Global longitudinal strain assessment by computed tomography in severe aortic stenosis patients - Feasibility using feature tracking analysis

Abstract

BackgroundGlobal longitudinal strain (GLS) detects subclinical myocardial changes in patients with aortic stenosis (AS). Although GLS is typically measured by transthoracic echocardiography (TTE), assessment by multiphasic gated computed tomography angiography (CTA) has become recently available. We sought to evaluate the feasibility of CTA-derived GLS assessment and compare its agreement with TTE using the same post-processing software in severe AS patients undergoing transcatheter aortic valve replacement (TAVR) evaluation. MethodsWe evaluated patients with severe AS, sinus rhythm and adequate image quality for GLS analysis by both CTA and TTE pre-TAVR using 2D CT-Cardiac Performance Analysis prototype software (TomTec). The 18-segment model was used for GLS analysis by averaging the three long-axis views in both CTA and TTE studies. Agreement was assessed using linear regression and Bland-Altman analysis. ResultsA total of 123 consecutive patients were included (mean age 84 ± 7 years, 45% female). The mean left ventricular ejection fraction (LVEF) by CTA and TTE were similar 53 ± 14% for both. On average, CTA-derived GLS was greater than by TTE (−20 ± 6.5% vs. −16 ± 4.9%, respectively, p < 0.001). There was a moderate correlation between GLS assessed by CTA vs. TTE (r = 0.62, p < 0.001), although variability between imaging methods existed. The correlation between GLS and LVEF was strong (r = −0.90, p < 0.001 for CTA, r = −0.88, p < 0.001 for TTE) using the same imaging modality. ConclusionCTA-derived GLS assessment is feasible in selected patients with sinus rhythm and adequate image quality. The agreement of GLS between TTE and CTA is moderate but not interchangeable suggesting a potential modality-specific GLS threshold.