FDA Radiology AI Devices

SwiftMR is a software medical device designed to automatically enhance brain MRI images by reducing noise and improving image sharpness. It processes MRI scans in DICOM format using advanced deep learning models, helping radiology professionals obtain clearer images for better diagnosis and clinical decision-making without interrupting their workflow.

VBrain-OAR is a software device from Vysioneer Inc. designed to assist trained radiotherapy personnel in brain tumor radiation therapy treatment planning by automatically generating contours of organs at risk in brain MRI images using AI deep learning. It also provides image registration functions to align multi-modality medical images. This device supports clinicians by providing initial contour outlines to improve the efficiency and consistency of radiation treatment planning, while still requiring final confirmation by qualified personnel.

SubtlePET is an advanced software tool designed to improve the quality of PET medical images by reducing noise. Utilizing deep learning algorithms, it enhances image clarity, aiding radiologists and nuclear medicine physicians in interpreting PET scans more accurately. This software integrates smoothly into existing radiology workflows by receiving, processing, and forwarding DICOM images, supporting multiple radiotracers used in PET imaging.

The Philips EPIQ and Affiniti Diagnostic Ultrasound Systems are advanced ultrasound machines that provide diagnostic imaging and fluid flow analysis for various body regions. They help clinicians visualize internal organs and blood flow in real time, assisting in the diagnosis and treatment of conditions related to the abdomen, heart, brain, musculoskeletal system, and more. The system also includes liver fat quantification tools to measure attenuation and hepato-renal index, supporting assessment of liver health.

Intelligent NR is a software function integrated with Canon's CXDI Control Software that uses machine learning-based noise reduction to enhance the quality of digital X-ray images. It helps clinicians obtain clearer images from conventional radiographic exams, improving diagnostic confidence especially in trauma, intensive care, and pediatric settings.

Deep Learning Image Reconstruction by GE Healthcare is a software that uses deep neural networks to reconstruct high-quality CT images from X-ray transmission data. It helps radiologists by producing images with improved noise reduction, spatial resolution, and artifact suppression, suitable for head, whole body, cardiac, and vascular CT scans at routine clinical throughput.

HealthCCSng is a software device that automatically analyzes non-cardiac-gated CT scans to estimate the amount of calcified plaque in coronary arteries, which can indicate coronary artery disease risk. It provides radiologists with calcium detection categories and annotated images to aid in clinical decision-making within their usual workflow.

Versana Premier is a general purpose diagnostic ultrasound system designed for use by trained healthcare professionals. It supports multiple clinical applications such as fetal, abdominal, gynecological, cardiac, vascular, musculoskeletal, and more. It includes various advanced imaging modes and features AI-based enhancements like Whizz Label to assist clinicians in acquiring and interpreting ultrasound images efficiently and accurately. The system is intended for hospitals and medical clinics to aid in the diagnosis and assessment of medical conditions.

The Aquilion Exceed LB is a whole-body multi-slice helical CT scanner that captures detailed cross-sectional images of the body, including the head. It incorporates AiCE, an AI-based noise reduction algorithm using deep convolutional neural networks, to improve image quality and reduce noise. This helps clinicians obtain clearer, more accurate diagnostic images for various body regions, including abdomen, pelvis, lung, cardiac, extremities, head, and inner ear.

ClariSIGMAM is a software tool designed to automatically analyze digital mammograms to estimate breast density by calculating the ratio of fibroglandular tissue to the total breast area. It categorizes breast density using BI-RADS groups to assist radiologists in assessing breast tissue composition, providing adjunctive information to support interpretation but is not a diagnostic aid.