This is the first multimodal imaging study in which 2 noninvasive and 2 invasive diagnostic techniques for the detection of extracranial venous anomalies, indicative of CCSVI were applied. The main finding of the study is that invasive techniques confirmed that noninvasive DS screening was a reliable approach for identifying patients eligible for further multimodal invasive imaging testing of the IJVs. In 19 of the 20 MS patients, the extracranial venous IJV anomalies indicative of CCSVI diagnosis were confirmed on CV or IVUS. However, it has to be noted that 50% of the screened MS population did not fulfill ≥2 extracranial VH DS criteria and were therefore not eligible to undergo invasive testing, which limited the study ability to investigate specificity of DS vs. invasive imaging diagnostic techniques. Nevertheless, the findings from this multimodal study are important, as they suggests that DS can be used reliably to select those patients who may present extracranial IJV venous anomalies, indicative of CCSVI, while in the same time, it can potentially exclude those patients who should not undergo an invasive testing of the IJVs. However, the noninvasive screening methods were inadequate to depict the total amount of VV anomalies that would indirectly reflect the pathology of the azygos vein identified with invasive testing. These findings are related to the fact that we were not able to directly non-invasively assess the azygos vein. In our opinion and experience
[6, 8, 22, 23], there are no reliable, non-invasive imaging modalities at this time that would directly image the azygous vein in vivo. Another important finding is related to the results from the invasive portion of the study, which confirmed the existence of severe extracranial venous anomalies, indicative of CCSVI that significantly impaired blood outflow from the brain.
A growing body of evidence suggests that the majority of CCSVI pathology is confined to the intra-luminal portion of extracranial veins, which requires high-resolution DS or IVUS B-mode imaging for the visualization of these anomalies
[8, 26, 27]. It has been shown that the presence and number of these anomalies may contribute to a higher number of collateral neck veins and functional abnormalities
[8, 27]. While CV is considered to be "the gold standard - benchmark" for detecting stenosis in blood vessels associated with altered blood flow, the PREMiSe study showed that CV may not be sensitive enough to reveal the exact nature of narrowed vein segments
. CV is a luminogram and brings little or no data regarding the vessel's intra-luminal structures because of dense opacification of the lumen with contrast, which obliterates subtle intra-luminal structures
. There are no consensus guidelines with respect to the use of angiographic contrast for extrcranial CV examination
. The recent position statement of the International Society for Neurovascular Disease on the use of angiographic contrast for the assessment of IJVs and the azygos vein on CV does not provide clear guidelines on this issue
. Angiographic contrast may be used diluted (1:1) or non-diluted. While the diluted contrast may allow a better visualization of endoluminal structures (valve leaflets, webs, etc.) non-diluted contrast allows a better opacification of epidural and other collaterals, as well as a better estimation of overall features of the veins
. In the PREMiSe study, non-diluted contrast was used. It could be that use of a diluted contrast could have produced different findings. The PREMiSe study also demonstrated the advantage of IVUS compared to CV in detecting intra-luminal abnormalities as well as the importance of including IVUS during CV examination, especially for the assessment of the azygos vein
. It is important to note that sensitivity of IVUS to depict extracranial venous anomalies on DS, indicative of CCSVI, was in better agreement than the CV findings, especially for the azygos vein/VVs territory. However, one of the important limitations of DS screening approach is that the azygos vein cannot be directly imaged. While the sensitivity for detecting VV anomalies on IVUS vs. DS was high, DS did not detect abnormal VV flow in 10 patients who had positive IVUS in the azygos vein. Similar limitations were observed for MRV. These results suggest that currently, available noninvasive indirect screening methods are inadequate in depicting the total amount of intra-luminal pathology of the azygos vein. The sensitivity of CV + IVUS to define total IJV pathology on DS was 100%. These findings support results from the 2 previous studies in which a higher sensitivity of DS to detect extracranial anomalies on IVUS compared to CV was found in the IJVs
[27, 30]. These results can also explain findings from some recent reports that found low correspondence between the DS screening assessment and the CV findings
Zamboni et al. proposed a set of 5 VH DS criteria by which MS patients were differentiated from healthy controls with 100% specificity and sensitivity. While the original publication did not provide the exact technical procedures for the protocol application in either a research or routine clinical setting, there were recent attempts to define the standardized CCSVI DS scanning protocol
[5, 32, 33]. These revised DS protocols propose the use of quantitative measures for the definition of functional anomalies such as blood flow velocity and volume that could be potentially more reliable in assessing the degree of venous outflow obstruction in the extracranial venous system
[32, 33]. They also refine originally proposed VH criteria
[5, 32, 34] and propose the use of the central blinded DS reading
. In this multimodal comparison of different noninvasive and invasive imaging techniques in phase 2 of the PREMiSe study, we read findings from multimodal techniques in a blinded manner by different experts and using a panel to reach a consensus when there were discrepancies by the readers. They also confirmed the correctness of the exam reading by a-priori comparing un-blinded reading results of the individual imaging modalities on a subject level. We did not consider the assessment of the second CCSVI VH criterion (reflux in deep cerebral veins) for several reasons: 1) the reproducibility of this criterion is lower compared to the other 4 VH criteria;
[6, 7] 2) there is no direct anatomical extracranial correlate for performing sensitivity and specificity comparisons with other multimodal imaging techniques; 3) use of this criterion contributes to the highest variability in making a CCSVI diagnosis and 4) 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
. Despite this, the results of the multimodal PREMiSe study further support the value of DS VH criteria for the screening of extracranial venous anomalies in territories of left and right IJVs.
When CV + IVUS findings were compared to MRV findings, sensitivity was high but the specificity was low, confirming our previous results
[23, 24]. Some other investigators used a slightly different grading system for the detection of extracranial venous anomalies on MRV and found similar sensitivity but better specificity compared to CV
. Therefore, the use of different MRV evaluation criteria may have yielded different sensitivity and specificity results compared to CV, IVUS and DS in the PREMiSe study. While there is still a lack of standardized guidelines for the detection of extracranial venous anomalies indicative of CCSVI on MRV, the findings from the PREMiSe study indicate that MRV should be incorporated in the armentorium of noninvasive screening techniques. Further work is needed to standardize MRV morphology criteria
[6, 22–24, 35–37] and incorporate flow and velocity information in determining subjects at risk for the detection of extracranial venous outflow anomalies with hemodynamic consequences
The combination of DS + MRV did not yield better reliability vs. invasive imaging techniques compared to the DS alone. However, phase 2 of the PREMiSe study included only MS patients with ≥2 VH extracranial criteria, which limited our ability to explore the additive value of MRV to DS in improving sensitivity and specificity vs. other invasive imaging techniques.
It was proposed that extracranial venous collateral circulation is a compensatory mechanism for impaired venous outflow because it bypasses blocked veins and thereby reduces resistance to drainage
[6, 8, 41]. The PREMiSe study showed an excellent correspondence between identifying collateral veins on MRV and CV. Approximately, 70%-85% of patients presented collateral veins on the right and 75%-80% on the left side of IJV on MRV and CV respectively. In addition, 80% of patients presented with collaterals of the azygos vein on CV. These findings confirm that the presence of collaterals on MRV and CV may represent an indirect compensatory mechanism for impaired venous outflow. In the previous study, we found high specificity for distinguishing MS vs. healthy controls based on >1 of collateral veins in the neck
PREMiSe was an endovascular angioplasty study that did not include healthy controls or MS patients without the presence of CCSVI diagnosis on DS. This selection bias of the included population was an important limitation of this diagnostic study, as the sensitivity and specificity findings of noninvasive vs. invasive techniques cannot be generalized to the prevalence of findings to other case–control studies. However, the main aim of this multimodal study was to define and reliably detect extracranial venous anomalies, indicative of CCSVI in the IJVs and azygos vein/VVs of patients using DS and to confirm the presence of these anomalies by using 2 invasive imaging techniques. Another potential limitation of the study is a relatively small sample size, which could skew our findings. Although PREMiSe was a limited pilot trial not powered to detect the prevalence of CCSVI in the general MS population and healthy individuals, it confirmed a general prevalence of extracranial venous anomalies, indicative of CCSVI that we have reported in large cohorts using DS and MRV
[5, 6, 8, 22].
Maybe the most important result of the PREMiSe study is that our multimodal imaging findings contradict the number of recent DS studies that reported a prevalence of CCSVI <10% in MS patients
[9, 11, 12, 16, 17, 33, 42–44]. In fact, the invasive diagnostic portion of PREMiSe confirmed that 19 of the 20 MS patients screened as CCSVI positive by DS had severe impairment of extracranial venous outflow with significant stenosis in the IJVs and azygos veins. Future, larger, case-controlled, multicenter, multimodal, noninvasive and invasive imaging studies that will include healthy controls, MS patients and patients with other neurological diseases should determine the real prevalence of CCSVI in these cohorts.