All Roads Lead to Rome: Diverse Etiologies of Tricuspid Regurgitation Create a Predictable Constellation of Right Ventricular Shape Changes

Figures

• 

  FIGURE 1. Examples of shape correspondence models of right ventricular shape obtained using ShapeWorks. The colored spheres on the surface of each RV are the optimized correspondence points/corresponding landmarks that fully parameterize the right ventricular shape based on the population data (see Methods: Statistical Shape Model Construction). The three examples showcase the complexity of the RV shapes in the cohort.

• 

  FIGURE 2. Image processing pipeline: The yellow boxes represent each step in the processing pipeline. The symbols along the arrows show the software tool used to go from one step to another. From cine-MRI images, volume stacks were created and segmented using Seg3D, then uploaded to ShapeWorks for particle-based shape modeling.

• 

  FIGURE 3. Summary of mean shapes. Columns, 1, 4, and 7 contain the mean shape of the right ventricle in healthy volunteers, patients with cardiac comorbidities, and patients with moderate or greater TR, respectively. Columns 3 and 6 show the difference between the group-mean shapes. The arrows indicate the direction of group differences, and the color represents the magnitude of the group difference. Columns 2 and 5 show the p-values of the group differences. The regions with red color showcase statistically significant group differences.

• 

  FIGURE 4. The four primary modes of shape variation between patients with TR and controls. The control group is shown in the leftmost column, patient group in the rightmost, and the overall mean in the center.

• 

  FIGURE 5. Receiver Operator Characteristic of the Classification model for Training and Testing splits.

• 

  FIGURE 6. The mean shape of the right ventricle in healthy volunteers, patients with pulmonary hypertension, and patients with pulmonary hypertension plus moderate or greater TR are shown in columns 1,4,7. Columns 3 and 6 show the difference between the group mean shapes. The arrows indicate the direction of group differences, and the color represents the magnitude of the group difference. Columns 2 and 5 show the p-values of the group differences. The regions with red color showcase statistically significant group differences.

• 

  FIGURE 7. Progression of RV shape change in congestive heart failure (CHF). The mean shape of the right ventricle in healthy volunteers, patients with CHF without TR, and patients with CHF and TR are shown in columns 1, 4, 7. Columns 3 and 6 show the difference between the group mean shapes. The arrows indicate the direction of group differences, and the color represents the magnitude of the group difference. Columns 2 and 5 show the p-values of the group differences. The regions with red color showcase statistically significant group differences.

• 

  FIGURE 8. Shape mapping to linear discrimination of variation between population means, for the subgroups of patients with CHF and pulmonary hypertension (pHTN). The group mean for patients with TR is set as −1, and for patients without TR is set as 1. PDF = probability density function. Selected RV shapes correspond to individual points on the graph, and are shown as seen from the free wall. The number below each shape denote the single scalar value (or a “shape-based-score”) that places a subject-specific anatomy on a group-based shape difference that is statistically derived from the shape population.