e-ASPECTS derived acute ischemic volumes on non-contrast-enhanced computed tomography images

Background and purpose Validation of automatically derived acute ischemic volumes (AAIV) from e-ASPECTS on non-contrast computed tomography (NCCT). Materials and methods Data from three studies were reanalyzed with e-ASPECTS Version 7. AAIV was calculated in milliliters (ml) in all scored ASPECTS regions of the hemisphere detected by e-ASPECTS. The National Institute of Health Stroke Scale (NIHSS) determined stroke severity at baseline and clinical outcome was measured with the modified Rankin Scale (mRS) between 45 and 120 days. Spearman ranked correlation coefficients (R) of AAIV and e-ASPECTS scores with NIHSS and mRS as well as Pearson correlation of AAIV with diffusion-weighted imaging and CT perfusion-estimated ischemic "core" volumes were calculated. Multivariate regression analysis (odds ratio, OR with 95% confidence intervals, CI) and Bland-Altman plots were performed. Results We included 388 patients. Mean AAIV was 11.6 +/- 18.9 ml and e-ASPECTS was 9 (8-10: median and interquartile range). AAIV, respectively e-ASPECTS correlated with NIHSS at baseline (R = 0.35, p < 0.001; R = -0.36, p < 0.001) and follow-up mRS (R = 0.29, p < 0.001; R = -0.3, p < 0.001). In subsets of patients, AAIV correlated strongly with diffusion-weighted imaging (n = 37, R = 0.68, p < 0.001) and computed tomography perfusion-derived ischemic "core" (n = 41, R = 0.76, p < 0.001) lesion volume and Bland-Altman plots showed a bias close to zero (-2.65 ml for diffusion-weighted imaging and 0.45 ml forcomputed tomography perfusion "core"). Within the whole cohort, the AAIV (OR 0.98 per ml, 95% CI 0.96-0.99) and e-ASPECTS scores (OR 1.3, 95%CI 1.07-1.57) were independent predictors of good outcome Conclusion AAIV on NCCT correlated moderately with clinical severity but strongly with diffusion-weighted imaging lesion and computed tomography perfusion ischemic "core" volumes and predicted clinical outcome.