Abstract 96: CTP-Mismatch Maps Improve Interobserver Agreement
Introduction: The use of mismatch between the ischemic core and penumbra to select patients who are likely to benefit from acute stroke therapy has gained popularity. Interpretation of the ischemic core and penumbra on standard CT-perfusion (CTP) maps is subjective. This may lead to variability among physicians in the decision if a patient is a good candidate for acute stroke therapy. A CTP-Mismatch map with outlines of the ischemic core and penumbra could limit this variability. The goal of this study was to determine if inter-observer agreement regarding a patient’s suitability for acute stroke therapy improves with the use of a CTP-Mismatch map. The figure shows a typical CTP-Mismatch map.
Methods: Ninety-six consecutive patients evaluated with CTP prior to intra-arterial therapy at St. Lukes Hospital in 2008-09 were included. 79 patients had adequate quality CTP for this analysis. Standard CTP maps (CBV, CBF, MTT, and Tmax) and a CTP-Mismatch map were generated with a fully automated program for processing of CTP source images (RAPID). RAPID assessed the ischemic core using a CBF threshold <30% of the contralateral hemisphere (rCBF<30%). The ischemic penumbra was defined by a Tmax threshold of >6 sec (Tmax>6s). The standard CTP maps and the CTP-Mismatch map were independently analyzed by two vascular neurologists in a blinded fashion. The raters assessed a patient's suitability for intra-arterial therapy based on the following mismatch criteria: (1) a ratio between (Tmax>6s) and (rCBF<30%) volumes >1.8 and (2) an absolute difference between (Tmax>6s) and (CBF<30%) volumes >15ml. Interobserver reliability was assessed with Cohen’s kappa.
Results: When assessment of suitability for intra-arterial therapy was based on interpretation of standard CTP maps, the two raters agreed in 58 of 79 patients (kappa=0.46; 95% CI=0.24-0.60). The agreement between observers improved when suitability was determined using CTP-Mismatch maps (agreement in 76 of 79 cases; kappa=0.92; 95% CI=0.75-0.92; p<0.001 for difference between kappa values). The 3 cases with inter-observer disagreement had artifact on the CTP-Mismatch map. Following concensus adjudication of these 3 cases, 40 of the 79 patients (51%) were deemed suitable candidates for acute stroke therapy.
Conclusion: CTP-Mismatch maps with estimates of ischemic core and penumbra volumes markedly improve inter-observer agreement regarding assessment of suitability for acute stroke therapy. Such maps, which can be generated automatically, may help standardize decision making algorithms for evaluation of potential intra-arterial therapy candidates.
- © 2012 by American Heart Association, Inc.