Effect of an advanced model-based iterative reconstruction algorithm on texture and visual impression of images privieded by a dual source CT scanner

TL;DR

This study evaluates how the advanced ADMIRE model-based iterative reconstruction algorithm affects image texture and visual quality in dual source CT scans, demonstrating potential for significant dose reduction while maintaining image quality.

Contribution

It introduces a comprehensive assessment combining texture analysis and visual similarity metrics to evaluate dose reduction capabilities of ADMIRE in CT imaging.

Findings

ADMIRE levels reduce texture contrast and entropy.

Comparable image quality achieved with up to 50% dose reduction.

Spatial resolution maintained up to 90% dose reduction.

Abstract

Purpose: To evaluate the influence of ADMIRE algorithm on image texture and image visual impression as a supplement to measurements of common image quality parameters such as noise levels and spatial resolution. Methods: An ACR phantom was examined at different radiation dose levels. To characterize the image texture, two Haralick texture parameters, contrast and entropy, for different dose level and reconstruction algorithms were assessed. The visual impression of images and their structural differences were evaluated using the structural similarity index (SSIM). Noise was determined for all dose and AMIRE levels and compared to those by filtered back projection (FBP). The spatial resolution was determined by the modulation transfer functions and the line spread function. Results: The Haralick texture parameters decreased with increasing ADMIRE levels I up to V. ADMIRE III, IV and V offered a comparable contrast and entropy to those calculated by FBP with a radiation dose reduction up to 50%. SSIM improved with increasing ADMIRE levels. SSIM calculated by ADMIRE IV and V showed similar values by the filter back projection with a dose decrease up to 50%. Spatial resolution was retained up to 90% dose reduction. With increasing ADMIRE as well as dose level the noise distribution shifted to a more narrow distribution, which was in accordance with the reconstructed images using ADMIRE. Conclusion: Texture analysis and SSIM allow a more realistic assessment of the dose reduction potential of iterative reconstruction algorithms than quality metrics only based on physical measurements of noise distribution or spatial resolution. This work presented that by means of the ADMIRE algorithm, a comparable image quality at reduced radiation doses can be reached.