Perfusion Computed Tomography Accurately Quantifies Collateral Flow After Acute Ischemic Stroke

Background and Purpose-This study aimed to derive and validate an optimal collateral measurement on computed tomographic perfusion imaging for patients with acute ischemic stroke. Methods-In step 1 analysis of 22 patients, the parasagittal region of the ischemic hemisphere was divided into 6 pial arterial zones to derive the optimal collateral threshold by receiver operating characteristic analysis. The collateral threshold was then used to define the collateral index in step 2. In step 2 analysis of 156 patients, the computed tomographic perfusion collateral index was compared with collateral scores on dynamic computed tomographic angiography in predicting good clinical outcome by simple regression. Results-The optimal collateral threshold was delay time >6 s (sensitivity, 88%; specificity, 92%). The computed tomographic perfusion collateral index, defined by the ratio of delay time >6 s/delay time >2 s volume, showed a significant correlation with dynamic computed tomographic angiography collateral scores (correlation coefficient, 0.62; P<0.001), with an optimal cut point of 31.8% in predicting good collateral status (sensitivity of 83% and specificity of 86%). When predicting good clinical outcome, the delay time collateral index showed a similar predictive power to dynamic computed tomographic angiography collaterals (area under the curve, 0.78 [0.67-0.83] and 0.77 [0.69-0.84], respectively; P<0.001). Conclusions-Computed tomographic perfusion can accurately quantify collateral flow after acute ischemic stroke.