Energetics of Cardiac Blood Flow in Hypertrophic Cardiomyopathy through Individualized Computational Modeling

Figure 1. Example cMR long-axis images and LV segmentation (blue). (A) End diastole (ED); (B) end systole (ES).

Figure 2. Example LV segmentation from a short-axis view. (A) Sample original contrasted image. (B) Cluster image representation. Each unique color refers to a different cluster (c = 10). (C) Image with closest traces to ventricle center from each cluster (enlarged). (D) Computed best fit trace for the region of interest, the LV (enlarged).

Figure 3. LV traces from the 3 cases studied. The figure shows traces at end diastole (ED, blue) and end systole (ES, orange) for 3 hearts: a normal heart, a heart with non-obstructive HCM (nHCM), and a heart with obstructive HCM (oHCM). The top boundary (y = 0) is the inlet-outlet (IO) boundary for each model. All three plots are on the same scale, and the scale bar is 20 mm (shown on first plot only).

Figure 4. Example of corresponding meshes used to impose wall motion in LV models. (A) End diastole (ED); (B) end systole (ES). By applying displacements to each node, the model simulates the LV wall motion during systole (ED to ES) and diastole (ES to ED).

Figure 5. Comparison of the spatial distribution of kinetic energy (KE) density at different time points over the cardiac cycle among different patients. Columns: T indicates normalized cardiac cycle period (from 0 to 1), with 0 corresponding to end diastole (ED). The second column (T = 0.25) corresponds to systolic ejection, and the third and fourth columns (T = 0.5 and 0.75) correspond to diastolic filling. Rows: Results from three patients are shown. nHCM: non-obstructive HCM; oHCM: obstructive HCM. Plots are color-coded by kinetic energy (KE) density values.

Figure 6. Simulation results obtained from 2D LV models for normal, nHCM, and oHCM patients. The plots show the variation of (A) LV area; (B) hemodynamic force; (C) flow rate (Q′); and (D) average kinetic energy density over the cardiac cycle. The vertical gray line indicates end systole (ES), and the beginning and end of the plot correspond to end diastole (ED).

Figure 7. LV blood flow vorticity from 2D models. (A) Average vorticity magnitude over the cardiac cycle depicts the magnitude of the vorticity vector (perpendicular to the longitudinal plane) averaged over the LV long-view surface. (B) Average vorticity accounting for the vorticity vector direction (positive or negative) to determine whether vorticity is balanced.

Figure 8. Patient-specific CFD modeling of HCM. Three-dimensional LV geometry and velocity vectors at a cut plane during diastole (A) and systole (B) for an nHCM patient. Velocity vectors (right plots of each panel) are color-coded by magnitude.

Figure 9. Comparison of simulation results from cMR-based CFD modeling of the LV. Three patient-specific models are shown: (top) a normal heart; (middle) a heart with non-obstructive HCM (nHCM); and (bottom) a heart with obstructive HCM (oHCM) at three phases during the cardiac cycle, with T indicating normalized (0–1) cardiac cycle period. T = 0.25 corresponds to systole, while T = 0.5 and T = 0.75 correspond to diastolic filling. Each pair of images shows blood velocities (in mm/s), represented by arrows. The left image of the pairs also shows the flow kinetic energy (KE, in mJ/m3) at long-view cut planes (color coded by value), while the right image shows blood flow velocity vectors at short-view planes (color-coded by magnitude). Model depictions are to scale, with the distance between short-view (cross-sectional) planes equal to 16 mm.

Figure 10. Simulation results for the three patients considered depicting the kinetic energy density (KE) and flow fields. The distinct graphs depict velocity vectors within the LV inside the volume of a KE isosurface (such that KE > isosurface value of 10 J/m3 or 5 J/m3). Three phases of the cardiac cycle (T ∈ [0,1]) are shown: T = 0.25 corresponds to systole, while T = 0.5 and T = 0.75 correspond to diastolic filling. Colors represent velocity magnitude, and the outer layer is the whole LV. Note that flow entering and leaving the heart is not symmetric but rather larger towards one side of the heart, particularly in the HCM hearts. Thus, natural ventricular flow asymmetry was partially captured by our models, even though our LV models were truncated and did not include the inlet and outlet within the heart base.