Abstract TP73: Heterogeneity of Occlusive Thrombus Investigated on Pretreatment Imaging in Acute Ischemic Stroke

Abstract

Background and Purpose: Investigation of occlusive thrombus in ischemic stroke is expected to clarify the etiology and mechanism of stroke fueling an optimal treatment strategy. However, in situ propagation due to secondary thrombosis after initial embolic occlusion could obscure the characteristics of initial emboli. Differentiation of the initial embolus and secondary thrombosis using retrieved specimens might be difficult due to fragmentation of whole thrombi. We aimed to investigate the heterogeneity of occlusive thrombi using pretreatment imaging of ischemic stroke. Methods: Among consecutive stroke patients with acute embolic occlusion of the anterior circulation eligible for endovascular reperfusion therapy, we retrospectively reviewed 81 patients who underwent both brain non-contrast CT (NCCT) and CT angiography (CTA) on admission from June 2015 to May 2018. Region of interest (ROI) was set on the whole thrombus on NCCT, which was identified referring to the filling defect on CTA, and the size of thrombus was measured. ROI was then divided equally into proximal and distal segments and Hounsfield Unit (HU) densities of those segments were measured respectively. Difference of HU between proximal and distal segments (ΔHU; HU in proximal segment - HU in distal segment) was calculated and analyzed. Results: HU density on NCCT in proximal segment was higher than in distal segment (mean difference: 3.75, 95% CI: 2.69-4.80, p<0.01). ΔHU was correlated positively with the length of thrombus (ρ=0.32, p<0.01), and correlated inversely with the time from last-known-well (LKW) to imaging (ρ=-0.29, p=0.01) and blood platelet level (ρ=-0.27, p=0.02). Stroke subtype and vascular risk factors had no significant correlation with ΔHU. On multivariate regression analysis, the length of thrombus (β=0.31, p=0.01) and the time from LKW to imaging (β=-0.30, p=0.01) were associated with ΔHU. Conclusions: Density of occlusive thrombi in the proximal segment is higher than in the distal segment. This difference increases as thrombi grow longer and decreases as thrombi get older after embolic occlusion. This time/length-dependent thrombus heterogeneity is suggestive of secondary thrombosis, occurring initially at the proximal side of occlusion.