AngioPy Segmentation: An open-source, user-guided deep learning tool for coronary artery segmentation.

Autor: Mahendiran T; Department of Cardiology, Lausanne University Hospital, Lausanne, Switzerland; Mathematical Data Science, EPFL, Lausanne, Switzerland., Thanou D; Mathematical Data Science, EPFL, Lausanne, Switzerland., Senouf O; Mathematical Data Science, EPFL, Lausanne, Switzerland., Jamaa Y; Center for Imaging, EPFL, Lausanne, Switzerland., Fournier S; Department of Cardiology, Lausanne University Hospital, Lausanne, Switzerland., De Bruyne B; Department of Cardiology, Lausanne University Hospital, Lausanne, Switzerland; Department of Cardiology, OLV Cardiovascular Center, Aalst, Belgium., Abbé E; Mathematical Data Science, EPFL, Lausanne, Switzerland., Muller O; Department of Cardiology, Lausanne University Hospital, Lausanne, Switzerland., Andò E; Center for Imaging, EPFL, Lausanne, Switzerland.
Jazyk: angličtina
Zdroj: International journal of cardiology [Int J Cardiol] 2024 Sep 26; Vol. 418, pp. 132598. Date of Electronic Publication: 2024 Sep 26.
DOI: 10.1016/j.ijcard.2024.132598
Abstrakt: Background: Quantitative coronary angiography (QCA) typically employs traditional edge detection algorithms that often require manual correction. This has important implications for the accuracy of downstream 3D coronary reconstructions and computed haemodynamic indices (e.g. angiography-derived fractional flow reserve). We developed AngioPy, a deep-learning model for coronary segmentation that employs user-defined ground-truth points to boost performance and minimise manual correction. We compared its performance without correction with an established QCA system.
Methods: Deep learning models integrating user-defined ground-truth points were developed using 2455 images from the Fractional Flow Reserve versus Angiography for Multivessel Evaluation 2 (FAME 2) study. External validation was performed on a dataset of 580 images. Vessel dimensions from 203 images with mild/moderate stenoses segmented by AngioPy (without correction) and an established QCA system (Medis QFR®) were compared (609 diameters).
Results: The top-performing model had an average F1 score of 0.927 (pixel accuracy 0.998, precision 0.925, sensitivity 0.930, specificity 0.999) with 99.2 % of masks exhibiting an F1 score > 0.8. Similar results were seen with external validation (F1 score 0.924, pixel accuracy 0.997, precision 0.921, sensitivity 0.929, specificity 0.999). Vessel dimensions from AngioPy exhibited excellent agreement with QCA (r = 0.96 [95 % CI 0.95-0.96], p < 0.001; mean difference - 0.18 mm [limits of agreement (LOA): -0.84 to 0.49]), including the minimal luminal diameter (r = 0.93 [95 % CI 0.91-0.95], p < 0.001; mean difference - 0.06 mm [LOA: -0.70 to 0.59]).
Conclusion: AngioPy, an open-source tool, performs rapid and accurate coronary segmentation without the need for manual correction. It has the potential to increase the accuracy and efficiency of QCA.
Competing Interests: Declaration of competing interest BDB reports receiving consultancy fees from Boston Scientific and Abbott Vascular, research grants from Coroventis Research, Pie Medical Imaging, CathWorks, Boston Scientific, Siemens, HeartFlow Inc., and Abbott Vascular, and owning equity in Siemens, GE, Philips, HeartFlow Inc., Edwards Life Sciences, Bayer, Sanofi, Celyad. All the other authors have nothing to disclose. The following are the supplementary data related to this article. Supplementary data to this article can be found online at https://doi.org/10.1016/j.ijcard.2024.132598.
(Copyright © 2024 The Author(s). Published by Elsevier B.V. All rights reserved.)
Databáze: MEDLINE
Externí odkaz: