Can a Simple Flip Make a Difference in MRI Safety in Patients with Cardiac Implantable Electronic Devices?

See also the article by Martinez et al in this issue.

Harold I. Litt, MD, PhD, is an associate professor of radiology and medicine and chief of the cardiothoracic imaging division at the Perelman School of Medicine of the University of Pennsylvania. He is the cardiac MRI deputy editor for Radiology: Cardiothoracic Imaging. Dr. Litt has published many articles in journals such as Radiology, The American Journal of Roentgenology, New England Journal of Medicine, Magnetic Resonance in Medicine, and Academic Radiology.

Introduction

The practice of MRI in patients with cardiac implantable electronic devices (CIEDs) has evolved from an absolute contraindication to routine at many institutions. There are now multisociety consensus statements for best practices (1), and Centers for Medicare and Medicaid Services (CMS) policy against reimbursement for studies performed with legacy or nonconditional CIEDs was reversed in April 2018 (2). While MRI conditional devices are increasingly available, many patients with legacy systems will require MRI in the future, and therefore it is important to work to find methods to minimize any potential safety risks in these studies. Both the consensus statements and the CMS guidelines recognize several potential safety issues when performing MRI in patients with legacy CIEDs related to the interaction of the device with the MR system, including direct effects on the device and risks related to the programming required in the MRI environment. In this issue, Martinez et al (3) focus on one of these risks, myocardial tissue damage at the lead tip related to heating caused by deposition of radiofrequency energy in device leads.

Recognizing that most CIED generators are implanted in the left upper chest, the authors hypothesized that inhomogeneity in the E-field would present an opportunity to reduce radiofrequency-induced lead heating by switching from a head-first (HF) to a feet-first (FF) patient positioning. They evaluated lead heating by using a titanium rod in two different 1.5-T MRI systems and found dramatic reductions in both derived specific absorption rate and measured heating for FF positioning. Using a phantom to simulate the properties of tissue, they re-created two actual CIED lead paths derived from cadavers with devices in place, and additionally studied a planar path in which the leads were all in the coronal plane. They studied nine different CIEDs at multiple table positions relative to isocenter for both HF and FF positioning. Their principal result was that lead heating is much lower in FF than HF positioning and that this effect was more important than other variables examined. Specifically, they found that 41% of the 378 combinations of devices, paths, and positions studied resulted in temperature increases of more than 2°C for the HF orientation, compared with only 13% for FF. However, they also redemonstrated previously known smaller dependencies of heating on specific lead path and position relative to isocenter, with higher heating for the planar lead path and at positions at which the lead tip was closest to isocenter.

The authors appropriately note several limitations of their results, specifically that only one phantom (and therefore only one body size) was tested and that only a limited number of devices, coils, lead lengths, and lead paths were tested. However, there are several other limitations, including use of only a single pulse sequence (turbo spin echo) that may not reproduce the conditions during actual scanning (specifically for cardiac MRI, in which steady-state free procession gradient-echo cine sequences comprise the majority of scans performed). Additionally, the actual device scanning took place only in one scanner at 1.5 T; although, at the moment, very few scans in patients with CIEDs are performed at 3.0 T, we can expect this to increase in the future. The scanner used has a long magnet (150 cm), with a 60-cm bore diameter—will the results be the same in a shorter, wider bore system in which the generator or leads may be in a region of higher field inhomogeneity?

They do note one more important limitation, which is that many types of examinations may not be able to be performed in the FF position; although in these brain, cervical spine, and shoulder examinations, the lead tip–myocardium interface is already further from isocenter. However, the authors do not mention what may be the most important limitation to their results, namely, that no study has every demonstrated a clinically relevant safety issue related to lead-tip heating. Prior studies have shown statistical elevations in both troponin (a biomarker of myocardial damage) and the number of patients with increases in pacing thresholds of more than 1 mV (a clinically relevant change) after MRI (4), but none of these studies measured these same variables in a control group of patients who did not undergo MRI. Also, none have ever demonstrated any safety effects of these changes, other than the theoretical risk of accelerated battery depletion if devices are programmed to have a higher output because of an increase in threshold measured after MRI.

Nevertheless, as Gimbel and Kanal noted in 2004 “…failing to identify an adverse event is not equivalent to demonstrating safety...” (5). Therefore, changes in the “controllable factors” described by Martinez et al, such as using FF orientation when possible and moving the lead tips further away from isocenter when not, provide easy ways to minimize the risks for patients with CIEDs undergoing MRI.