4D flow cardiovascular magnetic resonance recovery profiles following pulmonary endarterectomy in chronic thromboembolic pulmonary hypertension

Figures

• 

  Fig. 1. Semi-automated pipeline to extract velocity from 4D flow cardiovascular magnetic resonance (4D flow CMR) contained solely within the pulmonary arteries through image segmentation and shape analysis algorithms as well as automated computation of velocity metrics from branches and cross-sectional slices throughout the cardiac cycle. Visualization of the 4D flow CMR-derived metrics extracted from the pulmonary arteries (PAs) show illustrative explanations of acceleration time ratio, fractional area of reverse flow, vorticity, helicity, and helical flow index. * = metric was computed from a mid-slice located 50% down the length of the vessel. † = metric was computed from a proximal slice located 10% down the length of the vessel. All other metrics were computed from the volumes of the main, left, and right PA branches

• 

  Fig. 2. Hemodynamic and RV outcomes recovered to normal values in most patients with CTEPH post-pulmonary endarterectomy (PEA) for (a) mean pulmonary artery pressure (mPAP), (b) total pulmonary resistance (TPR), (c) % predicted right ventricular ejection fraction (RVEF), and (d) % predicted right ventricular end-systolic volume (RVESV) adjusted by sex, age, weight, and height [27]. Colored lines represent the San Diego classification of the PEA specimens for each patient – red = Type I, blue = Type II, green = Type III, grey = not recorded. Dotted lines in (a), (b), and (d) indicate healthy values of mPAP (< 20 mmHg), TPR (~ 3.5 WU), and % predicted RVESV (100%). Significant differences by t-test represented by *=p < 0.05, **=p < 0.01

• 

  Fig. 3. Volume and PA wall indices of patients with CTEPH changed post-PEA. (a) PA anatomies confined from the pulmonary valve to the first branch point in the left pulmonary artery (LPA) and right pulmonary artery (RPA) are shown at minimum volumes in the cardiac cycle for all patients with CTEPH (n = 20, first three rows, pre-PEA left, post-PEA right per box). Volumes of the (b) MPA decreased significantly post-PEA in patients with CTEPH. PA wall indices increased for the MPA (c) relative area change (RAC) and (d) compliance. The dotted line in (d) indicates healthy values of compliance (0.15 cm2/mmHg) [29]. Colored lines represent the San Diego classification of the PEA specimens for each patient – red = Type I, blue = Type II, green = Type III, grey = not recorded. *=p < 0.05, **=p < 0.01

• 

  Fig. 4. Flow waveforms averaged for all patients with CTEPH with high quality 4D flow CMR velocity (n = 15) pre-PEA (blue), post-PEA (orange) in the (a) MPA, (b) LPA, and (c) RPA where the solid line is the average, and the shaded region is the standard deviation. Centerline velocities in the (d) MPA, (e) LPA, and (f) RPA increased significantly post-PEA. Colored lines represent the San Diego classification of the PEA specimens for each patient – red = Type I, blue = Type II, green = Type III, grey = not recorded. Significant differences between groups denoted by: *=p < 0.05, **=p < 0.01

• 

  Fig. 5. (a) The mean systolic area fraction of reverse flow in a slice in the MPA decreased slightly post-PEA. (b) A cross-sectional slice in the middle of the PA trunk taken during mid-systolic downstroke shows more reverse flow in a representative patient (PH18) pre-PEA than post-PEA. (c) The mean systolic spatially averaged vorticity in the MPA increased post-PEA. (d) Visually this is shown in a representative patient (PH7) pre- and post-PEA with streamlines colored by vorticity during peak vorticity. (e) The mean fraction of positive helicity was approximately half of the MPA volume for the pre/post-PEA group. (f) A cross-sectional MPA slice taken during systole with surface vectors projecting forward moving velocity onto the slice and colored by helicity shows distinct counter rotating structures in the post-PEA (PH18) condition as opposed to the pre-PEA condition. Colored lines (a, c, & e) represent the San Diego classification of the PEA specimens for each patient –blue = Type II, green = Type III, grey = not recorded. *=p < 0.05, **=p < 0.01

• 

  Fig. 6. Longitudinal changes of 4D flow-derived PA area, velocity, and helicity were associated with mean PA pressure (mPAP), total pulmonary resistance (TPR), and percent predicted RV end-systolic volume (RVESV) from pre-PEA to post-PEA (negative ∆ values indicating a decrease in value from pre- to post-PEA). Decreases in minimum MPA area were associated with decreases in (a) mPAP and (b) TPR, but not with (c) % predicted RVESV. Increases in RPA centerline velocity was associated with decreases in (d) mPAP and (e) TPR, but not (f) % predicted RVESV. Increases in RPA helical flow index (HFI) were strongly associated with decreases in (g) mPAP and (h) TPR. Decreases in the MPA fraction of positive helicity were strongly associated with decreases in (i) % predicted RVESV. Correlations with area used n = 20 patients with clear anatomical resolution in the 4D flow images; velocity and helicity correlations used n = 15 patients to exclude 4D flow CMRs that did not maintain conservation of flow. Colored dots in (c)-(h) represent the San Diego classification of the PEA specimens for each patient – red = Type I, blue = Type II, green = Type III, grey = not recorded. Significance of the Spearman…