Association Between Thrombus Perviousness Assessed on Computed Tomography and Stroke Cause.

Kufner, Anna; Erdur, Hebun; Endres, Matthias; Nolte, Christian H; Scheel, Michael; Schlemm, Ludwig

BACKGROUND AND PURPOSE A recent study proposed that thrombus perviousness (TP)-the degree to which contrast agents penetrate the thrombus in an occluded vessel measured on noncontrast computed tomography (NCCT) and CT angiography-may be associated with cardioembolic stroke cause with high specificity. Our aim was to investigate which clinical and laboratory parameters affect measures of TP and to validate its diagnostic accuracy in an independent cohort of patients with acute ischemic stroke. METHODS Seventy-five patients from a prospectively maintained database with proximal occlusions of the middle cerebral artery (M1) were retrospectively analyzed. Thrombi were segmented on coregistered noncontrast computed tomography and CT angiography to determine the thrombus attenuation increase and void fraction (attenuation increase relative to contralateral side). RESULTS TP measures were significantly higher in patients with cardioembolic stroke compared to patients with stroke attributed to large artery atherosclerosis (median thrombus attenuation increase [interquartile range], 2.79 [-3.54 to 8.85] versus -5.11 [-11.23 to -1.47]; P=0.001). In linear regression analysis for TP including age, time to scan, prior medication with antiplatelets or anticoagulants, and selected laboratory parameters, only stroke cause was significantly associated with TP. In multivariable binary logistic regression analysis for dichotomized stroke cause (ie, cardioembolic versus noncardioembolic stroke), only thrombus attenuation increase was independently associated with cardioembolic stroke (odds ratio of 1.12 [95% CI, 1.04-1.22]; P=0.004). Receiver operating characteristic analysis indicated that TP can identify cardioembolic stroke with an area under the curve of 0.75 (95% CI, 0.63-0.87) for thrombus attenuation increase. With a cutoff value of 6.23 Hounsfield units, cardioembolic strokes were identified with 100% specificity. Results for void fraction were similar. CONCLUSIONS The assessment of TP on baseline noncontrast computed tomography/CT angiography in patients with M1 occlusion may aid in determining cardioembolic stroke cause and guide secondary prevention. Selected clinical and laboratory parameters other than stroke cause did not affect TP measures.