Clinical Evidence
SwiftMR™ 핵심 알고리즘에 대한 소개
에어스메디컬의 ‘정근우, SwiftMR™ 제품 연구 팀장’이 저술한 “All-in-One Deep Learning Framework for MR Image Reconstruction” 논문은 MRI 기술의 획기적인 혁신을 보여줍니다. 자사의 딥러닝 알고리즘은, 기존 모델들과는 달리, 모든 측면에서의 영상 품질 향상을 한번에 달성합니다. 즉, 다차원적인 영상 품질 향상을 통해 MR 이미지 재구성 과정을 간소화하고 전통적인 방법을 능가하는 혁신적인 성능을 제공합니다.
Retrospective study of 86 patients showing that accelerated single-breath-hold T2-weighted liver MRI reconstructed with SwiftMR achieved significantly higher sharpness and signal-to-noise ratio than the routine two-breath-hold protocol (p < 0.001), with higher lesion conspicuity in patients with focal liver lesions
Study of 173 patients showing improved image noise, capsule sharpness, and lesion conspicuity on dynamic contrast-enhanced prostate MRI, while preserving PI-RADS-based diagnostic performance for clinically significant prostate cancer (AUC unchanged; one reader's sensitivity for PI-RADS ≥4 lesions rising from 75.4%
Prospective, multi-institution study showing an approximately 80–86% reduction in lumbar spine MRI acquisition time using a SwiftMR-enabled ultrafast protocols. Diagnostic images were produced within 2 minutes, while maintaining diagnostic interchangeability with conventional MRI. A subsequent VLM-based automated reports reached substantial-to-almost-perfect
In this white paper, we compared SwiftMR™-processed MR exams to the standard of care. We analyzed 184 MR exam pairs from 12 anatomical locations, considering various pathologies, scanner vendors, and field strengths. Eighteen board-certified radiologists from six subspecialties assessed these
SwiftMR™, an AI-powered MRI reconstruction software, offers clear economic benefits to imaging centers. In an economic model, we estimated that a typical imaging center using SwiftMR™ could generate an additional $271,790 in revenue over one year, achieving a 453% return
This study investigated the improvement of image quality and diagnostic performance utilizing SwiftMR for time-resolved, contrast-enhanced magnetic resonance angiography (CE-MRA) in fast stroke imaging. Specifically, relatively low signal-to-noise ratio and spatial resolution of time-resolved angiography with interleaved stochastic trajectories (4D-TWIST-MRA)
The article, “All-in-One Deep Learning Framework for MR Image Reconstruction,” highlights a groundbreaking innovation in MRI technology. This deep learning framework is distinguished from previously developed models by integrating all essential functions into a single, comprehensive solution. By doing so,
The whitepaper by Geunu Jeong, MD, Head of SwiftMR Research at AIRS Medical Inc., introduces SwiftMR™, a deep learning-based technology designed to enhance MRI quality by improving the signal-to-noise ratio (SNR) and spatial resolution without increasing scan times or introducing
This review paper explores existing literature on the diverse applications of deep learning reconstruction (DLR) techniques for fast MR neuroimaging. It provides a comprehensive overview of how DLR is being utilized to reduce scan times while maintaining or even enhancing
This study explores the integration of SwiftMR in enhancing the image quality of brain epilepsy imaging. The research utilized SwiftMR to improve the image quality and detail of thin-section brain MRI – by comparing the standard-of-care (SOC) 3mm MRI with
This study investigated the application of SwiftMR in enhancing the efficiency and quality of MRI scans for degenerative lumbar spine diseases in a tertiary hospital environment. By comparing the standard-of-care (SOC) L-spine MRI with accelerated MRI using SwiftMR, the exam
AI applications can enhance imaging processes, reduce scan times, and improve image quality, thereby increasing efficiency. SwiftMR’s technology has the potential to significantly increase the annual revenue of each MR scanner by $900,000. This increase is achieved by enhancing scanner
This study evaluated the performance of SwiftMR in accelerating the knee imaging protocol. 45 participants with knee pain were scanned on three MRI scanners from three different vendors, where the findings suggested that the scan time for the standard-of-care protocol