Skip to main content
2024 BMES Annual Meeting Main banner
Icon Legend
This session is not in your schedule.
This session is in your schedule. Click again to remove it.

High School Poster

HS Poster

Poster 141 - Small Pixel Effect in Clinical Photon-Counting CT: Noise Power Spectrum and Low-Contrast Detectability

Saturday, October 26, 2024
10:00 AM - 12:00 PM EST
Location: Exhibit Hall E, F & G

Introduction: Clinical photon-counting computed tomography (PCCT), FDA-approved in 2021, has garnered significant attention for enhancing image quality, diagnostic performance, and dose efficiency compared to conventional CT. PCCT offers two scan modes: ultra-high-resolution (UHR) with native detector pixels and standard-resolution (SR) with 2x2 pixel binning. Phantom and clinical studies suggest that images acquired using UHR reconstructed with identical parameters to CT data collected in SR offer lower noise and potential dose reduction due to the small pixel effect compared to SR acquisition [1-3]. However, differences in noise power spectrum (NPS) and low-contrast detectability (LCD) between UHR and SR remain unexplored. This study aims to evaluate the impact of small pixel effects on NPS and LCD across various doses, reconstruction thicknesses, and convolutional kernels.

Materials and

Methods: A standard (American College of Radiology) ACR CT accreditation phantom was scanned on a clinical PCCT (NAEOTOM Alpha, Siemens) in SR mode (collimation: 144×0.4 mm) and UHR mode (120×0.2 mm) at 120 kV, pitch of 0.6, and four matched radiation doses [CT Dose Index volume (CTDIvol): 12, 6, 3, and 1 mGy]. Virtual monoenergetic images at 70 keV were reconstructed at 3 mm thickness with a body regular kernel (Br44) and quantum iterative reconstruction strength of 3, using a fixed field-of-view of 210 mm. The PCCT data acquired at 12 mGy were additionally reconstructed at slice thicknesses of 0.4, 1, and 5 mm, and with various convolutional kernels (Br36, 40, 48, 56, 60, 64, 68, 72, and 76). The reconstructed images were uploaded to CTPro (ctpro.net), a web-based software platform for CT image quality evaluation and radiation dose optimization for data processing [4, 5]. The ACR CT phantom’s uniformity module (i.e., the 3rd module) was used for noise evaluations, including noise magnitude and two-dimensional NPS using the NPS software tool in CTPro. The LCD module (i.e., the 2nd module) was used to determine the index of detectability (d’) and area under the ROC curve (AUC) using the channelized Hotelling observer (CHO) software tool in CTPro.

Results, Conclusions, and Discussions: The noise magnitude was consistently lower in images acquired with the UHR mode compared to the SR mode. More significant noise reduction was observed with decreased radiation dose from 12 mGy (4%) to 1 mGy (27%) and increased kernel sharpness from Br36 (2%) to Br76 (13%), but not with changes in slice thickness from 0.4 mm to 5 mm (3-4%). NPS analysis indicated that noise reduction is spatial frequency-dependent, showing minimal impact on low spatial frequencies but greater impact on medium to high spatial frequencies, with the average spatial frequency reduced by up to 37%, depending on slice thickness, radiation dose, and convolutional kernels (Figures 1-3). Both d’ and AUC were consistently improved in the UHR mode across all radiation dose levels, slice thicknesses, and kernel sharpness levels, demonstrating superior low-contrast detectability and overall image quality (Figures 4-5). Due to the small pixel effect, at matched radiation dose levels and reconstructed parameters, the UHR mode can suppress more medium and high spatial frequency noise, especially with reduced dose levels and increased kernel sharpness. It also consistently improves low-contrast detectability and enhances overall image quality.

Acknowledgements (Optional): Research reported in this abstract was supported by the RSNA Research & Education Foundation (Research Scholar Grant: RSCH2302). The content is solely the responsibility of the authors and does not necessarily represent the official views of the RSNA Research & Education Foundation.

References (Optional): 1. Fix Martinez, M., et al., Potential radiation dose reduction in clinical photon-counting CT by the small pixel effect: ultra-high resolution (UHR) acquisitions reconstructed to standard resolution. European Radiology, 2024. 34(7): p. 4484-4491.
2. Huflage, H., et al., Investigating the Small Pixel Effect in Ultra-High Resolution Photon-Counting CT of the Lung. Investigative Radiology, 2024. 59(4): p. 293-297.
3. Huflage, H., et al., The Small Pixel Effect in Ultra-High-Resolution Photon-Counting CT of the Lumbar Spine. Investigative Radiology, 9900: p. 10.1097/RLI.0000000000001069.
4. Fan, M., et al., A web-based software platform for efficient and quantitative CT image quality assessment and protocol optimization. SPIE Medical Imaging. Vol. 11595. 2021: SPIE.
5. Zhou, Z., et al., Automated web-based software for CT quality control testing of low-contrast detectability using model observers. SPIE Medical Imaging. Vol. 12925. 2024: SPIE.