Endothelial shear stress computed from coronary computed tomography angiography: A direct comparison to intravascular ultrasound

Introduction: Intravascular ultrasound (IVUS) studies have shown that biomechanical variables, particularly endothelial shear stress (ESS), add synergistic prognostic insight when combined with anatomic high-risk plaque features. Non-invasive risk assessment of coronary plaques with coronary computed tomography angiography (CCTA) would be helpful to enable broad population risk-screening. Aim: To compare the accuracy of ESS computation of local ESS metrics by CCTA vs IVUS imaging. Methods: We analyzed 59 patients from a registry of patients who underwent both IVUS and CCTA for suspected CAD. CCTA images were acquired using either a 64- or 256-slice scanner. Lumen, vessel, and plaque areas were segmented from both IVUS and CCTA (59 arteries, 686 3-mm segments). Images were co-registered and used to generate a 3-D arterial reconstruction, and local ESS distribution was assessed by computational fluid dynamics (CFD) and reported in consecutive 3-mm segments. Results: Anatomical plaque characteristics (vessel, lumen, plaque area and minimal luminal area [MLA] per artery) were correlated when measured with IVUS and CCTA: 12.7 ​± ​4.3 vs 10.7 ​± ​4.5 ​mm2, r ​= ​0.63; 6.8 ​± ​2.7 vs 5.6 ​± ​2.7 ​mm2, r ​= ​0.43; 5.9 ​± ​2.9 vs 5.1 ​± ​3.2 ​mm2, r ​= ​0.52; 4.5 ​± ​1.3 vs 4.1 ​± ​1.5 ​mm2, r ​= ​0.67 respectively. ESS metrics of local minimal, maximal, and average ESS were also moderately correlated when measured with IVUS and CCTA (2.0 ​± ​1.4 vs 2.5 ​± ​2.6 ​Pa, r ​= ​0.28; 3.3 ​± ​1.6 vs 4.2 ​± ​3.6 ​Pa, r ​= ​0.42; 2.6 ​± ​1.5 vs 3.3 ​± ​3.0 ​Pa, r ​= ​0.35, respectively). CCTA-based computation accurately identified the spatial localization of local ESS heterogeneity compared to IVUS, with Bland-Altman analyses indicating that the absolute ESS differences between the two CCTA methods were pathobiologically minor. Conclusion: Local ESS evaluation by CCTA is possible and similar to IVUS; and is useful for identifying local flow patterns that are relevant to plaque development, progression, and destabilization.