Abstract:Background: Estimates of regional left ventricular (LV) strains provide additional information to global function parameters such as ejection fraction (EF) and global longitudinal strain (GLS) and are more sensitive in detecting abnormal regional cardiac function. The accurate and reproducible assessment of regional cardiac function has implications in the management of various cardiac diseases such as heart failure, myocardial ischemia, and dyssynchrony. Purpose: To develop a method that yields highly reproducible, high-resolution estimates of regional endocardial strains from 4DCT images. Methods: A method for estimating regional LV endocardial circumferential ( ε c c ) $( {{\epsilon }_{cc}} )$ and longitudinal ( ε l l ${\epsilon }_{ll}$ ) strains from 4DCT was developed. Point clouds representing the LV endocardial surface were extracted for each time frame of the cardiac cycle from 4DCT images. 3D deformation fields across the cardiac cycle were obtained by registering the end diastolic point cloud to each subsequent point cloud in time across the cardiac cycle using a 3D point-set registration technique. From these deformation fields, ε c c and ε l l ${\epsilon }_{cc}\ {\rm{and\ }}{\epsilon }_{ll}$ were estimated over the entire LV endocardial surface by fitting an affine transformation with maximum likelihood estimation. The 4DCT-derived strains were compared with strains estimated in the same subjects by cardiac magnetic resonance (CMR); twenty-four subjects had CMR scans followed by 4DCT scans acquired within a few hours. Regional LV circumferential and longitudinal strains were estimated from the CMR images using a commercially available feature tracking software (cvi42). Global circumferential strain (GCS) and global longitudinal strain (GLS) were calculated as the mean of the regional strains across the entire LV for both modalities. Pearson correlation coefficients and Bland-Altman analyses were used for comparisons. Intraclass correlation coefficients (ICC) were used to assess the inter- and intraobserver reproducibility of the 4DCT-derived strains. Results: The 4DCT-derived regional strains correlated well with the CMR-derived regional strains ( ε c c ${\epsilon }_{cc}$ : r = 0.76, p < 0.001; ε l l ${\epsilon }_{ll}$ : r = 0.64, p < 0.001). A very strong correlation was found between 4DCT-derived GCS and 4DCT-derived EF (r = -0.96; p < 0.001). The 4DCT-derived strains were also highly reproducible, with very low inter- and intraobserver variability (intraclass correlation coefficients in the range of [0.92, 0.99]). Conclusions: We have developed a novel method to estimate high-resolution regional LV endocardial circumferential and longitudinal strains from 4DCT images. Except for the definition of the mitral valve and LV outflow tract planes, the method is completely user independent, thus yielding highly reproducible estimates of endocardial strain. The 4DCT-derived strains correlated well with those estimated using a commercial CMR feature tracking software. The promising results reported in this study highlight the potential utility of 4DCT in the precise assessment of regional cardiac function for the management of cardiac disease.

Comparison of left ventricular mass and wall thickness between cardiac computed tomography angiography and cardiac magnetic resonance imaging using machine learning algorithms

The Correlation Study of Left Ventricular Systolic Function Calculated by Automated Cardiac Motion Quantification

AI-Derived Left Ventricular Mass From Noncontrast Cardiac CT: Correlation With Contrast CT Angiography and CMR

Identifying left ventricular dysfunction using prospective electrocardiogram-triggered coronary computed tomography angiography

Prognostic value of left ventricular mass measured on coronary computed tomography angiography

AI-based, automated chamber volumetry from gated, non-contrast CT

Precision measurement of cardiac structure and function in cardiovascular magnetic resonance using machine learning

A comparison of visual and quantitative assessment of left ventricular ejection fraction by cardiac magnetic resonance

Comparative analysis of left ventricle function and deformation imaging in short and long axis plane in cardiac magnetic resonance imaging

Measurement accuracy of prototype non-contrast, compressed sensing-based, respiratory motion-resolved whole heart cardiovascular magnetic resonance angiography for the assessment of thoracic aortic dilatation: comparison with computed tomography angiography

Sixty-four-slice Multidetector Computed Tomography for Preoperative Evaluation of Left Ventricular Function and Mass in Patients with Mitral Regurgitation: Comparison with Magnetic Resonance Imaging and Echocardiography

Comparison of real-time three-dimensional echocardiography to magnetic resonance imaging for assessment of left ventricular mass in adult patients whose EF≥45%

AI Based CMR Assessment of Biventricular Function

Comparison of real-time three-dimensional transthoracic echocardiograpby with magnetic resonance imaging for assessment of left ventricular mass in elderly patients

Automatic quantification of the LV function and mass: a deep learning approach for cardiovascular MRI

Agreement in Left Ventricular Function Measured by Echocardiography and Cardiac Magnetic Resonance in Patients with Chronic Coronary Total Occlusion

Regional left ventricular endocardial strains estimated from low-dose 4DCT: Comparison with cardiac magnetic resonance feature tracking

Comparison of advanced echocardiographic right ventricular functional parameters with cardiovascular magnetic resonance in adult congenital heart disease

Automated analysis of 3D-echocardiography using spatially registered patient-specific CMR meshes

Differences Implying Clinical Decision

Automatic functional analysis of left ventricle in cardiac cine MRI.