High-resolution susceptibility-weighted imaging of clots in cerebral venous thrombosis

Purpose To distinguish cerebral venous clots from patent dural sinuses, cortical veins, and calvarium on high-resolution susceptibility-weighted imaging, since there is lack of a well-designed study in the literature. Methods A retrospective review of 51 consecutive cases and 27 controls was performed with susceptibility-weighted imaging independently by two investigators. MR venography along with MR sequences other than the susceptibility-weighted imaging served as the reference standard. Results There were 29 males and 49 females in the age range of 1 month to 70 years (mean 27 ± 19.8 years). Substantial (k 0.84 to 1.0) to good (k 0.60 to 0.75) inter-rater agreement was observed on the magnitude images for the demonstration of clots in all venous segments, excluding left sigmoid sinus, jugular bulb, and internal jugular vein (k 0.43 to 0.56). Comparatively magnitude images demonstrated better sensitivity (0.83 (0.54–1.13) to 1.0), specificity (0.92 (0.85–0.99) to 1.0), and negative predictive value (0.98 (0.94–1.02) to 1.0) for the detection of clots across all venous segments. Magnitude images showed positive predictive value ranging from 0.83 (0.66–1.0) to 1.0 for clots anywhere from the anterior aspect of superior sagittal sinus to bilateral transverse sinuses. For the detection of clots from bilateral sigmoid sinuses to internal jugular veins, magnitude images showed relatively better positive predictive value ranging from 0.57 (0.20–0.94) to 0.75 (0.45–1.06) in comparison to the processed magnitude and minimum intensity pixel images. Conclusion Susceptibility-weighted imaging can successfully distinguish majority of clots from patent dural sinuses, cortical veins, and calvarium with excellent inter-rater agreements.