Background

Cardiovascular magnetic resonance (CMR) fluoroscopy catheterization allows simultaneous measurement of chamber pressure and cardiac output, alongside characterization of cardiac tissue, cardiac anatomy and cardiac function in a single procedure [1].

To date, clinical CMR fluoroscopy heart catheterization has relied on polymer catheters that are visualized solely by the contents of distal balloons filled with gas or contrast agents [2,3,4,5,6]. As a result, only catheter tips and not shafts are visible during CMR fluoroscopy, which hampers catheter navigation, especially in patients with challenging anatomy and physiology, such as surgically corrected or uncorrected congenital heart disease, enlarged right sided heart structures and severe valvular regurgitation. Until now, metallic guidewires, an important adjunct for X-ray fluoroscopy catheterization, have not been used during CMR fluoroscopy catheterization out of concern that long conductive structures will heat by coupling electrically with radiofrequency (RF) excitation pulses [7, 8]. Glass and other non-metallic guidewires have been reported for CMR catheterization but suffer insurmountable fragility that will likely preclude wide adoption [9,10,11,12]. Non-conductive metallic guidewires have been reported that resist heating during CMR but that are not yet commercially available [13].

We hypothesized that standard commercial metallic guidewires could be safely used in patients by reducing CMR excitation energy (also known as low specific absorption rate (SAR) imaging) [14]. We defined conditions of use for a specific commercial nitinol guidewire having desirable insulation properties, used in combination with reduced excitation energy and prolonged readout times to reduce RF-induced heating. These consisted of reduced flip angle α (to 10o during gradient echo from 45o during balanced steady-state free precession (bSSFP), a long repetition time (TR) spiral readout instead of a short TR rectilinear readout (TR 10 ms vs 2.98 ms), and a specific 150 cm nitinol guidewire having fully-insulated proximal and distal tips (Glidewire 0.035″, Terumo). Using these parameters, we tested guidewire heating under exaggerated geometry conditions in vitro, in situ in post-mortem swine, and finally used this guidewire in consenting research subjects undergoing clinical CMR fluoroscopy catheterization.

Methods

CMR

Procedures were performed in a combined X-ray and CMR fluoroscopy suite (Artis Zee and Aera 1.5 T, Siemens Healthineers, Erlangen, Germany) for both animals and patients. The CMR fluoroscopy room incorporates LCD projectors, wireless sound-suppression communication headsets (IMROC IR, Optoacoustics, Or Yehuda, Israel), and an artifact-suppressing patient physiology interface (PRiME, nhlbi-mr.github.io/PRiME/ ) [15] for a standard high-fidelity hemodynamic recording system (Sensis, Siemens Healhtineers).

CMR fluoroscopy used one of two possible pulse sequences: low-SAR with metallic guidewire (spiral gradient echo (GRE), 16 spiral arms per frame, matrix 128 × 128, TR 10 ms, TE 0.86 ms, flip angle 10o, FOV 400 mm, slice thickness 8 mm, spatial resolution 3.125 mm × 3.125 mm, temporal resolution 160 ms per frame); or normal-SAR without metallic guidewire (rectilinear bSSFP, matrix 120 × 160, TR 2.98 ms, TE 1.49 ms, flip angle 45o, field-of-view (FOV) 400 mm, slice thickness 8 mm, GRAPPA acceleration rate 2–4, spatial resolution 3.33 mm × 2.5 mm, temporal resolution 89.4 ms - 178.8 ms per frame). Guidewire heating is proportional to the square of the flip angle and inversely proportional to TR [14] and therefore the parameters used for low-SAR protocol theoretically generated a 67 fold reduction in heating compared to the normal-SAR protocol. A saturation pre-pulse was used to improve visibility of gadolinium filled balloons interactively. The low-SAR sequences used a partial saturation pulse with flip angle 45° [16] and the normal-SAR sequence used a dark blood saturation preparation. Imaging was performed in normal SAR operating mode, limiting time-averaged whole body SAR to 2 W/kg.

Catheter and guidewire devices

We selected a polymer balloon-wedge endhole catheter with a rigid distal curve that in our experience is useful to navigate cardiac chambers (Swan-Ganz Flow-Directed Monitoring Catheter, T-tip, Dual-Lumen including inflation port, Model T111F7, 7Fr × 110 cm, 0.038″ lumen, Edwards Lifesciences, Irvine, California, USA) [2, 3], but had other shapes available on standby (Balloon Wedge Pressure Catheter, Model AI-07127, Arrow-Teleflex, Reading, Pennsylvania, USA). Balloon catheters were filled with gadobutrol (diluted 0.5% from stock) or air.

We selected a specific guidewire for its materials and insulation properties. The guidewire (Glidewire Model GR3506, Terumo Corporation, Tokyo, Japan, 0.035″ × 150 cm, standard stiffness, 3 cm flexible tip, angled tip configuration) has a single nitinol core and is completely insulated with expanded polytetrafluoroethylene (ePTFE) without exposed gaps. By contrast, a popular comparator PTFE-coated nitinol guidewire (Nitrex, Medtronic,Minneapolis, Minnesota, USA, Covidien EV3, for example 0.035″ × 145 cm Model N351452, Medtronic) has exposed and un-insulated tips, which we consider unsafe for use in CMR.

In vitro experiments

Confidence in freedom from heating was established with in vitro experiments using an ASTM 2182 gel phantom. This gel phantom exaggerates heating and in vivo we expect less heating because of blood flow cooling, the central position of the vessels and convective heat dissipation.

Guidewire temperature was evaluated for a range of guidewire configurations, as summarized in Table 1; catheter position relative to guidewire, guidewire looping, guidewire length and guidewire position within the bore were among conditions explored. A fiber optic temperature probes and signal conditioner with ±0.3 °C accuracy were used for temperature measurements (OpSens Solutions Inc., Quebec, Canada). The temperature probe was affixed to the guidewire using a polyimide channel fastened with heat shrink tubing [17], or positioned using a homebuilt modular positioning system that supports both the guidewire and temperature probe in contact.

Table 1 Guidewire configurations tested during in vitro experiments. Experiments were performed in the ASTM-2182 phantom using Terumo Glidewires and Medtronic Nitrex in a 70 cm bore 1.5 T MRI system (Aera, Siemens, Erlangen, Germany)
Full size table

Animal experiments

Animal experiments were approved by the Institutional Animal Care and Use Committee. Guidewire heating was assessed post-mortem in a 47 kg swine to replicate right-heart catheterization guidewire geometry without blood flow cooling. Temperature was measured at the tip of the selected Glidewire extended 1 cm from a 7F catheter (Goodale-Lubin, Medtronic).

Patient catheterization

The clinical research protocol was approved by the Institutional Review Board (NCT03152773). CMR fluoroscopy heart catheterization is a standard medical procedure at our institution. Consecutive patients undergoing medically necessary right heart catheterization between August and November 2017 were invited to participate. All gave written informed consent. Candidates were excluded for cardiovascular instability, pregnancy or nursing, or ineligibility for CMR. Those with an estimated glomerular filtration rate < 30 mL/min/1.73m2 were enrolled, but balloon catheters were filled with air rather than dilute gadolinium contrast.

Under moderate sedation, subjects underwent measurement of left ventricular pressure under X-ray guidance, followed by the research CMR fluoroscopy catheterization. Interventional procedures were performed before or after the research study, such as radial artery access or coronary intervention, and heparin anticoagulation was administered at the discretion of the physician. Baseline CMR was performed to assess anatomy, chamber dimensions and biventricular function.

For guidewire CMR fluoroscopy catheterization, time-to-enter each of four targeted chambers (superior vena cava, main pulmonary artery, right and left pulmonary arteries) was recorded. At their discretion, operators interactively retained or extended the guidewire beyond the catheter tip, to probe and enter the target chamber.

The guidewire was introduced following catheter positioning and extended 5-15cm from the catheter tip to engage each target chamber. Alternatively, the guidewire could be kept just inside of the catheter in order to stiffen it to advance the catheter into the chambers. The guidewire was oriented parallel to the z-axis in the cava and was interactively advanced and torqued to engage pulmonary artery branches to guide advancement of the catheter. Each chamber was sampled for hemoglobin oximetry.

As time (and subject patience) allowed, the procedure was repeated without a guidewire under conventional (normal-SAR bSSFP) imaging conditions. Subjects were observed for four hours after research catheterization, after which their research participation concluded.

Absent a safe method, guidewire tip heating was not directly measured in patients. Measures of systemic coagulation (plasma D-dimer immunoassay) and of focal red cell injury (spectrophotometric plasma free hemoglobin, Mayo Labs, Rochester, Minnesota, USA) were collected before and after CMR fluoroscopy catheterization, knowing they may be confounded by clinical conditions. Guidewire tips and catheter effluent were assessed for clot after every chamber sample, knowing that clot formation is common in guidewire catheterization without anticoagulation.

Data analysis

Two independent investigators assigned Likert-type ordinal scales to measure catheter visualization (0 = invisible, 1 = sometimes visible, 2 = most times visible, 3 = all times visible) and confidence in catheter position (0 = uncertain, 1 = probable and requires pressure confirmation, 2 = certain) [5]. Guidewire visibility and confidence in tip position was also scored when the guidewire was extended from the catheter to engage each chamber.

Results are presented as mean ± standard deviation (SD) or median (1st quartile, 3rd quartile) as appropriate and are compared using a Wilcoxon rank sum test. A p value < 0.05 was considered statistically significant.

Results

Guidewire heating

Heating of metallic guidewires was influenced by CMR pulse sequence, guidewire configuration and guidewire physical properties, as summarized in Table 2.

Table 2 Factors that modulate guidewire heating during CMR
Full size table

The selected Terumo Glidewire was tested under worst-case heating conditions extended 1 cm from catheter tip positioned at the edge of the ASTM 2182 phantom (45 cm insertion, > 5 cm depth, 12.7 cm off-isocenter) in a straight configuration. It generated < 4°C with normal-SAR imaging and < 0.1°C with low-SAR imaging during 2 min of continuous imaging (Fig. 1a). Under identical conditions, the comparator Nitrex guidewire generated 15.5°C of heating with normal-SAR imaging and 0.5°C heating with low-SAR imaging during 2 min of continuous imaging (Fig. 1b). The incomplete insulation exposes metal at the tip of the nitinol guidewire concentrates current density and contributes to heating of this guidewire model [8].

Fig. 1
figure 1

Heating of guidewires. Temperature at the tip of the fully insulated guidewire (Terumo Glidewire) (a) and uninsulated tip guidewire (Medtronic Nitrex) (b) during 2 min of continuous scanning (low-SAR and normal-SAR) in the ASTM 2182 phantom with homebuilt positioning system

Full size image

Animal post-mortem experiments found maximum heating of 0.3°C with normal-SAR imaging and 0°C with low-SAR imaging during 2 min of continuous scanning. Maximum heating was observed as the guidewire protruded from the femoral sheath into the vasculature.

Patients and clinical findings

Of 13 consecutive patients undergoing right heart catheterization during the study period, 8 consented and 7 underwent the study procedure. One was excluded because of hardware malfunction after vascular access was obtained; the operator felt the patient’s discomfort required swift conclusion of the procedure under X-ray. Two were excluded because they required emergency intervention, one declined CMR guidewire catheterization and underwent standard CMR fluoroscopy catheterization without a guidewire, and two required left but not right heart procedures.

Seven patients (46±20 years; 4 (57%) female) underwent guidewire CMR fluoroscopy catheterization. All had normal cardiac anatomy. All had suspected pulmonary hypertension. Underlying illness was sickle cell disease (n = 3); atrial septal defect (n = 2); cardiomyopathy (n = 1); and rheumatologic illness(n = 1). One had chronic renal failure and underwent catheterization using a gas-filled, instead of gadolinium contrast-filled, balloon catheter. One underwent nitric oxide vasodilator challenge under CMR. One underwent concomitant percutaneous coronary intervention and two underwent atrial sepal defect repair, all under X-ray guidance.

CMR fluoroscopy catheterization took a median of 30 (27, 48) minutes including catheter navigation, repeated procedure steps to measure chamber entry time and conspicuity using up to two imaging pulse sequences (low-SAR with guidewire and normal-SAR without guidewire), CMR flow, and in one case nitric oxide inhalation.

Hemodynamic and CMR findings are summarized in Table 3.

Table 3 CMR Catheterization Findings
Full size table

Clinical performance of CMR fluoroscopy catheterization with- and without a guidewire

CMR guidewire fluoroscopy catheterization was successful in entering intended cardiac chambers in 100% of attempts, compared with non-guidewire fluoroscopy catheterization, 94% (p = NS). The failures were exclusively branch pulmonary arteries, both left and right.

Time-to-enter each chamber was not significantly different between the two approaches, low-SAR with guidewire and normal-SAR without guidewire (35 ± 45 s vs 26 ± 38 s, p = NS).

Table 4 shows Likert-type scores of catheter and guidewire tip and shaft visibility compared using the two different imaging techniques. Overall visibility and operator confidence in catheter tip location was comparable among all modalities using saturation pre-pulses. The catheter shaft was rendered slightly more visible by a guidewire susceptibility artifact (Fig. 2), but we observed that catheter shaft visibility was reduced when saturation pre-pulses were applied to confirm contrast-filled balloon position. Guidewire tip visibility was poor throughout. Imaging during navigation from inferior to superior vena cava is shown in Additional file 1: Video 1 available online. As expected, GRE imaging (low-SAR pulse sequence) provided inferior blood-myocardium contrast compared to bSSFP imaging (normal-SAR pulse sequence) (comparison images provided in Fig. 3).

Table 4 Likert-type scoring of device visibility and confidence in catheter and guidewire position. Only catheter shaft visibility reached statistical significance (p = 0.001)
Full size table
Fig. 2
figure 2

Catheter shaft conspicuity imparted by guidewire. Example low-SAR spiral GRE images, with catheter and guidewire (arrowheads) positioned in the superior vena cava (a), main pulmonary artery (b) and left pulmonary artery (c); the guidewire imparts both hypo- and hyper-intense signal along the shaft. The dotted line indicates signal saturation caused by an orthogonal slice plane interleaved during imaging

Full size image
Fig. 3
figure 3

Comparator rectilinear bSSFP images. Image orientation comparable to Fig. 2 showing improved blood-tissue contrast with bSSFP rectilinear images (normal-SAR imaging mode) for inferior/superior vena cava view (a), main pulmonary artery view (b) and branched pulmonary arteries view (c). No devices were present during imaging

Full size image

Additional file 1: Video 1. (MP4 7886 kb)

Safety

There were no adverse events in this small cohort.

To test for subclinical sequelae of guidewire heating, and out of caution, we measured baseline and follow-up markers of coagulation (plasma D-dimer) and of erythrocyte destruction (plasma free hemoglobin). These were all unchanged except for one patient who had undergone percutaneous coronary stenting using heparin anticoagulation, in whom plasma free hemoglobin rose from 6.4 to 15.2 mg/dL, which is the upper limit of normal.

We also inspected guidewires and catheter aspirates for thrombus formation after use. Clot was observed on the guidewire or catheter aspirate in 15% of guidewire specimens versus 11% of non-guidewire specimens (p = NS), and none among subjects who received heparin anticoagulation (one for percutaneous coronary stenting, two for left atrial catheterization).

Discussion

Adoption of CMR fluoroscopy catheterization has been hampered by the unavailability of safe CMR-compatible clinical devices, especially guidewires. Metallic guidewires by definition are conductive and therefore prone to heating in the conditions required for CMR fluoroscopy catheterization in adults. Certain steps in right heart catheterization are difficult or impossible without a guidewire in some patients [2, 3, 5], especially selecting branch pulmonary arteries. The mechanical properties of commercial metallic guidewires are superior to non-metallic guidewires and we are not aware of any non-metallic guidewires with satisfactory mechanical properties for catheterization. To address concerns of metallic guidewire heating, we comprehensively tested and confirmed the safety of CMR catheterization using a specific guidewire under specific low-SAR imaging conditions, both in vitro and in animals before translating into human patients. Post-mortem animal experiments were chosen to exaggerate potential heating with no blood flow cooling. Heating of the selected guidewire with the low-SAR pulse sequence was negligible (< 0.07 °C) and less than the equipment accuracy of ±0.3 °C for all conditions tested in vitro.

This is the first in human report of CMR fluoroscopy cardiac catheterization with a standard commercially available metallic guidewire. The main findings of this study are that (1) a specific nitinol guidewire (Glidewire 0.035″ 150 cm angled-tip) can be used under specific low-SAR conditions to facilitate CMR fluoroscopy right heart catheterization at 1.5 T; (2) the guidewire imparts conspicuity to the shaft of otherwise-invisible polymer balloon-tipped catheters; (3) the incremental stiffness imparted to the polymer catheter facilitates procedure success under CMR fluoroscopy; (4) systemic measurements of coagulation (D-dimer) and erythrocyte injury (plasma free hemoglobin), although of unknown sensitivity, were not significantly perturbed using the guidewire; (5) clot accumulated on the guidewire and on guidewire-free catheters in patients undergoing CMR fluoroscopy when not anticoagulated, a known phenomenon under X-ray of unclear significance under these conditions; (6) Despite the reduced contrast-to-noise ratio, the low-SAR imaging mode was adequate to guide catheter navigation for CMR fluoroscopy right heart catheterization; and, (7) the visual performance of the specific nitinol guidewire used under CMR fluoroscopy is grossly inferior to X-ray in that the tip is not conspicuous, which is a common problem for passive devices in CMR [18] that must be considered carefully by operators.

Ours is not the first human interventional CMR procedure using a Terumo guidewire device. Paetzel et al. reported femoropopliteal balloon angioplasty in 15 patients under real-time CMR guidance at 1.5 T in 2005 [19]. They used low-SAR gradient echo pulse sequences and assured adequate insertion length and proximity to isocenter before CMR, and found no heating related events. The authors did not report on guidewire visibility, which presumably was as poor as in our experience.

Recommended workflow

Our approach would be simple to replicate at other medical centers. We describe a procedure for nitinol guidewire-assisted CMR fluoroscopy right heart catheterization using specific conditions at 1.5 T. We believe this approach should be only undertaken by expert catheter operators familiar with this operating environment to avoid (1) inadvertent guidewire scanning using normal-SAR conditions (high flip angle, short TR), (2) tissue injury or myocardial perforation caused by inadequate guidewire tip visualization despite tactile cues, (3) selection of an alternative commercial guidewire with unsafe heating properties.

Based on this experience, we intend to employ guidewires (with low-SAR CMR fluoroscopy) catheterization only when needed, much as they are employed during X-ray fluoroscopy catheterization. We recommend other sites adopt a workflow that clearly assures low-SAR imaging modes during guidewire steps, such as explicit team communications, checklists, and visual warnings on imaging displays. We also recommend routine low-dose anticoagulation which is widely employed by pediatric but not by most adult catheterization programs, in light of the low but appreciable incidence of clots recognized in all arms of this study.

We performed this work in a combined X-Ray/CMR suite, although this procedure would also be feasible in a standard 1.5 T CMR suite with an appropriate evacuation plan in case of emergency. A CMR compatible defibrillator [20] might be helpful for complex CMR-guided procedures.

Limitations

We have not yet tested CMR fluoroscopy guidewire catheterization during retrograde left heart catheterization in patients, and we caution against such procedure absent assurance about guidewire tip visualization during CMR fluoroscopy. We also warn against the likely danger of using guidewires other than the specific device described unless carefully examined for heating. The specific device has insulation and length properties and we were able to demonstrate freedom from heating under the conditions of use.

Visualization of commercial nitinol guidewires remains a challenging, though work in positive contrast imaging methods may improve the visibility at the expense of temporal resolution [14]. We chose an 8 mm slice thickness to avoid out-of-plane guidewire motion, but thinner slices could be explored to improve visualization of the signal void.

Conclusion

Metallic guidewires are an important adjunct for catheterization as they interactively control shaft stiffness, and help to select and enter target chambers. By lowering SAR (through reduced RF excitation flip angle and increased repetition time, implemented as spiral gradient echo CMR), and by selecting a specific fully-insulated nitinol guidewire, we safely employed a metallic guidewire during CMR fluoroscopy catheterization. The guidewire also improved shaft conspicuity, which we believe facilitated procedure success. These findings should not be applied to other nitinol guidewire devices or similar devices having different lengths without specific testing to assure freedom from heating. Though the guidewire visual performance is inferior to X-ray in that the tip is not conspicuous, the ability to safely use a metallic guidewire during CMR fluoroscopy could enable new and more complex CMR-guided procedures.