P869Predictive value of the endothelial shear stress distribution in three-dimensional quantitative coronary angiography models in detecting vulnerable plaques

Background Low Endothelial shear stress (ESS) is a well-known instigator of coronary atherosclerosis. Prospective intravascular ultrasound (IVUS)-based imaging studies with computational fluid dynamic analysis revealed its predictive merit in-vivo. However, whether coronary modelling derived from quantitative coronary angiography (QCA) is equally effective in detecting high-risk plaques remains to be established. Purpose To examine the value of endothelial shear stress (ESS) estimated in three-dimensional (3D) QCA models in detecting plaques that are likely to progress and cause events. Method We analysed the baseline intravascular ultrasound virtual histology (IVUS-VH) and angiographic data from 28 non-culprit lesions with a vulnerable phenotype (i.e., fibroatheroma or thin cap fibroatheroma) that caused major adverse cardiac events or required revascularization (nc-MACE-R) at 5-year follow-up and from a control group of 119 vulnerable plaques that remained quiescent. The segments studied by IVUS-VH at baseline were reconstructed using 3D-QCA software and in the obtained geometries blood flow simulation was performed and we estimated the resting Pd/Pa across the vulnerable plaque and the mean ESS values in 3mm sub-segments. A propensity score was built by the baseline plaque characteristics and the hemodynamic indices and its efficacy in detecting nc-MACE-R lesions was examined. Results Nc-MACE-R lesions were longer (32.5mm [18.0, 41.6], vs. 19.6mm [12.7, 31.3], p=0.03), had smaller minimum lumen area (MLA) (3.65mm2 [3.26, 4.36] vs. 5.03mm2 [3.98, 6.66], p<0.01), increased plaque burden (PB) (69.4% [63.5, 72.0] vs. 60.8% [53.7, 66.5], p<0.01), were exposed to higher ESS (9.40Pa [6.3, 12.5] vs. 4.1Pa [3.0, 6.9], p<0.01), and exhibited a lower resting Pd/Pa (0.97 [0.95, 0.98] vs. 0.98 [0.97, 0.99], p<0.01]. In multivariable analysis the only independent predictor of nc-MACE-R was the maximum 3mm ESS value (hazard ratio: 1.08 [1.02, 1.16], P=0.016). Lesions exposed to high ESS (>4.95Pa) with a high-risk anatomy (MLA<4mm2and PB>70%) had a higher nc-MACE-R rate (53.8%) than those with a low-risk anatomy exposed to high ESS (31.6%) or those exposed to low ESS that had high (20.0%) or low-risk anatomy (7.1%, P<0.001). Conclusion In the present study, 3D-QCA-derived local hemodynamic variables provided useful prognostic information and in combination with lesion anatomy enabled more accurate identification of nc-MACE-R lesions. Further research in a larger number of patients is need to confirm these findings before the conduction of large scale prospective studies that will combine intravascular imaging and 3D-QCA modelling to more accurately detect vulnerable plaques.