Original Research

MR-Guided Placement of a Temporary Vena Cava Filter: Technique and Feasibility

Christian Frahm, MD Hans-Bjorn Gehl, MD Heike Lorch, MD Martin Zwaan, MD Matthias Drobnitzky, BSc Gerhard A. Laub, PhD Hans-Dieter Weiss, MD

The purpose of this paper was to investigate the feas- ibility of MR-guided insertion of a temporary vena cava filter on an open low-field imager. In vivo procedures were performed on four anesthetized pigs using a com- mon nonferromagnetic temporary vena cava filter and a special prototype guidewire developed for vascular interventions guided by low-field MRI. Breath-hold spoiled gradient-echo sequences (fast low-angle shot [FLASH]) with flow compensation were used for posi- tion monitoring of the passively visualized intravas- cular devices. Using the described technique and equipment, all steps of the procedure were feasible in the MR unit. Practicability of the procedure seems to be sufficient for clinical purposes but was inferior com- pared to the conventional technique of filter place- ment.

Index terms: lnterventional MRI *Vascular interventions * Vena cava fil- ter * Passive \isualization . Low-field MRl

JMRl1998: 8 :105109

Abbreviations: 2DFT = two-dimensional Fourier transform. FLASH = fast

From the Department of Radiology. Medical University of Liibeck (C.F.. H -B.G.. H.L.. M.Z.. H -D.W.). Ratzeburger Allee 160, D-23538 Iiibeck. Ger- many. and Sitmens Medical Engineering (M.D.. G A.L.I. Erlangen, Germany. Received Janiiary 30. 1997 revision requested March 11: revision received May 30: accepted J u n e 18. Address reprint requests to C.F.

" ISMRM. 1998

THE TEMPORARY PLACEMENT of a removable filter into the inferior vena cava (WC) is applied to reduce the risk of pulmonary embolism in patients with known acute thrombosis of the iliac or iliofemoral veins and/or W C submitted to thrombolysis (1-4) or undergoing major ab- dominal and pelvic operations (2,4,5). The latter group also includes the cases with pregnancy thrombosis un- dergoing cesarean section (2,4).

We performed in vivo animal experiments on an open low-field magnet to investigate the feasibility of MR- guided placement of a temporary IVC filter using passive device visualization.

MATERIACS AND METHODS The procedures were performed on an open low-field

imager with vertical field axis (B, = .2 T, maximum gra- dient strength = 10 mT/m; Magnetom OPEN, Siemens Medical Systems, Erlangen, Germany).

The in vivo experiments were performed on four pigs (44-57 kg) under general anesthesia with artificial res- piration (intubation through a tracheal incision). The experimental protocol had been approved by the appro- priate institutional and governmental regulatory commit- tees.

For image-guided insertion of the guidewire and place- ment of the vena cava filter, we used two-dimensional Fourier transform (2DFT) spoiled gradient-echo se- quences (fast low-angle shot [FLASH]) with flow compen- sation suitable for breath-hold imaging of the large vessels of the body (imaging protocol I - for in-plane im- aging of the vessel in coronal, paracoronal, sagittal, or parasagittal slice direction: TR = 142 msec, slice-multi- plexed mode, TE = 18 msec, flip angle = 30-40°, matrix = 128 X 256 or 112 X 256, field of view [FOV] = 350 X 350 mm or 306 X 350 mm, one excitation, 6-mm slice thickness, 4 slices, acquisition time = 20 or 18 seconds, bandwidth = 39 Hz/pixel; imaging protocol 2 - for trans- verse slice direction, perpendicular to the vessel of inter- est: TR = 22 msec, single-slice mode, TE = 9 msec, flip angle = 40". matrix = 112 X 256, FOV = 306 X 350 mm, one excitation, slice thickness = 8 mm, 26-second ac- quisition time for 10 slices, bandwidth = 78 Hz/pixel). Imaging was directed within the scan room using a sec- ond mouse console with a shielded monitor.

The intravascular devices used in the procedure were visualized passively by their susceptibility artifacts.

105

Figure 1. Components of the filter kit (Antheor, Meditech/Bos- ton Scientific. Natick, MA, USA) completed by a MR-compatible prototype guidewire (Daum GmbH, Schwerin, Germany). (1) The self-expanding filter. permanently attached to the tip of a plastic catheter that permits the passage of a guidewire. (2) The long dilator with a metal ring within its tip. (3) The long filter sheath, fitting the length and the outer diameter of the filter catheter and the dilator. The filter is advanced as well as retracted through this sheath. Filter catheter and filter sheath must be left in the access vein during the temporary vena cava filtration.

Initial guidewire insertion into the IVC was performed with a new prototype guidewire (developed at our depart- ment in cooperation with Daum GmbH, Schwerin, Ger- many), designed for passive visualization in a low-field system such as our scanner (6). Applying the above-men- tioned imaging protocols at .2 T, and if the guide wire is positioned at an angle of approximately 90” to B,,, the wire causes artifacts of approximately 5 to 7 mni in diameter.

We used a temporary IVC filter kit, which is usually applied in the conventionally performed procedures in our department (Antheor, Meditech/Boston Scientific, Natick, MA) (Fig. 1). The filter is made of nonferromag- netic material (Phynox alloy with biocarbon coating) and described as MR-compatible within the manufacturer’s directions for use; the explanation of the manufacturer concerning the MRI compatibility of the filter agrees at least with “MRI compatibility of the first kind” as defined by Schenck (7). The filter kit comprises a self-expanding filter (at full expansion: diameter, 3 1 mm; length, 50 mm) fixed at a plastic catheter, and an introducer set consist- ing of a long sheath (outside diameter, 9 Fr) and a dilator (outside diameter, 7 Fr). Before the in vivo procedures, we performed in vitro trials using a simple flow-phantom and using imaging protocols 1 and 2 to check on the ar- tifact size of the filter and its introducer set (Fig. 2).

Placement Procedure The goal was to achieve infrarenal filter position within

the IVC, ie, a position below the leveI of the renal veins. The left or right internal jugular vein was accessed surgi- cally by phlebotomy and a usual short 9-Fr introducer sheath was inserted into the vein and secured by suture. The guidewire was passed through this introducer sheath and manipulated into the IVC via the internal jugular vein, brachiocephalic vein, superior vena cava, and the right atrium (Figs. 3a and 3b). Possible deviations of the guide- wire into the right ventricle, a hepatic vein, or a renal vein had to be recognized and corrected under MR guidance

b. Figure 2. In vitro MRI of the expanded Antheor filter using a simple flow phantom (inner tube diameter, 26 mm; water inside and outside the tube: flow rate within the tube approximately 15 cm/sec). The course of filter catheter and filter sheath is marked by passive visualization of the MR-compatible guidewire (arrows in a and b) passed through filter catheter and filter. (a) Coronal plane (imaging protocol 1) . b) Transverse plane (imaging proto- col 2).

(Fig. 4). The table position had to be readjusted at least once, because the region of sensitivity of the body coil is limited to approximately 35 cm in craniocaudal direction. When the desired position of the guidewire within the IVC had been reached, the introducer set of the filter (the long filter sheath containing the dilator) was passed over the guidewire (Fig. 3c). The tip position of the filter introducer could be checked on by a moderate artifact produced by

106 JMRl JanuaryIFebruary 1998

Figure 3. Passive device visualiza- tion during filter placement in vivo. (a) Breath-hold imaging of the large abdominal vessels (paracoronal plane, imaging protocol 1). (b) Inser- tion of the MR-compatible guidewire (straight arrows) into the N C and left common iliac vein (via right internal jugular vein, brachiocephalic vein, superior vena cava, and right atrium). (c) The filter introducer [long sheath and dilator) was passed over the guidewire. The position of the di- lator tip (or the end of the long filter sheath) can be checked on approxi- mately by a ball-shaped artifact caused by the metal ring within the dilator tip (curved arrow]. The end of the filter sheath is positioned below the level of the renal veins (arrow- head at the right renal vein). (a) The dilator was removed, and the filter catheter was advanced through the filter sheath: the expanded filter (curved arrow] is placed within the IVC in the infrarenal position [arrow- head at the right renal vein). The course of catheter or sheath can be delineated by the artifact of the MR- compatible guidewire introduced into the filter catheter [straight arrow at the J-tip of the guidewire).

C.

a metal ring incorporated within the dilator tip. That means that the filter sheath lacks a constituent suitable for passive visualization and is only localized by the di- lator metal ring (which was primarily designed a s a ra- diopaque marker to work in a comparable way in the conventional placement procedure). The course of the introducer was visualized by the guidewire (Fig. 3c). Af- ter positioning the filter introducer, the guidewire and the dilator were removed. Then the filter was passed into the IVC through the remaining long sheath (Fig. 3d). Leaving the end of the filter sheath, the filter expands. Advancing the filter through its long introducer sheath, the course of the filter catheter was visualized by intro- ducing the guidewire into the filter catheter. The posi- tion of the filter itself was recognizable due to its own susceptibility artifact (Fig. 3d).

RESULTS Using the technique and equipment described above,

the MR-guided placement of the temporary IVC filter was feasible: 1. The visualization of the MR-compatible guidewire was

reliable within the vessels relevant for the procedure, regarding the wire course as well as the tip position. Deviations of the guidewire into the right ventricle (all animals), a hepatic vein (one animal), or a renal vein (two animals) and the development of wire loops

within the IVC (two animals) could be recognized and corrected.

2. Passive visualization and localization of the filter in- troducer and the filter itself, respectively, were suffi- cient to achieve correct (infrarenal) filter placement within the IVC. The susceptibility artifact caused by the expanded filter depends on the degree of expan- sion. In vivo, with the filter positioned at 90" to B, within the IVC, the maximum length was approxi- mately 7.6 to 7.8 cm and the maximum width was 2.1 to 2.4 cm using imaging protocol 1 (coronal slice di- rection) or 1.8 to 2.4 cm applying imaging protocol 2. Imaging the more expanded filter within the flow phantom (inner tube diameter, 26 mm; filter posi- tioned at 90" to B,), we found a maximum artifact length of about 7.0 cm and a maximum width of ap- proximately 3.3 cm for protocol 1, and a maximum width of 2.6 cm (corresponding to actual expansion) for protocol 2. Additionally, clear delineation of the fil- ter struts was possible under these conditions (Fig. 2). In vitro and in vivo, the nonexpanded filter caused an artifact with a length of approximately 8.0 cm (proto- col 1) and a maximum width of 2.0 cm (protocol 1) or 1.7 cm (protocol 2). The metal ring incorporated into the dilator tip produced an artifact with a length of about 4.5 cm and a width of approximately 4.2 cm using protocol 1 with coronal slice direction and a maximum width of approximately 4.5 cm using pro-

Volume 8 * Number 1 - JMRl - 107

Figure 4. Deviation of the guidewire tip (straight arrow) into the right renal vein (paracoronal plane, imaging protocol 1).

tocol 2 (dilator positioned at 90" to B, within the NC or flow phantom, respectively).

3. Magnet design and receive coil design offered suitable access for the interventional manipulations. The ani- mals remained within the magnet during the entire procedure.

4. The time needed for filter insertion varied between 20 and 45 minutes (time for tracheotomy and phlebotomy not included).

DISCUSSION Using a low-field scanner without special hardware

modificat ions, a conventional nonferromagnetic temporary NC filter, and passive visualization of the intravascular devices, the feasibility of MR-guided filter placement was shown in vivo. Our investigations were particularly stim- ulated by some prior publications dealing with the suit- ability of usual intravascular devices (such as IVC filters, stents, and coils) for MRI (8-10). From these papers, it can be concluded that MRI of the IVC is feasible in pres- ence of nonferromagnetic filters even in high field units. Various aspects of passive device visualization (such as relevant influence factors of artifact size and their prac- tical implications, suitability of materials and devices) have been investigated and discussed elaborately with re- spect to nonvascular and vascular interventions (7, l l - 16). Therefore, the vertical field axis of low field magnets is advantageous for passive visualization of intravascular devices within the large vessels of the body because the angle between the device and B, - the most important factor influencing artifact size - does not vary consid- erably. On the other hand, the opportunities to accelerate the interventional imaging are relatively limited com- pared to mid or high field.

The spatial resolution (pixel size, 2.7 X 1.4 mrn) and especially the temporal resolution of the described im- aging technique for MR-guided filter insertion are clearly inferior compared to the conventionally performed pro- cedure monitored by x-ray fluoroscopy combined with

angiography. No real-time position monitoring is possible during the intravascular manipulations. This disadvan- tage becomes relevant mainly in the phase of initial guidewire positioning and might result in longer proce- dure times in difficult cases. The alternative experimental concept of active MR tracking - using intravascular de- vices equipped with small receive coils - offers higher tracking speeds but requires an incomparably higher technical expense and seems to be limited to road map techniques until now (17,18).

Furthermore, more accurate localization of the dilator tip (or better delineated ending of the filter sheath, re- spectively) is desirable and could be achieved by simple modifications of the incorporated metal ring to decrease the size of its susceptibility artifact (for example, by nar- rowing the ring or making the ring of another alloy). Due to the different echo times firstly, all artifacts differed in size depending on the imaging protocol applied (1 1). The position and deployment of the filter are also identified by a particular size and shape of susceptibility artifact. As in the animal experiments, the individual struts of the filter probably would not be clearly depicted applying our imaging protocols in man because the diameter of the IVC or degree of filter expansion would be very similar to those in the animal model (19). On the other hand, in- adequate expansion and damage during placement pro- cedure are very improbable due to the simple and robust filter design and did never occur in our patients under- going conventional placement of the temporary Antheor filter (until now, 57 cases). Thus, delineation of the in- dividual filter struts seems to be not essential for filter placement.

After all, the practicability of our technique appears just sufficient for clinical purposes. The limited localiza- tion accuracy (due to low spatial resolution as well as artifact sizes of the devices) is no considerable disadvan- tage in the particular procedure discussed here, because the relevant vessel sections have relatively large diame- ters, and adequate filter positioning requires no precision work. Although the usual fluoroscopy-guided filter inser- tion is undoubtedly preferable in most cases, a true clin- ical indication for the MR-guided procedure is imaginable in a particular group of patients: x-ray exposure and ad- ministration of contrast agent are avoided, and these as- pects might be decisive in giving preference to the MR-guided technique to place a temporary NC filter in pregnant patients with deep vein thrombosis before ces- arean section.

Acknowledgments: The authors thank Joachim Graessner (Sie- mens Medical Engineering, Hamburg, Gemany) for his valuable technical assistance and Keyvan Farahani (Department of Radiolog- ical Sciences, UCLA School of Medicine, Los Angeles, CA) and Jonathan Lewin (Department of Radiology, University Hospitals of Cleveland and CWRU, Cleveland, OH) for the extended and fruitful discussions via e-mail.

References 1. Kunisch M, Rauber K, Bachmann G, Rau WS. Temporarer Ka-

vafilter: Effektive Prophylaxe von Lungenembolien bei venosen Thrombosen im Bereich der Beckenstrombahn und der Vena cava inferior? Fortschr Geb Rontgenstr Neuen Bildgeb Verfahr

Linsenmaier U. Rieger J, Rock C. et al. Indications and man- agement of temporary and retrievable inferior vena cava filters (abstract). Cardiovasc Intervent Radio1 1996; 19(Suppl 2):S94.

3. Thery C, Asseman P, Amrouni N. et al. Use of a new removable vena cava filter in order to prevent pulmonary embolism in pa- tients submitted to thrombolysis. Eur Heart J 1990: 11:334-341. Zwaan M, Kagel C, Marienhoff N, et al. Erste Erfahrungen mit temporSren Vena-cava-Filtern. Fortachr Geb Rontgenstr Neuen Bildgeb Verfahr 1995; 163:171-176.

1995; 163~523-526. 2.

4.

108 - JMRl January/February 1998

5.

6.

7.

8.

9.

10.

11.

12.

Millward SF, Bormanis J. Burbridge BE, Markman SJ, Peterson RA. Preliminary clinical experience with the Gunther tempo- rary inferior vena cava filter. J Vasc lnterv Radiol 1994: 5:863- 868. Frahm C. Gehl HB. Melchert UH, et al. Passive visualization - a suitable approach for MR-guided vascular interventions in low-field systems? In vivo evaluation of a new guide wire. In: Proceedings of the 5th scientific meeting of the International So- ciety for Magnetic Resonance in Medicine. Berkeley, CA: Inter- national Society for Magnetic Resonance in Medicine, 1997; 1931. Schenck JF. The role of magnetic susceptibility in magnetic resonance imaging: MRI magnetic compatibility of the first and second kinds. Med Phys 1996: 23:815-850. Kim D, Edelman RR, Margolin CJ, et al. The Simon nitinol fil- ter: evaluation by MR and ultrasound. Angiology 1992; 43: 541-548. Teitelbaum GP, Bradley WG, Klein BD. MR imaging artifacts, ferromagnetism. and magnetic torque of intravascular filters. stents, and coils. Radiology 1988; 166:657-664. Teitelbaum GP, Ortega H V , Vinitski S, et al. Low-artifact intra- vascular devices: MR imaging evaluation. Radiology 1988;

Frahm C. GehI HB. Melchert UH. Weiss HD. VisuaIization of magnetic resonance - compatible needles a t 1.5 and 0.2 Tesla. Cardiovasc Intervent Radiol 1996; 19:335340. Kochli VD, McKinnon GC, Hofmann E, von Schulthess GK. Vascular interventions guided by ultrafast MR imaging:

168:7 13-7 19.

13.

14.

15.

16.

17.

18.

19.

evaluation of different materials. Magn Reson Med 1994; 31: 309-3 14. Ladd ME, Erhart P, Debatin JF, Romanowski EkJ, Boesiger P. McKinnon GC. Biopsy needle susceptibility artifacts. Magn Reson Med 1996; 36:64&651. Lewin JS, Duerk JL, Jain VR, Petersilge CA, Chao CP, Haaga JR. Needle localization in MR-guided biopsy and aspiration: effects of field strength, sequence design, and magnetic field orientation. Am J Roentgenol 1996; 166: 1337-1345. Lufkin R. Teresi L. Hanafee W. New needle for MR-guided as- piration cytology of the head and neck. h J Roentg&ol 1987: 1491380-382. Mueller PR. Stark DD. Simeone JF. et al. MR-guided asDira- tion biopsy: needle design and clinical trials. Rgdiology 1'986. 1 6 1 : 605609. Leung DA, Debatin JF. Wildemuth S. et al. Intravascular MR tracking catheter: preliminary experimental evaluation. A m J Roentgenol 1995: 164: 1265-1270. Ladd ME, Zimmermann G. Debatin JF, von Schulthess GK. McKinnon GC. Vascular guidewire visualization for MR fluo- roscopy. In: Proceedings of the 5th scientific meeting of the In- ternational Society for Magnetic Resonance in Medicine. Berkeley, CA: International Society for Magnetic Resonance in Medicine, 1997: 1937. Prince MR, Novelline RA, Athanasoulis CA, Simon M. The di- ameter of the inferior vena cava and its implications for the use of vena caval filters. Radiology 1983; 149:687-689.

Volume 8 Number 1 JMRl 109