AORTIC WALL STRAIN ANALYSIS IN PATIENTS WITH HEMODYNAMICALLY SIGNIFICANT AORTIC STENOSIS: APPLICATION OF 4D STRAIN FROM ECG-GATED CT ANGIOGRAPHY

Abstract

Aortic wall strain quantification has the potential to assist decision making for aortic root replacement at time of valve surgery among patients with aortopathy. However, the small tissue displacements experienced by the aorta are challenging for in-vivo assessment. While temporal resolution remains modest to date, the high isotropic spatial resolution of gated CTA offers a unique advantage for displacement tracking. The purpose of this study was to evaluate the clinical feasibility of 4D aortic strain analysis from routine pre-operative CT Angiographic studies prior to transcatheter aortic valve implantation (TAVI), and to correlate aortic wall strain with transvalvular pressure gradient to serve as validation for strain measurements through their anticipated physiologic relationship. Ten patients with pre-operative CT assessment prior to TAVI were studied. All patients had a mean aortic valve gradient ≥20 mmHg by echocardiography. Patients with significant left ventricular dysfunction (LV EF <40%) were excluded. Retrospective ECG-gated CT angiogram images through the thorax were acquired and reconstructed at 10% intervals throughout the cardiac cycle using a standard clinical protocol on a multidetector CT scanner (Discovery CT750 HD, GE Healthcare). Using ITK-SNAP, an end-diastolic aortic 3D mesh model was generated and smoothed (by fairing) for continuous curvature across the aortic wall. Aortic wall strain analysis was performed on the mesh model using GIUSEPPE, a custom Matlab-based (Mathworks, USA) feature-tracking algorithm. Deformed mesh elements were then used to compute wall dilation in its dominant direction, or peak principal strain amplitude (PPSA), for each phase of the cardiac cycle. The mean age was 79.6±5.3.The mean (AVMean) and peak (AVPeak) aortic valve gradients were 39.4±13.0 mmHg (range 21.8 to 64.2mmHg) and 71.3±23.6 (range 40.0 to 110.3mmHg), respectively. Aortic wall strain analysis was feasible in all subjects. End-systolic PPSA (PPSAES) for the ascending, arch and descending aortic segments were 14.83±6.81%, 8.33±5.07% and 5.19±4.01%, respectively. PPSAES in the ascending segment correlated well with AVMean (rho=-0.6818, p=0.0255) and AVPeak (rho=-0.6909, p=0.0231). A significant correlation between PPSAES and aortic valve area was also found (rho=0.9048, p=0.0046). In this pilot study we demonstrated clinical feasibility of 4D aortic wall strain analysis from gated CTA and confirmed anticipated associations with transvalvular gradients from echocardiography among patients with severe aortic stenosis. CTA-based aortic PPSA holds promise as a clinical tool for the in-vivo evaluation of aortic wall health, and may be of value for assisting in decision-making regarding aortic interventions.