Morphologically normalized left ventricular motion indicators from MRI feature tracking characterize myocardial infarction

Figures

• 

  Figure 1. Mean shapes occurring at end-diastole and end-systole after Parallel Transport (PT). Colors in end- systole refer to the difference from end-diastolic state (that is white = undeformed, red = deformed). The colormap ranges from white (=no deformation) to red (max) and refers to || xM-x || with xM the position at end-diastole and x its position at end-systole. As in SS the shapes are scaled to unit size and the deformations of epicardium and endocardium have different magnitudes, different color scales are used. The black line indicates the position of the inter-ventricular septum. See Supplementary Files 1 and 2 for dynamic visualizations.

• 

  Figure 2. Results of GPA + PT + PCA for epicardium and endocardium in SSS. For visualization purposes, only one per-group mean trajectory (300 Controls vs 300 MI) is shown. The colormap ranges from white (=no deformation) to red (=max) and refers to || xM-x || with xM the position at the Grand Mean and x its position at the PC axis extreme. The black line indicates the position of inter-ventricular septum. The first three PCs are used as “landmarks” in the trajectory analysis taking into account all 600 trajectories present in the dataset. See Supplementary Files 3 and 4 for dynamic visualization of trajectories and associated shape changes. Shapes are predicted at the positive and negative axes’ extremes.

• 

  Figure 3. Results of GPA + PT + PCA for epicardium and endocardium in SS. For visualization purposes, only one per-group mean trajectory (300 Controls vs 300 MI) is shown. The colormap ranges from white ( = no deformation) to red (max) and refers to || xM-x || with xM the position at the Grand Mean and x its position at the PC axis extreme. As in SS the shapes are scaled to unit size and the deformations of epicardium and endocardium have different magnitudes, different color scales are used. The black line indicate the position of inter-ventricular septum. The PC scores of first three PCs are used as “landmarks” in the trajectory analysis taking in to account all 600 trajectories present in the dataset. See Supplementary Files 5 and 6 for dynamic visualization of trajectories and associated shape changes. Shapes are predicted at actual positive and negative axes extremes.

• 

  Figure 4. The difference in trajectories size among groups.

• 

  Figure 5. Angle distributions in SSS and SS.

• 

  Figure 6. Temporal course of PC1 of the endocardial analysis in SSS (top) and SS (bottom). Asterisks indicate significance under ANOVA for the differences between Controls (green) and MI (black).

• 

  Figure 7. Temporal course of PC1 of the epicardial analysis in SSS (top) and SS (bottom). Asterisks indicate significance under ANOVA for the differences between Controls (green) and MI (black).

• 

  Figure 8. A configuration of one LV shape at end-diastole (a randomly chosen Control case). Red points (grey in printed version) are the epicardium, black points are the endocardium. Green lines (light grey in printed version) on the two layers indicate the middle of inter-ventricular septum. LV base is on the bottom, apex on top. Each of the 600 cases (300 MI and 300 Controls) comprises 30 shapes representative of one cardiac cycle for a total of 18,000 shapes.

• 

  Figure 9. Example of a finite element parameter in time. The blue dots (black in printed version) represent the time- registered points at each frame, the green line (where dots are placed) the cubic B-spline, and the red line (where dots are not placed) the Fourier partial sums with 5 harmonics.

• 

  Figure 10. The meaning of PT when applied to two different trajectories. Left: When classic GPA + PCA are applied PC axes explain concomitantly inter- and intra- individual variation. Right: PT applied to the same data with deformation transported toward the Grand Mean; in this case PC axes explains solely the deformation (thus without initial inter- individual shape differences) applied to the consensus.

• 

  Figure 11. Flowchart for the trajectory analysis using a platonic example.