AI & the FDA

| Raina Patel

The use of artificial intelligence (AI) in healthcare has been growing rapidly in recent years, and AI-enabled medical devices are playing a larger role in patient care. Such devices may use machine learning algorithms to analyze vast amounts of patient data and provide diagnoses as well as personalized treatment recommendations. In recent years, the FDA has approved various AI-enabled medical devices, some of which are discussed below.

IDx-DR is an AI-powered diagnostic system for detecting diabetic retinopathy (a complication of diabetes that can lead to blindness) through the analysis of retinal images. The system reportedly uses machine learning algorithms to analyze images and make a diagnosis, without the need for a specialist to interpret results. IDx-DR was the first FDA-approved autonomous AI diagnostic system for use in any field of medicine. LVO Stroke Platform is an AI-powered platform for analyzing CT scans of the brain to identify large vessel occlusions that can indicate a potential stroke. The system reportedly uses machine learning algorithms to analyze CT scan images and alert healthcare providers when a potential stroke is detected and that rapid treatment may be required, allowing for improved patient outcomes. LVO is the first FDA-approved AI platform for stroke detection and triage.

ProFound AI for Digital Breast Tomosynthesis (DBT) is an AI-based software from iCAD, Inc. for analyzing tomosynthesis images to assist radiologists in detecting breast cancer. The system reportedly uses machine learning algorithms to analyze mammograms and identify potential areas of concern (e.g., malignant soft tissue densities and calcifications), providing radiologists with a second opinion and improving diagnostic accuracy. ProFound AI was the first 3D tomosynthesis software using artificial intelligence (AI) to be cleared by the FDA, and the software’s current third generation has also received FDA clearance.

Eko DUO is an AI-enabled stethoscope for analyzing heart sounds and detecting potential cardiac abnormalities. The Eko DUO stethoscope, when used with the Eko App, reportedly uses machine learning algorithms to identify specific heart sounds and detect cardiac abnormalities including atrial fibrillation (AFib), murmurs, tachycardia, and bradycardia. Eko DUO is the first FDA-approved AI-enabled stethoscope for use in a clinical setting.

Caption Guidance is an AI-based software from Caption Health that assists in the acquisition of cardiac ultrasound (echocardiogram) images. The software reportedly uses AI to assess the diagnostic quality of such images in real-time, which helps guide healthcare providers in capturing echocardiogram images that are of sufficient diagnostic quality. The FDA authorized the marketing of Caption Guidance in February 2020, making Caption Guidance the first FDA-authorized software for guiding users through cardiac ultrasound image acquisition.