Automated stenosis estimation of coronary angiographies using end-to-end learning

被引:0
作者
Eschen, Christian Kim [1 ]
Banasik, Karina [1 ]
Dahl, Anders Bjorholm [2 ]
Chmura, Piotr Jaroslaw [1 ]
Bruun-Rasmussen, Peter [3 ]
Pedersen, Frants [4 ,5 ]
Kober, Lars [5 ,6 ]
Engstrom, Thomas [4 ,5 ]
Bottcher, Morten [7 ,8 ]
Winther, Simon [7 ,8 ]
Christensen, Alex Horby [4 ,5 ,6 ]
Bundgaard, Henning [4 ,5 ]
Brunak, Soren [1 ]
机构
[1] Univ Copenhagen, Novo Nord Fdn Ctr Prot Res, Fac Hlth & Med Sci, Copenhagen, Denmark
[2] Tech Univ Denmark, Dept Appl Math & Comp Sci, Sect Visual Comp, Lyngby, Denmark
[3] Univ Copenhagen, Fac Hlth & Med Sci, Dept Clin Immunol, Rigshosp, Copenhagen, Denmark
[4] Univ Copenhagen, Fac Hlth & Med Sci, Heart Ctr, Dept Cardiol,Rigshosp, Copenhagen, Denmark
[5] Univ Copenhagen, Fac Hlth & Med Sci, Dept Clin Med, Copenhagen, Denmark
[6] Univ Copenhagen, Herlev Gentofte Hosp, Fac Hlth & Med Sci, Dept Cardiol, Hellerup, Denmark
[7] Godstrup Hosp, Dept Cardiol, Herning, Denmark
[8] Aarhus Univ, Inst Clin Med, Aarhus, Denmark
关键词
Coronary angiography; Coronary artery disease; Ischemic heart disease; Deep learning; Quantitative coronary angiography; Myocardial infarction; FRACTIONAL FLOW RESERVE; MEDICAL THERAPY; PCI; REVASCULARIZATION;
D O I
10.1007/s10554-025-03324-x
中图分类号
R5 [内科学];
学科分类号
1002 ; 100201 ;
摘要
The initial evaluation of stenosis during coronary angiography is typically performed by visual assessment. Visual assessment has limited accuracy compared to fractional flow reserve and quantitative coronary angiography, which are more time-consuming and costly. Applying deep learning might yield a faster and more accurate stenosis assessment. We developed a deep learning model to classify cine loops into left or right coronary artery (LCA/RCA) or "other". Data were obtained by manual annotation. Using these classifications, cine loops before revascularization were identified and curated automatically. Separate deep learning models for LCA and RCA were developed to estimate stenosis using these identified cine loops. From a cohort of 19,414 patients and 332,582 cine loops, we identified cine loops for 13,480 patients for model development and 5056 for internal testing. External testing was conducted using automated identified cine loops from 608 patients. For identification of significant stenosis (visual assessment of diameter stenosis > 70%), our model obtained a receiver operator characteristic (ROC) area under the curve (ROC-AUC) of 0.903 (95% CI: 0.900-0.906) on the internal test. The performance was evaluated on the external test set against visual assessment, 3D quantitative coronary angiography, and fractional flow reserve (<= 0.80), obtaining ROC AUC values of 0.833 (95% CI: 0.814-0.852), 0.798 (95% CI: 0.741-0.842), and 0.780 (95% CI: 0.743-0.817), respectively. The deep-learning-based stenosis estimation models showed promising results for predicting stenosis. Compared to previous work, our approach demonstrates performance increase, includes all 16 segments, does not exclude revascularized patients, is externally tested, and is simpler using fewer steps.
引用
收藏
页码:441 / 452
页数:12
相关论文
共 34 条
[1]  
Austen W G, 1975, Circulation, V51, P5
[2]   CathAI: fully automated coronary angiography interpretation and stenosis estimation [J].
Avram, Robert ;
Olgin, Jeffrey E. ;
Ahmed, Zeeshan ;
Verreault-Julien, Louis ;
Wan, Alvin ;
Barrios, Joshua ;
Abreau, Sean ;
Wan, Derek ;
Gonzalez, Joseph E. ;
Tardif, Jean-Claude ;
So, Derek Y. ;
Soni, Krishan ;
Tison, Geoffrey H. .
NPJ DIGITAL MEDICINE, 2023, 6 (01)
[3]  
Cardoso M Jorge, 2022, ARXIV
[4]   Quo Vadis, Action Recognition? A New Model and the Kinetics Dataset [J].
Carreira, Joao ;
Zisserman, Andrew .
30TH IEEE CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR 2017), 2017, :4724-4733
[5]   State of the art: coronary angiography [J].
Collet, Carlos ;
Grundeken, Maik J. ;
Asano, Taku ;
Onuma, Yoshinobu ;
Wijns, William ;
Serruys, Patrick W. .
EUROINTERVENTION, 2017, 13 (06) :634-643
[6]   Real-time coronary artery stenosis detection based on modern neural networks [J].
Danilov, Viacheslav V. ;
Klyshnikov, Kirill Yu ;
Gerget, Olga M. ;
Kutikhin, Anton G. ;
Ganyukov, Vladimir, I ;
Frangi, Alejandro F. ;
Ovcharenko, Evgeny A. .
SCIENTIFIC REPORTS, 2021, 11 (01)
[7]   Fractional Flow Reserve-Guided PCI versus Medical Therapy in Stable Coronary Disease [J].
De Bruyne, Bernard ;
Pijls, Nico H. J. ;
Kalesan, Bindu ;
Barbato, Emanuele ;
Tonino, Pim A. L. ;
Piroth, Zsolt ;
Jagic, Nikola ;
Mobius-Winckler, Sven ;
Rioufol, Gilles ;
Witt, Nils ;
Kala, Petr ;
MacCarthy, Philip ;
Engstrom, Thomas ;
Oldroyd, Keith G. ;
Mavromatis, Kreton ;
Manoharan, Ganesh ;
Verlee, Peter ;
Frobert, Ole ;
Curzen, Nick ;
Johnson, Jane B. ;
Jueni, Peter ;
Fearon, William F. .
NEW ENGLAND JOURNAL OF MEDICINE, 2012, 367 (11) :991-1001
[8]   Training and validation of a deep learning architecture for the automatic analysis of coronary angiography [J].
Du, Tianming ;
Xie, Lihua ;
Zhang, Honggang ;
Liu, Xuqing ;
Wang, Xiaofei ;
Chen, Donghao ;
Xu, Yang ;
Sun, Zhongwei ;
Zhou, Wenhui ;
Song, Lei ;
Guan, Changdong ;
Lansky, Alexandra J. ;
Xu, Bo .
EUROINTERVENTION, 2021, 17 (01) :32-+
[9]   Classification of Left and Right Coronary Arteries in Coronary Angiographies Using Deep Learning [J].
Eschen, Christian Kim ;
Banasik, Karina ;
Christensen, Alex Horby ;
Chmura, Piotr Jaroslaw ;
Pedersen, Frants ;
Kober, Lars ;
Engstrom, Thomas ;
Dahl, Anders Bjorholm ;
Brunak, Soren ;
Bundgaard, Henning .
ELECTRONICS, 2022, 11 (13)
[10]   Comparison of interventional cardiology in two European countries: A nationwide internet based registry study [J].
Gudnason, T. ;
Gudnadottir, G. S. ;
Lagerqvist, B. ;
Eyjolfsson, K. ;
Nilsson, T. ;
Thorgeirsson, G. ;
Thorgeirsson, G. ;
Andersen, K. ;
James, S. .
INTERNATIONAL JOURNAL OF CARDIOLOGY, 2013, 168 (02) :1237-1242