Leticia C.S. Souza MD, Presenter: Nothing to Disclose

Michael H. Lev MD, Abstract Co-Author: Consultant, General Electric Company Consultant, CoAxia, Inc Consultant, Takeda Pharmaceutical Company Limited Medical Advisory Board, General Electric Company Medical Advisory Board, CoAxia, Inc Medical Advisory Board, Takeda Pharmaceutical Company Limited Research support, General Electric Company

Ana M. Franceschi MD, Abstract Co-Author: Nothing to Disclose

Jie Hi MD, Abstract Co-Author: Nothing to Disclose

Ramon Gilberto Gonzalez MD, PhD, Abstract Co-Author: Nothing to Disclose

Pamela Whitney Schaefer MD, Abstract Co-Author: Nothing to Disclose

The presence of a large (>70 ml) admission DWI lesion in acute stroke patients with a proximal vessel occlusion is an established marker for poor outcome, and hence considered to be a contraindication for endovascular therapy. Our aim was to establish the accuracy of CT perfusion (CTP) cerebral blood volume (CBV) and cerebral blood flow (CBF) maps for estimating infarct core and for determining patients likely to have a poor outcome regardless of treatment, compared to a DWI reference standard.

We analyzed the admission CTP and DWI of 62 stroke patients presenting with a large vessel occlusion on CTA. Infarct core was semi-automatically segmented on the admission DWI. We applied 85% and 70% reduction thresholds to CBF and CBV maps respectively, and subsequently determined lesion volumes both by semi-automatic segmentation and by visual estimation using the ABC/2 method. Linear regression was used to compare resulting lesion volumes. ROC curve analysis was conducted to evaluate the test characteristics of each set of volumes in prediction of clinical outcome, with modified Rankin score (mRS) < 2 considered good and > 2 considered poor outcome.

There was strong correlation between the thresholded CTP and DWI lesion volumes (R2=0.87, slope=1.11 for CBF; R2=0.82, slope=1.05 for CBV, all p<0.001); as well as between the segmented and ABC/2 estimated lesion volumes (R2=0.81, slope=1.03 for DWI; R2=0.883, slope=1.003 for CBF; R2= 0.946, slope=1.11 for CBV, all p<0.001). ROC curve analysis showed that there was no significant difference between CBV and CBF, or CBV and DWI, for overall prediction of clinical outcome (AUC=0.84 for DWI, AUC=0.74 for CBF, and AUC=0.74 for CBV, p=0.9 for CBV vs. CBF, and p=0.06 for CBV vs. DWI), although there was a significant difference between DWI and CBF (p=0.017 for DWI vs. CBF). All parameters were highly specific for poor outcome, but at different volume thresholds.

The 85% thresholded CT-CBF and 70% thresholded CT-CBV lesion volumes are strongly correlated with admission DWI lesion volumes. Large admission lesion volumes on these parameter maps were similarly highly specific for the prediction of poor long-term outcome.

The 85% thresholded CT-CBF and 70% thresholded CT-CBV lesion volumes are strongly correlated with admission DWI lesion volumes, and offer an alternative to those patients that cannot undergo DWI.

Souza, L, Lev, M, Franceschi, A, Hi, J, Gonzalez, R, Schaefer, P, Thresholded CTP Maps Can Accurately Determine Infarct Core When DWI Is Unavailable, and Have Similar Specificity in Identifying Patients Unlikely to Benefit from Thrombolysis.  Radiological Society of North America 2011 Scientific Assembly and Annual Meeting, November 26 - December 2, 2011 ,Chicago IL. http://archive.rsna.org/2011/11011477.html