Layer-specific strain in patients with heart failure using cardiovascular magnetic resonance: not all layers are the same

Figures

• 

  Fig. 1. Strain Method. a Registration of cine images is used to calculate pixel-wise displacement fields between each image frame and the reference frames. b The process of registration yields displacement images (∆xi, ∆yi) that contain the translation of each pixel from the reference frame to the target ith image frame. c The user specifies endocardial (red) and epicardial (green) points on the end-diastolic frame. A uniform grid of points is generated at end-diastole consisting of 20 equally spaced contours between the user-specified endocardial and epicardial tracings from, with 100 equally spaced points along each contour. d Each point, or equivalently, each series of points in a contour, is propagated to all other cardiac phases using the previously calculated displacement fields. Four sample contours are shown. Strain for each contour is calculated as the fraction change in the contour length to end-systole. For regions that contain multiple contours, as shown here for GLS_AVE and GRS, average strains are calculated as the average of the strains from the individual contours. GLS_AVE, average global longitudinal strain; GRS, global radial strain

• 

  Fig. 2. Layer-specific contours in long axis cine images, at endocardial (red), epicardial (green) and equally spaced intramyocardial contours (blue) at end-diastole and end-systole used for the calculation of layer-specific strain. A subset of the 20 total intramyocardial contours used for calculation of average strain are shown. Sample tracings for a four-chamber view are shown for a health control (top) and a patient with heart failure (HF) with preserved ejection fraction (HFpEF) (bottom). Strain values at each layer over the full cardiac cycle are shown on the right

• 

  Fig. 3. Layer-specific contours in short axis cine images, at endocardial (red), epicardial (green) and equally spaced intramyocardial contours (blue) at end-diastole and end-systole used for the calculation of layer-specific strain. A subset of the 20 total intramyocardial contours used for calculation of average strain are shown. Sample tracings for a mid-ventricular short-axis view are shown for a health control (top) and a patient with HFpEF (bottom). Strain values at each layer over the full cardiac cycle are shown on the right

• 

  Fig. 4. Scatter and box plots of LV ejection fraction (LVEF) (a) and epicardial global longitudinal strain (GLS_EPI) (b) in all five groups. For each group, box plots show the median value, 25th and 75th percentiles and the full extent of the data. Outliers are identified as unfilled circles in the box plot. Scatter and box plots for LVEF (c and e) and GLS_EPI (d and f) in the preserved LVEF groups are shown with grouping by sex. Individuals with reduced GLS_EPI values are contained within the dashed boxes in d (men: 11% of controls, 17% of the at risk group and 32% of HFpEF are within the box) and f (women: 9% of healthy controls, 14% of the at risk group and 45% of HFpEF are within the box). All subjects in the preserved LVEF groups have normal LVEF by definition

• 

  Fig. 5. Scatter and box plots for LVEF and all strain components for all five groups. Box plots show the median value, 25th and 75th percentiles and the full extent of the data. Outliers are identified as unfilled circles in the box plot

• 

  Fig. 6. Relationship between GRS and LV Mass / LVEDV (concentricity) in HFpEF patients (a and b) and in all subjects with preserved LVEF (c and d). See Table 2 for abbreviations