DeepFoqus-Accelerate is an AI-driven software solution designed to improve brain MRI scans by reconstructing accelerated MRI data up to 4 times faster, producing clinical-quality images from undersampled data. It helps radiologists and technologists by enhancing non-contrast brain MRI images from 1.5T or 3T Siemens and GE scanners, making MRI scanning more efficient while maintaining image quality.
DeepFoqus-Accelerate is a stand-alone software solution intended for acceptance, enhancement, and transfer of brain MRI images in DICOM format. It reconstructs non-contrast enhanced MRI images acquired on 1.5T or 3T Siemens and GE scanners, using Sagittal, Axial, or Coronal T1, T2, or FLAIR sequences. It is intended for adult scans only and is not for mobile devices.
DeepFoqus-Accelerate uses AI and machine learning models, including convolutional neural networks (CNN) and U-net architectures, to reconstruct MRI images from up to 4x accelerated scans. The pipeline includes ingestion of k-space and DICOM files, preprocessing, application of an ensemble of fine-tuned deep learning and signal processing models, post-processing, and final DICOM output. It accepts HDF5 and DICOM inputs and is trained on diverse datasets across magnet strengths and image orientations.
Performance evaluation included software verification and validation (code review, functional testing, penetration testing), analytical performance validation using SSIM, PSNR, and HaarPSI metrics, phantom bench testing according to ACR MRI guidelines, artifact assessment for reconstruction robustness, clinical performance validation by radiologist review of diverse scans, and user acceptance testing by radiologists and technologists. Results demonstrated comparable performance to the predicate device SwiftMR with no additional reconstruction errors.
No predicate devices specified
Submission
7/5/2024
FDA Approval
4/4/2025
Join hundreds of your peers who rely on RadAI Slice. Get the essential weekly briefing that empowers you to navigate the future of radiology.
We respect your privacy. Unsubscribe at any time.