scholarly journals Convolutional Neural Network for the Detection of End-Diastole and End-Systole Frames in Free-Breathing Cardiac Magnetic Resonance Imaging

2017 ◽  
Vol 2017 ◽  
pp. 1-10 ◽  
Author(s):  
Fan Yang ◽  
Yan He ◽  
Mubashir Hussain ◽  
Hong Xie ◽  
Pinggui Lei
Keyword(s):  
Neural Network ◽  
Cardiac Magnetic Resonance ◽  
Magnetic Resonance ◽  
Convolutional Neural Network ◽  
Free Breathing ◽  
Automatic Identification ◽  
Motion Signal ◽  
Examination Time ◽  
High Degree ◽  
Cmr Imaging

Free-breathing cardiac magnetic resonance (CMR) imaging has short examination time with high reproducibility. Detection of the end-diastole and the end-systole frames of the free-breathing cardiac magnetic resonance, supplemented by visual identification, is time consuming and laborious. We propose a novel method for automatic identification of both the end-diastole and the end-systole frames, in the free-breathing CMR imaging. The proposed technique utilizes the convolutional neural network to locate the left ventricle and to obtain the end-diastole and the end-systole frames from the respiratory motion signal. The proposed procedure works successfully on our free-breathing CMR data, and the results demonstrate a high degree of accuracy and stability. Convolutional neural network improves the postprocessing efficiency greatly and facilitates the clinical application of the free-breathing CMR imaging.

scholarly journals Fully Automatic Atrial Fibrosis Assessment Using a Multilabel Convolutional Neural Network

2020 ◽  
Vol 13 (12) ◽  
Author(s):  
Orod Razeghi ◽  
Iain Sim ◽  
Caroline H. Roney ◽  
Rashed Karim ◽  
Henry Chubb ◽  
...  
Keyword(s):  
Neural Network ◽  
Cardiac Magnetic Resonance ◽  
Magnetic Resonance ◽  
Convolutional Neural Network ◽  
Intensity Ratio ◽  
Blood Pool ◽  
Atrial Fibrosis ◽  
Data Set ◽  
Image Intensity ◽  
Fully Automatic

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.

scholarly journals Incremental value of feature-tracking cardiac magnetic resonance imaging in detecting myocardial dysfunction in patients with idiopathic ventricular arrhythmias

2021 ◽  
Vol 22 (Supplement_1) ◽  
Author(s):  
C Nikolaidou ◽  
C Kotanidis ◽  
J Leal-Pelado ◽  
K Kouskouras ◽  
VP Vassilikos ◽  
...  
Keyword(s):  
Cardiac Magnetic Resonance ◽  
Magnetic Resonance ◽  
Ventricular Arrhythmias ◽  
Feature Tracking ◽  
Myocardial Dysfunction ◽  
Global Longitudinal Strain ◽  
Left Ventricular ◽  
Volume Index ◽  
Control Subjects ◽  
Cmr Imaging

Abstract Funding Acknowledgements Type of funding sources: None. Introduction Cardiac magnetic resonance (CMR) imaging can identify the underlying substrate in patients with ventricular arrhythmias (VAs) and normal echocardiography. Myocardial strain has emerged as a superior index of systolic performance compared to ejection fraction (EF), with an incremental prognostic value in many cardiac diseases. Purpose To assess myocardial deformation using 2-D feature-tracking CMR strain imaging (CMR-FT) in patients with frequent VAs (≥500 ventricular premature contractions (VPC)/24 hours; and/or non-sustained ventricular tachycardia), and structurally normal hearts on echocardiography without evidence of coronary artery disease. Methods Sixty-eight consecutive patients (mean age 46 ± 16 years; 54% female) and 72 healthy controls matched for age and body surface area were included in the study. CMR imaging was performed on a 1.5T Magnetom Avanto (Siemens, Erlangen, Germany) scanner using a standard cardiac protocol. Results CMR showed normal findings in 30 patients (44%), while 16 (24%) had previous myocarditis, 6 (9%) had a diagnosis of non-ischaemic cardiomyopathy (NICM), 15 (22%) were diagnosed with VPC-related cardiomyopathy, and 1 patient had subendocardial infarction [excluded from strain analysis]. Mean left ventricular EF (LVEF) in patients was 62% ± 6% and right ventricular EF 64% ± 6% (vs. 65% ± 3% and 66% ± 4% in controls, respectively). Compared to control subjects, patients with VAs had impaired peak LV global radial strain (GRS) (28.88% [IQR: 25.87% to 33.97%] vs. 36.65% [IQR:33.19% to 40.2%], p < 0.001) and global circumferential strain (GCS) (-17.73% [IQR: -19.8% to -16.33%] vs. -20.66% [IQR: -21.72% to -19.6%], p < 0.001, Panel A). Peak LV GRS could differentiate patients with previous myocarditis from patients with NICM and those with VPC-related cardiomyopathy (Panel B). Peak LV GCS could differentiate patients with previous myocarditis from patients with NICM (Panel C). Peak LV GRS showed excellent diagnostic accuracy in detecting patients from control subjects (Panel D). In a multivariable regression model, subjects with a low GRS (<29.91%-determined by the Youden’s index) had 5-fold higher odds of having VAs (OR:4.99 [95%CI: 1.2-21.95]), after adjusting for LVEF, LV end-diastolic volume index, age, sex, BMI, smoking, hypertension, and dyslipidaemia. Peak LV global longitudinal strain (GLS) and RV strain indices were not statistically different between patients and controls. Conclusion Peak LV GRS and GCS are impaired in patients with frequent idiopathic VAs and can detect myocardial contractile dysfunction in patients with different underlying substrates. Our findings suggest that LV strain indices on CMR-FT constitute independent markers of myocardial dysfunction on top and independently of EF. Abstract Figure.

scholarly journals Identification and prognostication of the substrate of ventricular arrhythmia with cardiac magnetic resonance (CMR) imaging in patients with normal echocardiography

2021 ◽  
Vol 22 (Supplement_1) ◽  
Author(s):  
S Younus ◽  
H Maqsood ◽  
A Gulraiz ◽  
MD Khan ◽  
R Awais
Keyword(s):  
Heart Disease ◽  
Ventricular Tachycardia ◽  
Cardiac Magnetic Resonance ◽  
Chest Pain ◽  
Magnetic Resonance ◽  
Ventricular Arrhythmia ◽  
Heart Diseases ◽  
Sustained Ventricular Tachycardia ◽  
Structural Heart Diseases ◽  
Cmr Imaging

Abstract Funding Acknowledgements Type of funding sources: Other. Main funding source(s): Self Introduction Malignant ventricular arrhythmia contributes to approximately half of the sudden cardiac deaths. In common practice, echocardiography is used to identify structural heart diseases that are the most frequent substrate of VA. Identification and prognostication of structural heart diseases are very important as they are the main determinant of poor prognosis of ventricular arrhythmia. Purpose : The objective of this study is to determine whether cardiac magnetic resonance (CMR) may identify structural heart disease (SHD) in patients with ventricular arrhythmia who had no pathology observed on echocardiography. Methods : A total of 864 consecutive patients were enrolled in this single-center prospective study with significant ventricular arrhythmia. VA was characterized as >1000 ventricular ectopic beats per 24 hours, non-sustained ventricular arrhythmia, sustained ventricular arrhythmia, and no pathological lesion on echocardiography. The primary endpoint was the detection of SHD with CMR. Secondary endpoints were a composite of CMR detection of SHD and abnormal findings not specific for a definite SHD diagnosis. Results : CMR studies were used to diagnose SHD in 212 patients (24.5%) and abnormal findings not specific for a definite SHD diagnosis in 153 patients (17.7%). Myocarditis (n = 84) was the more frequent disease, followed by arrhythmogenic cardiomyopathy (n = 51), ischemic heart disease (n = 32), dilated cardiomyopathy (n = 17), hypertrophic cardiomyopathy (n = 12), congenital cardiac disease (n = 08), left ventricle noncompaction (n = 5), and pericarditis (n = 3). The strongest univariate and multivariate predictors of SHD on CMR images were chest pain (odds ratios [OR]: 2.5 and 2.33, respectively) and sustained ventricular tachycardia (ORs: 2.62 and 2.21, respectively). Conclusion : Our study concludes that SHD was able to be identified on CMR imaging in a significant number of patients with malignant VA and completely normal echocardiography. Chest pain and sustained ventricular tachycardia were the two strongest predictors of positive CMR imaging results. Abstract Figure. Distribution of different SHD

Sign in / Sign up

Export Citation Format

Share Document