Abstract WMP8: Thin-sliced Reformatted CT Imaging: Validation of a Safer and Faster Technique to Assess Clot Size and Vessel Diameter in Large Vessel Occlusions

Abstract

Background: Thin-sliced reformatted non-contrast CT head are not widely used to detect the thrombus. The purpose of this study was to investigate whether thin-sliced reformatted non-contrast CT scans could be reliably used to detect and measure size of the thrombus size in patients with acute ischemic stroke due to LVO and therefore serve as a potential substitute of CT Angiogram (CTA). Success of this paradigm could reduce stroke decision algorithm to nonenhanced CT scan before a code neurointervention is activated; hence saving time and contrast. Methods: Comprehensive prospective analysis of patients who underwent acute endovascular intervention for LVOs at a community based, university affiliated comprehensive stroke center during one year period (Jan 2015-Dec 2015) was done. The raw data of non-enhanced CT scans and CTAs were collected. All raw data were reconstructed with thin slices of 0.625 mm using standard GE software. Two observers independently evaluated the 5-mm maximum intensity projections of the thin slices and CTA in coronal and axial projections for best assessment of vessel diameter and thrombus length. Inter-observer agreement was measured using Cohen κ. Results: There were 749 patients who presented with acute ischemic stroke during the specified time period. Of those 67 were large vessel circulation strokes; of which, 22 had both CT and CTA done. Mean clot length measured was 14.75 (SD +/-4.95) on thin slices CT and 15.02 (SD +/-5.47) on CTA. Vessel diameter measured was 2.77 (SD +/-0.47) on thin sliced reformatted CT and 2.41 (SD +/-0.49) on CTA. There were no instances where clot size or vessel diameter could not be measured on thin slice reformatted CT. Inter-observer agreement was higher for both CTA (κ, 0.83) versus thin-slice nonenhanced CT reconstructions (κ, 0.80). Conclusions: Thin-sliced reconstructions of standard cranial nonenhanced CT raw data can be reliably used to detect and measure the thrombus size in LVOs. It also reliably measures the vessel diameter, making intervention planning possible. Larger multicenter trials are needed to validate our data.