The role of noninvasive and invasive diagnostic imaging techniques for detection of extra-cranial venous system anomalies and developmental variants

DS is clinically the most useful technique for detecting, localizing and evaluating peripheral venous obstruction and venous valvular incompetence [33, 109]. The sensitivity and specificity of venous DS for symptomatic proximal deep vein thrombosis exceeds 90% [110, 111]. Spectral analysis of the DS signal is used to confirm the presence or absence of flow and indicates its direction and the patterns. Spectral analysis of DS signal and color DS are used to confirm the presence of reflux. It has the advantage among other diagnostic techniques of being noninvasive, providing high-resolution images with real time dynamic information, such as flow and velocity, showing intra-luminal (Figure 1A) as well as extra-luminal anomalies and developmental variants (Figure 1B) and being considerably less expensive than other noninvasive imaging techniques. DS imaging can also be readily applied in the follow-up period of subjects undergoing endovascular treatment because it can recognize the associated complications (residual stenosis, restenosis or venous thrombosis) (Figure 1C) [28, 67, 68].

Recent findings suggest that the majority of CCSVI pathology is confined to the intra-luminal portion of extra-cranial veins, which requires high-resolution B-mode imaging for the visualization of these anomalies [31, 47]. The visible "stenoses" (Figure 1D) or extra-luminal venous anomalies most likely develop more frequently, merely with the progression of the disease or age [10].

Because of the advantages of DS in detecting intra-luminal venous pathology, it was initially promoted as a method of choice for the screening of extra-cranial venous anomalies and developmental variants, indicative of CCSVI [18, 27]. The diagnosis of CCSVI is both based on hemodynamic and imaging findings that utilize DS to study the deep cerebral veins, the IJVs and the vertebral veins (VVs) in both erect and supine positions. DS can also assess the hemodynamic consequences of outflow derangement while B-mode ultrasound detects structural venous intra-luminal anomalies (Figure 1E, F) [18, 27, 31, 33, 109, 112]. Zamboni et al. created a set of five DS VH criteria by which MS patients were differentiated from healthy controls with 100% specificity and sensitivity [18, 27] (Figure 1). However, in their original publication [18], they did not recommend exact technical procedures for the protocol application in either a research or routine clinical setting. The first attempt to define the standardized CCSVI scanning protocol was recently presented [98]. More recently, the International Society for the Neurovascular Disease (ISNVD) developed a more comprehensive consensus document that included the participation of more than 40 international experts in DS imaging. DS was proposed as a standardized screening tool for determining CCSVI status [33]. The protocol proposes the use of quantitative measures for the definition of functional anomalies, such as blood flow velocity and volume (Figure 2) that could be potentially more reliable in assessing the degree of venous outflow obstruction in the IJVs. It also refines originally proposed VH criteria. Even more recently, the European Society of Neurosonology and Cerebral Hemodynamics (ESNCH) expressed considerable concerns regarding the accuracy of the proposed criteria for CCSVI in MS [32], and proposed the central blinded DS reading as part of a recent multi-center Italian CoSMo study investigating the prevalence of CCSVI in MS patients, controls and patients with OND [113].

The main criticism of the recommended DS protocol is that its reproducibility depends on the training level and skills of the operator and it is not easy to be blinded and standardized in either a research or clinical setting [29‐33, 87]. Moreover, the value of the CCSVI VH criteria is controversial because they combine functional and structural intra- and extra-cranial venous anomalies/developmental variants in a single binary composite. Zamboni et al. used ≥2 abnormal DS VH criteria as a cutoff for CCSVI diagnosis classification [18, 27]. The dichotomous variable construct of the CCSVI diagnosis, based on the arbitrary decision biased toward characteristics of the originally studied population and on the obtained results without further testing and validation datasets [18, 27], may contribute to explaining major inconsistencies in the prevalent findings of CCSVI between different studies ranging from 0 to 100% [18, 27, 34, 78‐98, 100, 101, 114]. The assessment of the second CCSVI criterion (reflux in deep cerebral veins) (Figure 3) is particularly controversial because the direction of the blood flow in veins connecting cortical with deep veins may vary considerably as a consequence of the physiologic inter-individual variation of the cerebral venous anatomy [30, 32, 33, 87].

DS also has limits regarding extra-cranial vein characterization, since findings can be influenced by hydration status [10]. DS is a very time-consuming method and visualization of the central veins, particularly in the thorax and abdomen, is often limited and cannot give the global view of vein anatomy. Although it can detect extra-cranial collateral veins, which are probably associated with CCSVI, it is not technically feasible to follow the complete course of the collateral veins, which can be more easily visualized with use of MRV, CTV or CV [10]. Other pitfalls in DS imaging include the misidentification of veins. Additionally, overlying bone and muscle may prevent continuous imaging (cannot visualize suitably the confluence of the IJVs and the subclavian vein because clavicle commonly blocks direct visualization). Similarly, the cervical part of IJV and the jugular bulb cannot be visualized by DS because of the limited acoustic window resulting from the spine, mandible and skull [10, 112, 114].

So far, none of the recently published DS studies [30‐32, 34, 78‐101] have reproduced the originally reported CCSVI prevalence [18, 27], regardless of the diagnostic DS method utilized. Even those DS studies which detected a significant difference for CCSVI diagnosis between MS patients and the controls, reported a substantially lower prevalence than was originally reported [30, 31, 83, 88, 90, 92‐94, 98, 99].

The largest cohort published to date of MS patients and controls with DS examined in a blinded manner reported prevalence rates of 56.1% in MS patients, 42.3% in those with OND, 38.1% in clinically isolated syndrome and 22.7% in healthy controls [98]. There have been numerous additional DS studies that showed significant differences in CCSVI prevalence between MS patients and the controls [30, 31, 78, 83, 88, 90, 92‐95, 99]. However, an even higher number of DS studies have failed to show prevalence differences in CCSVI between MS patients and controls [34, 80‐82, 84‐87, 89, 91, 96, 97, 100],[101].

By using contrast-enhanced DS to assess cerebral circulation times (CCT) in MS patients and control subjects, Mancini et al. showed that MS patients had a significantly prolonged CCT and more frequent retrograde flow in IJVs [90].

Several studies have shown a relationship between IJV drainage anomalies, characterized by JVR and specific neurological diseases of undetermined etiology, such as transient global amnesia [14], transient monocular blindness [17], cough headache [13], primary exertional headache [16], idiopathic intra-cranial hypertension [115] along with a higher prevalence of white matter hyperintensities in older people [15]. JVR was also investigated in a large cohort of elderly subjects. An increased prevalence of JVR, dilated vessel lumen and slowed flow velocity in the left IJV, as well as decreased time-averaged mean velocity of bilateral IJV, was found in those over 70 years of age [116].

During the past decade, catheter-based digital subtraction angiography, as the preferred method for imaging of the intracranial venous anatomy, has been increasingly supplanted by MRV, usually performed with a two-dimensional time-of-flight (TOF) pulse sequence [118]. In the absence of better non-invasive techniques for the imaging of the dural venous sinuses, well-known and documented pitfalls associated with flow-sensitive MR techniques have been tolerated.

Furthermore, simple protocols that incorporate 2D-TOF acquisitions have already improved their accuracy for the diagnosis of deep venous thrombosis involving the femoral, popliteal or iliac veins [119]; however, experience with these techniques in the cervical veins is still limited. Thoracic central veins are largely inaccessible by DS, and MRV is an excellent technique for the assessment of axillary, jugular, subclavian, superior vena cava and pulmonary veins. TOF venography has the advantage of simplicity because no special pulse sequences are required and this technique is available on nearly every MRI system. TOF pulse sequences are spoiled gradient-echo or gradient echo acquisitions performed sequentially, that is, all phase encode steps are played out in a single slice before moving on to the next slice that results in much greater suppression of stationary tissue. It also has the advantage of avoiding the need for use of contrast agents and it remains the technique of choice in the evaluation of the pregnant patient with suspected dural sinus thrombosis. Furthermore, the accompanying conventional MR study is more sensitive in terms of the detection of cortical venous infarction than a CT [120]. Additionally, CTV always requires the use of intravenous contrast, while many non-contrast methods are available with MRV, making MRV the preferred technique in patients who also suffer from renal insufficiency or contrast allergy. CTV may also require two or more acquisitions to adequately capture contrast opacification of the veins, thereby increasing the radiation dose [103].

The axial orientation of the acquisition allows for high in-plane resolution, which is ideal for cross sectional area (CSA) measurements of the veins. However, the TOF sequence is easily affected by motion artifacts, especially from the patient’s breathing, swallowing, snoring or head motion [38, 41] (Figure 4). Relative insensitivity to in-plane flow is another limitation of the TOF technique. Regarding the direction of flow, the optimal acquisition plane lies orthogonal, which is inefficient from the standpoint of acquisition time and not always achievable. Although it has a higher spatial resolution 2D-TOF may overestimate stenosis in the setting of turbulent or slow flow [42].

All in all, standard conventional MRV techniques are more prone to artifacts than phase-contrast MRV and 3D-TOF angiography [10, 44]. These techniques can potentially alleviate some of the usual MRV artifacts and provide more detailed flow information. One obvious improvement is to image at higher field strength, such as 3T, because this increases signal-to-noise ratio and better characterizes slow flow.

In contrast to TOF techniques, which rely mainly on flow-related enhancement for producing vascular images, phase-contrast MR angiography (PC-MRI) uses velocity-induced phase shifts imparted upon the moving spins to distinguish flowing blood from the surrounding stationary tissue, thus providing information regarding both anatomy and flow (Figure 5). The major advantage of PC-MRI angiography is excellent background suppression as well as quantitative determination of blood velocities. However, it requires long imaging times and a prior estimate of blood flow velocity. Furthermore, it may also be more sensitive to signal loss due to turbulence or intravoxel dephasing [121, 122]. So far, to the best of our knowledge, there are only a few studies that used PC-MRI to quantify venous flow in MS patients. Sundström et al. studied the IJV flow normalized by the total arterial flow at the C2/C3 levels in 21 MS patients and 20 controls and found no statistically significant difference between the two [37]. On the other hand, Feng et al. characterized and compared the flow characteristics in a large cohort of the non-stenotic and stenotic MS patients and observed significantly reduced IJV flow in the stenotic group [41]. They concluded that a normalized total IJV flow of less than 50% of total arterial flow may be a potential biomarker for identifying significant stenoses in IJVs. Additionally, Haacke et al. showed that patients suffering from MS with structural venous anomalies on MRI exhibit an abnormal flow distribution of the IJVs [35]. In contrast to PC-MRI, Hartel et al. used very simple MRV protocol with T2FatSat and 2D-TOF sequences for the assessment of flow disturbances in IJVs and azygous vein [123]. They found that abnormal flow pattern in IJVs in MS patients is more common on the left side.

More studies are needed to validate the venous flow at the upper neck level on an adequate number of age- and gender-matched healthy controls with heterogeneous age groups.

Contrast-enhanced (CE) MRV, 3D time-resolved imaging of contrast kinetics (TRICKS) angiography is a noninvasive and safe method for the evaluation of head and neck veins, without the attendant risks of conventional angiography. It is preferred over TOF angiography because contrast medium reduces the T1 relaxation time of blood and virtually eliminates the effect of saturation [124, 125] (Figure 6).

CE MRV is probably the most widely-used technique and is essentially identical to 3D CE MR angiography, employing a 3D-spoiled gradient-echo sequence in conjunction with a bolus of gadolinium-based contrast. Vascular contrast results from the T1-shortening effects of gadolinium on adjacent water protons and has relatively little dependence on inflow effects. In contrast to MRA, the limitation of CE MRV is that maximal contrast enhancement achieved in veins is typically lower than arteries because the contrast bolus is more dilute by the time it reaches the venous system [126]. To improve background suppression and emphasize vascular signal, fat saturation can be added to a 3D spoiled gradient-echo sequence with a small increase in acquisition time. 3D reconstruction of CE MRV data is somewhat less straightforward than MR angiography reconstruction since the vein/background contrast is lower and there is usually arterial as well as venous enhancement.

Veins can have variable MR imaging signal intensity due to entry slice phenomenon, in-plane flow, flow turbulence effects and can have variable enhancement. The maximum intensity projection (MIP) volumetric reconstructions of these sequences often underestimate the vascular caliber, especially when there are segments with decreased flow (velocity or volume) [120].

Disadvantages of CE MRV include the expense of the contrast agent, as well as contrast toxicity and patient discomfort in obtaining antecubital venous access. In the case of dural sinus thrombosis, however, confident early diagnosis of this common and treatable disease can dramatically reduce patient morbidity.

Another promising MR technique is cine velocity-encoded phase-contrast 4D flow that may permit evaluation of not only anatomic stenoses but also their impact on venous waveforms. It is based on the principle that moving protons change phase in proportion to their velocity. By enabling a qualitative assessment of the presence and direction of collateral circulation, velocity-encoded cine MR imaging provides information about the presence and severity of obstruction. The technique has been most extensively used for the evaluation of patterns of blood flow in the thoracic aorta, including the characterization of abnormal flow patterns associated with pathologic disorders, such as ascending aortic aneurysm and dissection [127]. Recent studies have explored the use of 4D flow imaging for other areas of vascular anatomy and pathology, including intracranial arterial and venous blood flow [128]. With its detailed characterization of complex, dynamic blood-flow patterns and its ability to quantify flow, the technique could supplement both current noninvasive and invasive imaging of intra- and extra-cranial vascular pathologic disorders. The diagnostic and monitoring value of 4D flow imaging of venous flow anomalies, indicative of CCSVI, is currently lacking.

Finally, MRV suffers from its "snapshot" nature. An accurate depiction of these veins requires multiple views and maneuvers, such as inspiration and expiration, flexion and extension as well as rotation imaging of the neck. Its main disadvantages are the lack of MRV dynamism in real time, lower resolution compared with DS and CV (cannot evaluate intra-luminal pathology, such as the immobile valves, webs, septations, membranes and duplications) and it is affected by the nature of the veins themselves, which are prone to collapse under frequently encountered conditions, as opposed to arteries. MRV often detects spurious stenoses that are not confirmed by CV, especially in the lower parts of IJVs [42, 123]. These stenoses may represent transient phasic narrowings (functional) or may result from diminished flow above true stenoses commonly located at the confluent region of the veins [30, 31, 102, 123]. Additionally, it cannot satisfactorily evaluate the azygous and hemiazygous veins.

Unlike DS, with most MR scanners, data can only be collected in the supine position, although some scanners can do an upright scan as well. Niggemann et al. used positional MR imaging to describe the influence of positional changes on the cerebral venous outflow [129]. They found that IJV strictures are a common finding in healthy controls in the supine position without relevance in the erect position, which questions the validity of the DS VH criterion 5 (lack of collapse of the IJV in upright posture) for the diagnosis of CCSVI. It is obvious that this criterion (to study the change in flow in the IJVs from supine to sitting position) cannot be studied with the conventional MR system [130].

CV can be complemented by use of more sophisticated criteria such as time to empty contrast from the vein or wasting of the balloon across a stenosis [134]. Further, with the ability to perform pressure gradient measurements before and after the endovascular procedures it can indirectly give the information about hemodynamic significance of venous obstruction [28].

The advantages of IVUS compared with DS, among others, include the sonographic penetration from within the vessel by excluding extra-vascular soft tissues. It also assesses blood vessels not easily accessible by conventional DS, such as the lower part of the IJV (behind the clavicle), upper part of the IJV, intracranial sinuses and azygos vein. Additionally, it provides an image with a greater resolution of both lumen and wall (with additional 3D features), providing better vessel wall information. IVUS is superior in identifying intra-luminal venous anomalies/development variants compared to CV [107, 108, 134]. Moreover, CV is incapable of monitoring respiratory pulsatility which involves periods with reduced vessel diameter that can be investigated with IVUS. While values for stenosis definition used for CV (≥50%) rely on a ratio between the stenotic segment diameter and a pre-(non) stenotic vein, which is more variable, the IVUS definition is more strict (a lumen that embraces the IVUS probe for a critical stenosis) and does not refer to a non-stenotic segment [134]. It remains unclear at what level and with what criteria is there a significant hemodynamic effect of stenosis by either modality. Venous stenosis is currently measured using arterial criteria, which are clearly not optimal. The hemodynamics of venous flow remain a major area of investigation and better understanding will likely lead to a revision of stenosis criteria.

Ring-down artifacts produced by acoustic oscillations in the piezoelectric transducer that obscures the near field, results in an acoustic catheter size larger than its physical size and may adversely affect IVUS images [144]. Geometric distortion can result from imaging in an oblique plane (not perpendicular to the long axis of the vessel) [145]. Furthermore, visible distortion of the image can be due to another important artifact, "non-uniform rotational distortion", which arises from uneven drag on the drive cable of the mechanical style catheters, resulting in cyclical oscillations in rotational speed. The physical size of IVUS catheters (currently approximately 1.0 mm) constitutes an important limitation in the imaging of severe stenoses [146]. Further, depending on the probe there is a finite limit to IVUS resolution which rapidly degrades beyond this particular radius typically 10 to 12 mm. In summary, the frequency of the transducer, gain settings, depth of penetration and focal depth are some of the factors that affect the sensitivity of the IVUS imaging.

Further studies are needed to validate the role of IVUS in depicting extra-cranial venous anomalies and developmental variants, indicative of CCSVI. Protocol optimization and standardization are needed to make this imaging method more widely used. Preliminary IVUS studies that investigated extra-cranial venous anomalies and developmental variants have been extremely important in better understanding these structures [47, 107, 108, 134].

Apart from these early studies, little work has been done on the application of cervical plethysmography in the detection of extra-cranial venous anomalies and developmental variants. Further research is needed in identifying cutoff values, the reproducibility of the test along with assessing intra- and inter-observer variability. This methodology also shows great potential in monitoring postoperative patients after restorative endovascular procedures.