EUVIMEDEUVIMED entdecken
Open Access · CC BY-NC-ND 4.0

Pretrained subtraction and segmentation model for coronary angiograms

Yunjie Zeng, Han Liu, Juan Hu, Zhengbo Zhao, Qiang She

Scientific Reports · 2024

DOI: 10.1038/s41598-024-71063-5 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Scientific Reports, 2024. Verfasst von Yunjie Zeng und 4 weiteren Autor:innen. Beteiligte Einrichtung: Department of Cardiology, The Second Affiliated Hospital of Chongqing Medical University, Chongqing, 400010, China. Hauptthemen laut MEDLINE: Coronary Angiography, Coronary Vessels. Die Arbeit wurde bisher 7-mal zitiert (Crossref) und nennt 41 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-ND 4.0). Gefördert von: the Key project of Technology Innovation and Application Development in Chongqing, the Scientific and Technological Research Program of Chongqing Municipal Education Commission, the Chongqing Natural Science Foundation.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

This study introduces a novel self-supervised learning method for single-frame subtraction and vessel segmentation in coronary angiography, addressing the scarcity of annotated medical samples in AI applications. We pretrain a U-Net model on a large dataset of unannotated coronary angiograms using an image-to-image translation framework, then fine-tune it on a limited set of manually annotated samples. The pretrained model excels at comprehensive single-frame subtraction, outperforming existing DSA methods. Fine-tuning with just 40 samples yields a Dice coefficient of 0.828 for vessel segmentation. On the public XCAD dataset, our model sets a new state-of-the-art benchmark with a Dice coefficient of 0.755, surpassing both unsupervised and supervised learning approaches. This method achieves robust single-frame subtraction and demonstrates that combining pretraining with minimal fine-tuning enables accurate coronary vessel segmentation with limited manual annotations. We successfully apply this approach to assist physicians in visualizing potential vascular stenosis sites during coronary angiography. Code, dataset, and a live demo will be available available at: https://github.com/newfyu/DeepSA .

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Coronary AngiographyCoronary VesselsHumansImage Processing, Computer-AssistedAlgorithmsAngiography, Digital Subtraction

Stichwörter der Autor:innen

Coronary angiogramDeep learningDigital subtraction angiographyVessel segmentation

Autor:innen und Institutionen

  • Department of Cardiology, The Second Affiliated Hospital of Chongqing Medical University, Chongqing, 400010, China.
  • Department of Cardiology, The Affiliated Dazu's Hospital of Chongqing Medical University, Chongqing, 402360, China.
  • Department of Neurology, Jiulongpo District People's Hospital, Chongqing, 400050, China.
  • The First Affiliated Hospital of Chongqing Medical and Pharmaceutical College, Chongqing, 400060, China.
  • Department of Cardiology, The Second Affiliated Hospital of Chongqing Medical University, Chongqing, 400010, China. qshe98@cqmu.edu.cn.

Förderung

  • the Key project of Technology Innovation and Application Development in Chongqing
  • the Scientific and Technological Research Program of Chongqing Municipal Education Commission
  • the Chongqing Natural Science Foundation

Zitieren

APA

Zeng, Y., Liu, H., Hu, J., Zhao, Z., & She, Q. (2024). Pretrained subtraction and segmentation model for coronary angiograms. Scientific Reports. https://doi.org/10.1038/s41598-024-71063-5

Vancouver

Zeng Y, Liu H, Hu J, Zhao Z, She Q. Pretrained subtraction and segmentation model for coronary angiograms. Scientific Reports. 2024. doi:10.1038/s41598-024-71063-5

BibTeX
@article{Zeng2024,
  author = {Zeng, Yunjie and Liu, Han and Hu, Juan and Zhao, Zhengbo and She, Qiang},
  title = {Pretrained subtraction and segmentation model for coronary angiograms},
  journal = {Scientific Reports},
  year = {2024},
  doi = {10.1038/s41598-024-71063-5}
}
Kostenlos starten →