Interventional MRI - Devices & Cardiovascular ApplicationsKevan Anderson (kanders@sri.utoronto.ca)Physicists and clinicians interested in interventional MRI

1. Describe imaging strategies for passive and active device visualization

2. Recognize safety considerations for the use of cardiovascular devices in the MRI environment

Devices for cardiovascular applications consist mainly of guidewires and catheters. Guidewires are long thin wires that are inserted into an artery or vein and manipulated to form a path to a particular target through the vasculature. Guidewires are typically made of metals such as nitinol or stainless steel. Catheters for vascular applications are tubes with either a single of multiple lumens made of braided thermoplastic and are used to provide mechanical support for guidewires. For many non-vascular applications, such as electrophysiology procedures, catheters may be more sophisticated and could include additional features such as needles, ablation electrodes, and thermocouples.

The use of cardiovascular devices can be very challenging in the MRI environment. One must ensure that the materials used in their construction are MR-compatible to the extent that they will not experience any significant forces through interaction with the static magnetic field and that they will not significantly degrade the ability to image the anatomy of interest. Moreover, the use of long conductive structures, such as guidewires and conductors in catheters, can give rise to significant tissue heating. Techniques to visualize devices are also important, as one needs to know the position of the device with respect to the anatomy.

Broadly, devices developed for use in the MR scanner can be separated into two major groups: passive devices and active devices. Passive devices are broadly defined as devices that are not electrically connected to any other equipment. The use of passive devices is largely motivated by the greater simplicity and perceived safety compared to the collection of active devices which are described below. The visualization of passive devices is reliant strictly on image contrast. Several techniques have been proposed in order to improve the ability to visualize a passive device on an MR image. The use of susceptibility markers enables the visualization of devices through the presence of signal voids induced by magnetic susceptibility artifacts (1-4). Another mechanism involves filling a region of the catheter with a medium that creates either a MR signal void or an area of strong MR signal relative to that of the surrounding anatomy. Several other sources of contrast methods have been used. These include carbon dioxide gas (5) and gadolinium (6,7).

A significant limitation of passive visualization techniques is that the device must be manipulated within the scan plane (typically 5-10 mm thick) in order for it to be visualized. This can be difficult, especially during procedures in tortuous vessels and in the cardiac chambers. Moreover, if the device moves outside of the imaging slice, increasing the slice thickness will degrade contrast as more background signal from tissues will be present. Also, automatic scan plane prescription is not possible because catheter localization is performed visually and no quantitative information about catheter position is generated for feedback to the scanner. The time it takes to visualize the catheter is limited by the time required for image acquisition, image reconstruction, and image display. Slow reconstruction rates and display times can result in a decreased temporal resolution.

Unlike passive devices, active devices have an electrical connection to the MR scanner. Active devices are typically used by acquiring MR signal from small imaging coils located along their length or at their tip. One common method for visualizing active devices is to superimpose an image reconstructed from receive coils located on the device onto an anatomical image acquired from a larger surface coil (8-10). Using this method, the small sensitivity regions of the catheter-based receive coils indicate where the catheter is located. The images obtained from the device-based coils can be colour-coded to improve visualization. Due to the fact that imaging coils are able to receive MR signal from a significant volume around them, the device can appear blurred when it is visualized using this method. As a result, it is difficult to ascertain its exact position. Another limitation with this technique is that, like passive techniques, the device must be manipulated within the scan plane.

One popular catheter tracking technique involves projecting the MR signal obtained from a solenoid receive coil located on the distal tip of the catheter onto three orthogonal planes (11,12). This is accomplished by exciting the entire volume and generating gradient recalled echoes in three orthogonal directions while frequency encoding the position of the microcoil. The position of the coil can then be determined by locating the signal peak in each of the three projections. The calculated position of the catheter can then be displayed on a real-time anatomical image to visualize its position in an anatomical context.

Implementation of the projection technique into a rapid-acquisition sequence is done by intermittently acquiring projection data sets from the microcoil between rapid image acquisitions from a larger coil. Common types of acquisitions in which to intersperse the tracking sequence are spiral acquisitions and Cartesian steady state free procession (SSFP) acquisitions (13,14). Despite being a very rapid method for localizing a device, there are some limitations with the technique. The peaks of the projections that are acquired correspond to the location of maximum MR signal received by the coil. Depending on the structure of the surrounding medium, this region of maximum signal may not be located at the centre of the coil and the coil’s orientation within the main magnetic field can have the effect of creating dual-peaked signal projections. Another disadvantage with the projection method is that it is not able to provide any information about device orientation without the use of more than one device-based coil. Several types of active devices have been developed to address the problem of directionality. One utilizes more than one device-based microcoil placed in close proximity to each other (15,16). By determining the position of both micocoils using the projection method, and by assuming that the device does not bend significantly between the two coils, it is possible to determine the direction of the device. Another utilizes phase information in the MR signal to determine the orientation of the microcoil with respect to the static field (17).

Another class of techniques utilizes external coupling devices that inductively couple to conductive devices that pass through them (18-21). This makes it possible to receive MR signal from around the device without having a wired electrical connection between the intravascular device and the scanner. Using such a technique, one can actively visualize a conventional guidewire that could otherwise only be visualized using passive techniques.

The use of the MR signal received by active devices is not limited solely to device visualization and tracking purposes but can also be used for anatomical imaging. Using small device-based receive coils for imaging can be beneficial as high-resolution images with a small field-of-view can be produced.

The majority of intravascular imaging applications have focused on the imaging of the vessel wall and atherosclerotic plaques. Towards this goal, the majority of active devices developed for intravascular imaging have been designed to be ‘side-looking’. One example is the opposed-solenoid intravascular imaging catheter. This device consists of two solenoidal coils that are separated by 1-2cm and are concentric with the axis of the catheter (16,22-24). In addition to being able to image only beside the catheter, the opposed-solenoid design suffers from the limitation that the catheter must be positioned such that the anatomy of interest is centred between the two coils.

A novel active imaging device that overcomes this limitation is the loopless antenna, also commonly referred to as an imaging guidewire. Unlike the opposed-solenoid, the device does not contain any solenoidal coils and instead consists of a folded dipole antenna that is located on the tip of a coaxial cable (25). The sensitivity of the device to MR signal is much more distributed, and although the area of high sensitivity is still located to the side of the device, it is present for a greater distance over its length. This makes imaging much less sensitive to precise device positioning. Moreover, the imaging sensitivity of the device falls off much less rapidly with distance when compared to solenoidal designs. The loopless antenna can also be fabricated to be much thinner and more flexible than imaging devices that contain solenoidal coils.

One of the major limitations of active devices and guidewires is that significant safety concerns are associated with the use of long conducting structures in an MR scanner as tissue heating can occur due to coupling between the conductive structure and the electric component of radio-frequency (RF) excitation field (26-28). Tissue heating occurs because currents on the conductor can create large electric fields at the device tip thereby resulting in tissue heating.

Several solutions have been proposed to mitigate the heating of tissues around conductive devices and many have focused specifically on methods for reducing currents on the transmission lines that connect active devices to the MR scanner. Ladd et al. demonstrated that the amount of tissue heating at the tip of a transmission line could be reduced by incorporating quarter-wavelength coaxial current chokes onto the transmission line (29). Weiss et al. demonstrated that heating could be reduced through the use of transformers positioned at discrete intervals along the length of the transmission line (30,31). Both of these techniques, although shown to mitigate the heating effect, are restricted to use in thicker devices such as catheters as they require additional space in order to be implemented. For instance, the current choke as proposed by Ladd et al. requires the use of a tri-axial cable (which is thicker than ordinary coaxial cable) and transformers need to be of sufficient size to enable efficient signal propagation. As a result, they cannot be applied to thinner devices such as active and conventional guidewires.

For thinner devices, several strategies to mitigate heating have focused on the development of techniques to identify unsafe RF power levels used to excite spins in an MR scanner. Characterization studies have attempted to identify the power limits that can safely be used in the presence of interventional devices (32-34). Attempts have also focused on techniques to detect the presence of currents on devices before the tissue experiences a significant temperature rise. One such technique uses a low-power reverse-quadrature excitation (35). Another detects currents through the use of an inductively-coupled pick-up coil positioned over top of the wire (36). Current research is also focusing on excitation techniques using surface coil arrays to minimize the coupling with conductive devices (37,38).