Sixteen consecutive relapsing-remitting (RR) MS patients and eight age- and sex-matched healthy controls (HC) were enrolled in this study as previously described [6]. Briefly, the inclusion criteria required clinically definitive MS, RR MS disease course, an Expanded Disability Status Scale (EDSS) score between 0 and 5.5, age 18 to 65 years, disease duration between 5 and 10 years, being treated with currently U.S. Food and Drug Administration-approved, disease-modifying treatments and having normal renal function (creatinine clearance >58 ml/min). Exclusion criteria included an acute relapse and/or steroid treatment within the 30 days preceding study entry, preexisting medical conditions associated with brain pathology and abnormal renal function.
All investigators conducting assessments were blinded to the clinical, demographic, and subject group (MS or HC) characteristics. We aimed to ensure proper blinding by instructing subjects not to reveal their disease status during the Doppler examination and including RR MS patients with low disability or walking difficulties. The Italian research group conducted the Doppler assessment, and the Buffalo research group conducted clinical and magnetic resonance imaging (MRI) examinations. The clinical, Doppler and MRI assessments were conducted on the same day for each subject.
The study was approved by the institutional review board, and written informed consent was obtained from all study subjects.
MRI scan acquisition and analysis
All subjects were examined on a 3-T GE Signa Excite scanner (General Electric, Milwaukee, WI, USA). The following sequences were acquired: two-dimensional (2-D) multiplanar dual fast spin-echo proton density and T2-weighted images, fluid-attenuated inversion recovery (FLAIR) images, a 3-D high-resolution (HIRES) fast spoiled gradient echo (FSPGR) with magnetization-prepared inversion recovery (IR) pulse- and perfusion-weighted imaging (PWI).
One average was used for all pulse sequences. With the exception of PWI, all sequences were acquired with a 256 × 192 matrix (frequency × phase) and field of view (FOV) of 25.6 cm × 19.2 cm (256 × 256 matrix with phase FOV = 0.75) for an in-plane resolution of 1 mm × 1 mm. For all 2-D scans (PD/T2, FLAIR and SE T1), 64 slices were collected with a thickness of 2 mm and no gap between slices. For the 3-D HIRES IR-FSPGR, 184 locations were acquired, 1 mm thick, providing for isotropic resolution. Other relevant parameters were as follows. For dual FSE PD/T2, echo and repetition times (TE and TR) TE1/TE2/TR = 9/98/5300 ms; flip angle (FA) = 90; echo train length (ETL) = 14; and acquisition time (AT) = 5:08 (min:sec). For FLAIR images, TE/TI/TR = 120/2100/8500 ms (TI inversion time), FA = 90, ETL = 24 and AT = 6:49. For SE T1-WI, TE/TR = 16/600 ms, FA = 90 and AT = 6:11. For 3-D HIRES T1-WI, TE/TI/TR = 2.8/900/5.9 ms, FA = 10 and AT = 9:18.
Dynamic susceptibility contrast-enhanced PWIs were acquired during and after injection of 15 ml of 0.1 mM/kg gadolinium-diethylenetriamine penta-acetic acid with an MRI-compatible power injector at a speed of 5 ml/s. The HC were also injected with the contrast agent. Single-shot, gradient-echo, echo planar imaging was used with the following parameters: TR 2275 ms, TE 45 ms, FOV 26 × 26 cm, matrix 96 × 96 (resulting in in-plane voxel sizes of 2.71 × 2.71 mm), 20 slices (7 mm thick) with no gap. Forty time points were acquired per slice.
PWI characteristics of the gray matter (GM) and white matter (WM) tissue compartments were assessed by using SIENAX [7]. Subcortical gray matter (SGM) structures were assessed by using FMRIB's FIRST software to segment high-resolution, 3-D, T1-weighted images http://www.fmrib.ox.ac.uk/fsl/first/index.html and included the thalamus, pulvinar nucleus of thalamus, caudate, putamen, globus pallidus, hippocampus, amygdala, nucleus accumbens, red nucleus and substantia nigra. Briefly, FIRST is a model-based segmentation/registration program that uses shape/appearance models constructed from manually segmented images. The manual labels are parameterized as surface meshes and modeled as a point distribution model. Deformable surfaces are used to automatically parameterize the volumetric labels in terms of meshes; the deformable surfaces are constrained to preserve vertex correspondence across the training data. Normalized intensities along the surface normals are sampled and modeled. The shape and appearance model is based on multivariate Gaussian assumptions. Shape is then expressed as a mean with modes of variation (principal components). On the basis of the learned models, FIRST searches through linear combinations of shape modes of variation for the most probable shape instance, given the observed intensities in the input image.
Calculation of perfusion cerebral blood flow (CBF), cerebral blood volume (CBV) and mean transit time (MTT) was conducted by blinded operators using a previously described method [8]. Briefly, we used the Java Image Manipulation software package (Xinapse Systems, Thorpe Waterville, UK) with an automated additive interval-finding algorithm (searching 500 "artery-like" candidate voxels and retaining the 40 best fitting voxels) and singular value decomposition (cutoff at 20% of maximum singular value) for perfusion curve fitting [9]. CBF and CBV values were relative, based on estimated tissue relaxivity and hematocrit parameters (arterial relaxivity 1.0 L/s/M, tissue relaxivity 1.0 L/s/M, arterial hematocrit 0.45, tissue hematocrit 0.45). Correction for patient motion prior to perfusion analysis was performed using FMRIB's Linear Image Registration Tool for Motion Correction (MCFLIRT). Using the first steady-state volume before contrast arrival time, we applied FMRIB's Linear Image Registration Tool (FLIRT) to derive an affine transformation matrix, providing a transformation from the native perfusion acquisition space to the high-resolution FLAIR space. This matrix was then used to coregister perfusion MTT, CBF and CBV maps into the subject-specific upsampled FLAIR space. These maps were then overlaid onto all relevant region-of-interest masks to calculate mean values for MTT, CBF and CBV within each tissue compartment (Figure 1).
Assessments of cerebral venous hemodynamics
Cerebral venous return was examined using echo-color Doppler equipped with 2.5 and 7.5 to 10 MHz transducers (Esaote-Biosound My Lab 25, Genoa, Italy), with the subject positioned on a bed tilted at 90° and 0° as previously described [2].
All subjects were scanned in a blinded manner following the established protocol for diagnosis of CCSVI [2], consisting of transcranial and extracranial echo-color Doppler to measure the following five venous hemodynamic (VH) parameters indicative of CCSVI: (1) reflux in the IJVs and/or in the vertebral veins (VVs) in sitting and in supine positions (90° and 0°), with reflux defined as flow directed toward the brain for a duration of >0.88 seconds; (2) reflux in the intracranial veins with reflux defined as reverse flow for a duration of 0.5 seconds in one of the insonated veins (superior and inferior petrosus sinus, and/or Rosenthal vein); (3) B-mode abnormalities causing absence of flow or significant flow disturbances (vestigial valves, septum, immobile valve leaflets, see Figure 2), or stenoses in IJVs. IJV stenosis was defined as a cross-sectional area of this vein ≤0.3 cm2 ; (4) flow that is not Doppler-detectable in IJVs and/or VVs despite multiple deep breaths; and (5) a wider cross-sectional area of the IJVs in the upright positions respect to supine. A subject was considered CCSVI-positive if at least two VH criteria were fulfilled as previously proposed [2].
We also calculated a VH insufficiency severity score (VHISS), as previously described [6]. The VHISS is an ordinal measure of the overall extent and number of VH flow pattern anomalies, with a higher value of VHISS indicating a greater severity of VH flow pattern anomalies. For each of the five VH criteria, a "VHISS contribution score" was assigned using the scheme described below. These scores combined gave an overall severity measure: the VHISS. The minimum possible VHISS value is 0 and the maximum is 16.
As regards criterion VH1, there are eight venous segments that can potentially exhibit reflux in the two postures, and one point was assigned for each one at which reflux was found to be present. Consequently, VH1 had a VHISS contribution score that could range from a minimum of 0 to a maximum of 8.
Criterion VH2 was assigned a VHISS contribution score of 1 if reflux was present in the intracranial veins in only one posture and a VHISS contribution score of 2 if it was present in both postures. The VHISS contribution score for this criterion was additionally weighted with a factor of 2 if reflux toward the subcortical GM could be detected. Consequently, the VHISS contribution score for VH2 could range from a minimum of 0 to a maximum of 4.
The VHISS contribution score for VH3 ranged from 0 to 2, depending on whether B-mode anomalies disturbing outflow were present in none, one or both of the IJVs, respectively (Figure 1). VH3 was assigned a contribution score of 0 if either VH1 or VH4 was positive for the presence in either posture of reflux or obstruction in the IJV of interest.
The scoring scheme for the contribution of VH4 to the VHISS was the same as that for VH1, with the difference being that only blocks were considered. No points were assigned for segments and postures in which reflux had previously been detected under VH1.
The VH5 criterion had an overall VHISS contribution score between 0 and 4, calculated by assigning 0 to 2 points for each IJV. A -ΔCSA value was assigned a score of 2, whereas a ΔCSA value <7 mm2, corresponding to the 25th percentile of ΔCSA distribution in healthy controls, was assigned a score of 1. ΔCSA >7 mm2 was assigned a score of 0.
The overall VHISS score was defined as a weighted sum of the scores contributed by each individual abnormal venous haemodynamics (A-VH) criterion. The formula for VHISS calculations was as follows:
The subscripts in this formula indicate the subscores for the five VH criteria.
Statistical analysis
Statistical analysis was performed using the SPSS version 16.0 (SPSS, Inc., Chicago, IL, USA). The ages and proportions of females and males in the MS and HC groups were assessed with the Student's t-test and Fisher's exact test, respectively. The nonparametric Mann-Whitney U test was used to assess the VH differences between the MS and HC groups. Spearman correlation analysis was used to assess the relationship between PWI measures and the severity of CCSVI.
Since this was a preliminary exploratory study, we used a false discovery rate (FDR) correction [10] rather than a family-wise error rate (FWER) correction to correct for multiple comparisons on our outcome measures. FDR provides a statistical bound on the total percentage of incorrectly rejected null hypotheses rather than on the probability of any error occurring and is therefore considerably more powerful while still providing strong control. For the current work, we used an FDR threshold of 0.05, so we expect a 5% error rate for findings we consider significant. In the results, we report both uncorrected (P) and FDR-corrected (Q) statistics.
Comments
View archived comments (2)