rectangular outline around the area on the endoscopic monitor, serving as an informational visual aid to enhance the clinician's observational capabilities without replacing their medical judgment. The system's AI algorithm, which forms the core of its detection science, remains consistent with its predicate version. Key operational features include the automatic pausing of polyp detection when surgical tools are present or if lighting is inadequate, preventing procedural interference. The device features an updated single-box hardware architecture for improved component longevity and image quality, alongside software enhancements for better image handling, error management, and overall usability. Non-clinical testing results successfully validated the software, confirmed electrical safety and electromagnetic compatibility, and demonstrated positive human factors outcomes. Performance improvements include a significant reduction in video delay to near-zero (0.0ms for SDI) and no introduction of pixel-level degradation to the endoscopic image.

AI Regulatory Concerns

Powered by

Jev reads the full FDA letter and returns calibrated concern probabilities per regulatory category. These are AI estimates, not regulatory advice.

• Q1. How does the AI algorithm's sensitivity and specificity for polyp detection compare to unassisted colonoscopy, and were there any changes in these metrics due to the hardware and software updates (despite the algorithm itself being unchanged)?; What specific improvements in "image handling" and "error handling" were implemented in the software update, and how do these translate to tangible benefits in clinical workflow or diagnostic accuracy support?; The document states the algorithm is unchanged, implying clinical performance is unchanged. Were any bench or simulated clinical studies performed to confirm that the hardware changes (e.g., single box, updated components) and associated software updates do not inadvertently affect the AI's performance on diverse polyp morphologies or challenging visual conditions?; Given the significant reduction in video delay to 0.0ms (SDI), what technological advancements in the hardware or software architecture enabled this improvement, and how does this near-zero latency enhance the user experience for gastroenterologists?; The system pauses detection when endoscopic tools are present. What is the threshold or method for detecting "inadequate lighting conditions" that also triggers a pause in polyp detection, and how does the system ensure rapid reactivation once conditions are optimal?