Aortic stenosis (AS) is the most prevalent valvular heart disease and is associated with a high mortality [1]. The diagnosis of severe AS is based on echocardiographic measures including Mean Pressure Gradient (MPG) and aortic valve effective orifice area (AVA) [2]. Class I indications for valve replacement are severe AS with symptoms or reduced LVEF [2]. However, the symptomaticity of AS is highly subjective and can be confounded by various other diseases. Moreover, AVA and MPG do not account for post-stenotic pressure recovery and are therefore prone to misclassify AS severity [3], in particular since aortic flow characteristics and valve geometry are not factored in.

In contrast, the measurement of Turbulent Kinetic Energy (TKE) by Phase-Contrast MRI (PC-MRI) enables direct investigation of the mechanisms responsible for energy dissipation [4]. TKE is largely dissipated into heat and thereby allows probing energy losses due to AS. In the present study we hypothesized that TKE provides additional information for the assessment of AS severity beyond echocardiographic measures and thereby has the potential to enhance future classification and stratification of AS patients.

For this cross-sectional study, 55 patients with aortic stenosis (67±15 years, 20 female) and 10 healthy age-matched controls (69±5 years; 5 female) were prospectively recruited. All subjects underwent time-resolved 3D Phase-Contrast MRI (4D Flow MRI) in addition to a routine cardiac MRI protocol and an echocardiography examination.

Data were acquired on a clinical 3T system (Philips Healthcare, Best, The Netherlands) using a 4D Bayesian MultiPoint PC-MRI sequence [5] with three velocity encoding steps in each direction. Prospective cardiac triggering and respiratory navigator-based gating allowed for acquisition during free-breathing with an isotropic spatial resolution of 2.5 x 2.5 x 2.5 mm³ and a heart rate dependent temporal resolution of 22 to 44 ms. The acquisition was accelerated using 8-fold k-t PCA [6] with a net acceleration factor of 7.1, resulting in a total scan time of 15 to 30 min depending on navigator gating efficiency. Voxelwise TKE values were computed using a Bayesian approach [5] and were integrated over the ascending aorta and the aortic arch. For analysis, peak systolic values (Peak TKE) are reported. For testing differences, one- or two-way ANOVA was performed. All data are expressed as mean ± SD.

Data acquisition and evaluation was successful in 51 out of the 55 enrolled patients and in all controls. According to MPG, 27 patients had a severe aortic stenosis (MPG ≥ 40 mmHg) and 24 patients had mild/moderate aortic stenosis (MPG < 40 mmHg). Dilatation of the ascending aorta (AAo) was present in 15 patients and 11 patients had a bicuspid aortic valve (BAV).

Peak TKE was found to be significantly elevated in patients with aortic stenosis compared to controls (25±10 mJ vs. 4.8±1.0 mJ, p<0.001). The relation between MPG and Peak TKE is illustrated in Figure 1. A significant but weak correlation between TKE and MPG was found in the entire study population (Peak TKE vs MPG: R2 = 0.26). However, when excluding the healthy controls, no significant correlation was present anymore. Figure 2 illustrates flow fields and TKE distributions of patients with comparable MPG indicating moderate-to-severe AS but highly different TKE levels.

The comparison of patients with BAV vs patients with tricuspid aortic valves and patients with dilated AAo vs those with normal AAo showed that in both populations Peak TKE is significantly elevated (BAV vs tricuspid valve: p<0.001, dilated AAo vs normal AAo: p<0.001). In contrast, no significant difference of MPG was found between patients with bicuspid and tricuspid aortic valves (p=0.14) or between patients with dilated and normal AAo (p=0.41). The results are displayed in Figure 3.

TKE was found to be significantly higher in patients with aortic stenosis compared with healthy controls. However, only weak correlation between TKE and MPG was detected, implying that TKE accounts for other AS characteristics than echocardiographic measures. Peak TKE was significantly influenced by aortic geometry and valvular morphology. Dilation of the ascending aorta and BAV significantly increase TKE while having no effect on MPG. The increased energy loss may be related to a higher degree of turbulent flow mixing at the borders of the jet, thereby resulting in higher TKE levels. These features of AS are not detectable by echocardiography but can be assumed to influence the cardiac effects of AS and long-term outcome. In summary, Turbulent Kinetic Energy derived from Phase-Contrast MRI allows assessing the influence of valve and aortic geometry on the hemodynamic burden of AS and thereby provides a future perspective for better stratification of AS patients.