Cartesion Prime (PCD-1000A/3) V10.8

The Cartesion Prime (PCD-1000A/3) V10.8 is a hybrid PET/CT imaging system, integrating a high-throughput PET with digital SiPM sensors and a high-end multi-slice CT, designed for comprehensive diagnostic assessment by providing fused anatomical and metabolic/functional information. It features AiCE-i, a deep learning-based image reconstruction technology that enhances PET image quality from FDG whole-body data by statistically differentiating signal from noise, thereby improving image clarity, diagnostic confidence, and quantitative accuracy.

The Canon Cartesion Prime (PCD-1000A/3) V10.8 is a diagnostic imaging system integrating a high-throughput Positron Emission Tomography (PET) scanner with a high-end multi-slice helical Computed Tomography (CT) scanner. The CT component, featuring a 780 mm gantry aperture and 700 mm maximum scan field of view, generates cross-sectional anatomical images. The PET component, equipped with digital SiPM sensors and achieving a temporal resolution typically around 263 ps, images the distribution of injected PET radiopharmaceuticals to provide metabolic and functional information. These PET images utilize CT data for attenuation correction and precise anatomical referencing in fused PET/CT images, aiding healthcare professionals in evaluating, diagnosing, and managing oncological, cardiovascular, and neurological conditions; the system also functions as a standalone whole-body CT scanner. A key technological advancement in the V10.8 model is 'AiCE-i for PET' (Advanced Intelligent Clear-IQ Engine-integrated), a reconstruction technology employing deep learning artificial neural networks. This methodology enhances PET image quality by analyzing statistical properties of signal and noise from FDG whole-body data, significantly reducing noise and improving signal differentiation. Performance testing, including NEMA NU 2-2018 standards and phantom studies, demonstrated AiCE-i's ability to improve signal-to-noise ratio, enhance quantification accuracy (including Contrast Recovery Coefficient and Background Variability), allow for reduced count rates or scan times while preserving noise levels, and maintain the integrity of SUV and activity concentration without introducing artifacts. Clinical evaluations by experienced nuclear medicine physicians further confirmed AiCE-i processed images are of diagnostic quality, showing improvements in overall image quality, sharpness, and typically reduced noise compared to traditional Gaussian methods.

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-i for PET vary with different PET radiopharmaceuticals beyond FDG, and across a wider range of patient demographics or disease states not detailed in the provided summary?; What are the specific computational requirements (e.g., processing time, hardware specifications) for the AiCE-i deep learning reconstruction, and how do these impact clinical workflow efficiency compared to traditional methods?; Can the AiCE-i algorithm be customized or further fine-tuned by users for specific, less common clinical applications or research protocols to optimize its noise reduction and image quality benefits?; Beyond improved image quality metrics like SNR and physician-rated quality, what is the direct impact of AiCE-i on diagnostic accuracy, lesion detectability, particularly for small or low-contrast lesions, or changes in patient management strategies?; What were the key characteristics and diversity of the training dataset used for the AiCE-i deep learning model, and what steps were taken to ensure its generalizability and mitigate potential biases across varied patient populations or rare pathologies?