Neuroradiology/Head and Neck Imaging
Review
Bone Subtraction 3D CT Venography for the Evaluation of Cerebral Veins and Venous Sinuses: Imaging Techniques, Normal Variations, and Pathologic Findings
OBJECTIVE. Varying anatomic characteristics and clinical and radiologic manifestations are diagnostic challenges in the evaluation of the cerebral vein and of venous sinus diseases. The purpose of this article is to introduce bone subtraction CT venography and review normal variations and diseases involving the cerebral veins and venous sinuses.
CONCLUSION. Knowledge of the normal variations and pathologic findings will be helpful for the accurate diagnosis of diseases involving the cerebral venous system. Bone subtraction CT venography offers complete 3D visualization of the cerebral venous system and can be useful for the evaluation of the cerebral vein and venous sinus diseases.
Keywords: bone subtraction, cerebral venous sinus, CT venography
Varying anatomic characteristics and clinical and radiologic manifestations are diagnostic challenges in the evaluation of the cerebral vein and venous sinus diseases. Digital subtraction angiography (DSA) has been widely used for evaluation of the cerebral venous system. However, it is an invasive technique with risks [1]. Furthermore, compared with imaging of the cerebral arteries, 3D visualization of the cerebral veins and venous sinuses is not easy with DSA because the starting point of scanning and maintenance of sinus opacification during scanning should be individually adapted.
The cerebral veins and venous sinuses are complex in their anatomy. Complete 3D visualization of the cerebral venous system is helpful for accurate diagnosis and appropriate treatment of cerebral venous and venous sinus diseases. With rapid scan time and the wide range of scan coverage, MDCT can depict the entire cerebral venous system with high spatial resolution. However, 3D visualization of the cerebral venous system with conventional CT venography (CTV) is not always possible because most of the cerebral veins and venous sinuses are located close to the skull, and separating the cerebral veins and venous sinuses from the skull is usually difficult. Bone subtraction technique may be an effective, rapid, and semiautomated procedure that improves differentiation between the bones and vessels [2]. In this article, we introduce bone subtraction CTV technique and illustrate various cases, including normal variations and diseases involving the cerebral veins and venous sinuses.
For bone subtraction CTV there are two kinds of bone removal technique. One is conventional bone subtraction CTV [3], and the other is dual-energy bone removal CTV [4]. Conventional bone subtraction CTV requires both unenhanced and enhanced CTV datasets. Additional radiation exposure for unenhanced CT is one of the limitations of conventional bone subtraction CTV. By use of one-half or one-fourth the tube current of unenhanced CT, radiation exposure can be reduced without substantial loss of image quality [5]. Another problem of conventional bone subtraction CTV is that patient movement between the two acquisitions can confound the subtraction process, resulting in incomplete bone removal. However, through the registration process, motion artifact due to minor patient movement can be corrected because most commercially available software has a motion correction feature.
By exploiting the differences in attenuation of iodine and calcium at different x-ray energies due to their different atomic numbers, dual-energy CT has theoretic potential for differentiating vessels filled with contrast medium from bones or calcified plaques. The CT values for vessels filled with iodine nearly double with the decrease from 140 to 80 kV, whereas bony structures have a markedly smaller increase in CT values with lower tube voltage [6]. Dual-energy CT allows simultaneous acquisition of low- and high-energy data in one examination and thus allows simultaneous imaging without inter-scan motion and with the use of only a small additional radiation dose [4]. For the acquisition of CTV images of diagnostic quality, it is important to select adequate scanning parameters, injection rate, and total volume of contrast medium. Adequate scan delay after the contrast injection is also important. Our scan protocol is shown in Table 1, and we obtain CTV images of diagnostic quality in most cases.
The superior sagittal sinus (SSS) originates near the crista galli in the anterior aspect and extends posteriorly to its confluence with the straight and transverse sinuses [7]. In occasional cases the anterior SSS is absent. In this variant, the SSS forms in a more posterior location by the union of several prominent superior cortical draining veins [8] (Fig. 1). This anomaly is seen in 6–7% of cases and should not be mistaken for pathologic occlusion [9].
![]() View larger version (308K) | Fig. 1A —33-year-old man with segmental aplasia of superior sagittal sinus and developmental venous anomaly (DVA). A, Contrast-enhanced T1-weighted MR image shows DVA (arrow) in left sylvian fissure. |
![]() View larger version (182K) | Fig. 1B —33-year-old man with segmental aplasia of superior sagittal sinus and developmental venous anomaly (DVA). B, CT venogram shows segmental aplasia of superior sagittal sinus (short arrows) and DVA (long arrow). |
Isolated absence or hypoplasia of part or all of a transverse sinus is common [10]. This normal variant can usually be differentiated from pathologic occlusion by the lack of enlarged collateral pathways and associated brain parenchymal abnormality. Hypoplasia of the ipsilateral jugular foramen also is important evidence of a hypoplastic dural sinus.
A jugular bulb located above the bony annulus of the temporal bone is referred to as a high-riding jugular bulb. A jugular bulb may have an upward-pointing dilatation, called a jugular diverticulum (Fig. 2). A jugular diverticulum and a high-riding bulb may cause pulsatile tinnitus and present clinically as a vascular-appearing retrotympanic mass [8].
![]() View larger version (347K) | Fig. 2A —70-year-old woman with jugular diverticulum and meningioma. A, Contrast-enhanced T1-weighted MR image shows strongly enhancing mass at right parietal falx (arrow). |
![]() View larger version (293K) | Fig. 2B —70-year-old woman with jugular diverticulum and meningioma. B, CT venogram shows focal narrowing of superior sagittal sinus by tumor invasion (short arrow) and upward pointing dilatation of left jugular bulb (jugular diverticulum) (long arrow). |
Duplication or fenestration of the SSS is a rare condition (Fig. 3). Although variations of the proximal SSS are common and have been described, only a few cases of a duplicated SSS with or without an associated parietal encephalocele have been published [11, 12]. Duplication of the straight sinus has rarely been reported in spite of its relatively frequent occurrence. Goto et al. [13] reported that they observed two varieties of duplication of the straight sinus in 7.3% of 700 cases. In 40 cases, two straight sinuses were arranged vertically in the midline (5.7%), whereas the straight sinuses were arranged horizontally or side by side in 11 cases (1.6%).
![]() View larger version (283K) | Fig. 3 —67-year-old man with fenestration of superior sagittal sinus. CT venogram shows fenestration in distal portion of superior sagittal sinus (arrow). |
Arachnoid granulations (AG) are CSF-filled meningothelium-lined protrusions that extend into the venous sinuses through openings in the dura [14]. AG are observed in two thirds of cadaveric specimens and are most commonly found in the SSS and transverse sinus [15–17]. The reported frequency of AG on images is variable, ranged from 0.3% to 55% [16–18]. Despite the common finding of AG within the SSS in previous anatomic studies [17, 18], most of the filling defects seen at CT and MRI examinations are present within the transverse sinuses [16, 18] (Fig. 4). Occasionally, AG can exceed 1 cm in diameter (so-called giant AG) and should not be mistaken for thrombus. Thrombosis usually involves an entire segment of a sinus or multiple sinuses and can extend into cortical veins [19]. AG produce focal well-defined defects. The attenuation of AG is CSF-like on CT images, and the signal intensity parallels that of CSF signal intensity on both T1- and T2-weighted images [18].
![]() View larger version (426K) | Fig. 4A —55-year-old woman with arachnoid granulation and mixed (cavernous-venous) angioma. A, Contrast-enhanced T1-weighted MR image shows filling defect (arachnoid granulation) (arrow) in right transverse sinus. |
![]() View larger version (403K) | Fig. 4B —55-year-old woman with arachnoid granulation and mixed (cavernous-venous) angioma. B, T2-weighted MR image shows hemorrhagic lesion with peripheral hemosiderin rim (arrow) in right frontal lobe. |
![]() View larger version (263K) | Fig. 4C —55-year-old woman with arachnoid granulation and mixed (cavernous-venous) angioma. C, Volume-rendered basal (C) and lateral oblique (D) CT venograms show arachnoid granulation (short arrow). Venous anomaly (long arrow) in right frontal region drains into superior sagittal sinus. |
![]() View larger version (306K) | Fig. 4D —55-year-old woman with arachnoid granulation and mixed (cavernous-venous) angioma. D, Volume-rendered basal (C) and lateral oblique (D) CT venograms show arachnoid granulation (short arrow). Venous anomaly (long arrow) in right frontal region drains into superior sagittal sinus. |
The occipital sinus is a small venous channel situated near the attachment of the falx cerebelli [20]. The occipital sinus is found in 37.7% of patients at contrast-enhanced MR venography (MRV) [21] and in 64.5% of patients at autopsy [22]. The occipital sinus may be solitary, duplicated, or composed of a mesh of venous collaterals [23] and usually communicates with the confluence of sinuses in the cranial aspect and with the vertebral venous plexus and the marginal sinus in the caudal aspect [20]. It occasionally also communicates with the jugular bulb, sigmoid sinus, transverse sinus, SSS, or straight sinus [22]. An occipital sinus that drains into the sigmoid sinus is termed an oblique occipital sinus [24] (Fig. 5).
![]() View larger version (321K) | Fig. 5 —68-year-old woman with occipital sinus. Volume-rendered CT venogram shows hypoplastic right transverse sinus (short arrows). Occipital sinus is prominent and drains into bilateral distal sigmoid sinus (long arrow). |
The falcine sinus is a midline dural venous structure located in the falx cerebri connecting the vein of Galen or the inferior sagittal sinus with the SSS. The falcine sinus is an embryonic vessel that normally involutes after birth [25]. In previous reports, most persistent falcine sinuses were associated with congenital disorders, such as malformation of the vein of Galen, absence of the corpus callosum, acrocephalosyndactyly, and Chiari II malformation. In adults, falcine sinus can be associated with acquired occlusion of the straight sinus [26].
Autopsy results have shown that the falcine plexus is common in the falx cerebri and that the inferior sagittal sinus communicates with the SSS via the falcine venous plexus [27]. Recanalized falcine sinuses probably represent an increase in caliber of one of the channels of the normal falcine venous plexus secondary to acquired occlusion of the venous sinus, to preserve venous drainage [28]. Falcine sinuses are known to be extremely rare in adults [26]. However, a 2010 study of a large series of patients examined with CTV showed that persistent falcine sinus was not rare in an adult population (2.1%). Most of these cases were not associated with congenital anomalies or acquired sinus occlusion [28] (Fig. 6).
![]() View larger version (314K) | Fig. 6 —70-year-old man with persistent falcine sinus. Sagittal 2D reformation CT venogram shows persistent falcine sinus (arrows). Size of straight sinus is normal. |
AVMs are a complex network of abnormal vascular connections that consists of one or more arterial feeders, a nidus, and enlarged venous outflow channels [29]. Although they are found incidentally during neuroimaging, AVMs are an important cause of nontraumatic intracranial hemorrhage in both adults and children. Imaging modalities such as dynamic MR angiography and dynamic CT angiography are increasingly being used in the detection of intracranial pathologic vessels [29]. Bone subtraction CTV can be a useful noninvasive imaging tool for the evaluation of AVMs (arterial feeders, nidus, and draining veins) (Fig. 7). However, in many cases in which surgical or endovascular treatment is considered, DSA is needed for the evaluation of AVMs because it has higher spatial and time resolution than CTV.
![]() View larger version (304K) | Fig. 7A —10-year-old boy with arteriovenous malformation (AVM). A, Volume-rendered anterior oblique (A) and lateral (B) CT angiograms and anterior oblique (C) and lateral (D) digital subtraction angiograms show AVM in left frontoparietal region. For delineation of AVM nidus, feeding arteries, and draining veins, CT angiography is comparable to digital subtraction angiography. |
![]() View larger version (304K) | Fig. 7B —10-year-old boy with arteriovenous malformation (AVM). B, Volume-rendered anterior oblique (A) and lateral (B) CT angiograms and anterior oblique (C) and lateral (D) digital subtraction angiograms show AVM in left frontoparietal region. For delineation of AVM nidus, feeding arteries, and draining veins, CT angiography is comparable to digital subtraction angiography. |
![]() View larger version (281K) | Fig. 7C —10-year-old boy with arteriovenous malformation (AVM). C, Volume-rendered anterior oblique (A) and lateral (B) CT angiograms and anterior oblique (C) and lateral (D) digital subtraction angiograms show AVM in left frontoparietal region. For delineation of AVM nidus, feeding arteries, and draining veins, CT angiography is comparable to digital subtraction angiography. |
![]() View larger version (289K) | Fig. 7D —10-year-old boy with arteriovenous malformation (AVM). D, Volume-rendered anterior oblique (A) and lateral (B) CT angiograms and anterior oblique (C) and lateral (D) digital subtraction angiograms show AVM in left frontoparietal region. For delineation of AVM nidus, feeding arteries, and draining veins, CT angiography is comparable to digital subtraction angiography. |
Development venous anomalies (DVAs), previously known as venous angiomas, are the most common of all intracranial vascular malformations. They are considered extreme variations of the medullary venous drainage rather than a true malformation [30]. They may be associated with other cerebral vascular malformations, including AVM, cavernous angioma, and capillary telangiectasia, especially cavernous angioma [31]. DVAs consist of radially arranged dilated veins (so-called caput medusae) that drain into an enlarged transcortical or subependymal collecting vein. DVAs are usually solitary and asymptomatic and are found incidentally during routine brain imaging [30]. At DSA, arterial phase findings are normal in nearly all cases of the DVA. Opacification during the venous phase of angiography concomitantly with the normal cerebral vein is the pathognomic angiographic appearance of a DVA. The anatomic features of a DVA, including both collecting vein and the caput medusae, is well visualized with bone subtraction CTV [31] (Fig. 1).
Mixed angiomas can be defined as a combination of two or more of the cerebrovascular malformations histologically distinguishable in separate regions of the same lesion. Mixed cavernous-venous angioma is the most common combination (10–30% of cases) [32]. At gross examination, cavernous angiomas are discrete multilobulate lesions that contain blood-filled cysts with hemorrhage in various stages of evolution [33]. Because cavernous angiomas are slow-flow lesions with no large arterial feeders, arterialized veins, or large venous outflow vessels, they are angiographically normal [34]. Cavernous angiomas have therefore been classified in the so-called angiographically occult or cryptic vascular malformations category [35]. A mixed cavernous-venous angioma may have the typical imaging findings of both DVA and cavernous angioma: enhancing vascular structure on contrast-enhanced images as a finding of DVA, a popcorn ball appearance with a hypointense hemosiderin rim on T2-weighted MR images, and high-attenuation hemorrhagic lesions on unenhanced CT images as a finding of cavernous angioma (Fig. 4).
Dural arteriovenous fistulas (DAVFs) are composed of numerous abnormal direct communications between dural arteries and dilated dural veins without an intervening capillary bed [36]. Although DAVFs can arise anywhere in the dura mater covering the brain, they occur most commonly in the transverse-sigmoid and cavernous sinuses, followed by the SSS and anterior skull base (Fig. 8). Patients may have no symptoms or may experience a range of conditions from mild symptoms to fatal hemorrhage. Venous drainage patterns are most predictive of the risk of aggressive symptoms [37, 38] and provide a foundation for the classification schemes of Borden et al. [39] and Cognard et al. [37]. The type of DAVF correlates well with clinical outcome. Stenosis, thrombosis, or ectasis of venous drainage and cortical venous drainage can be predictive of poor prognosis because they may cause venous congestion and eventually cause venous infarction or hemorrhage [40]. Although bone subtraction CTV usually depicts stenosis, thrombosis, and ectasis of the venous drainage system, it has limitations for the evaluation of feeding arteries because DAVFs are usually fed by multiple small arteries (Fig. 8).
![]() View larger version (334K) | Fig. 8A —53-year-old man with ethmoidal-dural arteriovenous fistula (AVF). A, Unenhanced CT image shows hematoma in right frontal lobe and intraventricular hemorrhage. |
![]() View larger version (268K) | Fig. 8B —53-year-old man with ethmoidal-dural arteriovenous fistula (AVF). B, CT angiogram (B) and digital subtraction angiogram (C) show dural AVF in right frontal region, which is fed by ethmoidal branch of ophthalmic artery (arrow) and drains into superior sagittal sinus. C clearly shows both feeding artery and draining vein. B, however, is limited for evaluation of feeding artery. |
![]() View larger version (366K) | Fig. 8C —53-year-old man with ethmoidal-dural arteriovenous fistula (AVF). C, CT angiogram (B) and digital subtraction angiogram (C) show dural AVF in right frontal region, which is fed by ethmoidal branch of ophthalmic artery (arrow) and drains into superior sagittal sinus. C clearly shows both feeding artery and draining vein. B, however, is limited for evaluation of feeding artery. |
Cerebral venous sinus thrombosis (CVST) is difficult to diagnose if it is not considered in the differential diagnosis as a possible cause of neurologic deterioration. CVST is an important cause of stroke among young adults [41]. In rare instances, subarachnoid hemorrhage can occur [42]. Direct visualization of thrombosis may show a hyperdense clot sign in the dural sinus and the cord sign, which is seen as a linear area of high attenuation of cortical vein on unenhanced CT images. On contrast-enhanced CT images, thrombus appears as a filling defect with surrounding dural enhancement, the so-called empty delta sign. On MR images, thrombus can be seen as hyperintensity at the subacute stage of hemorrhage on T1-weighted images [19]. Thrombus may also become evident as absence of a flow void in the involved dural sinus on T2-weighted images. On DSA, MRV, and CTV images, thrombus appears as luminal filling defects, whereas a completely occluded sinus appears as an empty channel in the expected location of the normal sinus (Fig. 9). The differential diagnosis of CVST includes giant AG and subdural hematoma [43]. Aplasia or hypoplasia of the sinus and flow-related artifacts, such as loss of flow void and signal loss, on both conventional MR and MRV images can also mimic CVST.
![]() View larger version (265K) | Fig. 9A —48-year-old woman with extensive venous sinus thrombosis. A, Unenhanced CT image shows multiple hematomas in right parietal lobe (arrow). |
![]() View larger version (255K) | Fig. 9B —48-year-old woman with extensive venous sinus thrombosis. B, Volume-rendered lateral oblique (B) and basal (C) CT venograms show multiple segmental occlusions due to venous sinus thrombosis in superior sagittal (large arrows, B), inferior sagittal (small arrows, B), and right transverse (arrows, C) sinuses. |
![]() View larger version (280K) | Fig. 9C —48-year-old woman with extensive venous sinus thrombosis. C, Volume-rendered lateral oblique (B) and basal (C) CT venograms show multiple segmental occlusions due to venous sinus thrombosis in superior sagittal (large arrows, B), inferior sagittal (small arrows, B), and right transverse (arrows, C) sinuses. |
Isolated cortical vein thrombosis (ICVT) is extremely rare and difficult to diagnose. The presenting features may be different from those of CVST. Headache and impairment of consciousness are less prominent in ICVT, presumably because of the absence of intracranial hypertension [44, 45]. The imaging findings of ICVT are characterized by localized signal-intensity abnormalities of the cortical and subcortical locations on MR images. Focal intracerebral hematoma or isolated subarachnoid hemorrhage in the cerebral sulci may also be a primary or secondary finding of ICVT (Fig. 10). Because the cortical veins are variable in number, size, and location, conventional MRI and MRV may miss ICVT. Although DSA findings may confirm the diagnosis as a filling defect in the cortical vein, ICVT usually has indirect signs, such as collateral venous pathways, tortuous veins, and delayed local venous drainage. ICVT may show a filling defect on CTV images (Fig. 10). However, if the thrombosed cortical vein is not opacified, ICVT can be missed at CTV. In that case, T2*-weighted or susceptibility-weighted MRI can be helpful for the detection of the thrombus, which exhibits a magnetic susceptibility effect at the site of ICVT [44, 45].
![]() View larger version (399K) | Fig. 10A —59-year-old man with cortical subarachnoid hemorrhage (SAH) due to cortical vein thrombosis. A, FLAIR MR image shows cortical SAH in right cerebral sulci. |
![]() View larger version (367K) | Fig. 10B —59-year-old man with cortical subarachnoid hemorrhage (SAH) due to cortical vein thrombosis. B, CT venogram (B) and digital subtraction angiogram (C) show short segmental cortical vein thrombosis (arrow). |
![]() View larger version (256K) | Fig. 10C —59-year-old man with cortical subarachnoid hemorrhage (SAH) due to cortical vein thrombosis. C, CT venogram (B) and digital subtraction angiogram (C) show short segmental cortical vein thrombosis (arrow). |
Meningiomas can invade a venous sinus and cause stenosis or occlusion of the sinus. Intravenous extension of intracranial meningiomas occurs in approximately 20% of cases [46]. Meningiomas are usually benign (90%), but recurrence after surgical resection is not rare. Incomplete resection of infiltrated surrounding tissue, especially at the wall of a dural sinus is considered a major cause of recurrence [47]. Primary or metastatic cortical tumors can also invade cortical veins and venous sinuses. The presence of sinus invasion by the tumors is critical for treatment planning. In most cases, CTV is sufficient for the evaluation of sinus invasion and the relation between the tumor and venous structure [48] (Fig. 2).
The cerebral cortical veins and venous sinuses are variable in size, number, and anatomic locations, and they have many normal variations, which can mimic pathologic conditions. The clinical and radiologic manifestations of the cerebral vein and venous sinus diseases are also variable. Acknowledgment of the normal variations and pathologic features and clinical suspicion are helpful for the accurate diagnosis of diseases involving the cerebral venous system. Bone subtraction CTV affords complete 3D visualization of the cerebral venous system and can be useful for the evaluation of the cerebral vein and venous sinus diseases.
WEB
This is a web exclusive article.

Audio Available | 























