Cardiovascular Fluid Mechanics Laboratory

Predicting post-TAVR Thrombosis

Parametric Study

Sub-clinical thrombosis has been identified in transcatheter aortic valves (TAVs). Previous studies in our lab identified flow stasis in the neo-sinus of TAVs to be a key mechanistic factor contributing to this phenomena. The objective of this research project is to develop custom TAV devices of varied leaflet design, and assess the effects of these design features on the neo-sinus and aortic sinus flow characteristics using experimental fluid mechanics. 

Hypoattenuated Leaflet Thickening (HALT)

Thrombosis in post-transcatheter aortic valve replacement (TAVR) patients has been correlated with flow stasis in the neo-sinus. This study investigated the effect of the post-TAVR geometry on flow stasis. Computed tomography angiography of 155 patients who underwent TAVR using a SAPIEN 3 were used to identify patients with and without thrombosis, and quantify thrombus volumes. Six patients with 23-mm SAPIEN 3 valves were then selected from the cohort and used to create patient-specific post-TAVR computational fluid dynamic models. Regions of flow stasis (%Volstasis, velocities below 0.05 m/s) were identified. The results showed that all post-TAVR anatomical measurements were significantly different in patients with and without thrombus, but only sinus diameter had a linear correlation with thrombus volume (r = 0.471, p = 0.008). A linear correlation was observed between %Volstasis and thrombus volume (r = 0.821, p = 0.007). The combination of anatomy and valve deployment created a unique geometry in each patient, which when combined with patient-specific cardiac output, resulted in distinct flow patterns. While parametric studies have shown individual anatomical or deployment metrics may relate to flow stasis, the combined effects of these metrics potentially contributes to the biomechanical environment promoting thrombosis, therefore hemodynamic studies of TAVR should account for these patient-specific factors.

Transcatheter aortic valve replacement (TAVR) is a minimally invasive method of treating aortic stenosis. Hypoattenuated Leaflet Thickening (HALT) happens in 4% to 40% of TAVR cases and may reduce leaflet motion and increase aortic valve pressure gradient. Leaflet thrombosis, even if resolved with anticoagulants, may lead to early structural valve degeneration and calcification of the leaflets. Thus, a prediction model of leaflet thrombosis may help pre-planning and exploring alternative options with varying valve deployment parameters or a different valve choice. Computational Fluid Dynamics (CFD) and Fluid-Structure Interaction (FSI) have been applied to predict post-TAVR thrombose formation. CFD, and FSI methods are computationally expensive and time-consuming. Thus, these methods are not appropriate options for fast patient-specific thrombus formation prediction applications. As such, a quick patient-specific method is required to predict the likelihood of thrombus formation after TAVR. To this goal, a dimensionless parameter called Normalized Circulation (NC) has been developed by our group, and it has been demonstrated that this parameter is a criterion that indicates the possibility of post-TAVR thrombosis. This study aimed to evaluate the Machine Learning (ML) algorithms combined with dimensionless normalized circulation parameter (NC) ability to predict thrombus formation post-TAVR implantation. The current results show the combined machine learning and scale analysis method can be used to predict the post-TAVR thrombose formation with an accuracy higher than 90%.

Flow stasis post-TAVR

THV deployment and anatomic characteristics were found to influence hemodynamics and flow stasis on an individual neosinus-specific basis. Twenty-three patients were included in this study. Thrombus volume (quantified via CT data) was found to be significantly correlated (rho = 0.621, P < .0001) with washout time (a measure of flow stasis). Measurement of neosinus flow stasis may guide strategies to improve outcomes in TAVR.

Selected Publications

  1. Hatoum H, Singh-Gryzbon S, Esmailie F, et al. Predictive Model for Thrombus Formation After Transcatheter Valve Replacement [published correction appears in Cardiovasc Eng Technol. 2022 Jan 7;:]. Cardiovasc Eng Technol. 2021;12(6):576-588. doi:10.1007/s13239-021-00596-x
  2. Trusty PM, Bhat SS, Sadri V, et al. The role of flow stasis in transcatheter aortic valve leaflet thrombosis [published online ahead of print, 2020 Nov 26]. J Thorac Cardiovasc Surg. 2020;S0022-5223(20)33099-3. doi:10.1016/j.jtcvs.2020.10.139
  3. Singh-Gryzbon, S., Ncho, B., Sadri, V. et al. Influence of Patient-Specific Characteristics on Transcatheter Heart Valve Neo-Sinus Flow: An In Silico Study. Ann Biomed Eng 48, 2400–2411 (2020). https://doi.org/10.1007/s10439-020-02532-x
  4. Ncho, B., Sadri, V., Ortner, Kollapaneni, S., and Yoganathan, A.. In-vitro Assessment of the Effects of Transcatheter Aortic Valve Leaflet Design on Neo-sinus Geometry and Flow. ABME https://doi.org/10.1007/s10439-020-02664-0
  5. Raghav, V., C. Clifford, P. Midha, I. Okafor, B. Thurow, and A. Yoganathan. Three-dimensional extent of flow stagnation in transcatheter heart valves. J. R. Soc. Interface 16:20190063, 2019.
  6. Midha, P. A., V. Raghav, R. Sharma, J. F. Condado, I. U.Okafor, T. Rami, G. Kumar, V. H. Thourani, H. Jilaihawi, V. Babaliaros, R. R. Makkar, and A. P. Yoganathan. The fluid mechanics of transcatheter heart valve leaflet thrombosis in the neosinus. Circulation 136:1598–1609, 2017.