Stroke Evaluation in the Interventional Suite Using Dual-Layer Detector Cone-Beam CT: a First-in-human Prospective Cohort Study (the Next Generation X-ray Imaging System Trial)

Abstract

Abstract Purpose Cone-beam CT in the interventional suite could be an alternative to CT to shorten door-to-thrombectomy time. However, image quality in cone-beam CT is limited by artifacts and poor differentiation between gray and white matter. This study compared non-contrast brain dual-layer cone-beam CT in the interventional suite to reference standard CT in stroke patients. Methods A prospective single-center study enrolled consecutive participants with ischemic or hemorrhagic stroke. The hemorrhage detection accuracy, per-region ASPECTS accuracy and subjective image quality (Likert scales for gray-white matter differentiation, structure perception and artifacts) were assessed by three neuroradiologists blinded to clinical data on dual-layer cone-beam CT 75 keV monoenergetic images compared to CT. Objective image quality was assessed by region-of-interest metrics. Non-inferiority for hemorrhage detection and ASPECTS accuracy was determined by the exact binomial test with a one-sided lower performance boundary prospectively set to 80% (98.75% CI). Results 27 participants were included (74 years ± 9; 19 female) in the hyperacute or acute stroke phase. One reader missed a small bleeding, but all hemorrhages were detected in the majority analysis (100% accuracy, CI lower boundary 86%, p = 0.002). ASPECTS majority analysis showed 90% accuracy (CI lower boundary 85%, p < 0.001). Sensitivity was 66% (individual readers 67%, 69%, and 76%), specificity was 97% (97%, 96%, 89%). Subjective and objective image quality were inferior to CT. Conclusion In a small single-center cohort, dual-layer cone-beam CT showed non-inferior hemorrhage detection and ASPECTS accuracy to CT. Despite inferior image quality, the technique may be useful for stroke evaluation in the interventional suite. Trial Registration Number NCT04571099 (clinicaltrials.gov). Prospectively registered 2020-09-04.