Fully Automatic Atrial Fibrosis Assessment Using a Multilabel Convolutional Neural Network

Overview

Journal Circ Cardiovasc Imaging

Specialties Cardiology & Vascular Diseases
Radiology

Date 2020 Dec 15

PMID 33317334

Citations 12

Authors

Orod Razeghi

Iain Sim

Caroline H Roney

Rashed Karim

Henry Chubb

John Whitaker

Louisa ONeill

Rahul Mukherjee

Matthew Wright

Mark ONeill

Steven E Williams

Steven Niederer

Affiliations

Soon will be listed here.

Abstract

Background: Pathological atrial fibrosis is a major contributor to sustained atrial fibrillation. Currently, late gadolinium enhancement (LGE) scans provide the only noninvasive estimate of atrial fibrosis. However, widespread adoption of atrial LGE has been hindered partly by nonstandardized image processing techniques, which can be operator and algorithm dependent. Minimal validation and limited access to transparent software platforms have also exacerbated the problem. This study aims to estimate atrial fibrosis from cardiac magnetic resonance scans using a reproducible operator-independent fully automatic open-source end-to-end pipeline.

Methods: A multilabel convolutional neural network was designed to accurately delineate atrial structures including the blood pool, pulmonary veins, and mitral valve. The output from the network removed the operator dependent steps in a reproducible pipeline and allowed for automated estimation of atrial fibrosis from LGE-cardiac magnetic resonance scans. The pipeline results were compared against manual fibrosis burdens, calculated using published thresholds: image intensity ratio 0.97, image intensity ratio 1.61, and mean blood pool signal +3.3 SD.

Results: We validated our methods on a large 3-dimensional LGE-cardiac magnetic resonance data set from 207 labeled scans. Automatic atrial segmentation achieved a 91% Dice score, compared with the mutual agreement of 85% in Dice seen in the interobserver analysis of operators. Intraclass correlation coefficients of the automatic pipeline with manually generated results were excellent and better than or equal to interobserver correlations for all 3 thresholds: 0.94 versus 0.88, 0.99 versus 0.99, 0.99 versus 0.96 for image intensity ratio 0.97, image intensity ratio 1.61, and +3.3 SD thresholds, respectively. Automatic analysis required 3 minutes per case on a standard workstation. The network and the analysis software are publicly available.

Conclusions: Our pipeline provides a fully automatic estimation of fibrosis burden from LGE-cardiac magnetic resonance scans that is comparable to manual analysis. This removes one key source of variability in the measurement of atrial fibrosis.

Citing Articles

The Use of Artificial Intelligence for Detecting and Predicting Atrial Arrhythmias Post Catheter Ablation.

Lallah P, Laite C, Bangash A, Chooah O, Jiang C Rev Cardiovasc Med. 2024; 24(8):215.

PMID: 39076714 PMC: 11266764. DOI: 10.31083/j.rcm2408215.

Novel Domain Knowledge-Encoding Algorithm Enables Label-Efficient Deep Learning for Cardiac CT Segmentation to Guide Atrial Fibrillation Treatment in a Pilot Dataset.

Ganesan P, Feng R, Deb B, Tjong F, Rogers A, Ruiperez-Campillo S Diagnostics (Basel). 2024; 14(14).

PMID: 39061675 PMC: 11276420. DOI: 10.3390/diagnostics14141538.

FK-means: automatic atrial fibrosis segmentation using fractal-guided K-means clustering with Voronoi-clipping feature extraction of anatomical structures.

Firouznia M, Henningsson M, Carlhall C Interface Focus. 2023; 13(6):20230033.

PMID: 38106915 PMC: 10722213. DOI: 10.1098/rsfs.2023.0033.

Segmenting computed tomograms for cardiac ablation using machine learning leveraged by domain knowledge encoding.

Feng R, Deb B, Ganesan P, Tjong F, Rogers A, Ruiperez-Campillo S Front Cardiovasc Med. 2023; 10:1189293.

PMID: 37849936 PMC: 10577270. DOI: 10.3389/fcvm.2023.1189293.

Artificial Intelligence in the Image-Guided Care of Atrial Fibrillation.

Lyu Y, Bennamoun M, Sharif N, Lip G, Dwivedi G Life (Basel). 2023; 13(9).

PMID: 37763273 PMC: 10532509. DOI: 10.3390/life13091870.

References

Whitaker J, Rajani R, Chubb H, Gabrawi M, Varela M, Wright M . The role of myocardial wall thickness in atrial arrhythmogenesis. Europace. 2016; 18(12):1758-1772. PMC: 5841556. DOI: 10.1093/europace/euw014. View

Harrison J, Jensen H, Peel S, Chiribiri A, Grondal A, Bloch L . Cardiac magnetic resonance and electroanatomical mapping of acute and chronic atrial ablation injury: a histological validation study. Eur Heart J. 2014; 35(22):1486-95. PMC: 4048535. DOI: 10.1093/eurheartj/eht560. View

Quail M, Grunseich K, Baldassarre L, Mojibian H, Marieb M, Cornfeld D . Prognostic and functional implications of left atrial late gadolinium enhancement cardiovascular magnetic resonance. J Cardiovasc Magn Reson. 2019; 21(1):2. PMC: 6317232. DOI: 10.1186/s12968-018-0514-3. View

Hwang S, Oh Y, Lee D, Shim J, Park S, Kim Y . Evaluation of quantification methods for left arial late gadolinium enhancement based on different references in patients with atrial fibrillation. Int J Cardiovasc Imaging. 2014; 31 Suppl 1:91-101. DOI: 10.1007/s10554-014-0563-0. View

Marrouche N, Wilber D, Hindricks G, Jais P, Akoum N, Marchlinski F . Association of atrial tissue fibrosis identified by delayed enhancement MRI and atrial fibrillation catheter ablation: the DECAAF study. JAMA. 2014; 311(5):498-506. DOI: 10.1001/jama.2014.3. View

Kostin S, Klein G, Szalay Z, Hein S, Bauer E, Schaper J . Structural correlate of atrial fibrillation in human patients. Cardiovasc Res. 2002; 54(2):361-79. DOI: 10.1016/s0008-6363(02)00273-0. View

King J, Azadani P, Suksaranjit P, Bress A, Witt D, Han F . Left Atrial Fibrosis and Risk of Cerebrovascular and Cardiovascular Events in Patients With Atrial Fibrillation. J Am Coll Cardiol. 2017; 70(11):1311-1321. DOI: 10.1016/j.jacc.2017.07.758. View

Platonov P, Ivanov V, Ho S, Mitrofanova L . Left atrial posterior wall thickness in patients with and without atrial fibrillation: data from 298 consecutive autopsies. J Cardiovasc Electrophysiol. 2008; 19(7):689-92. DOI: 10.1111/j.1540-8167.2008.01102.x. View

Wong G, Nalliah C, Lee G, Voskoboinik A, Prabhu S, Parameswaran R . Dynamic Atrial Substrate During High-Density Mapping of Paroxysmal and Persistent AF: Implications for Substrate Ablation. JACC Clin Electrophysiol. 2019; 5(11):1265-1277. DOI: 10.1016/j.jacep.2019.06.002. View

10.

Benito E, Carlosena-Remirez A, Guasch E, Prat-Gonzalez S, Perea R, Figueras R . Left atrial fibrosis quantification by late gadolinium-enhanced magnetic resonance: a new method to standardize the thresholds for reproducibility. Europace. 2016; 19(8):1272-1279. DOI: 10.1093/europace/euw219. View

11.

Boldt A, Wetzel U, Lauschke J, Weigl J, Gummert J, Hindricks G . Fibrosis in left atrial tissue of patients with atrial fibrillation with and without underlying mitral valve disease. Heart. 2004; 90(4):400-5. PMC: 1768173. DOI: 10.1136/hrt.2003.015347. View

12.

Tao Q, Ipek E, Shahzad R, Berendsen F, Nazarian S, van der Geest R . Fully automatic segmentation of left atrium and pulmonary veins in late gadolinium-enhanced MRI: Towards objective atrial scar assessment. J Magn Reson Imaging. 2016; 44(2):346-54. DOI: 10.1002/jmri.25148. View

13.

January C, Samuel Wann L, Alpert J, Calkins H, Cigarroa J, Cleveland Jr J . 2014 AHA/ACC/HRS guideline for the management of patients with atrial fibrillation: executive summary: a report of the American College of Cardiology/American Heart Association Task Force on practice guidelines and the Heart Rhythm Society. Circulation. 2014; 130(23):2071-104. DOI: 10.1161/CIR.0000000000000040. View

14.

Trayanova N . From genetics to smart watches: developments in precision cardiology. Nat Rev Cardiol. 2018; 16(2):72-73. PMC: 6550458. DOI: 10.1038/s41569-018-0149-y. View

15.

Piccinelli M, Veneziani A, Steinman D, Remuzzi A, Antiga L . A framework for geometric analysis of vascular structures: application to cerebral aneurysms. IEEE Trans Med Imaging. 2009; 28(8):1141-55. DOI: 10.1109/TMI.2009.2021652. View

16.

Chubb H, Karim R, Mukherjee R, Williams S, Whitaker J, Harrison J . A comprehensive multi-index cardiac magnetic resonance-guided assessment of atrial fibrillation substrate prior to ablation: Prediction of long-term outcomes. J Cardiovasc Electrophysiol. 2019; 30(10):1894-1903. DOI: 10.1111/jce.14111. View

17.

Ho S, Cabrera J, Tran V, Farre J, Anderson R, Sanchez-Quintana D . Architecture of the pulmonary veins: relevance to radiofrequency ablation. Heart. 2001; 86(3):265-70. PMC: 1729909. DOI: 10.1136/heart.86.3.265. View

18.

Platonov P, Mitrofanova L, Orshanskaya V, Ho S . Structural abnormalities in atrial walls are associated with presence and persistency of atrial fibrillation but not with age. J Am Coll Cardiol. 2011; 58(21):2225-32. DOI: 10.1016/j.jacc.2011.05.061. View

19.

Khurram I, Beinart R, Zipunnikov V, Dewire J, Yarmohammadi H, Sasaki T . Magnetic resonance image intensity ratio, a normalized measure to enable interpatient comparability of left atrial fibrosis. Heart Rhythm. 2013; 11(1):85-92. PMC: 3900295. DOI: 10.1016/j.hrthm.2013.10.007. View

20.

Sim I, Razeghi O, Karim R, Chubb H, Whitaker J, ONeill L . Reproducibility of Atrial Fibrosis Assessment Using CMR Imaging and an Open Source Platform. JACC Cardiovasc Imaging. 2019; 12(10):2076-2077. DOI: 10.1016/j.jcmg.2019.03.027. View