Longitudinal Quantitative Assessment of 40 keV Photon-Counting CT Artifacts for System Stability and Detector Performance Evaluation
Abstract
Purpose
Ring artifacts were observed on our photon-counting CT (PCCT) daily quality control images, with 40 keV virtual monoenergetic images (VMIs) exhibiting the most pronounced artifacts. Because these artifacts may reflect degraded system stability and detector performance, longitudinal monitoring is essential. However, visual assessment alone is challenging for consistent and objective artifact tracking. This study aims to perform a longitudinal quantitative assessment of artifacts on 40 keV images and to evaluate correlations with visual assessment.
Methods
An ACR phantom was scanned daily on a PCCT (NAEOTOM Alpha, Siemens) as an established QC program. Both helical and axial acquisitions were performed at 120 kV and a fixed dose (CTDIvol: 20 mGy). Since 09/23/2024, both 40 keV VMI and low-energy images (T3D) have been reconstructed for artifact quantification. Twenty slices from the uniformity module were used to generate a maximum intensity projection (MIP) image after calculating the absolute Hounsfield Unit value for each pixel. A large ROI covering 80% of the phantom cross-section was placed on the MIP to compute the artifact index, defined as the mean value within the ROI. Artifact indices and corresponding images acquired before and after a major system software upgrade (VA50A_sp1 to VB10_sp1) were compared.
Results
The artifact index showed a strong correlation with artifact severity. Artifact index for T3D images remained stable over 15 months (minimum: 8.3; maximum: 9.5 HU). In contrast, 40 keV images showed a gradual increase in artifact index from 25.2 HU (mean of first 30 measurements) to 36.0 HU (mean of the last 30 measurements) before the software upgrade, after which the artifact index decreased to approximately 16.2 HU and remained stable.
Conclusion
The artifact index serves as an effective metric for quantifying global artifact severity and may serve as a valuable tool for longitudinal monitoring of detector stability and system performance.