Susceptibility artifacts from metallic markers and cardiac catheterization devices on a high-performance 0.55 T MRI system

Abstract

Introduction:

Visualization of passive devices during MRI-guided catheterizations often relies on a susceptibility artifact from the device itself or added susceptibility markers that impart a unique imaging signature. High-performance low field MRI systems offer reduced RF-induced heating of metallic devices during MRI-guided invasive procedures, but susceptibility artifacts are expected to diminish with field strength, reducing device visualization. In this study, field strength and orientation dependence of artifacts from susceptibility markers and metallic guidewires were evaluated using a prototype high-performance 0.55 T MRI system.

Materials and methods:

Artifact volume from nitinol and stainless steel passive susceptibility markers was quantified using histogram analysis of pixel intensities from three-dimensional gradient echo images at 0.55 T, 1.5 T and 3 T. In addition, visibility of commercially available clinical catheterization devices was compared between 0.55 T and 1.5 T using real-time bSSFP in phantoms and in vivo.

Results:

A low-tensile strength stainless-steel marker produced field strength- and orientation-dependent artifact size (1.7 cm³, 1.95 cm³, 2.21 cm³ at 0.55 T, 1.5 T, 3 T, respectively). Whereas, a high-tensile strength steel marker, of the same alloy, produced field strength- and orientation-independent artifact size (3.35 cm³, 3.41 cm³, 3.42 cm³ at 0.55 T, 1.5 T, 3 T, respectively). Visibility of commercially available nitinol guidewires was reduced at 0.55 T, but imaging signature could be maintained using high-susceptibility stainless steel markers.

Discussion and conclusion:

High-susceptibility stainless-steel markers generate field-independent artifacts between 0.55 T, 1.5 T and 3 T, indicating magnetic saturation at fields <0.55 T. Thus, artifact size can be tailored such that interventional devices produce identical imaging signatures across field strengths.

1. Introduction

Real-time MRI-guidance of cardiovascular catheterization procedures offers improved tissue contrast and accurate measurements of cardiac function and flow. During catheterization procedures, devices such as guidewires and catheters are navigated through chambers of the heart. RF-induced heating of standard metallic guidewires and catheters has limited the clinical translation of these procedures at 1.5 T due to the risk of thermal injury [1,2]. To enable freedom from radiofrequency (RF)-induced heating of conductive devices, while maintaining imaging speed for interventional and cardiac imaging applications, we modified a commercial MRI scanner to operate at 0.55 T [3]. Since heating is quadratically related to field strength, a 0.55 T system reduces heating by 7.5-fold compared with 1.5 T. We have demonstrated the use of commercial metallic guidewires for cardiac catheterizations in patients using this system [3].

Our goal to perform MRI-guided procedures depends on device conspicuity. During MRI-guided invasive procedures, “passive visualization” of devices uses the signal voids created by metallic devices, such as stainless-steel and nitinol guidewires, for procedural guidance. Moreover, some interventional devices are designed with discrete passive markers, such as iron oxide bands [4–6], to impart a unique imaging signature in MRI. Passive markers improve the specificity of the imaging signature and therefore confidence in device positioning.

Interventional devices are commonly manufactured from nitinol, stainless steel 316, or stainless steel 304, which are all paramagnetic. There is a common assumption that metal susceptibility artifacts are reduced at lower field because absolute off-resonance, in Hertz (Hz), scales linearly with field strength. However, medical stainless steel contains iron, and therefore exhibits a hysteresis of magnetization versus magnetic field strength [7,8]. This results in saturation at magnetic fields within the range of clinical MRI systems, such that artifact size does not change with increased field strength for some devices. Moreover, changes to the crystal structure, for example by coldworking or welding, cause austenitic stainless steel alloys (316 and 304) to exhibit ferromagnetic properties [9]. Practically, this results in a change to both the ultimate tensile strength and the magnetic permeability of stainless steel alloys, and thus modifies magnetic field saturation points. Previous work evaluating MRI susceptibility artifacts for orthopedic implants [10,11], aneurysm clips [12,13], and other ferromagnetic materials [7,8] have demonstrated a complex relationship between field strength, orientation, and artifact size dependent on the material properties.

We sought to evaluate the conspicuity of clinical metal guidewires and susceptibility markers using our high-performance prototype 0.55 T system. We evaluated the field strength- and orientation-dependence of susceptibility artifacts at 0.55 T, 1.5 T, and 3 T for small metallic markers composed of nitinol and stainless-steel, with samples coldworked to modify the magnetic field saturation point. We also performed in vivo comparisons of susceptibility artifacts from commercial guidewires and susceptibility markers at 0.55 T and 1.5 T.

2. Materials and methods

2.1. Imaging systems

Imaging was performed on three MRI systems: a commercial MRI system that was modified to operate at 0.55 T (prototype MAGNETOM Aera, Siemens Healthcare, Erlangen, Germany), a 1.5 T system (MAGNETOM Aera, Siemens Healthcare, Erlangen, Germany), and a 3 T system (MAGNETOM Skyra, Siemens Healthcare, Erlangen, Germany).

2.2. Passive marker artifact quantification

The susceptibility artifacts from three small metallic markers were quantified at 0.55 T, 1.5 T and 3 T in a phantom to demonstrate the saturation points of the three materials. We assessed one nitinol marker and two stainless-steel 304 samples modified by cold-working to generate ultimate tensile strength (UTS) = 172ksi and UTS = 197ksi. Susceptibility markers were fabricated by cutting stainless steel 304 and nitinol tubes to 1 mm (inner diameter: 0.022”, outer diameter: 0.026”) (K-Tube Technologies, Cook Medical, Poway CA). The markers were embedded in an agarose gel doped with 1 g/L copper sulphate (CuSO₄), according to ASTM-F2119 [14]. Markers were imaged on each MRI system parallel to B0 and perpendicular to B0 to determine the dependence of artifact size on both field strength and marker orientation.

We quantified the artifact size produced by passive device markers from three-dimensional images with identical imaging parameters (spoiled gradient echo (GRE) acquisition, TE/TR = 3.64/102 ms, flip angle = 30°, FOV = 120 mm × 150 mm, isotropic imaging resolution = 0.7 mm³). GRE imaging generates a signal void caused by intravoxel dephasing which is directly proportional to susceptibility (T2* effect). By comparison, in spin echo imaging, the 180° pulse corrects the susceptibility-induced dephasing and metallic objects cause only geometric distortion; and balanced stead-state free-precession (bSSFP) imaging of a metallic point source generates artifacts with a dipole-appearance that combine signal voids and banding artifacts. Therefore, we selected GRE imaging to specifically quantify artifact size caused by susceptibility. The images were SNR-matched by increasing the number of averages at the 0.55 T and 1.5 T systems.

The artifact size was calculated using MATLAB R2018a. We selected a 3D volume of 60 × 90 × 96 voxels (42 mm × 63 mm × 68 mm) in the phantom, and generated a histogram of signal intensity from those voxels. The distribution of the voxel signal intensities was approximately Gaussian (Fig. 1). The gradient echo artifact contains hyperintense signal caused by signal pileup in addition to the susceptibility-induced signal void, and we used a threshold of ±3 standard deviations from the mean to label artifact voxels. The largest spatially-connected region was used to isolate the marker signal from other signal perturbations such as air bubbles.

Fig. 1. Three-dimensional (3D) artifact volume calculation pipeline.

Pixel intensities from 3D images were normalized to a zero-mean signal with unit standard deviation. A threshold was applied to select pixels outside ±3 standard deviations from the background signal. The resultant signal was reconstructed using a connected region algorithm to isolate the marker.

2.3. Phantom imaging of commercial guidewires with/without passive markers

Commercial guidewires were imaged in a phantom at 0.55 T and 1.5 T using the clinically-relevant bSSFP sequence applied during patient heart catheterization. These images were used to evaluate guidewire conspicuity across field strengths, and to measure bSSFP artifacts from passive susceptibility markers affixed to metallic guidewires. Two clinical workhorse guidewires were used (nitinol Glidewire 260 cm × 0.035”, Terumo, Japan and stainless-steel 304 InQwire, 260 cm × 0.035”, Merit Medical, UT, USA). In addition, the two stainless-steel markers (UTS = 172 ksi and UTS = 197 ksi) were fixed to the tip of the nitinol Glidewire in order to improve its conspicuity, replicating a possible use-case. Each device was placed in a polyacrylic acid gel prepared according to ASTM-F2182 [15] and imaged parallel to B0 and perpendicular to B0, as well as a looped configuration which may be encountered during clinical application. A bSSFP sequence was used to replicate patient application conditions (TE/TR 1.92/3.84 ms, flip angle 45°, matrix = 256 × 256, FOV 313 mm × 313 mm, slice thickness = 8 mm, receiver bandwidth = 560 Hz/Px). Slice thicknesses of 16 mm and 24 mm were also assessed for the looped configuration, to simulate interactive modifications of slice thickness during real-time procedural guidance. Marker artifact area was measured on 2D images using Osirix DICOM Viewer (Pixmeo, Geneva, Switzerland) to evaluate field-dependent artifact size from bSSFP imaging.

2.4. In vivo guidewire and marker imaging

To assess artifact visibility for catheterization procedures, a nitinol guidewire (Glidewire 150 cm × 0.035”, Terumo, Japan), and a guidewire with high-UTS passive tip marker (modified Glidewire 260 cm × 0.035”, Terumo, Japan) were imaged in the descending aorta and left ventricle of a swine model at 0.55 T and 1.5 T. We used real-time bSSFP for guidewire imaging with matched parameters for like-to-like comparison (TE/TR 1.4/2.8 ms (1.5 T), TE/TR = 1.2/2.4 ms (0.55 T), flip angle 45°, resolution = 2.5 mm × 2.5 mm, slice thickness 6 mm, receiver bandwidth = 1490 Hz/Px, GRAPPA acceleration factor = 2).

3. Results

3.1. Passive marker artifact quantification

Stainless-steel and nitinol marker images from identical acquisitions at all three field strengths are provided in Fig. 2. Fig. 3 illustrates the differences in measured artifact sizes from the metallic markers at each field strength and orientation relative to B0. Low-susceptibility nitinol markers generated field-strength-dependent artifacts, as expected. The nitinol marker artifact size exhibited more orientation-dependence at higher field strength (Fig. 3B; Parallel to B0 artifact size: 1.7mm³ at 0.55 T, 2 mm³ at 1.5 T and 2.3mm³ at 3 T; Perpendicular to B0 artifact size: 1.8mm³ at 0.55 T, 2.6mm³ at 1.5 T and 3.5mm³ at 3 T).

Fig. 2. Example slices from the three-dimensional marker images.

Artifacts from the nitinol marker (A), low UTS (172 ksi) stainless-steel marker (B), and high UTS (197ksi) stainless-steel marker (C) placed parallel and perpendicular to B_o at 0.55 T, 1.5 T and 3 T.

Fig. 3. Passive-marker artifact volume.

(A) 3D artifact volume plotted relative to field strength comparing all three markers on the same axes. Each marker is also plotted independently to highlight field-strength- and orientation-dependence for (B) the high UTS marker, (C) the low UTS marker and (D) the nitinol marker. The two stainless-steel markers produce artifacts of differing sizes despite their equivalent alloy, mass and dimension. The high UTS (197 ksi) marker creates a larger artifact at all field strengths, and its artifact volume is unchanged between 1.5 T and 3 T. The low UTS marker produces an artifact that increases in volume with field strength, and shows increased orientation dependency at lower field strengths.

For two stainless-steel markers with identical alloy and material grade, but with different UTS, the field strength dependence of artifact size differed (Fig. 3C and D). The high-susceptibility steel marker (UTS = 197 ksi) produced consistent artifact size at 0.55 T, 1.5 T and 3 T (3.35 cm³, 3.41 cm³, and 3.42 cm³, respectively), and the low-susceptibility stainless-steel marker (UTS = 172 ksi) produced field-strength-dependent artifact size (1.7 cm³, 1.95 cm³, and 2.21 cm³ at 0.55 T, 1.5 T, and 3 T respectively, in parallel orientation). In addition, the orientation dependence on artifact size was negligible for the high UTS marker and more prominent for the low UTS stainless-steel marker. This suggests that the low UTS stainless-steel is not fully saturated at 0.55 T, but coldworking decreases the saturation point to ≤0.55 T. Therefore, the selection of a high UTS marker may create a unique imaging signature which is field-strength independent.

3.2. Phantom imaging of commercial guidewires with/without passive markers

Nitinol guidewires produced small signal voids at 1.5 T that were further reduced at 0.55 T (Fig. 4A, B). This indicates reduced paramagnetic device conspicuity using standard real-time bSSFP imaging at 0.55 T. The imaging signature varied between parallel, perpendicular and looped configurations. Imaging slice thickness did not impact device conspicuity in a looped configuration. Whereas, stainless-steel 304 guidewires produced large blooming artifacts at both field strengths that may still interfere with procedural navigation (Fig. 4C), with only modest decrease in artifact size at 0.55 T.

Fig. 4. Susceptibility artifacts of commercial guidewires.

(A) A nitinol guidewire (Glidewire 260 cm/0.035”, Terumo) produces artifacts that are smaller at 0.55 T compared with 1.5 T. (B) The nitinol guidewire artifact size is accentuated when samples are placed in a looped configuration, imaged with three different slice thicknesses. (C) The stainless-steel 304 guidewire (InQwire, 260 cm/0.035”, Merit Medical, UT, USA) produces large artifacts at 0.55 T and 1.5 T, that are exaggerated when perpendicular to the main field.

Passive stainless steel tip markers added to nitinol guidewires generated a dipole artifact using the real-time bSSFP imaging sequence in a phantom (Fig. 5). Slice alignment was optimized for marker visualization; therefore, appearance of the nitinol guidewire shaft varies across the two fields. Using real-time bSSFP imaging, modest field-strength-dependence of both stainless-steel markers was observed, but was more prominent for low-UTS stainless steel.

Fig. 5. Real-time imaging artifacts from guidewire with affixed passive markers.

Stainless-steel markers were added to the tip of a commercial nitinol guidewire (Glidewire 260 cm/0.035”, Terumo, Japan) and imaged using a standard bSSFP sequence. (A) Artifacts from the low-susceptibility (UTS = 172ksi) stainless-steel marker increase with field strength, which is more prominent when perpendicular to B0. Whereas, no observable difference is present for the high- susceptibility stainless-steel marker (UTS = 197ksi) across fields. (B) Area of the artifact size from each marker during a bSSFP sequence scan is plotted against field strength.

3.3. Commercial guidewire imaging in vivo

In vivo imaging demonstrated the reduced visibility of commercially-available nitinol guidewires at 0.55 T compared with 1.5 T (Fig. 6A, Supplemental Movie 1). At 1.5 T visibility of the nitinol guidewire is improved due to increased susceptibility artifact in combination with the bright signal from RF amplification. This bright signal is associated with RF-induced heating, and is not visible at 0.55 T. The diminished guidewire visibility is important for procedural safety at 0.55 T. However, as expected, the artifacts created by the high-UTS stainless steel marker were constant across the two field strengths during in vivo device navigation (Fig. 6B, Supplemental Movie 2). Therefore, choosing high susceptibility marker material (e.g. high UTS stainless steel) enables an in vivo imaging signature that is consistent across field strengths.

Fig. 6. In vivo imaging of guidewires and passive markers at 0.55 T and 1.5 T.

(A) Images of a nitinol guidewire (Glidewire 150 cm/0.035”, Terumo, Japan) in the aorta of a swine using real-time bSSFP imaging. Red arrow points to the guidewire, but the location of the guidewire tip is ambiguous. Reduced susceptibility artifact can be appreciated at 0.55 T, compared with 1.5 T. (B) Nitinol guidewire (Glidewire 260 cm/0.035”, Terumo, Japan) with a high-UTS stainless-steel marker affixed at its tip during left heart catheterization at 0.55 T and 1.5 T. No observable change in artifact size is evident across the two field strengths. Red arrows point to the susceptibility marker at the guidewire tip. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

4. Discussion

We have reported on the metal susceptibility artifacts for passive device markers and commercial guidewires to enable MRI-guided catheterization procedures on a high-performance 0.55 T system. We demonstrated reduced conspicuity at 0.55 T for commercial paramagnetic interventional devices (nitinol guidewires). We found that small stainless-steel markers, if at a sufficiently high susceptibility, may generate identical artifacts between 0.55 T, 1.5 T and 3 T, indicating saturation at fields ≤0.55 T, which is lower than previously observed [7]. We demonstrated that imaging signatures of passive markers can be varied by altering susceptibility by cold-working metal for the same alloy and marker dimensions. Importantly, for device manufactures and researchers pursuing interventional devices with passive susceptibility markers [4,16–19], high susceptibility materials can be chosen such that artifact size is independent of field strength, meaning the same device design can be used across field strengths.

Guidewires and catheters used for invasive cardiovascular procedures are typically stainless steel or nitinol, and RF-induced heating of these devices has been a roadblock for MRI-guided procedures. At 0.55 T the reduced RF-induced heating creates opportunities to use commercial devices for MRI-guided cardiovascular procedures, and we have demonstrated MRI-guided right heart catheterization in patients using off-the-shelf nitinol guidewires [3]. However, but the burden of safety has shifted to device conspicuity. At low field strengths, low- susceptibility materials such as nitinol are inconspicuous, which can be unsafe for MRI-guidance. By comparison, stainless steel guidewires produced large blooming artifacts that are incompatible with MRI- guided invasive procedures even at 0.55 T.

The susceptibility artifacts caused by a metallic device depend on many factors including material composition and magnetic susceptibility; size and geometry of the metallic object; orientation of the metallic object with respect to the direction of the main magnetic field; the strength of the applied magnetic field; and the pulse sequence details. Although higher-field MR systems generate images with higher signal- to-noise ratio, the susceptibility artifacts are also generally larger at higher fields. For 3D artifact volume calculation, we chose to use GRE acquisition rather than spin echo, as GRE provides signal differences from susceptibility (T2*), whereas spin echo does not. Similarly, GRE offers an advantage over bSSFP for this purpose, because bSSFP banding artifacts near metallic objects can interfere with accurate quantification of the artifact size. Using a GRE acquisition with identical parameters at all field strengths, we were able to quantify the artifact size caused solely by susceptibility.

It is known that magnetic behavior of stainless steel alloys can be changed by applying mechanical stress to it [20]. Specifically, certain manufacturing processes such as machining or coldworking can change to the crystal structure, which alters the UTS of the alloys. We showed two example stainless-steel markers with different UTS, hence modified magnetic susceptibility. These markers generated different artifact size for the same stainless-steel alloy by changing the saturation magnetization. The high-susceptibility stainless-steel markers with a high UTS (197 ksi) demonstrated no dependence of artifact size on field strengths and we conclude that the marker saturates below 0.55 T. Whereas, low- susceptibility markers with lower UTS (172 ksi) did not reach magnetic saturation. Low-susceptibility markers generated susceptibility artifacts that diminished with decreasing field strength.

We performed exploratory in vivo imaging of nitinol guidewires with and without high-susceptibility markers at the guidewire tip at 1.5 T and 0.55 T. We observed changes in the nitinol guidewire appearance but no changes in high susceptibility marker artifact appearance at lower field strength. RF-amplification at 1.5 T creates bright signal around the metallic guidewire which is not observed at 0.55 T where RF-induced heating is negligible.

There are a few limitations to the study presented here. We used matched imaging protocol parameters for our study, corresponding to our standard 1.5 T real-time imaging. For clinical application at 0.55 T, further parameter optimization may alter the artifact appearance. We did not investigate all possible guidewire configurations in vitro, rather assessed the clinically applicable configurations in order to compare the conspicuity of nitinol guidewires with and without markers at 0.55 T and 1.5 T. We limited our study to evaluate field dependency of stainless steel and nitinol susceptibility markers, and did not include all possible materials used as markers for interventional devices (eg. iron oxide). Additional work is required to characterize the clinical implications of our findings.

5. Conclusion

A high-performance low-field MRI system may enable progress in interventional MRI due to improved device safety with retained imaging performance. Metallic artifacts from paramagnetic interventional devices, including nitinol guidewires, were reduced at lower field. High-susceptibility materials used for passive device markers generate constant artifact size across field strengths, indicating that the same passive device markers can be used at lower field strength. Susceptibility artifacts of stainless-steel passive markers can be tuned such that artifact size is either constant or changes proportionally to field strength for different applications.