A Review of: The Acceptability of Iterative Reconstruction Algorithms in Head CT
An Assessment of Sinogram Affirmed Iterative Reconstruction (SAFIRE) vs. Filtered Back Projection (FBP) Using Phantoms - Martine A Harris 1, John Huckle 2, Denis Anthony 2, Paul Charnock 3
21 January, 2024 by
A Review of: The Acceptability of Iterative Reconstruction Algorithms in Head CT
ROTHBAND, Paul Dixon

In this study, the primary focus is on optimizing CT scans for neurologic conditions, a procedure often repeated for patients, leading to cumulative radiation doses. With the introduction of dose-reduction technologies, the study examines how these advancements, particularly iterative reconstruction (IR), can enhance image quality while minimising radiation exposure.

Background: Computed tomography (CT) is a key diagnostic tool for various neurological conditions, often leading to patients receiving cumulative radiation doses. Iterative reconstruction (IR) offers a way to optimise these doses. However, head CT poses specific challenges regarding image quality, potentially influencing radiologist preferences.

Objectives: This study aims to examine the relationship between image quality metrics, texture, and the radiation dose in IR head CT protocol optimisation, using a simulated patient environment. Additionally, it explores how these optimised protocols affect diagnostic confidence, employing a specialised phantom for this purpose.

Methods and Setting: A three-stage phantom study was conducted to evaluate different reconstruction methods across a spectrum of exposures, comparing them to the local standard. Measurements included CT numbers, pixel noise, noise uniformity, noise power spectrum, and contrast-to-noise ratio (CNR), alongside high- and low-contrast resolution assessments. Optimised images were evaluated by reviewers against established criteria.

Results: Stronger IR application noticeably reduced pixel noise and noise variance while enhancing CNR. At standard reference levels, image noise dropped by 1.5 standard deviations and CNR improved by 2.0. Image quality remained consistent with up to 24% reduction in relative dose. Except for image sharpness, there were no significant quality differences between IR and filtered back projection reconstructions.

Conclusions: IR can positively affect pixel noise, CNR, and noise variance (image texture), and carefully optimised IR protocols can maintain the image quality seen in filtered back projection methods. This research has informed the adoption and acceptance of lower dose head CT protocols locally.

Rothband Comment:
As the supplier of the phantoms used in this study, we are extremely pleased with our collaboration with Martine and her team. Their innovative approach in optimising head CT protocols using iterative reconstruction (IR) has highlighted the vital role of high-quality phantoms in advancing medical imaging research. Our company prides itself on providing tools that enable such groundbreaking work, and seeing our phantoms contribute to significant improvements in image quality and dose reduction is truly rewarding. We are excited about the potential implications of this study for the medical community and look forward to future collaborations that push the boundaries of diagnostic imaging.

A Review of: The Acceptability of Iterative Reconstruction Algorithms in Head CT
ROTHBAND, Paul Dixon 21 January, 2024
Share this post

To install this Web App in your iPhone/iPad press and then Add to Home Screen.