Abstract
Purpose: We test if assessment of attenuation on NCCT identifies similar patients with acute ischemic stroke as the MRI DWI-FLAIR mismatch paradigm.
Methods: Data are from the Keimyung Stroke Registry. Patients with anterior circulation occlusion on baseline CT-angiography, known symptom onset time and MRI within 60 minutes of baseline CT were included. All patients received revascularization therapy. Baseline MRI DWI-FLAIR mismatch, hypoattenuation on baseline NCCT and parenchymal hemorrhage (PH) on follow-up imaging were assessed by consensus. Ratio of ipsilateral/contralateral NCCT Hounsfield Unit (HU) (rCT) within baseline DWI lesion was calculated. Statistical methods were used to assess if CT hypoattenuation was a reliable biomarker of time from stroke symptom onset and if it compared well with DWI-FLAIR mismatch in predicting PH at 24 hours.
Results: Of 127 patients included [median age 68 (IQR-15), 53.5% male, median onset to MR time 158 (94) minutes], DWI-FLAIR mismatch was seen in 85/127 (67%). NCCT hypo-attenuation was seen in 111/127 (87%). A statistically significant negative correlation was noted between rCT and stroke symptom onset to MRI time (Spearman’s r = -0.33, p<0.001). A rCT > 0.87 best predicted the presence of DWI-FLAIR mismatch [c statistic = 0.84 (95% CI 0.77-0.91), sensitivity 73.75% (95%CI 62.71%-82.95%); specificity 76.92% (95%CI 60.67-88.87%)]. Models with CT hypoattenuation were similar to a model with DWI-FLAIR mismatch in ability to discriminate PH (p>0.5).
Conclusion: Degree of hypoattenuation on NCCT can be used in place of DWI-FLAIR mismatch to identify patients with wake up strokes or unknown time of onset.