Cerebral venous sinus thrombosis: Comparison of multidetector computed tomography venogram (MDCTV) and magnetic resonance venography (MRV) of various fi eld strengths

Objectives: To investigate the accuracy of multidetector computed tomography venography (MDCTV) and magnetic resonance venography (MRV) of differing fi eld strength in diagnosing cerebral venous sinus thrombosis (CVST). To assess the visualization rate of the cerebral veins and dural sinuses between these imaging modalities. Methods: A retrospective review of 27 patients with clinical suspicion of CVST who underwent both MDCTV and corresponding MRI/MRV brain scans at 0.35T, 1.5T or 3.0T fi eld-strength were performed. Results were compared with the defi nitive fi nal diagnosis. In the nonthrombotic group of patients, a descriptive study of the anatomical visualization of cerebral veins and dural sinuses was also undertaken. Results: Ten of the 27 patients had a fi nal diagnosis of CVST. The Neuroradiologists’ consensus reading identifi ed these 10 cases of CVST in both the MDCTV and MRV scans with a sensitivity rate of 100%. There was one false positive in MDCTV and three false-positives in the MRV group (1.5T) with positive predictive values of 90.9% and 76.9%, respectively. MDCTV and MRV demonstrated a specifi city of 94.1% and 82.4%, respectively in diagnosing CVST. No false negative images were seen demonstrating a negative predictive value of 100% in both modalities. There was no statistical difference in the anatomical visualization rate of the cerebral veins or the dural sinuses among the various strengths of MRI. However, statistically signifi cant (p<0.05) fewer visualised vessels in MRV compared to MDCTV, for the SSS, ISS, straight sinus, ICV and VOG. Conclusion: MDCTV is equal to MRV of various strength in its sensitivity for CVST diagnosis. MDCTV also provides better anatomical visualization of the dural sinuses and cerebral veins. Neurology Asia 2012; 17(4) : 281 – 291 Address correspondence to: Assoc Professor Dr Kartini Rahmat, Department of Biomedical Imaging, Faculty of Medicine, University of Malaya, 50603 Kuala Lumpur, Malaysia. Tel: +603-79492091, Fax: +603-79494603, e-mail: [email protected] INTRODUCTION Cerebral venous sinus thrombosis (CVST) often presents with non-specifi c symptoms, hence, the diagnosis can be elusive. The main presenting symptom in 70 to 90% of cases is headache, which is often misdiagnosed as migraine or tension headaches.1 Propagation of thrombus would cause venous infarcts resulting in more specifi c focal neurological defi cits, such as hemiparesis, paraesthesia, seizures or altered level of consciousness (in 33-66% of patients). Generally, by the time diagnosis is made, an interval of 7 days has elapsed from the onset of symptoms.1 The gold standard for diagnosing dural sinus thrombosis has been cerebral angiogram.1-3 The commonly accepted method, possibly considered the gold standard for non-invasive techniques in diagnosing dural venous thrombosis is magnetic resonance venography (MRV).1,2,4,5 However, the newer MDCT scans, which are capable of performing helical scans at high speed coinciding with peak venous enhancement, is also able to provide good diagnostic image quality. Thus, the idea of using MDCTV to rule out dural venous thrombosis is compelling as it is fast and more readily available. Moreover, MRV is prone to motion artefact due to its long acquisition time and susceptible to fl ow artefact in the 2D-TimeOf-Flight (TOF) sequences. Another imaging technique of MRV, which is the 2D-phase contrast sequence, displays limited 03 Cerebral venous.indd 281 12/14/12 12:07:54 PM Neurology Asia December 2012 282 view as it only acquires a single midline image of sagittal plane as compared to the full volume acquisition in MDCTV. To our knowledge there has been no study to compare the sensitivity and specifi city in diagnosing CVST between 16-sliced MDCTV against various fi eld-strengths of 0.35T, 1.5T and 3.0T MRV as undertaken in the present investigation This study also includes a descriptive component to compare the visualisation rate of dural venous sinus anatomy between these modalities (Figures 1 and 2).