Determining the Diagnostic Value of Venous Sinus Density Indices in Non-Contrast Brain CT Scan for Early Diagnosis of Cerebral Venous Sinus Thrombosis

Background and Aim

A non-contrast brain CT Non-contrast computed tomography (NCCT) scan is a valuable and cost-effective way to detect cerebral venous sinus thrombosis (CVST) during its acute phase. The goal of this study was to evaluate how effective this diagnostic approach is, including its various density indices, to enable a more precise and timely diagnosis of this debilitating condition.

Method

This retrospective case-control study was conducted on 88 patients with suspected acute CVST. We analyzed NCCT scans of patients with suspected CVST using a Bayesian regression model with a 95% confidence level. We also conducted ROC analysis in R4.1.0 to determine the optimal cut-off point.

Result

We discovered a significant variance in the average sinus attenuation when comparing patients with acute CVST (p < 0.001). By utilizing an optimal cutoff of 61 HU (Hounsfield unit), we achieved sensitivities of 77.1% and specificities of 92.5% for average sinus attenuation. In addition, the optimal cutoff for standardized parameters included the ratios of HU/H (Hematocrit), HU/ICA (internal carotid artery), HU/BA (basilar artery), HU/FRONTAL lobe, HU/TEMPORAL lobe, and HU–BA, which were 1.41, 1.52, 1.63, 1.6, 1.6, and 23, respectively (p < 0.001). Area under the ROC curve for absolute venous Hounsfield was 0.88.

Conclusion

NCCT is a reliable diagnostic approach for acute cerebral venous sinus thrombosis in emergency scenarios. Standardized parameters of absolute Hounsfield unit venous sinus thrombosis increase diagnosis accuracy. It is suggested to use these parameters as a complement to each other for more accurate diagnosis.