Poster Poster Program Diagnostic and Interventional Radiology Physics

Evaluation of a Deep Learning Image Reconstruction Algorithm for Improved Contrast and Decreased Image Noise In Abdominal CT Imaging

Abstract
Purpose

To examine the subjective and objective image quality (IQ) of a deep learning image reconstruction algorithm (Precise Image Quality Engine (PIQE)) in abdominal CT.

Methods

Sixty-nine adult patients who underwent routine, contrast-enhanced abdominopelvic imaging on the Aquilion ONE/INSIGHT CT system (Canon Medical Systems, Japan) were retrospectively enrolled in the IRB-approved trial. Images were reconstructed using Advanced Intelligent Clear-IQ (AiCE) (strength level (L) 1, 512x512 matrix) and PIQE (L1 and L2, 1024x1024 matrix). Four radiologists were blinded to the identify of the reconstruction and assessed the image noise, contrast, small structure visibility, image sharpness, artifacts, and image preference with Likert scales. Reader agreement was assessed with linearly weighted Cohen’s kappa. Circular regions of interest (ROIs) were placed on five slices to include the left and right liver, portal vein, aorta, subcutaneous fat, and bilateral psoas muscles. CT numbers, noise, signal-to-noise (SNR) and contrast-to-noise (CNR) ratios were determined for each ROI. Significant differences between reconstruction algorithms were assessed via the Friedman test with post-hoc Dunn-Sidak corrections for both subjective and objective image quality metrics.

Results

Reader agreement was fair (k = 0.21). PIQE L2 was preferred for image contrast and noise (p < 0.02). PIQE L1 was preferred for image sharpness (p < 0.02). No statistical difference was found for small structure visibility, artifacts, or image preference scores. CT numbers were significantly different between the AiCE and PIQE reconstructions (p < 0.05). Noise was statistically lowest in PIQE L2 compared to AiCE (p < 0.05). Mean SNR and CNR were superior for PIQE L2 compared to AiCE and PIQE L1 for all structures assessed (p < 0.003).

Conclusion

The best subjective IQ metrics for image contrast, noise, and sharpness were obtained with PIQE reconstructions. PIQE L2 demonstrated the best objective IQ metrics (SNR and CNR) across multiple abdominopelvic anatomic structures.

People

Related

Similar sessions

Poster Poster Program
Jul 19 · 07:00
B-Trac – Breast Tissue Rotation and Compression Apparatus for Calibration

Mammography (compressed 2D) and MRI (uncompressed 3D) capture breast tissue under different conditions, complicating tumor localization across modalities. To bridge this gap, we developed a customizable physical platform to simul...

Dayadna Hernandez Perez
Diagnostic and Interventional Radiology Physics 0 people interested
Poster Poster Program
Jul 19 · 07:00
Comprehensive Medical Physics Assessment of Digital Mammography Equipment: A Three-Year Multi-Site Evaluation of Technical Performance and Radiation Safety at 24 Saudi Arabian Healthcare Institutions (2022–2024)

To conduct a comprehensive multi-center audit evaluating the technical performance, image quality, and radiation safety of digital mammography systems across 24 unique healthcare facilities in Saudi Arabia. This study aims to est...

Sami Alshaikh, PhD
Diagnostic and Interventional Radiology Physics 0 people interested
Poster Poster Program
Jul 19 · 07:00
Starting Small: Implementing a CT Protocol Optimization Program

This talk describes our organization’s CT optimization program, and how we implemented it to make efficient use of limited physicist time.

Robert J. Cropp, PhD
Diagnostic and Interventional Radiology Physics 0 people interested