Diagnostic and Interventional Imaging (2014) 95, 1145—1150

CONTINUING EDUCATION PROGRAM: FOCUS. . .

Imaging of cerebral venous thrombosis F. Bonneville Department of Neuroradiology, Hôpital Pierre-Paul-Riquet, CHU Purpan, 1, place du Dr-Baylac, 31000 Toulouse, France

KEYWORDS Cerebral venous thrombosis; CT scan; MRI

Abstract Cerebral venous thrombosis (CVT) is a potentially life-threatening emergency. The wide ranging of clinical symptoms makes the use of imaging in ‘‘slices’’ even more important for diagnosis. Both CT and MRI are used to diagnose the occlusion of a venous sinus, but MRI is superior to CT for detecting a clot in the cortical or deep veins. CT can show the hyperintense clot spontaneously and CT angiography the intraluminal defect. MRI also detects this thrombus, whose signal varies over time: in the acute phase, it is hypointense in T2*, whilst T1 and T2 can appear falsely reassuring; in the subacute phase, it is hyperintense on all sequences (T1, T2, FLAIR, T2*, diffusion). MRI easily shows the ischemic damage, even hemorrhagic, in the cerebral parenchyma in cases of CVT. Finally, imaging may reveal pathology at the origin of the CVT, such as a fracture of the skull, infection, tumor, dural fistula, or intracranial hypotension. © 2014 Éditions franc ¸aises de radiologie. Published by Elsevier Masson SAS. All rights reserved.

Cerebral venous thrombosis (CVT) is a rare and potentially life-threatening pathology. It is therefore an emergency that requires the rapid implementation of anticoagulant treatment, but whose diagnosis can be challenging. The range of clinical symptoms and the variability of MRI or CT findings may explain certain diagnostic errors. The radiologist therefore has three objectives: confirm the diagnosis by showing direct signs of occlusion of the venous structure by a clot, assess any damage to the cerebral parenchyma secondary to this thrombosis by looking for signs of venous cerebral ischemia and finally, attempt to demonstrate an origin or a pathology associated with this thrombosis. The aim of this article is to illustrate the various aspects of CVT on CT and MRI.

E-mail address: [email protected] http://dx.doi.org/10.1016/j.diii.2014.10.006 2211-5684/© 2014 Éditions franc ¸aises de radiologie. Published by Elsevier Masson SAS. All rights reserved.

1146

Direct signs of venous occlusion The thrombus The occlusive thrombus itself may be detected on CT and MRI. Its appearance is directly linked to the time that has elapsed between the formation of the clot and imaging. Thus, when it is visible on a CT scan without contrast at the start of the disease, the clot appears as a spontaneous hyperintensity in the form of a venous structure, in general tubular in shape; this is the ‘‘cord’’ sign. This sign is visible in around 20 to 25% of cases [1] and disappears within 1 to 2 weeks [2]. However, spontaneous hyperintensity of the veins is not specific and may be seen in young patients with a high hematocrit, in cases of polycythemia, or in dehydrated patients [3]. In such conditions, the entire intracranial vascular network — both arterial and venous — will appear hyperintense. MRI is more sensitive for revealing the clot due to the combination of sequences and high sensitivity to the magnetic susceptibility of blood degradation products, notably on gradient echo T2-weighted images (T2*) [4]. As with CT scan, the intensity of the signal of the clot at MRI depends on its age. Thus, in the first few days, during the deoxyhemoglobin stage, the clot appears isointense on T1weighted images (WI) and hypointense on T2-WI and FLAIR, thus almost mimicking the normal venous flow signal and creating a potential for diagnostic error (Fig. 1). A venous MRA sequence, but also a T2*, improves diagnostic accuracy,

F. Bonneville since the thrombus is frankly hypointense, artefactual, on this last sequence. Later, during the 2nd week and the presence of methemoglobin, the clot appears hyperintense on all sequences, i.e. T1-WI, T2-WI, and FLAIR, but also in T2* and on diffusion-WI [5] (Fig. 2). It is important to know that dural venous sinuses with normal circulation have an opposite signal on FLAIR and T2*. In the event of thrombosis, they abnormally show the same signal intensity, irrespective of the timing of the MRI exam. In the chronic phase, the clot signal is very variable and depends on the degree of organization of the clot. It is typically isointense on T1-WI, iso-/hyperintense on T2-WI, and hypointense on T2*. It is not always easy to see the clot on MRI, venography may therefore be necessary to confirm the diagnosis and visualize the extension of the venous occlusion (Fig. 3).

Venous occlusion Between CT and MRI, there are numerous ways of performing a non-invasive venography [2,3]. To avoid flow problems, enhanced images are now recommended. CT angiography (Figs. 3 and 4) is a remarkable method for demonstrating the occlusion on raw slices (empty delta sign), on multiplanar reconstructions in thick slices (MIP), and even in 3D (volume rendering), notably for cortical veins [2]. There is no single acquisition protocol for venous CT angiography, but the latter should include an inframillimetric helical acquisition, from the vertex to the foramen magnum, around

Figure 1. Acute venous thrombosis. (a) sagittal T1; (b) Axial FLAIR; (c) Axial T2*. Acute thrombosis may go unnoticed in T1 and FLAIR, as the sinus has a near-normal signal on these sequences. The axial T2 slice in gradient echo helps to narrow-down the diagnosis by showing an unusual frank hypointensity in the occluded venous structures, and notably the superior sagittal sinus. This hypointensity is artefactual, linked at this stage to the presence of deoxyhaemoglobin, and appears larger than the actual size of the sinus itself.

Figure 2. Evolution of the signal of the venous thrombosis at day 7 in the same patient as Fig. 1. (a) sagittal T1; (b) axial FLAIR; (c) Axial T2*. Due to the transformation in extracellular methaemoglobin, the clot is now clearly hyperintense on all of the sequences, including T2*. Thus, in the subacute stage, if the diagnosis does not pose any problems on T1 and FLAIR, T2* can be misleading as the signal of the clot is close to the normally circulating sinus. Note the appearance of capsulo-lenticular parenchymal suffering in FLAIR.

Imaging of cerebral venous thrombosis

1147

Figure 3. Isolated deep vein thrombosis. (a) Axial FLAIR; (b) axial diffusion; (c) axial T2*; (d) CT angiography in sagittal MIP reconstruction. Despite the venous thrombosis of two internal cerebral veins, clearly visible as a hypointensity moulding these structures in T2*, the parenchymal repercussions can only be seen unilaterally. This isolated hyperintensity of the left thalamus, which appears heterogeneous in diffusion, is not easy to diagnose for a venous thrombosis. The latter was confirmed on CT angiography, which clearly shows a thrombosis limited to the deep network, the superior sagittal sinus being patent.

45 seconds after the intravenous injection of 80 to 100 mL of an iodine-based contrast medium. At MRI, the venous MRA sequences with injection of gadolinium are now preferred to 2D TOF sequences or MRA in phase contrast, which are too susceptible to flow artifacts thus resulting in too much signal loss in a small diameter venous sinus or one that is circulating slowly, as it is often the case in the left transverse sinus (Fig. 5). The semiology of enhanced venous MRA is the same as that of CT angiography, showing a filling defect of the occluded venous structure (Fig. 4) and thus producing an empty delta sign. Watch out for Paccioni granulations, which also produce a focal filling defect in a venous sinus, but the latter is generally small, less than 2 cm long, oblong, and preferentially located along the superior sagittal sinus and close to the junction between the transverse and sigmoid sinus [3]. Although short and suspended occlusions are rarely pathological, they can be seen in cases of associated dural fistula. Dynamic 4D MRA, which perfectly demonstrates the entire intracranial venous network, opacifies arteriovenous shunts before the veins [6].

Signs of cerebral venous ischemia CVT is life-threatening because of the cerebral ischemia that it can cause. The venous ischemia starts with a vasogenic

edema that can be accompanied by cytotoxic edema [7]. Hemorrhage via venous stasis and rupture of the blood-brain barrier occurs in around 30 to 50% of venous ischemias [8]. MRI is superior to CT for detecting these different processes described in venous ischemia thanks to the combination of FLAIR, T2*, and diffusion-WI (Fig. 3). Given the pathophysiology, the signal diffusion anomalies are very variable within a same parenchymal lesion of venous ischemia, but note that an eventual hyperintensity with a reduction in the apparent diffusion coefficient has no prognostic value [9]. The lesions therefore often appear hyperintense on T2-WI/FLAIR, hypointense on T1-WI, with no gadolinium uptake, with a cortico-sub-cortical topography, but with no arterial distribution, and often present a clearly visible hypointense hemorrhagic component on T2*, often in the form of flaky, multinodular, sub-cortical zones (Fig. 4). These intraparenchymal hemorrhages may be more compact and appear as a veritable lobar hematoma, causing a space occupying effect with signs of engagement. These forms may be acute and clinically obvious, causing neurological deficits, or even loss of consciousness. Thus faced with a lobar hematoma, one should always remember to examine the venous structures and look for a CVT. A particular form should be noted, whether in terms of clinical presentation or imaging, in cases of thrombosis of the deep venous network and notably of the internal

1148

F. Bonneville

Figure 4. Acute thrombosis of the transverse and left sigmoid sinuses. (a) Unenhanced CT scan; (b) axial slice of a CT angiography; (c) axial FLAIR; (d) axial T2*; (e) venous MRA with contrast. The CT scan shows a typical left superficial temporal venous infarction. Note the occasional foci of sub-cortical haemorrhage within this large area of vasogenic oedema of the white matter. Confronted with this characteristic image, highly evocative of a thrombosis of the underlying transverse and sigmoid sinuses, CT angiography and MRA confirm the diagnosis whilst the FLAIR and T2* slices, despite passing through the plane of the clot, can be falsely interpreted as normal.

Figure 5. Pseudo-occlusion of the left transverse sinus by slow flow, in (a) phase contrast MRA, but clearly permeable on (b) venous MRA with contrast, here in dynamic MRA.

cerebral veins. It should be considered with a bilateral thalamic edematous lesion, bearing in mind that the thrombosis of a single internal cerebral vein is possible (Fig. 3).

Pathologies associated with CVT The causes of CVT are very numerous, including prothrombotic factors (C or S protein deficit), prothrombotic

conditions such as pregnancy and the peripartum period, a neoplastic or infectious condition, dehydration, or even certain drugs. Although focal, intracranial, causes that are visible on cerebral imaging are actually relatively rare, they should be known so that appropriate treatment can be envisaged. Therefore, one needs to look for signs of local infection (sinusitis, meningitis), trauma (fracture passing opposite the venous sinus and in particular the sigmoid sinus), and

Imaging of cerebral venous thrombosis

1149

systemic diseases such as Behcet’s disease, tumor, or even intracranial hypotension.

Conclusion Venous CT angiography enables the diagnosis of CVT in the vast majority of cases, but MRI is still superior to CT for detecting isolated cortical venous thromboses and parenchymal damage. TAKE-HOME MESSAGES • Venous CT angiography is an excellent examination for diagnosing CVT. • MRI is superior to CT for detecting the exceptional isolated thromboses of cortical veins, but also for assessing parenchymal damage in the event CVT. • The signal of the clot in the thrombosed vein changes over time and appears differently on all of the sequences.

Figure 7.

Axial FLAIR.

Figure 8.

Venous MRA in 2D TOF.

Clinical case study This young, 37-year-old woman, with no previous history, has been complaining for several weeks of unusual incapacitating headaches, which only cease once she is lying down. For the past 3 days, the headaches have become more intense, constant and not relieved by simple analgesics. An MRI was requested as an emergency.

Questions 1) Describe the anomalies visible on these different sequences that are in favor of a CVT (Figs. 6—10). 2) List the signs indicative of a triggering cause. 3) What is your diagnosis (topography of the CVT and triggering cause)?

Answers

Figure 6.

Sagittal T1.

1) The MRI clearly shows the occlusion of the transverse venous and right sigmoid sinuses, not enhanced by gadolinium and with no venous TOF signal. 2) The MRI in particular shows signs of intracranial hypotension, as the clinical presentation suggests, with a slightly squashed appearance to the mesencephalo-diencephalic junction, a convex hypophysis, a small ventricular system, continuous thickening of the pachymeninges, a sub-dural collection, and an abnormally turgescent and rounded appearance to the permeable venous sinuses. 3) Thrombosis of the transverse and right sigmoid sinuses with intracranial hypotension. The onset of a venous thrombus in cases of intracranial hypotension is classic but rare, reported in barely 2% of cases. Its pathophysiology is multifactorial, but could be a combination of venous stasis, vascular distortion due to cerebral displacement

1150

Figures 9 and 10.

F. Bonneville

Axial T1 after gadolinium.

in the posterior fossa and increased blood viscosity secondary to depletion of the cerebrospinal fluid (CSF).

Disclosure of interest The author declares that he has no conflicts of interest concerning this article.

References [1] Virapongse C, Cazenave C, Quisling R, Sarwar M, Hunter S. The empty delta sign: frequency and significance in 76 cases of dural sinus thrombosis. Radiology 1987;162: 779—85. [2] Rodallec MH, Krainik A, Feydy A, Hélias A, Colombani JM, Jullès MC, et al. Cerebral venous thrombosis and multidetector CT angiography: tips and tricks. Radiographics 2006;26(Suppl 1):S5—18.

[3] Leach JL, Fortuna RB, Jones BV, Gaskill-Shipley MF. Imaging of cerebral venous thrombosis: current techniques, spectrum of findings, and diagnostic pitfalls. Radiographics 2006;26(Suppl. 1):S19—41. [4] Linn J, Michl S, Katja B, Pfefferkorn T, Wiesmann M, Hartz S, et al. Cortical vein thrombosis: the diagnostic value of different imaging modalities. Neuroradiology 2010;52:899—911. [5] Wasay M, Azeemuddin M. Neuroimaging of cerebral venous thrombosis. J Neuroimaging 2005;15:118—28. [6] Yi˘ git H1, Turan A, Ergün E, Kos¸ar P, Kos¸ar U. Time-resolved MR angiography of the intracranial venous system: an alternative MR venography technique. Eur Radiol 2012;22:980—9. [7] Star M, Flaster M. Advances and controversies in the management of cerebral venous thrombosis. Neurol Clin 2013;31:765—83. [8] Pongmoragot J1, Saposnik G. Intracerebral hemorrhage from cerebral venous thrombosis. Curr Atheroscler Rep 2012;14:382—9. [9] Mullins ME, Grant PE, Wang B, Gonzalez RG, Schaefer PW. Parenchymal abnormalities associated with cerebral venous sinus thrombosis: assessment with diffusion-weighted MR imaging. AJNR Am J Neuroradiol 2004;25:1666—75.

Imaging of cerebral venous thrombosis.

Cerebral venous thrombosis (CVT) is a potentially life-threatening emergency. The wide ranging of clinical symptoms makes the use of imaging in "slice...
2MB Sizes 0 Downloads 9 Views