Abstract

Accurate assessment of dural sinus, deep and cortical venous thrombosis on MR imaging is challenging. The aim of this study is to evaluate the accuracy of 3D-T1 turbo spin echo (T1S), sequences in detecting venous thrombosis and comparing it with susceptibility-weighted imaging (SWI), magnetic resonance venography (MRV) and post contrast T1 magnetization-prepared rapid acquisition gradient echo (T1C). A blinded retrospective observational analysis of 71 consecutive patients evaluated for cerebral venous thrombosis (CVT) and 30 control patients was performed. Multimodality reference standard adopted included T1C, SWI with MRV. Sub-analyses in superficial, deep and cortical venous segments were performed in addition to correlation of signal intensity of thrombus with the clinical stage. A total of 2222 segments in 101 complete MRI examinations were evaluated. Sensitivity/specificity/positive predictive value/negative predictive value/accuracy and precision of T1S for detection of cortical vein thrombosis was 0.994/1/1/0.967/0.995/1, 1/0.874/0.949/1/0.963/0.950 for detection of superficial venous sinus thrombosis and 1/1/1/1/1/1 for deep venous thrombosis. The AUC yield for T1S was 0.997 for cortical, 1 for deep and 0.988 for superficial venous segments. T1S paralleled the accuracy of conventional sequences in the overall detection of CVT but showed superior accuracy in the detection of cortical venous thrombosis. It makes a fitting addition to the CVT MRI protocol in scenarios demanding negation of gadolinium administration.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call