ASGE International Sampler (On-Demand)-5

Pdf Summary

In this document, Dr. Nicholas Petrick introduces the concept of medical artificial intelligence (AI) and machine learning (ML). He outlines the potential benefits of AI/ML in improving diagnostic accuracy, streamlining clinical workflows, improving quality, and personalizing healthcare.<br /><br />Dr. Petrick provides a brief history of AI/ML, starting from the 1950s when machines were developed to emulate human thinking and learning. He mentions the advent of probabilistic approaches in the 1990s, driven by computer science and statistics, and the rise of deep learning in the 2010s, enabled by faster GPUs.<br /><br />The document discusses the traditional AI/ML pipeline, which involves feature identification and classifier development/training, as well as the deep learning pipeline that utilizes deep convolution neural networks.<br /><br />Dr. Petrick mentions the expected growth of healthcare AI/ML, with a projected 37% year-by-year growth rate from 2022 to 2030.<br /><br />He emphasizes the need for best practices in AI/ML, including standardizing nomenclature and definitions. Various organizations and collaborative communities are working towards developing these best practices, such as the IEEE AI Medical Device Working Group and the AI World Consortium Collaborative Community.<br /><br />The role of the Office of Science and Engineering Labs (OSEL) at the U.S. Food and Drug Administration (FDA) is highlighted, as they conduct regulatory science research to support the development of safe and effective medical devices. OSEL focuses on various areas, including AI/ML, and aims to accelerate patient access to innovative devices through regulatory science.<br /><br />Dr. Petrick briefly mentions relevant FDA guidances, such as those regarding software as a medical device, clinical decision support software, premarket submissions for device software functions, and marketing submission recommendations for AI/ML-enabled device software functions.<br /><br />The document goes on to discuss specific applications of AI/ML in medical imaging, particularly in endoscopy. These applications include quantitation, synthetic images, quality control, and automatic report generation. Dr. Petrick highlights the advancements in these areas and the potential for further improvement through AI/ML.<br /><br />In conclusion, medical AI/ML tools are expected to increase in number, and there are plenty of opportunities for AI/ML to enhance endoscopy and other medical practices. Efforts are being made to establish standards and guidance to streamline AI/ML development, assessment, and regulatory review.

Meta Tag

Concept Artificial Intelligence in Medicine

Concept Machine Learning

Concept Diagnostic Accuracy

Concept Deep learning

Concept Medical Imaging

Keywords

medical artificial intelligence

machine learning

diagnostic accuracy

clinical workflows

deep learning

healthcare AI growth

AI best practices

FDA regulations

medical imaging

endoscopy applications

Artificial Intelligence in Medicine

Machine Learning

Diagnostic Accuracy

Deep learning

Medical Imaging