Purpose of the study.
Stroke is a leading cause of mortality in patients supported by Left Ventricular Assist Devices (LVAD) and a contraindication for cardiac transplantation in several centers.1–3 Early detection of stroke is crucial in the management of these patients, which may potentially improve their clinical outcomes. Portable low-field brain MRI (pMRI) has been shown to be safe for extracorporeal membrane oxygenation (ECMO) patients, even with an Intra-aortic balloon pump (IABP) or Swan Ganz catheter (SGC); however, its safety in LVAD patients is untested.4,5 In this study, we aimed to determine the compatibility of this innovative technology with the HeartMate3 (HM3) LVADs.
Main Results.
We present the first known demonstration of portable low-field brain magnetic resonance imaging (pMRI) use with a phantom model HM3 LVAD (Abbott Laboratories, Pleasanton, CA). pMRI was performed using a 64 mT Swoop® MR imaging system (Hyperfine, Inc., Guilford, CT). The MR system was positioned with controllers outside the magnet’s 5 Gauss (5G) safety line (Figure 1). Following this, the site quality assurance (SQA) phantom, used to mimic the brain and allow image quality assessment, was placed into the head coil of the scanner.
Figure 1:
Map of the Static Magnetic Field and the locations of HM3 Left Ventricular Assist Device (LVAD) placement.
We placed the LVAD into a container filled with approximately one liter of water in it, completely submerging the pump. To accurately determine the effect of the magnetic field on the device, we placed the electrical controller next to the container (Supplemental Figure 1). We then placed this setup at three distinct distances from the center of the scanner. The three locations (Figure 1) were at 180 cm (outside the 5G boundary) to record the baseline parameters, at 80 cm (at the 5G boundary) (Supplemental Figure 1), and at 40 cm (inside the 5G boundary), a distance that mimics in vivo LVAD positioning. We selected these locations to assess a) the impact on the LVAD of the increasing magnetic field strength as we moved the device towards the center of the scanner, and b) the effect of LVAD on the quality of the SQA scan. At each distance, we assessed predetermined LVAD parameters of function and performed scans of the SQA phantom to determine image quality. The primary outcomes were the safety and feasibility of pMRI at clinically relevant distances of an HM3 LVAD phantom from the center of the scanner. LVAD function was assessed by significant changes in pump speed, pump flow, power, and printed circuit board temperature (PCBT). Imaging was assessed by evaluating SQA imaging quality with both Diffusion Weighted Imaging (DWI) and Transverse relaxation time (T2) sequences. We chose DWI and T2 protocols as they have been shown to provide clear delineation of alterations in brain tissue and hence, their applicability in the clinical setting. We eliminated T1 sequence as this decreased scanning time without loss of image sensitivity. Images generated at 180cm were used as the baseline dataset against which the images obtained at the other locations were compared for radiofrequency noise artifact. Quantitative analysis was not performed as the LVAD was deemed to be unsafe.
At 180 cm, the magnetic field was less than 5G, within an order of magnitude of the Earth’s field strength. At this point, we recorded the baseline LVAD parameters in the absence of exposure to radiofrequency waves i.e., with the scanner turned off (Table 1) followed by generation of DWI and T2-W images of the SQA phantom. No LVAD parameters were affected and the image quality of the SQA was perfect with no distortions.
Table 1:
Device parameters recorded at different locations.
| Distance (from the center of scanner) | Pump speed (rpm) | Pump flow (L/min) | Power (Watts) | Printed circuit board temperature (PCBT) |
|---|---|---|---|---|
| 180 cm (scanner off) | 5400 | 3.3 | 3.4 | 29 |
| 180 cm (on) | 5400 | 3.3 | 3.4 | 30 |
| 80 cm (on) | 5400 | 3.3 | 3.4 | 31 |
| 40 cm (on) | 0 | 0 | 21 | 35 |
We then moved the device to 80 cm (at the 5G boundary) (Figure 1). The 5G boundary, roughly 10x the Earth’s magnetic field, is considered the transition zone to a potentially significant magnetic field. At this location, LVAD parameters and the SQA imaging were unperturbed.
We considered 40 cm, with a magnetic field of 200G (Figure 1), as the position corresponding to the location in the human body where the HM3 LVAD would be expected, representing a true reflection of the impact of the magnetic field on the HM3. At this point, the HM3 ceased to function, with pump speed and pump flow dropping to zero. The power, which represents the current required to maintain the motor’s RPM’s and fluid flow, increased, from 3.4 W at 80 cm to 21 W at 40 cm, representing the increased workload of the device as its mechanism was impacted by the stronger magnetic field. The device PCBT rose from 31, the temperature seen at 80 cm to 35 with the HM3 at 40 cm (Table 1). This rise in PCBT was not significant as per the company’s guidelines for the device as the normal PCBT for patients can be seen anywhere from the mid-40s to mid-50s. True heating is when PCBT is in the range of 70s. Therefore, circuit heating was excluded as a cause of device malfunction. Importantly, the SQA images obtained at this location showed no distortions or artifacts.
We anticipate that patients with a decommissioned (non-functioning) device, i.e., in the case of cardiac remission, or even an LVAD stopped temporarily for the pMRI procedure for patients who might tolerate such, may be considered for a pMRI as the image quality was unaffected by the proximity of the device. However, because the device would be implanted, there could be a trauma risk due to unintended movement of the implanted device and the magnetic field might interact with the integrity of the pump.
Conventional MRI relies on high-strength magnetic fields (1.5–3T) and is incompatible with the ECMO circuit and LVADs due to magnetic torque on the pump leading to resistance to spinning, heating, and, possible device migration. It was hypothesized that pMRI, with a field 1/50th as strong as the conventional MRI, would be compatible with the LVAD, allowing sensitive brain imaging of these patients, even in a critical care setting.4 However, our study shows that the HM3 is incompatible even with this lower field strength. The fully magnetically levitated HM3 rotor most certainly represents a key vulnerability, increasingly misaligned with increasing external magnetic field exposure. The magnetic interference pulled the rotor to the side of the pump, and the rotor was unable to keep itself centered and levitated to spin at the set pump speed, which led to the device’s malfunctioning. However, because this configuration was experimental and the pump was not inside a human with a patient’s heart, other organs, or chest wall to act as a shield, this may have changed how the magnetic field affected the device’s parameters. Therefore, future research is required to ascertain this interaction using translational research trials or ex-vivo heart models.
In conclusion, positioned at what would be a real-life distance from the center of this pMRI magnet, the HM3 did not function, exhibiting an incompatibility with the magnet. However, imaging did not seem to be impacted by the device, with scans showing good image quality unaffected by LVAD proximity. Through this experiment, we demonstrated that patients with HM3 cannot be assessed for neurological injury even with pMRI; as a result, brain computed tomography (CT) scan, although being a less sensitive modality to identify strokes, is hitherto the only reliable neurological imaging test in LVAD patients.
Supplementary Material
Supplementary Figure 1: Demonstration of HM3 Left Ventricular Assist Device (LVAD) placed at 80 cm from the scanner center.
Sources of Funding Statement:
This study was funded by Hyperfine, Inc. S.M.C. is funded by NHLBI 1K23HL157610.
Footnotes
Conflict of Interest: All other authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
References.
- 1.Patlolla V, Mogulla V, DeNofrio D, Konstam MA, Krishnamani R: Outcomes in patients with symptomatic cerebrovascular disease undergoing heart transplantation. J Am Coll Cardiol. 58(10):1036–1041, 2011. [DOI] [PubMed] [Google Scholar]
- 2.Cho SM, Mehaffey JH, Meyers SL, et al. Cerebrovascular Events in Patients With Centrifugal-Flow Left Ventricular Assist Devices: Propensity Score-Matched Analysis From the Intermacs Registry. Circulation. 144(10):763–772, 2021. [DOI] [PubMed] [Google Scholar]
- 3.Plecash AR, Byrne D, Flexman A, Toma M, Field TS: Stroke in Patients with Left Ventricular Assist Devices. Cerebrovasc Dis. 51(1):3–13, 2022. [DOI] [PubMed] [Google Scholar]
- 4.Cho SM, Wilcox C, Keller S et al. : Assessing the SAfety and FEasibility of bedside portable low-field brain Magnetic Resonance Imaging in patients on ECMO (SAFE-MRI ECMO study): study protocol and first case series experience. Crit Care 26, 119, 2022 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Wilcox C, Acton M, Rando H, et al. : Safety of Bedside Portable Low-Field Brain MRI in ECMO Patients Supported on Intra-Aortic Balloon Pump. Diagnostics (Basel) 12(11):2871, 2022. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplementary Figure 1: Demonstration of HM3 Left Ventricular Assist Device (LVAD) placed at 80 cm from the scanner center.