Abstract:Introduction Myocardial scar or fibrosis has significant implications in patient management and prognosis. Currently, cardiovascular magnetic resonance (CMR) is the reference technique for detection and quantification of scar though other imaging modalities have been proven to be useful as well. Purpose This study aims to investigate the usefulness of 3D strain echocardiography to identify myocardial scar in patients with ischaemic and non-ischaemic myocardial fibrosis using late gadolinium enhancement (LGE) CMR as a reference. Methods 149 individuals with ischaemic (89), non-ischaemic scar (26) and no scar (34) on LGE CMR underwent 3D transthoracic echocardiography. The 3D data sets were processed with the 4D LV analysis tool by TOMTEC Imaging Systems GmbH, Germany. The longitudinal, circumferential, radial, and principal tangential strain were calculated. Using a 16 LV-segment model the areas of LGE were assessed on CMR images, and a correlation between 3D strain values and LGE was investigated. Results A multivariate binary logistic regression model was built with the conditional backward elimination method using LGE as the dependent variable. The predictors made a significant contribution to the model (Wald test: 49.985, p<0.001). The Cox & Snell R-square was 0.141, and Nagelkerke R-square was 0.232, suggesting that the final model explains 14.1% or 23.2% of LGE variability. By logistic regression analysis, less negative peak principal tangential strain (P-PTS), peak circumferential strain (P-SC), and end-systolic circumferential strain (ES-CS) increased the odds, whereas increased peak radial strain (P-RS) decreased the likelihood of LGE (Table 1). The LGE extent was calculated as the average LGE transmurality of all 16 segments and was used as the dependent variable in linear regression models. The LGE extent was significantly correlated with all strain indices (Table 2). By stepwise linear regression analysis, P-PTS (B=0.029, 95%CI:0.019-0.040, P<0.001) outperformed the remaining strain variables in terms of linear association with LGE extent. Association between strain indices and the presence or absence of myocardial fibrosis in each of the 16 segments was tested by point biserial correlation (Table 3). PTS and CS indices were associated with myocardial fibrosis in 10/16 segments; RS indices were associated with LGE in 9/16 segments, and LS was not related to the presence of LGE in 15/16 segments. Septal (basal-mid-apical) and mid/apical anterior and inferior segmental strain showed more consistent associations with corresponding myocardial fibrosis. Conclusion 3D strain echocardiography is a promising and easily accessible technique, that can be used to identify areas of myocardial scar as it correlates well with LGE in many LV segments. Echocardiography is a bedside and widely available technique, and it could be used in patients with contraindications or no access to CMR or myocardial perfusion scan.Table 3

A deep-learning approach for myocardial fibrosis detection in early contrast-enhanced cardiac CT images

Identification of Fibrosis in Hypertrophic Cardiomyopathy: a Radiomic Study on Cardiac Magnetic Resonance Cine Imaging.

Anatomically informed deep learning on contrast-enhanced cardiac magnetic resonance imaging for scar segmentation and clinical feature extraction

Anatomically-Informed Deep Learning on Contrast-Enhanced Cardiac MRI for Scar Segmentation and Clinical Feature Extraction

Deep learning of echocardiography distinguishes between presence and absence of late gadolinium enhancement on cardiac magnetic resonance in patients with hypertrophic cardiomyopathy

Strong Diagnostic Performance of Single Energy 256-row Multidetector Computed Tomography with Deep Learning Image Reconstruction in the Assessment of Myocardial Fibrosis

Deep learning to diagnose cardiac amyloidosis from cardiovascular magnetic resonance

Cardiac Fibrosis Automated Diagnosis Based on FibrosisNet Network Using CMR Ischemic Cardiomyopathy

Deep Learning for Diagnosis of Chronic Myocardial Infarction on Nonenhanced Cardiac Cine MRI.

A Radiomic Approach to Predict Myocardial Fibrosis on Coronary CT Angiography in Hypertrophic Cardiomyopathy.

Computer-Assisted Algorithm for Quantification of Fibrosis by Native Cardiac CT: A Pilot Study

Joint Deep Learning for Improved Myocardial Scar Detection from Cardiac MRI

Machine Learning-Based Segmentation of Left Ventricular Myocardial Fibrosis from Magnetic Resonance Imaging

Recent Advances in Fibrosis and Scar Segmentation From Cardiac MRI: A State-of-the-Art Review and Future Perspectives

Three-dimensional deformation imaging to detect myocardial scar

Deep learning automatically distinguishes myocarditis patients from normal subjects based on MRI

Deep Learning for Multi-Level Detection and Localization of Myocardial Scars Based on Regional Strain Validated on Virtual Patients

Novel Self-Calibrated Threshold-Free Probabilistic Fibrosis Signature Technique for 3D Late Gadolinium Enhancement MRI

Deep learning for automatic volumetric segmentation of left ventricular myocardium and ischaemic scar from multi-slice late gadolinium enhancement cardiovascular magnetic resonance

Identification of High-Risk Imaging Features in Hypertrophic Cardiomyopathy using Electrocardiography: A Deep-Learning Approach

Deep learning-based segmentation of left ventricular myocardium on dynamic contrast-enhanced MRI: a comprehensive evaluation across temporal frames