Diagnostic accuracy of computed tomography in acute ischemic stroke: A retrospective evaluation of sensitivity, specificity, and time-dependent performance

Abstract

Computed tomography (CT) is a widely accessible and rapid imaging modality for acute ischemic stroke diagnosis, but its sensitivity in early stages remains suboptimal compared to diffusion-weighted MRI (DWI-MRI). This study aimed to evaluate the diagnostic performance of CT in comparison with DWI-MRI in acute ischemic stroke. A retrospective, single-center study was conducted, including 80 patients with suspected acute ischemic stroke and 30 controls who underwent both CT and DWI-MRI within 24 hours of presentation. Gray-white matter density differences and the hyperdense middle cerebral artery (HD-MCA) sign were analyzed to enhance CT’s diagnostic accuracy. Sensitivity, specificity, and the Area Under the Curve (AUC) were calculated for CT across four time intervals (0-6h, 6-12h, 12-24h, 24-48h). The overall diagnostic accuracy of CT (AUC=0.717) was moderate, with sensitivity and specificity values of 69.6% and 50% in the 0-6-hour interval. CT performance improved in the 12-24-hour interval, with sensitivity and specificity reaching 68.8% and 66.7%, respectively. The HD-MCA sign was observed in 18.8% of patients and showed 100% specificity but low sensitivity (30%). Gray-white matter density measurements significantly improved the diagnostic objectivity of CT, particularly in early stages. CT remains a crucial tool for the early evaluation of stroke due to its accessibility and rapid imaging capabilities. However, integrating objective measurements such as gray-white matter density ratios and complementary imaging modalities like DWI-MRI can enhance its diagnostic performance. Future studies should focus on prospective, multi-center designs with larger patient cohorts, potentially integrating artificial intelligence, to refine these strategies. Keywords: Acute ischemic stroke; Computed tomography; Diffusion-weighted MRI; Diagnostic accuracy