Aquilion Lightning (TSX-036A/7) V10.2 with AiCE-i

The Aquilion Lightning (TSX-036A/7) V10.2 system is a computed tomography scanner that incorporates AiCE-i, a Deep Convolutional Neural Network-based image reconstruction technology. This methodology is designed to enhance diagnostic capabilities by significantly reducing image noise, improving spatial resolution, and enabling substantial reductions in patient radiation dose compared to conventional reconstruction algorithms.

The Aquilion Lightning (TSX-036A/7) V10.2 with AiCE-i is an 80-row, whole-body, multi-slice helical computed tomography (CT) scanner designed to acquire and display cross-sectional volume data sets. Its primary purpose is to facilitate specialized medical studies by providing detailed images of the entire body, including the head. A significant technological advancement in this system is the integration of AiCE-i (Advanced intelligent Clear-IQ Engine - integrated), a sophisticated image reconstruction technology. AiCE-i utilizes Deep Convolutional Neural Network (DCNN) methods, which are trained to analyze the statistical properties of the imaging signal and differentiate anatomical structures from noise, thereby aiming for rapid and high-quality CT image reconstructions for every patient. The AiCE-i algorithm is specifically indicated for improving image quality and reducing noise in CT scans of the abdomen, pelvis, lungs, cardiac regions, extremities, head, and inner ear. Performance evaluations using phantoms have demonstrated notable benefits: the AiCE-i technology can achieve double the high-contrast spatial resolution for body imaging when compared to the previous AIDR 3D reconstruction method. Furthermore, studies indicate that AiCE-i supports substantial radiation dose reduction, potentially up to 82.9% relative to traditional Filtered Backprojection (FBP) techniques. When compared to AIDR 3D at the same radiation dose, AiCE-i offers a 15% improvement in low-contrast detectability and a 29% reduction in image noise for body applications, all while maintaining a noise texture that is similar to that produced by FBP.

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 the AiCE-i DCNN vary across the different indicated anatomical regions (e.g., abdomen vs. head vs. inner ear) in terms of the balance between noise reduction, spatial resolution, and low-contrast detectability?; What specific Deep Convolutional Neural Network architecture is utilized in AiCE-i, and what were the key characteristics of the dataset used for its training to ensure robust performance across diverse patient anatomies and pathologies?; The document states up to 82.9% dose reduction relative to FBP. What is the quantified dose reduction potential of AiCE-i when compared directly to the AIDR 3D technology at equivalent image quality levels for various body regions?; Are there any specific image artifacts uniquely introduced or potentially amplified by the AiCE-i DCNN reconstruction, and how do these compare, qualitatively and quantitatively, to artifacts observed with AIDR 3D or FBP, particularly in challenging imaging scenarios?; What is the computational overhead and reconstruction time for AiCE-i compared to AIDR 3D and FBP, and how might this impact workflow efficiency in a busy clinical setting, especially for complex or large volume datasets?