Deep-learning-based spectral motion artifact correction on photon-counting cardiac CT images
Physics in Medicine & Biology · 2026
DOI: 10.1088/1361-6560/ae3eee ↗Kurz gefasst
Fachartikel in der Zeitschrift Physics in Medicine & Biology, 2026. Verfasst von Ruihan Huang und 6 weiteren Autor:innen. Beteiligte Einrichtung: Department of Physics, KTH Royal Institute of Technology, Stockholm, Sweden. Hauptthemen laut MEDLINE: Artifacts, Tomography, X-Ray Computed, Deep Learning, Heart. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: The Swedish Research Council, Göran Gustafssons Stiftelse för Naturvetenskaplig och Medicinsk Forskning, China Scholarship Council.
Abstract (Originalsprache)
Objective: While photon-counting CT (PCCT) improves image quality and reduces radiation dose, artifacts induced by cardiac and respiratory motion is still a challenge. The purpose of this work is to evaluate the potential of an image-domain motion-artifact-correction method based on a deep-learning model that incorporates spectral information (material basis images).
Approach: We simulated PCCT imaging of five XCAT phantoms, and used these for training two deep neural networks—one with and one without spectral information—to map two motion-corrupted virtual monoenergetic images to corresponding motion-free images. Using images from another simulated XCAT phantom, we calculated the CT number error on five regions of interest and 10 segmented organs. The method was also evaluated visually on clinical cardiac PCCT images. Stretch quantification of endocardial engraved zones was used to calculate regional wall motion and mechanical delay. The results were compared with the motion-free image using a paired t-test.
Main results: Out of 45 regions and organs, the CT number accuracy is improved in 41 regions(91%). Among these, the best accuracy is obtained with spectral information in 25 regions(61%). Both models, in particular the one with spectral information, improves visual image quality in simulated and clinical images. The model significantly (P
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Department of Physics, KTH Royal Institute of Technology, Stockholm, Sweden.
- MedTechLabs, BioClinicum, Karolinska University Hospital, Solna, Sweden.
- Department of Clinical Neuroscience, Karolinska Institutet, Stockholm, Sweden.
- Department of Neuroradiology, Karolinska University Hospital, Stockholm, Sweden.
- Department of Radiology and Imaging Sciences, Emory University, Atlanta, GA, United States of America.
- Department of Biomedical Engineering, Georgia Institute of Technology, Atlanta, GA, United States of America.
ORCID-Profile: Ruihan Huang, Dennis Hein, Staffan Holmin, Thomas Wesley Holmes, Mats U Persson
Förderung
- The Swedish Research Council
- Göran Gustafssons Stiftelse för Naturvetenskaplig och Medicinsk Forskning
- China Scholarship Council
Zitieren
APA
Huang, R., Larsson, K., Hein, D., Holmin, S., Holmes, T. W., Pourmorteza, A., & Persson, M. U. (2026). Deep-learning-based spectral motion artifact correction on photon-counting cardiac CT images. Physics in Medicine & Biology. https://doi.org/10.1088/1361-6560/ae3eee
Vancouver
Huang R, Larsson K, Hein D, Holmin S, Holmes TW, Pourmorteza A, et al. Deep-learning-based spectral motion artifact correction on photon-counting cardiac CT images. Physics in Medicine & Biology. 2026. doi:10.1088/1361-6560/ae3eee
BibTeX
@article{Huang2026,
author = {Huang, Ruihan and Larsson, Karin and Hein, Dennis and Holmin, Staffan and Holmes, Thomas Wesley and Pourmorteza, Amir and Persson, Mats U},
title = {Deep-learning-based spectral motion artifact correction on photon-counting cardiac CT images},
journal = {Physics in Medicine & Biology},
year = {2026},
doi = {10.1088/1361-6560/ae3eee}
}
EUVIMED