Value of MR Venography for Detection of Internal Jugular Vein Anomalies in Multiple Sclerosis: A Pilot Longitudinal Study

Discussion

This pilot study investigated the value of extracranial MRV for the assessment of IJV anomalies in patients with MS and healthy controls. In patients with MS and healthy controls, TRICKS and TOF showed no agreement with Doppler sonography findings of the IJVs both at baseline and follow-up. There was low specificity and PPV and no agreement between the 2 MRV techniques and catheter venography findings in patients with MS at baseline. On the other hand, Doppler sonography showed high specificity and PPV, as well as strong agreement with catheter venography findings at baseline. These results indicate that conventional MRV has limited value for the detection of venous abnormalities cross-sectionally and longitudinally. Given the low accuracy of conventional MRV for depicting cross-sectional and longitudinal IJV anomalies, MRV findings should be interpreted with caution, both for diagnosis and monitoring treatment outcome.

A recently published longitudinal study in patients with MS diagnosed with CCSVI and treated with PTA used Doppler sonography and catheter venography for a diagnosis of CCSVI.³ The present study was based on a subcohort of patients with MS who participated in the EVTMS⁸ and underwent MRV at baseline, in addition to Doppler sonography and catheter venography. At 6-month follow-up, they were assessed with MRV and Doppler sonography. We did not perform a follow-up catheter venography in patients with MS to confirm the Doppler sonography 6-month findings. In addition, a group of 6 healthy controls was enrolled to increase confidence in the detection of IJV venous anomalies and to test the longitudinal stability of MRV and Doppler sonography.

At baseline, as per inclusion criteria, all patients with MS had to present ≥2 venous hemodynamic criteria and meet a diagnosis of CCSVI to be enrolled in the EVTMS.¹ Catheter venography confirmed the venous anomalies identified by Doppler sonography in the IJVs in 100% of patients with MS on the left and 90% on the right, resulting in a strong agreement between the 2 imaging techniques (κ = 0.737–1). The sensitivity, specificity, PPV, and NPV were also very high (Table 3). The venous anomalies consisted of the annulus, septum, membrane, and malformed valve. PTA was performed exclusively at the levels of the azygous vein and IJVs when these anomalies were detected. It consisted of dilation of the veins by using compliant-type balloon catheters. Catheter venography was repeated at the end of the therapeutic procedure to document the immediate outcome of the treatment. Of the 10 patients with MS who underwent the PTA treatment, 3 presented with ≥2 venous hemodynamic criteria at the 6-month follow-up and were diagnosed with restenosis. However, we detected at least 1 venous anomaly on Doppler sonography in 50% of the treated patients. This suggests that the PTA approach for treatment of IJV anomalies has a somewhat limited value. The B-mode abnormalities (flaps, membrane, malformed valve, septa, and web) that do not cause apparent extraluminal stenoses but contribute to altered intraluminal flow are difficult to treat with compliant-type balloon catheters. All 5 patients with MS who presented with Doppler sonography abnormalities at follow-up had these types of anomalies. Use of specifically developed stents for IJVs or cutting balloon angioplasty may be a possible solution for treatment of these types of venous anomalies. The safety and efficacy data from the EVTMS study will be the subject of future reports.

A major limitation of all studies of diagnostic accuracy is the necessity to refer to a criterion standard. In our case, we are referring to catheter venography (Table 3). However, in this particular field, catheter venography can, on occasion, also present limitations. For instance, if malformed and/or reversed valve cusps are crossed by the catheter, the malformed valve can be kept open artificially, thereby preventing documentation of stenosis. Moreover, recognition of intraluminal B-mode abnormalities requires experience and proper training.

We previously reported that the inter-rater variability rate between trained and untrained Doppler sonography operators substantially increased after training was performed.¹¹ The interobserver agreement between trained operators was much more reliable (κ = 0.80) compared with the agreement between untrained operators (κ = 0.47). Therefore, the Doppler sonography protocol requires appropriate training before it can be applied.

The MRV techniques easily and globally depict the extracranial venous system compared with Doppler sonography and catheter venography. However, in contrast to high-resolution Doppler sonography and catheter venography, MRV techniques do not have sufficient resolution to show vessel wall or intraluminal B-mode abnormalities such as the annulus, flaps, webs, and so forth. This is one of the main limitations of comparing conventional MRV with Doppler sonography and catheter venography. We considered pinpoint and absent flow on TOF and TRICKS to be abnormal findings in IJVs. We did not consider flattening to be a pathologic finding because, in a previous study, it had the same appearance and approximate frequency in healthy controls as in patients with MS.⁶ Flattening was quite a common finding at the level of the lateral masses of the atlas and at the thyroid gland level

In patients with MS, TOF and TRICKS showed high sensitivity and NPV but low specificity and PPV relative to catheter venography. No agreement was found between TOF and catheter venography in 70% of patients with MS and between TRICKS and catheter venography in 60% of patients with MS. More than 50% of the healthy controls did not show agreement between MRV and Doppler sonography findings, confirming results observed in patients with MS pretreatment. Both in patients with MS and healthy controls, no agreement was found between MRV and Doppler sonography at the 6-month follow-up.

The lack of dynamism in real-time of conventional MRV is an important disadvantage when studying veins because, in contrast to arteries, veins are prone to morphologic and hemodynamic changes under multiple circumstances. MRV is performed only in the supine position, and consequently, it misses all the dynamic information provided by Doppler sonography in the upright position. For this reason, a static image of a vein in a certain moment can be misleading and it should be considered with caution when contemplating stenosis. Veins have a tendency to collapse and change their morphology and size depending on hydration status, position (gravitational variability), intrathoracic pressure (respiration, Valsalva), cardiac status, and compression from adjacent structures.^12–15 In particular, a blank jugular vein with a block of the flow due to a stenosing lesion of the outlet has reduced transmural pressure. In this condition, it becomes easily compressed by adjacent structures. Consequently, when one conducts MRV, changes in the morphology of the IJVs could be attributed to several factors: different respiratory phases during sequence acquisition, different positioning of the head and neck coil, changes in the contact points (extrinsic compression) with the coil, changes due to swallowing movements, physiologic stenosis of the left brachiocephalic vein during regular breathing in the supine position,^16–20 and different cerebral drainage patterns, as described in the horizontal position in healthy controls (jugular drainers versus nonjugular drainers).^13,21 This may explain why we found variability in the morphology of the IJVs between TOF and TRICKS. Both examinations were conducted at the same scanning session, and this finding applied to healthy controls as well as patients with MS. MRV variability was also detected at follow-up examinations in healthy controls.

The MR imaging techniques we used are relatively standard, and other more advanced approaches are possible. These include phase-contrast MRV,²² which is capable of simultaneous flow quantification in multiple directions, and multidirectional TOF MRV, which could potentially alleviate some of the usual MRV artifacts and provide more detailed flow information. Although they still lack the positional dynamism of Doppler sonography, MRV techniques like these might provide more sensitivity and/or specificity for detection of venous anomalies. However, these techniques are complex and more technically challenging, perhaps too much so to be applied to routine clinical practice. The focus of our current investigation was to understand the value of conventional MRV. Our results speak to that and leave open the possibility that more nonconventional MRV approaches may lead to different results.

Apart from the anomalies described in the IJVs, extracranial venous collateral circulation was previously described as a compensatory mechanism of CCSVI in patients with MS.^1,3 We were not able to compare collateral circulation between the 3 imaging techniques because our original protocol did not systematically evaluate asymmetries and prominence.

Although the azygous vein is another important target for the diagnosis and treatment of CCSVI,^1,3 we did not analyze this venous segment in the present study. During the development of the MRV protocol, we tried to image the azygous vein, but the quality of the protocol was very low and did not reliably assess the morphology of the azygous vein. Primarily, this was due to the fact that we did not use cardiac gating and the FOVs were centered on the neck. Moreover, higher concentrations of contrast than we applied are needed for proper visualization of the azygous vein. Therefore, the diagnostic value of MRV for assessment of the azygous vein needs further technical improvement.

The differences between various imaging modalities in our study emphasize the need for a multimodal approach for assessment of CCSVI. At this time, no standardized consensus definitions are available that describe venous pathology. There is also a lack of diagnostic reference guidelines related to the evaluation of the extracranial cerebrospinal venous system and collateral circulation. Use of intraluminal Doppler methods²³ and pathologic approaches can provide more evidence on whether CCSVI exists in MS.