|Home | About | Journals | Submit | Contact Us | Français|
To design a deflectable guiding catheter that omits long metallic components yet preserves mechanical properties to facilitate therapeutic interventional MRI procedures.
The catheter shaft incorporated Kevlar braiding. 180° deflection was attained with a 5 cm nitinol slotted tube, a nitinol spring, and a Kevlar pull string. We tested three designs: passive, passive incorporating an inductively-coupled coil, and active receiver. We characterized mechanical properties, MRI properties, RF induced heating, and in vivo performance in swine.
Torque and tip deflection force were satisfactory. Representative procedures included hepatic and azygos vein access, laser cardiac septostomy, and atrial septal defect crossing. Visualization was best in the active configuration, delineating profile and tip orientation. The passive configuration could be used in tandem with an active guidewire to overcome its limited conspicuity. There was no RF-induced heating in all configurations under expected use conditions in vitro and in vivo.
Kevlar and short nitinol component substitutions preserved mechanical properties. The active design offered the best visibility and usability but reintroduced metal conductors. We describe versatile deflectable guiding catheters with a 0.057” lumen for interventional MRI catheterization. Implementations are feasible using active, inductive, and passive visualization strategies to suit application requirements.
The promise of interventional cardiovascular MRI — to offer radiation-free guidance of invasive procedures based on the superb soft-tissue contrast of MRI — remains unmet largely because of the unavailability of safe and conspicuous catheter devices (1,2). Most preclinical and clinical work to date has been conducted using MRI-compatible non-braided catheters, which are both flimsy and inconspicuous (3).
Complex structural heart and cardiovascular interventional procedures require a range of devices having different geometry, which is expensive to realize in a research environment requiring application-specific active MRI catheters. Deflectable MRI catheters would allow a single custom device to be applied to a range of applications, because the geometry can be adjusted interactively. Moreover, deflectable devices might facilitate enhanced navigation of complex structures, such as ventricular septal defects.
Commercial deflectable catheters are constructed using metal (stainless steel or nitinol) braided tubing in the catheter shaft to provide pushability, torqueability, flexibility and kink resistance. Most also use metallic pull wires that run along the entire catheter shaft as their deflection mechanism. Long, metallic components have the potential for heating during MRI radiofrequency (RF) excitation, and should be avoided, if possible, in the construction of dedicated interventional MRI devices (4). We designed a deflectable guiding catheter with an open lumen intended to direct the insertion of interventional devices during real-time MRI. Our deflectable catheter replaces traditional metallic elements, catheter shaft braiding and pull wire, with a MR-safe synthetic polymer (Kevlar) to achieve pushability, torqueability, flexibility and kink resistance. We created a deflection mechanism using short nitinol components to achieve up to 180° deflection. We implemented three MRI visualization strategies with different iterations of the deflectable catheter: passive, active, and inductively-coupled. We used these catheters during representative pre-clinical invasive procedures including hepatic vein access to guide intrahepatic portosystemic shunting, transcatheter repair of atrial septal defect, laser cardiac septostomy and catheter access of the azygos vein.
We manufactured three iterations of the deflectable catheter: a passive deflectable catheter, a passive deflectable catheter incorporating an inductive coil, and an actively visualized deflectable catheter incorporating loop receiver coils.
The deflectable catheter design consists of three sections: a distal deflectable tip, the proximal catheter shaft, and a custom handle.
The proximal catheter shaft was constructed using Kevlar-braided Pellethane tubing providing columnar support and critical torque responsiveness (Figure 1A). The inner lumen of the deflectable catheter contained a 1.45 mm polytetrafuoroethylene (PTFE) liner to allow smooth guidewire passage. Thermoplastic elastomer (Pebax) covering sealed the entire device and components.
The distal deflectable section contained a short (less than 5 cm) nitinol laser-cut slotted tube and nitinol spring in order to provide kink resistance during deflection and to help restore the catheter shape upon release of tension on the pulling thread (Figure 1B and 1C). Kevlar thread (WT 0.38 mm) was used as the pulling thread and was attached to the nitinol slotted tube with medical grade glue (Loctite, Westlake, OH). The pulling thread was housed in polyimide tubing along the entire catheter shaft. Polyolefin heat shrink tubing covered the nitinol slotted tube.
The custom handle was designed and manufactured using Pro-Engineer computer aided design (CAD) software (PTC, Needham, MA) and a three dimensional prototyping machine (uPrint, Eden Prairie, MN) (Figure 2). The design incorporated a hub for guidewire insertion into the catheter lumen and a deflection control mechanism (Figure 2A). It also incorporated a locking mechanism to fix the deflection angle.
The maximum outer diameter of the deflectable catheter was 3.8 mm (compatible with a 12Fr introducer sheath). The maximum inner diameter of the deflectable catheter was 1.45 mm. The stages of deflection are shown in Figure 2(B–D). The maximum deflection possible was 180°. The working length of the deflectable catheter (from handle to tip) was 100 cm.
The inductive deflectable catheter iteration added an inductive coil to the tip of the otherwise passive deflectable catheter design. The intent was to couple wirelessly with the surface receiver coil and to enhance visibility of the catheter. The tip coil was constructed of tightly wound copper magnet wire (0.2 mm diameter) and an MR-compatible, ceramic capacitor (51 pF, American Technical Ceramics, Huntington Station, NY) was connected to the wire ends to form a resonant circuit at 63.86 MHz.
The active deflectable catheter iteration added a loop antenna to the passive design. Two coils were wound with a single 0.25 mm diameter copper wire proximal and distal to the slotted nitinol tube, connecting by running the copper wire through the inside of the slotted nitinol tube (Figure 3). The copper wire was then soldered to a 50 Ω coaxial transmission line (0.31 mm diameter) that was run along the catheter shaft, underneath the Pebax outer covering. The active loop channel was tuned and matched to 50 Ω at 63.86 MHz using an external circuit box connected to the coaxial transmission line at the proximal end of the catheter. The circuit box was then attached to the RF receiver of the MRI system using 50 Ohm coaxial cable.
A 0.89 mm passive nitinol guidewire (Glidewire, Terumo, Somerset, NJ) and a 0.89 mm active guidewire incorporating a modified loopless antenna were used in tandem with the three deflectable catheter iterations during real-time interventional procedures (5).
All mechanical testing was conducted with the passive deflectable guiding catheter iteration.
The shaft of the passive deflectable catheter was secured to a flat surface and the deflectable catheter was deflected 25 times to 180°, then released. The curve reach of the deflectable tip was also measured at 90° and 180° deflection to assess curve performance.
The force necessary to deflect the tip of the deflectable catheter was measured with a hand-held force gauge (IMADA, Inc., Northbrook, IL). The force to deflect to 90° and 180° of catheter distal tip was measured and the same measurement was performed with a “rigid” 0.97 mm nitinol guidewire (Radiofocus stiff shaft angled glidewire, Terumo, Somerset, NJ) inserted in the lumen and the guidewire tip advanced 5 cm past the tip of the catheter.
This test simulates the torque necessary to rotate the deflectable catheter in vivo when the deflected tip meets resistance from a tissue structure or vessel wall. The torque necessary to rotate the 90°-deflected distal tip from 0° to 45°, 90°, 135° and 180° was measured with the hand-held torque gauge (IMADA, Inc., Northbrook, IL) by fixing the catheter shaft to a flat surface and attaching the torque gauge to the 90° deflected tip of the catheter.
All iterations were tested in a water-filled phantom and imaged on a 70×120cm bore 1.5T MRI scanner (Espree, Siemens Healthcare Systems, Erlangen, Germany) with and without an active guidewire. Orientation dependence of the susceptibility artifact from the nitinol slotted tube was investigated by aligning the passive deflectable catheter parallel and perpendicular to B0 and imaging the catheter with a truFISP sequence (TR/TE 836.8/1.8 msec, slice thickness 5 mm, flip angle 45, field of view 170×340, matrix 128×256). All iterations were also imaged using a real-time balanced steady-state free precession (bSSFP) sequence (TR/TE 3.23/1.67 msec, slice thickness 6mm, flip angle 45, field of view 340×340 mm, matrix 192×144).
The deflectable catheter was tested to detect radiofrequency induced heating during MRI. Fiberoptic temperature sensors were attached externally to the deflectable catheters at locations of maximal expected heating on the catheters (Opsens, Quebec, CA). On the passive deflectable catheter, fiberoptic temperature sensors were fixed on the distal, middle and proximal edges of the nitinol slotted tube. Temperature sensors were fixed to the capacitor in the inductive coil on the inductive deflectable catheter and covered with heat shrink to protect the sensors. On the active deflectable catheter, temperature sensors were placed on the two loop coils and the solder point connecting the copper solenoid coil and coaxial transmission line. All heating tests were performed in a gel phantom in accordance with ASTM standard F2182-09 (7).
The deflectable catheters were tested at the center of phantom (0 cm horizontal offset) and at 12.7 cm horizontal offset with a fixed depth of 6 cm (measured from the bottom of the phantom). Each deflectable catheter was tested at insertion lengths ranging from 30–55 cm to determine a critical insertion length causing maximum heating. The catheters were then tested at the critical insertion length while deflected 90° and rotated to both patient left and patient right to simulate use in vivo.
All temperature measurements reflect the change from baseline during MR scanning. The reported values are the average over the last minute of a 15 minute MR scan.
Animal experiments were approved by the institutional animal care and use committee and followed contemporary NIH guidelines. Pigs underwent general anesthesia and mechanical ventilation with inhaled isoflurane. Percutaneous jugular or femoral vein access was obtained using ultrasound or X-ray before transfer to the MRI head-first and supine. Representative experiments included hepatic vein access to guide intrahepatic portosystemic shunting (n=2), transcatheter repair of atrial septal defect (n=3), cardiac laser septostomy (n=1) and catheter access of the azygos vein (n=1). Catheter experiments were performed by single operators who were cardiologists or, for hepatic vein access, students. Catheter navigation was guided completely by MRI using a real-time workstation (Interactive Front End, Siemens). Active devices were connected to the scanner on independent receive-only channels depicted in color.
In vivo mechanical performance was assessed qualitatively through physician feedback during technical development experiments. This included the ability of the deflectable catheter to access difficult target anatomies, mechanical handling, and ease of visualization. Technical development was iterated until working prototypes were generated.
In vivo heating tests of the inductive deflectable catheter and the active deflectable catheter were conducted with fiberoptic temperature sensors attached as described for in vitro heating tests. Tests were performed with the catheters inserted in the inferior vena cava from the right femoral vein, with insertion lengths ranging from 40 to 60 cm.
Continuous parameters are reported as mean ± standard deviation.
During 25 repetitions the tip deflected smoothly on a single plane to 180° and when released, returned to under 20° of deflection. The curve reach of the distal deflectable tip was measured to be 3.2 cm at 90° deflection and 1.9 cm at 180° deflection (Figure 2D).
Used alone or as intended with a 0.97 mm Glidewire, full 180° deflection remained attainable (Table 1). The force required to break the connection between the Kevlar pulling string and the nitinol slotted tube was approximately 22 N. The force required to deflect the catheter alone, and with a Glidewire, to 180° was well within the limits of the pull string connection to the nitinol slotted tube (Table 1).
The force required to deflect the catheter to 90° and 180° was linearly proportional to the angle of deflection, indicating that the Kevlar pulling string does not elongate under stress.
The torque measured was linearly proportional to the angle of rotation and the torque necessary to achieve 180° rotation was 0.103 N-m.
The passive catheter design is shown in figure 4. The appearance of the susceptibility artifact is influenced by deflection and by orientation. For example, when the tip is deflected perpendicular to the main magnetic field, the artifact is large and blurry and may make it difficult for the operator to visualize the exact location of the catheter tip. The inductively coupled catheter design is shown in figure 5(A–C), used in conjunction with a coaxially-inserted active guidewire. Panel 5A indicates signal from the tip solenoid coupled with the active guidewire. Panel 5B shows the active guidewire extending beyond the tip of the catheter, which remains distinctly evident, even in the projection image (panel 5C).
The active catheter design is depicted in figure 5 panels D–F. A distinct “imaging signature” is evident in this design, which depicts the shape and orientation of the tip and shaft in both straight and deflected configurations.
The critical insertion length for maximum heating for the passive deflectable catheter was 40 cm at 12.7 cm horizontal offset. The observed temperature rise was 0.1 ± 0.1 °C, 0.3 ± 0.1 °C, 0.4 ± 0.1 °C at the distal edge, middle, and proximal edge of the nitinol slotted tube, respectively. The observed heating in the various catheter configurations (90° deflection rotated to patient right, 90° deflection rotated to patient left) did not differ significantly from the observed heating while straight.
Heating of the passive deflectable catheter with an inductive coil was measured at 12.7 cm horizontal offset and insertion lengths of 40 and 50 cm. The average temperature increase during the last minute of a 15 minute scan was 0.5 ± 0.4°C for 40 cm insertion length and 0.1 ± 0.4°C for 50 cm insertion length.
For the active deflectable guiding catheter, at 12.7 cm horizontal offset, the critical insertion length causing maximum heating was 55 cm. The observed temperature rise was 6.4 ± 0.1°C, 4.5 ± 0.1°C, and 2.5 ± 0.5°C, for the distal coil, proximal coil and solder point, respectively. The active deflectable catheter was also tested at 0 cm horizontal offset at the critical insertion length of 55 cm. Under this condition, the temperature rise was 0.2 ± 0.1°C, 0.5 ± 0.1°C and 0.3 ± 0.3°C at the distal coil, proximal coil, and solder point, respectively. The effect of position of the distal deflectable portion was also tested in order to simulate use during in vivo experiments. At 0 cm horizontal offset with an insertion length of 55 cm, the active deflectable catheter was deflected 90° and rotated to patient right and patient left. Varying the direction of the deflected tip did not significantly affect the heating in either case (Table 2).
All iterations of the deflectable guiding catheter were able to achieve and maintain the desired deflection for the interventions performed, with and without the guidewire advanced past the catheter tip.
The pushability and torqueability of the deflectable guiding catheter was sufficient to access the hepatic veins from femoral and jugular access, cross atrial septal defects and position the tip of the deflectable catheter on the ventricular septum in order to use an active laser to create a ventricular septal defect. The deflectable catheter’s primary limitation was its large size, which limited its trackability. However, the large wall thickness of the catheter was a positive feature for interventions requiring increased backup support; for example, positioning the active laser catheter in order to burn through the ventricular septum.
Using an active guidewire with the passive deflectable catheter allowed operators to access a wide range of anatomical targets under real-time MRI. Figure 6A–B shows the passive deflectable catheter directing the guidewire into the hepatic vein from both transjugular and transfemoral approaches. We also used the passive deflectable catheter with an active guidewire to access the azygos vein using transfemoral access (image not shown). These examples demonstrate the clear visibility of the curvature and orientation of the passive deflectable catheter using the contrast from the active guidewire signal, which distinguished the catheter tip susceptibility artifact from the dark appearance of the surrounding vessels.
The passive deflectable catheter was also used in conjunction with an active guidewire to cross an atrial septal defect. Figure 6C shows the passive deflectable catheter entering the right atrium and Figure 6D shows the active guidewire being directed into the left ventricle after crossing the defect. This shows that the passive deflectable catheter, in tandem with the active guidewire, was able to direct a guidewire into an anatomical target (atrial septal defect) whose properties may not be defined pre-intervention and are often highly variable between patients.
The inductive deflectable catheter showed coupling to the active guidewire which enhanced tip visibility compared with the passive deflectable catheter (Figure 7). This visualization strategy could be especially useful in certain in vivo applications in which it is more critical to differentiate the catheter tip from the guidewire tip, such as when tracking the catheter over the guidewire into an anatomical target.
We used the active deflectable guiding catheter to deliver a laser catheter to the interventricular septum in swine, in order to create an experimental communication between heart chambers. Figure 8A shows the catheter tip positioned against the ventricular septum. The active signal from the loop coil clearly depicted the deflectable segment’s curvature and location of catheter tip. Furthermore, the presence of an active device channel enabled the use of “projection mode”, an imaging tool that projects the active device signal from its orientation in the three-dimensional imaging volume onto a two-dimensional plane, resembling X-ray fluoroscopy (6, 8). Projection mode was useful for locating the active device when out of the slice planes (inset of Figure 8A).
We also used the active deflectable catheter to deliver a passive nitinol guidewire across an atrial septal defect under real-time MRI guidance (Figure 8B). The active deflectable catheter was easily visualized, allowing the operator to cross the defect without difficulty. Figure 8B shows the deflectable catheter in 90° deflection, with signal from the tip and distal coils, as well as the susceptibility artifact, visible in a single plane, providing clear visualization of the catheter tip orientation relative to the surrounding anatomy.
At critical insertion lengths for maximum heating, the inductive deflectable catheter heated 0.4 ± 0.3°C while the active catheter heated up to 0.5 ± 0.3°C while inserted into the inferior vena cava. The passive catheter showed only minor heating in vitro and therefore was not tested in vivo.
We developed a deflectable guiding catheter that uses Kevlar braiding for the catheter shaft and for the pull string in place of stainless steel or nitinol. The catheter is deflected using a 5 cm nitinol slotted tube and a nitinol spring that work together to provide up to 180° deflection in a single plane. These elements also provide kink resistance during deflection, a constant radius of curvature and restoring force upon release of deflection. The passive design eliminated all long metallic components from the deflectable guiding catheter while preserving the desired mechanical properties. As expected, we did not observe any RF heating greater than 1°C of the passive deflectable guiding catheter and inductive deflectable guiding catheter under any conditions in vitro or in vivo. Using the active receiver coil design, we observed RF heating up to 6°C at the maximum horizontal offset in vitro. The presence of a coaxial cable running along the catheter shaft most likely contributed to the formation of standing waves and eddy currents along the catheter, increasing the potential for heating at the tip coils. However, the heating was acceptable (less than 1.0°C at 15 minutes) under expected use position near the center of the scanner bore in vitro and during in vivo heating tests using all catheter iterations.
We demonstrated three different catheter designs having three different visualization strategies. The passive deflectable guiding catheter — the simplest design — could be used in tandem with an active guidewire to overcome its limited conspicuity. The active deflectable guiding catheter design offered the most versatility and provided the best visibility. Specifically, the active deflectable guiding catheter could be used in tandem with a commercial nitinol guidewire to allow catheter exchange while preserving guidewire access to an anatomic target. This allowed the operator to exchange the active deflectable guiding catheter over the commercial guidewire for another catheter or sheath once access was gained into the anatomical target. The inductive deflectable guiding catheter had enhanced tip visibility compared with the passive design, but was inferior to the active catheter design. The small level of heating observed at maximum horizontal offset in vitro shows that the active deflectable catheter heats more than the other catheter iterations. However, heating was not observed in vitro in the center of the bore or in vivo with the catheter advanced into the inferior vena cava from the femoral vein. Application-specific in vivo heating tests should be conducted on the active deflectable catheter for applications requiring use closes to the edge of the scanner bore.
Karmakar and colleagues created a deflectable catheter incorporating the metallic catheter shaft braiding and metal pull wire into a loopless antenna, connected to de-tuning circuitry at the proximal end to reduce heating during RF transmit (9). Another group has designed a magnetically-assisted remote control steerable catheter, but having limited deflection ability at certain orientations with respect to the main magnetic field (10,11). Other groups have developed catheters with ferromagnetic beads and use pulse sequences to apply MR gradient forces that navigate the catheter tip during scanning (12,13); however, the ferromagnetic materials create large artifacts that obscure small vessel branches and make navigation difficult.
Delineating the shape and orientation of the catheter tip is crucial to navigating vessel branches and tortuous anatomy. Other groups have incorporated active tracking coils onto fixed curve catheters in order to test the ability to navigate difficult targets such as vessel branching points (14,15). Such catheters are easily visualized under MR but lack control of deflection to actively navigate into anatomical targets.
The primary limitation of our prototype is the large outer diameter. Future iterations will employ lower profile Kevlar thread that will allow a thinner catheter wall, a reduced deflection force necessary, and a lower-profile pulling string. Our prototype pullstring breaks at twice the maximum required deflection force. This safety margin appears to suit our main application of target vessel selection, wherein pullstring failure would create little hazard apart from requiring device substitution. Alternative nitinol spring designs might enhance the restoring force of the single-thread Kevlar pull-string deflection mechanism.
In conclusion, we have developed a deflectable guiding catheter that is a significant addition to the interventional MRI catheter armamentarium. Kevlar braiding in the catheter shaft, Kevlar pulling string, and a short (less than 5 cm) nitinol deflection mechanism preserved important mechanical properties while improving the RF safety profile of the catheter. We have shown that this deflectable guiding catheter, using different visualization techniques, has appropriate mechanical characteristics, including range of deflection and torqueability, to access challenging anatomy using guidewires and therapeutic devices. Our deflectable guiding catheter is a versatile tool that can enable a wide range of catheter-based procedures under real-time MRI guidance.
We thank Katherine Lucas, Joni Taylor, Victor Wright and Bill Schenke for technical assistance.
NIH and Siemens have a collaborative research and development agreement. Kanishka Ratnayaka serves without compensation on a Siemens Pediatric Advisory Council.
This work was supported by the Division of Intramural Research (Z01-HL005062-08, Z01-HL006041-01), National Heart Lung and Blood Institute, National Institutes of Health.