Multicenter validation of natural shear wave imaging for non-invasive assessment of left ventricular filling pressure compared with guideline-based algorithms
A Youssef, L Wouters, K Papangelopoulou, A Araujo Andrade Sousa, K Gumiezna, W Desmet, G Castaldi, P Frederiks, J D'hooge, J U VoigtAbstract
Background
Diastolic dysfunction plays a pivotal role in many cardiovascular diseases, yet its assessment remains challenging. Current guidelines rely on complex decision trees integrating multiple echocardiographic parameters to predict elevated left ventricular filling pressure (LVFP) as a surrogate of diastolic dysfunction. Recently, high–frame rate shear wave (HFR-SW) imaging has emerged as a promising single, non-invasive biomarker to predict LVFP. Natural SWs—mechanical vibrations induced, for example, by mitral valve closure (MVC)—propagate at velocities directly related to myocardial stiffness, a key determinant of LVFP. However, multicenter validation of HFR-SW imaging using commercial scanning systems against invasive LV pressure measurements has not yet been performed.
Purpose
To compare the diagnostic accuracy of natural SW imaging with both existing and newly proposed guideline algorithms for predicting elevated LVFP in a multicenter setting.
Methods
At the Leuven site, we prospectively enrolled 44 patients (mean age 67.7 ± 10.8 years) undergoing clinically indicated left heart catheterization, irrespective of diastolic function status. Comprehensive conventional and HFR echocardiography (ca. 1200 frames/s) was performed within 1 hour after catheterization. In the parasternal long-axis view, SWs after MVC were tracked using an anatomical M-mode positioned along the interventricular septum from base to apex and color-coded for tissue acceleration. SWs appeared as tilted green bands, and their propagation velocity was measured semi-automatically (Figure 1). Data analysis at other centers are still in process.
Results
In our cohort, LV mid-diastolic pressure (LVMDP) ranged from 2 to 19.3 mmHg, and LV end-diastolic pressure (LVEDP) from 7.3 to 29.3 mmHg. Elevated filling pressures were defined as LVMDP > 15 mmHg and LVEDP > 16 mmHg. SW propagation velocities were significantly higher among patients with elevated LVEDP regardless of age (Figure 2A) and correlated strongly with invasive diastolic pressures (r = 0.74 for LVEDP, and r = 0.67 for LVMDP; p < 0.001 for both) (Figure 2 B, C). A SW speed cut-off ≥ 5.8 m/s accurately identified elevated LVMDP (AUC = 0.92; sensitivity = 88%; specificity = 83%; accuracy = 84%), outperforming both the former and updated guideline algorithms (80% and 73%, respectively) (Figure 2 D). The same cut-off also discriminated elevated LVEDP with excellent performance (AUC = 0.99; sensitivity 93%; specificity 100%; accuracy 97%), again exceeding guideline-based approaches (77% and 75%; Figure 2 E).
Conclusions
Our preliminary data showed that SW velocities after MVC are strongly related to LV filling pressures and can predict elevated LV filling pressures more accurately than the existing guidelines algorithms. This ongoing study shows the potential of natural shear wave imaging as a promising single, non-invasive tool for assessment of diastolic function.Figure 1Figure 2