Original Research

Contrast Enhanced MR Venography WithGadofosveset Trisodium: Evaluation of theIntracranial and Extracranial Venous System

Larry A. Kramer, MD,1* Alan M. Cohen, MD,1 Khader M. Hasan, PhD,1

Jared H. Heimbigner, DO,1 Andrew D. Barreto, MD,2 Staley A. Brod, MD,2

Ponnada A. Narayana, PhD,1 and Jerry S. Wolinsky, MD2

Purpose: To demonstrate the efficacy of contrastenhanced magnetic resonance venography (CEMRV) usinggadofosveset trisodium in the comprehensive evaluationof the intracranial and extracranial venous system.

Materials and Methods: Temporal signal decay, in-plane saturation and flow artifacts were assessed in aninstitutional review board approved, HIPAA compliantCEMRV study of 99 subjects. In a 39 subject subset, per-cent diameter narrowing of the internal jugular (IJ), bra-chiocephalic and azygous veins were coded according tothe following ordinal grades for both catheter venography(CV) and CEMRV: grade 0�50%, grade 1 >50%and�75%, grade 2 >75% and <100% and grade3¼100% and compared with pressure gradient measure-ments obtained during CV.

Results: There was no significant signal decay, in-planesaturation or flow artifacts identified on CEMRV or hemo-dynamically significant pressure gradients identified onCV. All brachiocephalic and azygous veins had matchedgrade 0 narrowing on both modalities. Discrepancybetween modalities occurred in the IJ veins at the level ofthyroid gland where 15% of IJ veins had CEMRVgrade�1 narrowing compared with 4% for CV or belowthe thyroid gland where 5% of IJ veins had CEMRVgrade�1 narrowing compared with 20% for CV. Therewas fair agreement (k¼0.24) between modalities for gradeof narrowing in the combined data set of all coded veins.

Conclusion: CEMRV using gadofosveset trisodium isaccurate in the evaluation of the venous system.

Key Words: MRV; contrast; gadofosveset trisodium; mul-tiple sclerosis; GadoliniumJ. Magn. Reson. Imaging 2013;00:000–000.VC 2013 Wiley Periodicals, Inc.

MR VENOGRAPHY (MRV) is a well-established modal-ity. Although MRV studies can be acquired withoutcontrast using two-dimensional (2D) time-of-flight(TOF) gradient-echo pulse sequences due to theirinherent sensitivity to slow flow (1), this approach haslimitations such as in-plane flow saturation, signalloss due to turbulence (2,3), limited through-planespatial resolution and spatial misregistration artifactsdue to physiologic motion. All these issues can makeimage interpretation problematic. To minimize in-plane saturation effects, TOF MRV is optimallyobtained perpendicular to the course of the vein (4).However, in areas of multidirectional venous anatomy,multiple acquisitions are often required to fully inter-rogate a region of interest. The presence of complexflow patterns surrounding valves, coalescing veins orvenous tortuosity may be associated with vortical orturbulence induced intravoxel phase dispersion (4,5)resulting in a pseudo-stenosis or a false intraluminalfilling defect (6), which is not readily correctible.

In-plane saturation effects can be reduced withcontrast-enhanced MRV (CEMRV) using intravenousgadolinium-based contrast agents (7–10) enablingacquisition of 3D volumetric data in any orientationwithout significant signal loss (11). With 3D volumet-ric data sets there is improved through-plane spatialresolution, reduced motion-induced misregistrationartifacts and shorter echo-times minimizing signalloss due to intravoxel phase dispersion seen in com-plex flow patterns (12). A possible limitation ofCEMRV is that the commonly available extracellularcontrast agents have a relatively short-lived venousphase and thus imaging strategies to acquire multipleanatomic locations with consistent image qualityacross a variable patient population may be challeng-ing (13).

1Department of Diagnostic and Interventional Imaging, University ofTexas Health Science Center at Houston, Houston, Texas, USA.2Department of Neurology, University of Texas Health Science Centerat Houston, Houston, Texas, USA.

Contract grant sponsor: National Multiple Sclerosis Society; Contractgrant number: RC 1019-A-5; Contract grant sponsor: the NationalCenter for Research Resources; Contract grant number: UL1RR024148.

*Address reprint requests to: L.A.K., Department of Diagnostic andInterventional Imaging, University of Texas Health Science Center atHouston 6431 Fannin Street, MSB 2.100, Houston, TX 77030. E-mail: larry.a.kramer@uth.tmc.edu

Received June 20, 2013; Accepted August 9, 2013.

DOI 10.1002/jmri.24409View this article online at wileyonlinelibrary.com.

JOURNAL OF MAGNETIC RESONANCE IMAGING 00:00–00 (2013)

CME

VC 2013 Wiley Periodicals, Inc. 1

Gadofosveset trisodium, with its prolonged venousenhancement and steady state properties (14) has thepotential to maximize the range of anatomic coverage,provide consistent intraluminal signal intensity overmultiple acquisitions and allow sufficient time for sus-pended respiration in multiple anatomic compart-ments such as the chest, abdomen and pelvis.Additional imaging time to repeat sequences degradedby motion artifacts is also available without the needfor a repeat contrast injection.

A promising application is the comprehensive evalu-ation of the extracranial venous outflow system,reported by Zamboni et al (15) to be abnormal at highincidence in patients with multiple sclerosis (MS)compared with controls. Although catheter venogra-phy (CV) was used to evaluate patterns of venousobstruction of the internal jugular (IJ) veins throughthe lumbar venous plexus it was not routinely appliedto all subjects due to ethical considerations thusunavoidably introducing selection bias. Furthermore,Doppler ultrasound combined with CV did not fullyinterrogate the dural sinuses or proximal internal jug-ular veins that are known sites of venous stenosis(16,17). The goal of this study was to evaluate theutility of CEMRV with gadofosveset trisodium to pro-vide an analysis of the intracranial and extracranialvenous outflow system.

MATERIALS AND METHODS

This prospective Health Insurance Portability andAccountability Act (HIPAA) compliant study wasapproved by our local institutional review board aspart of a larger effort evaluating the relationship ofaltered venous outflow from the central nervous sys-tem to MS. That study enrolled 276 adult volunteers(206 with MS; 11 healthy controls and 59 subjectswith cerebrovascular and other neurological disor-ders), all of whom underwent ultrasound of the major

intracranial and extracranial venous outflow (18).Imaging analysis reported here was performed onCEMRV and CV data acquired from 2011 through2012 on a subset of the MS subjects from the parenttrial (19). The decision to invite only MS subjects foradditional imaging was made by the trial principalinvestigator (JSW) and was not communicated toother radiologists until after the trial and analyseswere concluded. All subjects from the parent studyprovided written consent and written reconsent ifinvited to participate on the CEMRV and CV studiesafter the nature of the study and each of the subse-quent procedures were fully explained. A previous his-tory of allergic reaction to gadolinium-based contrastmaterial and/or renal insufficiency was criteria forexclusion.

Data Acquisition

The study was performed on a 3 Tesla (T), PhilipsIntera scanner using various combinations of the six-teen and fifteen channel SENSE compatible neurovas-cular and spine coils respectively with a maximumslew rate of 200mT/ms/m and maximum gradientamplitude of 80mT/m (Philips Medical Systems, Best,The Netherlands). Pulse sequence parameters for theMRV are shown in Table 1.

CEMRV was performed on 99 subjects (average age47.6 years, 69% females, average duration of MSsymptoms 14.2 years, average Expanded DisabilityStatus Scale score 2.3) using the intravenous admin-istration of gadofosveset trisodium (AblavarVR , Lan-theus Medical Imaging, N. Billerica, MA) administeredby a pressure injector at a dose of 0.03 mmol/kgbody weight at a rate of 0.28 ml/second followed by30 cc of saline flush.

A subset of the CEMRV subjects (n¼41, averageage 44.7 years, 69% females, average duration of MSsymptoms 11.1 years, average Expanded Disability

Table 1

Pulse sequence parameters. Sequence 1 is a pre-contrast 2D time-of-flight acquisition. Sequences 2–5 are post-contrast acquisitions and

Sequence 2 is obtained dynamically during contrast injection with a total of four continuous phases. Sequence 3 and 4 were acquired dur-

ing suspended respiration.

Parameter 2D axial T1 FFE

3D coronal T1

FFE 4 phases

3D coronal and

sagittal T1 TFE

3D coronal/sagittal

T1 TFE

3D sagittal

T1 TFE

Sequence 1 2 3 4 5

Repetition Time (ms) 6.1 4.6 3.3 3.3 8.1

Echo Time (ms) 3.0 1.3 1.5 1.5 3.7

Flip Angle 30 27 10 10 6

Field of View (cm) 35 30 37 37 25

Options Flow Compensation SPAIR SPAIR

Slice Thickness (mm) 3 3 4 4 1

Slice Spacing (mm) 2 1.5 2 2 1

Matrix 232 � 232 300 � 300 187 � 187 187 � 187 256 � 256

Repetitions 1 1 1 1 1

Time(min:sec) 1:39 1:23/5:33 00:19 00:19 5:56

Location Skull base/Neck Skull base/Neck Chest Abdomen/Pelvis Brain

Coil SENSE NV-16 SENSE NV-16 SENSE NV-16

SENSE Spine-15

SENSE Spine-15 SENSE NV-16

Reconstructed Pixel Size (mm) 0.4 � 0.4 � 3.0 0.7 � 0.7 � 1.5 1.5 � l.5 � 2.0 1.5 � l.5 � 2.0 1.0 � 1.0 � 1.0

Abbreviations: TFE¼Turbo Field Echo; FFE¼Fast Field Echo; SPAIR¼SPectral Attenuated Inversion Recovery; SENSE¼SENSitivity

Encoding for fast MRI; NV¼NeuroVascular.

2 Kramer et al.

Status Scale score 2.2) underwent CV performed byan interventional radiologist (AMC) with 36 years ofexperience by means of a common femoral veinapproach using an 18 gauge needle and Seldingertechnique. A 90-cm-long 6 french sheath (Cook Inc,Bloomington, IL) was placed over a 0.035-in Bensonguidewire (Boston Scientific, Natick, MA) and throughit a 5 French 125 cm vertebral catheter (Cook Inc,Bloomington. IL) selectively placed in the right IJ vein.The sheath was then passed over the catheter into thevein to be studied, except the azygous vein. A pres-sure transducer (ICU Medical, San Clemente, CA) con-nected by means of a y-connector was placed as closeto the right atrium level as possible and then zeroedto air. Pressure measurements were repeated at leasttwice for each vessel including pull back to the rightatrium. Up to five repeated measurements were madeat each site until considered reproducible. The repro-ducible measurement was recorded. The mean pres-sure for the right atrium was used. Venograms andsubsequent pressure measurements were thenobtained in the right IJ vein, followed by pullbackpressures in the right brachiocephalic vein, superiorvena cava and right atrium. This process wasrepeated on the left side starting at the left IJ vein.The venogram and subsequent pressure measurementof the azygous vein was obtained using a C-2 125-cm-long 5 french catheter (Cook, Bloomington, IL). Ante-roposterior projections were obtained 20 to 38 degreesapart at each station to estimate the degree of steno-sis. Because venography in either IJ vein allowed theentire outflow to the right heart to be visualized, injec-tions were not usually needed more centrally. Iohexol320 mg of iodine per ml (VisipaqueTM, General Elec-tric, Princeton, NJ) contrast material was handinjected at each site, from 3–5 cc per injection in theazygous to up to 10–20 cc per injection in the IJveins. A final injection and pressure measurementthrough the sheath was performed from the externaliliac vein on the side of catheterization. The cathetervenogram did not interrogate the proximal IJ veinsdue to intentional avoidance of catheter placementwithin the dural sinuses. All catheter venograms wereperformed on separate days and at varying intervalsfollowing the CEMRV.

Pressure gradients obtained on CV were codedaccording to the following two ordinal grades: grade0<3 mmHg and grade 1�3 mmHg. Hemodynamicallysignificant venous stenosis was defined by a pressuregradient�3 mmHg corresponding to a threshold valueof>50% diameter luminal narrowing (20).

Image Analysis

In-plane Saturation, Temporal Decay, and FlowArtifact Analysis

CEMRV in all subjects was subjectively evaluated by adiagnostic radiologist (LAK) with 26 years of magneticresonance imaging experience for evidence of signifi-cant loss of intraluminal signal due to temporal decayand in-plane saturation of the dural sinuses, IJ veins,brachiocephalic veins, superior and inferior vena cava,azygous vein, iliac veins, and lumbar venous plexus.

A total of 39 CEMRV subjects from the CV subsetunderwent flow artifact analysis by the same observer.The IJ veins were quantitatively evaluated for intralu-minal filling defects on 2D time-of-flight sequencesinitially according the following 3 ordinal grades:grade 0¼no defects, grade 1¼defect less than 50% oflumen, and grade 2¼defect greater than 50% oflumen. The highest grade of defect found between thepaired IJ veins was recorded for each subject andthen compared in a nonblinded manner at the samelocation to the CEMRV. Subjects were excluded ifthere was evidence of occlusion or filling defect on CV.

Catheter Venography Analysis

The percentage of narrowing of the IJ, brachioce-phalic, and azygous veins were estimated relative toan adjacent venous segment with stable maximaldiameter according to the following four ordinalgrades: grade 0�50%, grade 1� 51% and�75%,grade 2�76% and�99% and grade 3¼100%(occluded). The grade of narrowing was independentlycoded in a blinded manner by the same interventionaland diagnostic radiologists to determine inter-observer correlation.

CEMRV Analysis

Using the same CV analysis method and four ordinalgrades of narrowing, the corresponding CEMRV coro-nal maximum intensity projections (MIP) of the IJ,brachiocephalic and azygous veins were coded by twodiagnostic radiologists having 26 (L.A.K.) and 5(J.H.H.) years of MRI experience to determine inter-observer correlation. In addition, the smallest cross-sectional area (CSA) was determined for each IJ veinusing the 3D CEMRV dataset reconstructed into theaxial plane and then hand traced to determine CSAusing the open-source image processing software(Osirix [v.3.8.1]). A CSA�25 cm2 was considered ste-notic (21). The venous phase following the first passof contrast enhancement with the least motion arti-facts was preferentially evaluated.

To evaluate the hemodynamic significance of previ-ously reported abnormal cross-sectional morphologiccharacteristics of the IJ veins at the level of the thy-roid gland (22,23), CEMRV grade�1 narrowing wasadditionally coded (LAK) using axial reconstructionsaccording to the following classification: flattened withmaintained lumen (F), pinpoint narrowing (P) orfocally absent flow with distal reconstitution (A).Although category F is found in MS subjects inapproximate frequency to normal controls (24) andnot considered a definite abnormality, it is included inthe data for completeness. Results were comparedwith pressure gradient measurements, grade narrow-ing on CV and CSA.

Statistics

The nonparametric Wilcoxon test for paired data wasused to compare correlated measurements. Statisticalsignificance was considered at values of P�0.05 (25).

Contrast Enhanced MR 3

Kendall correlation was applied for intermodality com-parison at the level of the thyroid gland (26). Kappainterpretation of interobserver and intermodalityagreement was based on publication by Viera andGarrett (27). Statistical analyses were performedusing JMPVR software (version 10.0.2, SAS InstituteInc., Cary, NC) and MATLAB R12.1 Statistical Toolboxv 3.0 (The Mathworks Inc, Natick, MA).

RESULTS

CEMRV of brain, neck, chest, abdomen, and pelviswas successfully obtained on 98/99 (99%) consecu-tive MS patients. A localized intravenous catheterextravasation of contrast occurred in one subject, all

other studies were without complications. This exami-nation and corresponding CV study were excludedfrom further analysis because of inadequate enhance-ment levels seen on the dynamic CEMRV sequence.Thirty-nine of 41 subjects completed the CV compo-nent; 1 subject experienced a vasovagal response oninitial cannulation of the inguinal vein resulting inaborting the procedure and catheterization of the azy-gous vein could not be achieved in another subject.Of these 39 CV subjects, 39 had successful CEMRVstudies for correlation, however complete CV datawere only consistently available for approximately theinferior half of the IJ veins, which included the thy-roid level and distal IJ veins (Fig. 1a). Grade 0 pres-sure gradients were detected in all IJ, brachiocephalicor azygous veins.

Figure 1. Representative contrast enhanced magnetic resonance venography (CEMRV) images. a: Composite image of twoseparate coronal CEMRV acquisitions in the head and neck region obtained approximately 10 min apart. The top image rep-resents the delayed sequence acquired without subtraction. Note the signal intensity in the left IJ vein (*) is relatively main-tained on the delayed image. The bottom image is subtracted from a precontrast data set minimizing background signal. Thewhite line represents the midpoint of the IJ veins. b: Composite of two separate coronal CEMRV acquisitions that include thechest, abdomen and pelvis showing overlapping venous and arterial structures in the steady state of contrast enhancement.These images are displayed with an inverted gray scale. Precontrast data sets were used to subtract out the background tis-sue. Note that signal intensity of the veins is relatively maintained in all regions. c: Sagittal CEMRV image through the azy-gous vein. Both the horizontal (arrow) and the vertical portions (arrowheads) of the azygous vein show no in-plane saturationeffects. d: Magnified coronal CEMRV image showing excellent detail of the ascending lumbar vein (arrowheads) and left com-mon iliac vein (arrow). e: Axial CEMRV reconstructed image of the transverse (arrows) and sigmoid (arrowheads) sinusesobtained 12 min after injection of contrast showing excellent intraluminal signal retention.

4 Kramer et al.

ARTIFACTS

In 98 CEMRV, studies there was no significant signalloss in the venous system attributable to temporaldecay or in-plane saturation effects over an imagingspan of 12–14 min from the start of injection (Figs. 1and 2a,b). One of the subjects in the flow artifactanalysis study was excluded because of chronicvenous occlusion of the right IJ vein on CV. Withinthe remaining subset, there were no intraluminal clotsidentified on CV. The 2D TOF sequences averaged amaximal intraluminal filling defect grade of0.37 6 0.76 (min¼0, max¼2, median¼0, n¼15)compared with 0.04 6 0.19 (min¼0, max¼1,median¼0, n¼3) (P¼0.002) for CEMRV. A compari-son between 2D TOF MRV and CEMRV filling defectsis shown in Figures 2c,d.

Interobserver Correlation

There was almost perfect and substantial agreementin masked independent coding for grade of narrowingof the IJ, subclavian, and azygous veins between twoobservers for CEMRV (k¼0.86) and CV (k¼0.70),respectively.

Intermodality Comparison at the Level of theThyroid Gland

Fifteen percent of the IJ veins (12/78) had CEMRVgrade�1 narrowing with various combinations of flat-tening, pinpoint narrowing or absent flow with distalreconstitution of the thyroid level IJ veins on CEMRVcompared with 4% (3/78) grade�1 narrowing on CVand indicating a weak association between modalities

(Kendall’s tau-b¼0.26) (Table 2) (Fig. 3). Although allthese IJ veins had CSA measurements less than 0.25cm2 there were no hemodynamically significant pres-sure gradients observed (Table 2).

Intermodality Comparison Below the Level of theThyroid Gland

There was fair agreement between observers for cod-ing of CV and CEMRV for the distal jugular veinsbelow the thyroid gland (k¼0.25) (Table 3) with 53 of66 (80%) instances of grade 0 narrowing matched onboth CV and CEMRV and a single case of matchedGrade 3. In comparison 3 of 66 (5%) instances of IJveins below the thyroid gland had CEMRV grade�1narrowing compared with 13 of 66 (20%) instances forCV (Table 3) (Figs. 4, 5, and 6). Although 2 IJ veinsbelow the thyroid gland had CSA measurements lessthan 0.25 cm2, there were no hemodynamically signif-icant pressure gradients observed (Table 3).

Intermodality Comparison of the Brachiocephalicand Azygous Veins

Grade 0 narrowing was found in 78 of 78 (100%) bra-chiocephalic and 39 of 39 (100%) azygous veins onboth modalities with perfect intermodality agreement(k¼1.0) and confirmed with grade 0 pressure gradientin all cases.

Proximal Internal Jugular Veins

Narrowing of the proximal IJ veins at the level of C1was observed on CEMRV (Fig. 7) but was outside the

Figure 2. Examples of in-plane saturation and intraluminal artifacts on noncontrast 2D time-of-flight (TOF) MRV comparedwith CEMRV. Note complete saturation of the left transverse sinus (arrow) on reconstructed coronal 2D TOF MRV (a) com-pared with the left transverse sinus of the reconstructed coronal CEMRV image (b) (arrow). Similar but less pronounced effectis seen in the right transverse sinus (arrowheads). In the left jugular bulb on axial 2D TOF MRV (c) there is a grade 2 intralu-minal filling defect (arrow) that cannot be appreciated on corresponding axial CE MRV (d) (arrowhead).

Contrast Enhanced MR 5

range of the CV acquisition and thus coding was notperformed.

Summary Intermodality Comparison

Comparing modalities, grade of narrowing for allcoded veins were in fair agreement (k¼0.24).

DISCUSSION

Gadofosveset trisodium is a gadolinium based con-trast agent with a serum half-life of 2–3 h (28). Theprolonged intravascular half-life is achieved throughimmediate reversible noncovalent protein binding toalbumin mediated by a diphenylcyclohexyl group,minimizing first pass extraction of the gadoliniumchelate in the capillaries, and decreasing renal excre-tion. Bound albumin increases T1 relaxivity due to aslower molecular tumbling rate (29). The combinationof these physical properties suggests a potentialadvantage in CEMRV.

Our results confirm that CEMRV using a singledose intravenous injection of gadofosveset trisodiumin 98 patients with MS proved effective in the compre-hensive evaluation of the venous system without sig-nificant signal loss due to temporal decay or in-planesaturation. Intraluminal flow artifacts were reduced in

both severity and frequency compared with 2D TOFsequences. Although the IJ veins demonstrate signifi-cant variability in caliber throughout the neck, inter-observer agreement was substantial or better for bothCEMRV and CV grade of stenosis suggesting excellentquality of both data sets. Although there was perfectagreement of the grade of narrowing of the brachioce-phalic and azygous veins between CEMRV and CV,discrepancies in the grade of narrowing in the IJ veinsat or below the thyroid gland suggested modality spe-cific differences rather than a limitation of dataquality.

The 2D TOF MRV has been successfully applied inthe evaluation of deep venous thrombosis (DVT) of thepelvis and lower extremity (30,31), in part due to therelatively perpendicular orientation of these veins inthe axial acquisition plane, minimizing in-plane satu-ration. In contrast, CEMRV using gadofosveset triso-dium, can readily diagnose lower extremity DVT by3D acquisition in the time efficient but saturationprone coronal plane (14). Our study demonstratesthat CEMRV with gadofosveset trisodium is also effec-tive in characterizing even more complex multidimen-sional venous anatomy with either coronal or sagittalacquisitions.

Zivadinov et al considered the cross-sectionalCEMRV appearance of pinpoint or absent flow in the

Table 2

Morphologic characteristics of CEMRV grade>1 narrowing at the level of the thyroid gland compared to grade of narrowing on catheter

venography (CV), with corresponding pressure grade and cross-sectional area measurements (CSA). Grade 0¼� 50% narrowing,

A¼absent flow with distal reconstitution, P¼pinpoint flow and F¼ flattening. Shaded CSA measurements represent hemodynamically sig-

nificant values.

Subject

CEMRV

Type Right

CSA

Right (cm2)

CV Grade

Right

Pressure

Grade Right

CEMRV

Type Left

CSA

Left (cm2)

CV Grade

Left

Pressure

Grade Left

1 A 0 0 0 A 0 0 0

2 F 0.22 0 0 A 0 1 0

3 P 0.07 0 0 A 0 0 0

4 A 0 1 0 P 0 1 0

5 P 0 0 0 P 0 0 0

6 F 0 0 0 P 0.13 0 0

Abbreviations: A¼absent flow with distal reconstitution; P¼pinpoint flow; F¼ flattening.

Figure 3. Example of false positive narrowing on CEMRV. a: Coronal CEMRV demonstrates segmental absence of the distalright IJ vein (arrows) at the level of the thyroid gland (T). b: By contrast, the corresponding catheter venogram shows widepatency of the distal left IJ vein and no pressure gradient (0 mmHg) (arrows). c: Axial CEMRV reconstruction shows relation-ship of the IJ veins to the sternocleidomastoid muscle (SCM) at the level of the thyroid gland (T). Note absence of the left IJvein posterolateral to the left internal carotid artery (arrowhead) and flattened right IJ vein (arrow) immediately deep to theSCM.

6 Kramer et al.

Table 3

Comparison of grade of narrowing on catheter venography (C V) versus CEMRV in the IJ veins below the level of the thyroid gland with

corresponding pressure grade and CEMRV cross sectional area (CSA) measurements. Shaded CSA measurements represent hemody-

namically significant values.

Subject

CV

Grade (R)

CEMRV

Grade (R)

CSA (R)

(cm2)

Pressure

Grade (R)

CV

Grade (L)

CEMRV

Grade (L)

CSA (L)

(cm2)

Pressure

Grade (L)

1 0 0 0.86 0 0 0 0.37 0

2 3 3 NA NA 0 0 0.46 0

3 2 0 0.75 0 0 0 0.62 0

4 0 0 1.3 0 0 0 0.81 0

5 1 0 0.67 0 0 0 0.87 0

6 0 0 0.79 0 0 0 0.57 0

7 0 0 1.8 0 0 0 0.3 0

8 2 0 0.81 0 0 0 0.83 0

9 1 0 1.13 0 0 0 0.3 0

10 0 0 0.74 0 0 0 0.23 0

11 1 0 0.86 0 0 0 0.84 0

12 0 0 0.93 0 0 0 1.06 0

13 0 0 1.2 0 0 0 0.93 0

14 0 0 0.88 0 0 0 0.63 0

15 0 0 0.34 0 0 0 1.2 0

16 0 0 1.6 0 0 0 0.59 0

17 1 0 0.68 0 0 0 0.15 0

18 0 0 0.67 0 0 0 0.26 0

19 0 0 0.85 0 0 0 0.53 0

20 0 0 1.3 0 0 0 0.4 0

21 0 0 1.77 0 0 0 0.71 0

22 1 0 1.46 0 0 0 0.73 0

23 2 0 1 0 0 0 0.19 0

24 0 0 1.43 0 0 0 1.38 0

25 2 1 0.28 0 1 0 0.29 0

26 0 0 1.46 0 0 0 0.59 0

27 0 0 0.87 0 0 0 0.33 0

28 2 1 0.29 0 1 1 0.23 0

29 0 0 0.9 0 0 0 1.01 0

30 0 0 1.1 0 0 0 0.28 0

31 0 0 1.1 0 0 0 0.4 0

32 0 0 0.43 0 0 0 0.46 0

33 0 0 0.49 0 0 0 0.46 0

Abbreviations: (R)¼ right; (L)¼left; (NA)¼not applicable

Figure 4. Example of false positive narrowing on catheter venogram possibly due to localized vascular spasm. a: Cathetervenogram of the right IJ vein (i) demonstrates abrupt focal narrowing (grade 2) of the distal right IJ vein (arrow) proximal tothe confluence of the right subclavian vein (arrowhead). This was thought to represent a web at the time of study, howeverno pressure gradient (0 mmHg) was observed. b: By contrast, the corresponding CEMRV shows wide patency of the distalright IJ vein (arrow) in the same region.

Contrast Enhanced MR 7

IJ veins at the level of the thyroid glands as abnormaland identified this pattern in greater frequency in MSpatients compared with healthy controls although thehemodynamic significance was not confirmed withdirect pressure measurement (22). Our study demon-strates that pinpoint or absent flow of the IJ veins atthe level of the thyroid gland, although associated insome cases with an abnormality on CV, did not corre-spond to any detectable hemodynamically significantpressure gradient indicating a false positive finding inall cases. We hypothesize that severe narrowing of theIJ vein at the thyroid gland level is related to extrinsiccompression by the sternocleidomastoid muscle andsurrounding soft tissues relative to highly collapsibleIJ veins. We also hypothesize that a transient change

in fluid volume after local contrast injection couldresult in mild distention of the IJ vein, whereasperipherally injected contrast volume would be rapidlydissipated in the systemic venous system.

In the distal segment of the IJ vein, CV identifiedgrade�1 narrowing at significantly higher frequencythan CEMRV. A previous study comparing CEMRVand CV in the IJ veins also noted this discrepancy,but attributed the result to the low specificity andpositive predictive value of MR venography technique(23). In that study, no concomitant pressure measure-ments were obtained. In comparison, grade�1 nar-rowing identified by CV in our study corresponded tograde 0 pressure gradients in all cases indicating lessthan 50% diameter stenosis and alternatively

Figure 5. Laminar flow characteristics. a: CV of the right IJ vein (i) demonstrates dilute contrast from the right subclavian (s)and external jugular (*) veins converging with the right IJ vein (arrow) resulting in visualization of the postvalve recess (arrow-heads) and delineating the valve leaflet in the distal IJ vein. Laminar flow arising from the center of the valve is seen as asolid column of contrast (arrow) creating the appearance of focal narrowing. b: The corresponding CEMRV demonstrates com-plete opacification of the external jugular, subclavian and brachiocephalic veins without narrowing at the distal IJ vein. Thepost valve recess is estimated (arrowheads). c: In another subject, the external jugular and subclavian veins are not opacifiedand thus there is no delineation of the valve. The hypothetical location of the postvalve recess or nonopacified collateral flow(arrowheads) is estimated. Laminar flow arising from the center of the vein is seen as a solid column of contrast (arrow) creat-ing the appearance of focal narrowing (grade 1), however the pressure gradient is only 1 mmHg. d: The corresponding CEMRVdemonstrates a well opacified external jugular, subclavian and brachiocephalic veins without distal IJ vein narrowing (arrow-heads). Compared with the first case the external jugular vein branch (*) joins the IJ vein proximal to the subclavian vein pro-viding an additional column of nonopacified laminar flow contributing further to apparent narrowing.

8 Kramer et al.

suggesting a diagnostic limitation of CV at this level.Although the exact explanation for this phenomenonis not well understood, possibilities include transientcatheter or contrast-induced vasospasm (32,33),incomplete luminal opacification immediately down-stream to a valve leaflet (34), or from the confluenceof nonopacified venous flow. We hypothesize that acolumn of opacified lumen distal to the IJ vein valveleaflet is maintained discretely from nonopacifiedlumen secondary to the laminar flow characteristicsin veins (35). It is important to note that the valve inthe IJ vein is on average within 3 mm of the junctionwith the brachiocephalic vein and present bilaterallyin 87% of cases (36). All abnormalities in the IJ veinbelow the thyroid level we identified with CV werenear the brachiocephalic vein and likely associatedwith a valve in the majority of cases. Thus, in somestudies apparent narrowing is related to fluidmechanical patterns and not anatomic restriction.

In comparison, with intravenous gadofosveset triso-dium, complete vascular distribution of contrast,referred to as the steady state, is achieved within 4–5min of injection (14,37), eliminating any differences inopacification across the vein lumen, postvalverecesses or with adjacent collateral flow. This steady-state property may represent a distinct advantage ofan intravascular contrast agent in a region of complex

flow where first-pass enhancement is potentiallyincomplete. Further studies, however will be neededto confirm this hypothesis.

CSA of the IJ vein�25 cm2 is considered hemody-namically significant as it represents a third of thenormal average CSA (38). This is less than 30 mm2

used by Zamboni et al (15) to determine functionalsignificance of IJ stenosis based on neurosonographyassessment. Although 8 of 33 (24%) of distal IJ veinshad CSA�30 cm2, no significant pressure gradientwere observed even below CSA�25 cm2 indicatingfunctional significance based on CSA is just as limitedin the IJ veins as percent stenosis. The thin wall andeasy collapsibility of the IJ veins make assessmentbased on CSA or percent stenosis prone to error.

A major disadvantage of the gadofosveset trisodiumin the evaluation of the venous system is persistentarterial enhancement resulting in overlapping vascu-lar structures on reconstructed images. Ideally, withdynamic acquisitions, a peak arterial phase can besubtracted from the peak venous phase resulting invenous only signal (39). This approach is applicableto a single region of interest, but is less effective tosubsequent stations as residual steady state venousenhancement is also subtracted out. This disadvant-age can be partially obviated by the use of a worksta-tion to create subvolumetric maximum intensityprojections of the veins minimizing overlapping ofvenous and arterial structures. Background signalsuppression using precontrast masks in the neck,chest, abdomen, and pelvis is also useful in improvingdelineation of the extracranial venous system.

There were several limitations of our study. First,all subjects included in our analysis had a diagnosisof MS, which could potentially limit generalizing ourresults. Second, CV and CEMRV were not obtained

Figure 6. Illustration of laminar flow characteristicshypothesized at the confluence of the external jugular, IJ,subclavian and brachiocephalic veins. Note the narrowedcolumn of laminar flow (arrow) distal to the right IJ veinvalve leaflet (arrowhead). Countercurrent flow is indicated bycircular flow pattern immediately downstream to valvemargin.

Figure 7. Volume rendered CEMRV coronal image showingcompression of the IJ veins (arrows) at the level of the trans-verse process of C1 (*). O¼odontoid process; i¼ internal jug-ular vein.

Contrast Enhanced MR 9

concurrently, therefore, subject hydration or cardio-vascular status at the time of the study could influ-ence the diameter of veins between examinations.Additionally, methodological differences relative tosubject position or changes in intrathoracic pressureduring suspended respiration or valsalva maneuvercould also influence degree of perceived stenosis (23).Third, we did not directly compare gadofosveset triso-dium with other gadolinium-based contrast agentsthat have weak protein binding characteristics suchas gadobenate dimeglumine (40) and gadobutrol (41).A study of CCSVI with gadobutrol only resulted in suf-ficient image quality to assess 35 of 40 azygous veinsand did extend imaging to the abdomen or lumbarvenous plexus (13). Finally CV comparison of theproximal IJ veins and lumbar venous plexus eval-uated with CEMRV was beyond the scope of the studyand limits understanding of venous narrowing com-monly identified at the C1 level reported to be relatedto compression by the styloid process or posteriordigastric belly (17).

In conclusion, CEMRV using a single dose of gado-fosveset trisodium can provide evaluation of intracra-nial and extracranial venous system. Steady stateCEMRV provides continuity over multiple anatomiccompartments that cannot be duplicated with neuro-sonography and CV combined. Although CV is consid-ered the reference modality, overestimation ofhemodynamically significant stenoses in the distal IJveins by morphologic assessment alone is a potentiallimitation also demonstrated with CEMRV andemphasizing the importance of direct pressure mea-surement to confirm physiologically relevant disease.

ACKNOWLEDGMENTS

Drs. J. William Lindsey, Flavia Nelson, and John A.Lincoln were kind enough to refer several patients forthese volunteer studies. The authors thank Vipulku-mar S. Patel, RTR for his technical assistance withdata acquisition. The study was supported by theNational Multiple Sclerosis Society with limited per-sonnel support provided through our local Center forClinical and Translational Sciences funded by theNational Center for Research Resources.

REFERENCES

1. Gao JH, Gore JC. NMR signal from flowing nuclei in fastgradient-echo pulse sequences with refocusing. Phys Med Biol1994;39:2305–23218.

2. Oshinski JN, Ku DN, Pettigrew RI. Turbulent fluctuation velocity:the most significant determinant of signal loss in stenotic vessels.Magn Reson Med 1995;33:193–199.

3. Evans AJ, Blinder RA, Herfkens RJ, et al. Effects of turbulenceon signal intensity in gradient echo images. Invest Radiol 1988;23:512–518.

4. Ayanzen RH, Bird CR, Keller PJ, McCully FJ, Theobald MR,Heiserman JE. Cerebral MR venography: normal anatomy andpotential diagnostic pitfalls. AJNR Am J Neuroradiol 2000;21:74–78.

5. Siegel JM Jr, Oshinski JN, Pettigrew RI, Ku DN. Computationalsimulation of turbulent signal loss in 2D time-of-flight magneticresonance angiograms. Magn Reson Med 1997;37:609–614.

6. Widjaja E, Shroff M, Blaser S, Laughlin S, Raybaud C. 2D time-of-flight MR venography in neonates: anatomy and pitfalls. AJNRAm J Neuroradiol 2006;27:1913–1918.

7. Kirchhof K, Welzel T, Jansen O, Sartor K. More reliable noninva-sive visualization of the cerebral veins and dural sinuses: com-parison of three MR angiographic techniques. Radiology 2002;224:804–810.

8. Rollins N, Ison C, Reyes T, Chia J. Cerebral MR venography inchildren: comparison of 2D time-of-flight and gadolinium-enhanced 3D gradient-echo techniques. Radiology 2005;235:1011–1017.

9. Bosmans H, Marchal G, Lukito G, et al. Time-of-flight MR angiog-raphy of the brain: comparison of acquisition techniques inhealthy volunteers. AJR Am J Roentgenol 1995;164:161–167.

10. Klingebiel R, Bauknecht HC, Bohner G, Kirsch R, Berger J,Masuhr F. Comparative evaluation of 2D time-of-flight and 3Delliptic centric contrast-enhanced MR venography in patientswith presumptive cerebral venous and sinus thrombosis. Eur JNeurol 2007;14:139–143.

11. Blatter DD, Parker DL, Robison RO. Cerebral MR angiographywith multiple overlapping thin slab acquisition. Part I. Quantita-tive analysis of vessel visibility. Radiology 1991;179:805–811.

12. Lev MH, Romero JM, Gonzalez RG. Flow voids in time-of-flightMR angiography of carotid artery stenosis? It depends on the TE!AJNR Am J Neuroradiol 2003;24:2120.

13. Doepp F, Wurfel JT, Pfueller CF, et al. Venous drainage in multi-ple sclerosis: a combined MRI and ultrasound study. Neurology2011;77:1745–1751.

14. Hadizadeh DR, Kukuk GM, Fahlenkamp UL, et al. SimultaneousMR arteriography and venography with blood pool contrast agentdetects deep venous thrombosis in suspected arterial disease.AJR Am J Roentgenol 2012;198:1188–1195.

15. Zamboni P, Galeotti R, Menegatti E, et al. Chronic cerebrospinalvenous insufficiency in patients with multiple sclerosis. J NeurolNeurosurg Psychiatry 2009;80:392–399.

16. Farb RI, Vanek I, Scott JN, et al. Idiopathic intracranial hyperten-sion: the prevalence and morphology of sinovenous stenosis. Neu-rology 2003;60:1418–1424.

17. Jayaraman MV, Boxerman JL, Davis LM, Haas RA, Rogg JM.Incidence of extrinsic compression of the internal jugular vein inunselected patients undergoing CT angiography. AJNR Am J Neu-roradiol 2012;33:1247–1250.

18. Barreto AD, Brod SA, Bui TT, et al. Chronic cerebrospinal venousinsufficiency: case-control neurosonography results. Ann Neurol2013;73:721–728.

19. Brod SA, Kramer LA, Cohen AM, et al. Chronic cerebrospinalvenous insufficiency: masked multimodal imaging assessment.Mult Scler 2013 [Epub ahead of print].

20. Labropoulos N, Borge M, Pierce K, Pappas PJ. Criteria for defin-ing significant central vein stenosis with duplex ultrasound. JVasc Surg 2007;46:101–107.

21. Feng W, Utriainen D, Trifan G, et al. Characteristics of flowthrough the internal jugular veins at cervical C2/C3 and C5/C6levels for multiple sclerosis patients using MR phase contrastimaging. Neurol Res 2012;34:802–809.

22. Zivadinov R, Lopez-Soriano A, Weinstock-Guttman B, et al. Useof MR venography for characterization of the extracranial venoussystem in patients with multiple sclerosis and healthy controlsubjects. Radiology 2011;258:562–570.

23. Zivadinov R, Galeotti R, Hojnacki D, et al. Value of MR venogra-phy for detection of internal jugular vein anomalies in multiplesclerosis: a pilot longitudinal study. AJNR Am J Neuroradiol2011;32:938–946.

24. Hojnacki D, Zamboni P, Lopez-Soriano A, et al. Use of neck mag-netic resonance venography, Doppler sonography and selectivevenography for diagnosis of chronic cerebrospinal venous insuffi-ciency: a pilot study in multiple sclerosis patients and healthycontrols. Int Angiol 2010;29:127–139.

25. Glantz SA. Primer of biostatistics. 6th ed. New York: McGraw-HillMedical; 2005.

26. Leach C. Introduction to statistics: a nonparametric approach forthe social sciences. New York: Wiley; 1979.

27. Viera AJ, Garrett JM. Understanding interobserver agreement:the kappa statistic. Fam Med 2005;37:360–363.

28. Parmelee DJ, Walovitch RC, Ouellet HS, Lauffer RB. Preclinicalevaluation of the pharmacokinetics, biodistribution, and

10 Kramer et al.

elimination of MS-325, a blood pool agent for magnetic resonanceimaging. Invest Radiol 1997;32:741–747.

29. Lauffer RB, Parmelee DJ, Dunham SU, et al. MS-325: albumin-targeted contrast agent for MR angiography. Radiology 1998;207:529–538.

30. Carpenter JP, Holland GA, Baum RA, Owen RS, Carpenter JT,Cope C. Magnetic resonance venography for the detection ofdeep venous thrombosis: comparison with contrast venographyand duplex Doppler ultrasonography. J Vasc Surg 1993;18:734–741.

31. Cramer SC, Rordorf G, Maki JH, et al. Increased pelvic veinthrombi in cryptogenic stroke: results of the Paradoxical Embolifrom Large Veins in Ischemic Stroke (PELVIS) study. Stroke2004;35:46–50.

32. Duan X, Ling F, Shen Y, Yang J, Xu HY. Venous spasm duringcontrast-guided axillary vein puncture for pacemaker or defibril-lator lead implantation. Europace 2012;14:1008–1011.

33. Barber CJ. Central venous catheter placement for intravenousdigital subtraction angiography: an assessment of technical prob-lems and success rate. Br J Radiol 1989;62:599–602.

34. Singh RN, Salvoza MI. Laminar flow due to venous valves mas-querading as vein graft spasm. Cathet Cardiovasc Diagn 1983;9:569–575.

35. Helps EP, Mc DD. Observations on laminar flow in veins. J Phys-iol 1954;124:631–639.

36. Harmon JV Jr, Edwards WD. Venous valves in subclavian andinternal jugular veins. Frequency, position, and structure in 100autopsy cases. Am J Cardiovasc Pathol 1987;1:51–54.

37. Nikolaou K, Kramer H, Grosse C, et al. High-spatial-resolutionmultistation MR angiography with parallel imaging and blood poolcontrast agent: initial experience. Radiology 2006;241:861–872.

38. Mayer CA, Pfeilschifter W, Lorenz MW, et al. The perfect crime?CCSVI not leaving a trace in MS. J Neurol Neurosurg Psychiatry2011;82:436–440.

39. Lebowitz JA, Rofsky NM, Krinsky GA, Weinreb JC. Gadolinium-enhanced body MR venography with subtraction technique. AJRAm J Roentgenol 1997;169:755–758.

40. Cavagna FM, Maggioni F, Castelli PM, et al. Gadolinium chelateswith weak binding to serum proteins. A new class of high-efficiency, general purpose contrast agents for magnetic reso-nance imaging. Invest Radiol 1997;32:780–796.

41. Platzek J, Blaszkiewicz P, Gries H, et al. Synthesis and structureof a new macrocyclic polyhydroxylated gadolinium chelate usedas a contrast agent for magnetic resonance imaging. Inorg Chem1997;36:6086–6093.

Contrast Enhanced MR 11