Abstract W P3: Core Infarct Size Agreement Between NCCT And DWI: Implications For Patient Selection For IAT
Background: Pre-treatment infarct volume is an important determinant of outcome after intra-arterial therapy (IAT). Ongoing IAT trials are using MRI DWI and noncontrast CT (NCCT) to quantify core infarction on the Alberta Stroke Program Early CT Score (ASPECTS) for patient selection. However, the degree to which these imaging approaches agree remains uncertain.
Methods: Consecutive acute ischemic stroke patients over a 2-year period were included in this analysis if they had occlusion of the intracranial ICA or MCA M1 segment and underwent CT and MRI within 1 hour of each other and within 8 hours of onset. Two raters independently scored ASPECTS on both NCCT and DWI in separate reading sessions. Modified DWI-ASPECTS (including only >20% regional involvement) was also scored. Differences were resolved by consensus, and the consensus scores were compared across the three scales using Bland-Altman analysis. Agreement was also assessed based on a dichotomized ASPECTS threshold of 0-5 vs. 6-10 using the kappa statistic.
Results: There were 74 patients with mean age 69.6 years and median NIHSS 17. Occlusions involved the ICA in 27 (36 %) patients. Mean interval between CT and MRI was 37±15 minutes. The median (IQR) ASPECTS scores were 7 (4-8) on NCCT, 5 (3-7) on DWI (P<0.0001 vs. NCCT), and 7 (4-8) on modified DWI (P=0.42 vs. NCCT). In Bland-Altman analysis, NCCT ASPECTS was mean 1.4 points (95%CI: -1.9 to 4.7 points) higher than DWI ASPECTS, and was mean 0.2 points (95%CI: -3.4 to 3.9 points) higher than modified DWI ASPECTS. For NCCT, DWI, and modified DWI, ASPECTS 6-10 scores were seen in 69%, 41%, and 60% of patients, respectively. The inter-scale agreement for ASPECTS >5 was only moderate between DWI and NCCT (kappa=0.47) and good between modified DWI and NCCT (kappa=0.68). However, differences between modified DWI and NCCT for dichotomized ASPECTS were found in 15% of patients.
Conclusion: Using ASPECTS, modified DWI has better agreement with NCCT than standard DWI evaluation. However, there is a sizeable proportion of patients where dichotomized ASPECTS differs even with the use of modified DWI. This has important implications for patient selection in clinical trials.
Author Disclosures: A.J. Yoo: Research Grant; Significant; Penumbra Inc.. Other Research Support; Modest; Dutch Heart Foundation. S. Kamalian: None. S.A. Sheth: None. F. Mehrkhani: None. M.H. Lev: Research Grant; Significant; GE Healthcare. Consultant/Advisory Board; Modest; Millenium Pharmaceuticals. D.S. Liebeskind: Consultant/Advisory Board; Modest; Stryker, Covidien. Research Grant; Significant; NIH-NINDS.
- © 2015 by American Heart Association, Inc.