Vantage Orian 1.5T, MRT-1550, V6.0 with AiCE Reconstruction Processing Unit for MR

The Vantage Orian 1.5T, MRT-1550, V6.0, is a diagnostic magnetic resonance imaging system that produces detailed anatomical images by detecting nuclear magnetic resonance signals based on proton density, relaxation times, flow, and chemical shift. This system incorporates the AiCE Reconstruction Processing Unit, which utilizes Deep Convolutional Neural Networks, trained to identify and remove Gaussian noise, to enhance image quality and reduce thermal noise, particularly for brain and knee applications, thereby improving signal-to-noise ratio while maintaining image contrast.

The Vantage Orian 1.5T, MRT-1550, V6.0 is a 1.5 Tesla Magnetic Resonance Imaging (MRI) system designed for diagnostic purposes, generating cross-sectional images (transaxial, coronal, sagittal, and oblique) of anatomical structures within the head or body. It is also capable of non-contrast enhanced imaging techniques, such as Magnetic Resonance Angiography (MRA). The system's imaging methodology is based on detecting nuclear magnetic resonance (NMR) signals from protons, utilizing variations in proton density (PD), spin-lattice relaxation time (T1), spin-spin relaxation time (T2), flow dynamics, and chemical shifts to construct images. Depending on the diagnostic requirements for specific regions of interest, contrast agents may be administered to enhance image clarity and provide additional diagnostic information to a trained physician. A significant enhancement to this system is the introduction of the AiCE (Advanced Intelligent Clear-IQ Engine) Reconstruction Processing Unit, an optional feature focused on improving image quality. AiCE employs Deep Convolutional Neural Network (DCNN) methods to reduce thermal noise, with a specific design emphasis on brain and knee imaging. The DCNN is trained using high-signal-to-noise ratio (SNR) images to which Gaussian noise of varying amplitudes has been artificially added, enabling the network to learn and effectively remove thermal noise contributions from MR images. Performance evaluations, including model observer studies and a human observer study involving board-certified radiologists, demonstrated that AiCE can maintain or improve low-contrast detectability, leads to an increase in SNR, and successfully preserves image contrast when compared to conventional imaging protocols and other noise reduction filters.

AI Regulatory Concerns

Powered by

Jev reads the full FDA letter and returns calibrated concern probabilities per regulatory category. These are AI estimates, not regulatory advice.

• Q1. How does the performance of AiCE vary across different MRI sequences (e.g., T1-weighted, T2-weighted, FLAIR) for brain and knee imaging, beyond overall SNR improvement?; What are the specific computational requirements, such as processing time and hardware specifications, for the AiCE Reconstruction Processing Unit, and how might these impact clinical workflow efficiency?; Are there particular types of imaging artifacts or image distortions that AiCE might introduce or inadvertently amplify, especially in complex pathological cases or when applied to anatomical regions beyond the trained brain and knee areas?; Given AiCE is trained to remove Gaussian distributed noise, how robust is its noise reduction capability when faced with non-Gaussian noise or structured noise patterns that may occur in clinical MRI acquisitions?; What was the quantitative range of signal-to-noise ratio (SNR) improvement observed with AiCE in the sample clinical brain and knee images, and how does this improvement potentially translate to scan time reductions while maintaining or enhancing diagnostic image quality?