Diagnostic accuracy of using Alberta Stroke Program Early Computed Tomography Score on CT perfusion map to predict a...

To assess the diagnostic accuracy of using Alberta Stroke Program Early Computed Tomography Score (ASPECTS) on CT perfusion (CTP) map to predict a volumetric target mismatch in patients with acute ischemic stroke (AIS).

Three-hundred and seven AIS patients with an onset time within 24h or unclear onset time who underwent CTP evaluation for large vessel occlusion of anterior circulation were enrolled. CTP ASPECTS was evaluated on cerebral blood flow (CBF) and time-to-maximum (Tmax) colored maps, respectively. Automated perfusion analysis software was used to calculate the volumes of ischemic core (volume_CBF<30%) and tissue at risk (volume_Tmax>6s). Target mismatch was defined as volume_{CBF< 30%}<70ml, volume_mismatch≥15ml, and volume_{Tmax >6s}/volume _{CBF< 30%}≥1.8. Spearman correlation and receiver operating characteristic curves were used for statistical analyses.

Strong correlations were found between CBF ASPECTS and volume_CBF<30%, and between Tmax ASPECTS and volume_Tmax>6s for overall population (ρ=-0.872, -0.757) and late-arriving patients (ρ=-0.900, -0.789). Mismatch ASPECTS moderately correlated with mismatch volume for overall population (ρ=0.498) and late-arriving patients (ρ=0.407). A CBF ASPECTS≥5 optimally predicted an ischemic core volume<70ml in overall population (sensitivity, 94.4%; specificity, 80.4%) and late-arriving patients (sensitivity, 89.5%; specificity, 90.5%). A CBF ASPECTS≥6 combined with a Mismatch ASPECTS≥1 optimally identified a target mismatch in overall population (sensitivity, 84.5%; specificity, 77.0%) and late-arriving patients (sensitivity, 83.7%; specificity, 90.0%).

CTP ASPECTS might be useful in predicting target mismatch derived from automated perfusion analysis software, and assisting in patient selection for endovascular therapy.