Estimation of the Ischemic Penumbra Based on CT Perfusion: A Pilot Study

Rationale and Objectives Ischemic penumbra (IP), the target of thrombolytic therapies, could be estimated by the mismatch region between magnetic resonance imaging (MRI) diffusion- and perfusion-defined abnormalities; however, the accuracy of this method has been challenged recently. In this study, we try to establish a method for calculating IP size based on computed tomography perfusion (CTP) and to observe the early evolution of IP in detail. Materials and Methods The middle cerebral artery occlusion (MCAO) model in monkey was used to compare the accuracy in estimating the IP between CTP and MRI methods. A receiver operating characteristic (ROC) curve was performed to calculate the IP threshold of the different CTP parameters, and then the best parameter was obtained. The dynamic evolutions of estimated size of IP by these two methods were compared. Results Among the three CTP parameters, relative cerebral blood flow (rCBF) had the highest sensitivity (83.3%) and specificity (98.5%) in estimating the IP. The optimal cutoff threshold of rCBF was 0.203. During the first 15 hours of the MCAO model, the estimated size of IP by the rCBF was larger than that of the MRI method; however, this relationship was reversed 15 hours later. Conclusion This study suggests that the rCBF method is more accurate in estimating the IP since previous studies have reported that the MRI method underestimated the exact IP in the early stage of ischemia and overestimated the exact IP in the later stages. Further experimental and clinical studies are needed to validate the conclusion.