Gastroenterology and Artificial Intelligence: 4th Annual Global Artificial Intelligence Summit (On-Demand | August 2022-Regulations for Platform-based "Suites of AI Technology"

Pdf Summary

This document discusses the regulations for platform-based "Suites of AI Technology" in the medical device industry. Medical devices are classified into three classes based on their risk level, with Class III having the highest risk. The pre-market requirements for each device class are outlined, including general controls, special controls, premarket notification, de novo classification, and premarket approval. <br /><br />The document specifically focuses on the classification and premarket requirements for gastrointestinal (GI) AI/ML products, such as the Gastrointestinal Lesion Software Detection System (QNP). Three unique devices cleared under QNP are mentioned, which aid endoscopists in detecting colonic mucosal lesions.<br /><br />For QNP devices, there are special controls in place, including clinical performance testing, non-clinical performance testing, usability assessment, and performance data verification. The labeling requirements for these devices are also outlined.<br /><br />The document discusses other potential types of AI/ML aids for endoscopy, including computer-aided detection, computer-aided diagnosis, and computer-aided acquisition/optimization. These types of aids may be eligible for a De Novo designation based on risk and impact on the standard of care.<br /><br />The assessment studies for endoscopic CADe (Computer-Aided Detection) are explained, including clinical testing and standalone performance testing. The importance of an MRMC (Multiple Reader, Multiple Case) study design for clinical testing is emphasized.<br /><br />Standalone performance testing involves evaluating the algorithm's performance alone, measuring object-based and frame-based performance. The importance of including a range of acquisition systems and demographic populations in the testing is highlighted.<br /><br />In summary, the document provides an overview of the regulations and requirements for platform-based AI technology in the medical device industry, with a specific focus on gastrointestinal AI/ML products and endoscopic CADe devices. The importance of clinical and standalone performance testing is emphasized in the regulatory review process.

Asset Subtitle

Nicholas Petrick, PhD

Keywords

regulations

platform-based AI technology

medical device industry

device classification

pre-market requirements

gastrointestinal AI/ML products

QNP devices

clinical performance testing

standalone performance testing

regulatory review process