Author + information
- Received December 5, 2016
- Revision received March 7, 2017
- Accepted March 13, 2017
- Published online September 18, 2017.
- Huma Samar, MDa,b,
- June A. Yamrozik, BSc,
- Ronald B. Williams, BAc,
- Mark Doyle, PhDc,
- Moneal Shah, MDa,c,
- Christopher A. Bonnet, MDc,d and
- Robert W.W. Biederman, MDa,c,∗ ()
- aDivision of Cardiology, Centre for Cardiac MRI, Allegheny General Hospital, Pittsburgh, Pennsylvania
- bLoma Linda Veterans Affairs Hospital, Loma Linda, California
- cGerald McGinnis Cardiovascular Institute, Pittsburgh, Pennsylvania
- dDivision of Electrophysiology, Gerald McGinnis Cardiovascular Institute, Pittsburgh, Pennsylvania
- ↵∗Address for correspondence:
Dr. Robert W.W. Biederman, Division of Cardiology, Centre for Cardiac MRI, Allegheny Health Network, Allegheny General Hospital, 320 East North Avenue, Pittsburgh, Pennsylvania 15212.
Objectives The objective of this study was to assess the diagnostic usefulness of thoracic and nonthoracic magnetic resonance imaging (MRI) imaging in patients with implantable cardiac devices (permanent pacemaker or implantable cardioverter-defibrillators [ICDs]) to determine if there was a substantial benefit to patients with regard to diagnosis and/or management.
Background MRI is infrequently performed on patients with conventional pacemakers or ICDs. Multiple studies have documented the safety of MRI scans in patients with implanted devices, yet the diagnostic value of this approach has not been established.
Methods Evaluation data were acquired in 136 patients with implanted cardiac devices who underwent MRIs during a 10-year period at a single institution. Specific criteria were followed for all patients to objectively define if the diagnosis by MRI enhanced patient care; 4 questions were answered after scan interpretation by both MRI technologists and MRI physicians who performed the scan. 1) Did the primary diagnosis change? 2) Did the MRI provide additional information to the existing diagnosis? 3) Was the pre-MRI (tentative) diagnosis confirmed? 4) Did patient management change? If “Yes” was answered to any of the preceding questions, the MRI scan was considered to be of value to patient diagnosis and/or therapy.
Results In 97% (n = 132) of patients, MR added value to patient diagnosis and management. In 49% (n = 67) of patients, MRI added additional valuable information to the primary diagnosis, and in 30% (n = 41) of patients, MRI changed the principle diagnosis and subsequent management of the patient. No safety issues were encountered, and no adverse effects of undergoing the MRI scan were noted in any patient.
Conclusions MRI in patients with implanted pacemakers and defibrillators added value to patient diagnosis and management, which justified the risk of the procedure.
An increasing number of pacemakers (PMs) and implantable cardioverters-defibrillators (ICDs) are implanted as a result of expanding indications (1). Approximately 1.8 million people in the United States have a cardiac PM or ICD (1), and it is estimated that up to 75% of them will require a clinically indicated magnetic resonance imaging (MRI) scan during their lifetime (2), with the rate of MRI use increasing at >5%/year (3). This is reflected in the increasing number of MR conditional PMs and ICDs that are commercially available. However, because a large number of patients are presently encountered in the clinical setting with conventional devices (non-MR conditional), imaging is problematic. Non-MR conditional PM and ICD labeling currently cautions physicians against using MRI, and MRI manufacturers contraindicate MRI in patients with PMs and ICDs. Presently, the U.S. Food and Drug Administration does not recommend the routine use of MRI in this population, and it is still considered “off-label” (4). This is despite the fact that there are numerous studies in the literature that include the results of the recent 1,500 patient multicenter MagnaSafe Registry (6), which documents the safety of MRI scans in these patients when performed with a strict protocol to minimize risk (5–8). Based on these studies, there are an increasing number of centers now performing MRI in these patients. However, there remains a much larger population who may benefit from MRI, but who are denied this option due to a widespread lack of recognition of the risk-to-benefit ratio or due to lack of training and experience. Despite MRI being established as the gold standard for many indications, the usefulness of MRI is often overshadowed by less precise and accurate imaging tools, with the notion that the added value is not worth the incremental risk. Therefore, in particular, this concept was a key focus and motivation for this study. Previously, attention was overwhelmingly focused on the safety aspect, with little data on the diagnostic usefulness of MRI (7–9). In the current environment, in which a large and growing number of patients with traditional devices require MRI scans, we addressed the issue of the potential for benefit, with the understanding that safety aspects were addressed by an established protocol (1,5,6). Stated alternatively, once safety was established, the equally critical aspect of medical usefulness could then be assessed.
The objective of the present study was to determine the diagnostic usefulness of thoracic and nonthoracic MRI performance in patients with implantable cardiac devices (permanent PMs or ICDs) and to determine if there was a substantial benefit to the patient with regard to diagnosis and/or management. Critical to this was the recognition by our Center that safety issues following an established protocol had been well defined by us and others over the preceding decades, permitting the current focus on the diagnostic yield of the MRI scan. To our knowledge, this is the first study to advance the notion of “value added for risk assumed.”
This prospective study was reviewed and approved by the Allegheny General Hospital Institutional Review Board. All patients with implantable devices (non-MR conditional PMs and ICDs) who underwent an MRI scan between September 2005 and February 2015 (97% performed after 2009) at Allegheny General Hospital (Pittsburgh, Pennsylvania) were offered inclusion in the study. Most patients (>99%) underwent previous imaging with either a computed tomography scan or myelography for neurological and/or orthopedic cases, and transthoracic echocardiography, transesophageal echocardiography, nuclear examination, or computed tomography for cardiac indications. All patients referred for MRI had a valid clinical indication, and >95% of the referred patients underwent the requested imaging study. Only patients with devices and leads that were implanted after 2004 were imaged, in concert with the MagnaSafe criteria. No restrictions were imposed on different transvenous leads, models, or generators. All patients gave written informed consent witnessed by a family member after a detailed discussion with the cardiologist performing the study.
Candidates with an implantable cardiac device were referred from a variety of inpatient (69%) and outpatient (31%) services to Allegheny General Hospital Cardiac MRI Laboratory. All patients had a clinical indication for MRI because previous imaging studies had either not been definitive, or in some cases, patients could not tolerate certain imaging procedures (e.g., myelography). Patients were enrolled into the study prospectively without regard to whether the device was found to be safe by previous in vitro phantom and in vivo animal testing. Due to extremis in <3% of scans, patient were accepted under the following circumstances: device implantation <6 weeks before MRI (as per the package insert, a 6-week duration after implantation is usually needed for endothelialization of the leads; this was done in 1 patient due to the emergent nature of the central nervous system scan); and those with nontransvenous epicardial leads, no fixation (such as superior vena cava coil), or abandoned leads. In all cases, no restrictions were imposed on different transvenous lead models.
MRI device preparation protocol
The protocol for device monitoring and reprogramming before and after the scan followed the protocol of the MagnaSafe Registry (5). All devices were interrogated and reprogrammed by an electrophysiology nurse under guidance from an MR trained cardiologist with experience in imaging patients with implantable devices. The cardiologist was present throughout the entire study. All MRI personnel were certified in Advanced Cardiac Life Support, and a code cart was present at all times in the MR suite.
All patients were imaged on a 1.5-T magnet MRI scanner (1.5-T GE, Milwaukee, Wisconsin). Continuous electrocardiography and oximetry were monitored. Noninvasive blood pressure measurements were obtained every 5 min. MRI was performed according to institutional protocol for the region of interest and suspected diagnosis. Estimated whole-body averaged specific absorption rate (SAR) for each sequence was targeted to be <2.0 W/kg. The sequences used for noncardiac MRIs were T2 fluid-attenuated inversion recovery and T2 short tau inversion recovery fast spin echo. Cardiac function was assessed using gradient-recalled echo sequences in place of steady-state free precession sequences to limit SAR. Late gadolinium enhancement imaging was performed as part of the cardiac protocol as needed (MultiHance, Bracco, Princeton, New Jersey). MRI to include angiograms was performed if indicated, with alteration of sequence parameters to reduce SAR. We noted that in the MagnaSafe Trial (5) a SAR of <4.0 was permitted, and previous studies reported safety at varying maximal estimated SAR levels. However, the value of 2.0 was used in the present study to allow an adequate safety margin below the worst-case scenario conditions previously tested in in vitro and in vivo animal studies. To accomplish this, adjustments were made to the estimated SAR by increasing repetition time, the reducing scan matrix, or reducing the radiofrequency flip angle while maintaining the diagnostic capability of MRI, as assessed by the cardiologist in attendance. With this protocol, >98% of all sequences were performed with SAR <2.0. Two scans (<2%) were performed with SAR of 2.01 to 2.25. In contradistinction to most other reports, we did not specifically exclude patients who were deemed PM dependent, but instead placed them in an asynchronously paced mode at a lower rate limit of 30 to 40 beats/min depending on the comfort of the patients. However, in general, deactivation of the magnet mode was performed. Tachyarrhythmia detection and all therapeutics were turned off on all ICDs.
All cardiac and noncardiac scans were performed in a dedicated cardiac MRI suite, and although noncardiac scans were interpreted by radiologists with advanced MR training, all cardiac studies were interpreted by a level III MRI-trained cardiologist. All scanning and decision making (noncardiac or cardiac) were performed with the supervision of an advanced trained cardiac magnetic resonance (CMR) cardiologist (National Institutes of Health−sponsored CMR fellowship cardiologist with >18 years of experience). Interrogation of PMs and ICDs before and after scanning was performed by a member of the electrophysiology team.
A total of 132 of the 136 patients (97%) underwent imaging between 2010 and 2015. The results of all 136 scans were noted, and a detailed and extensive review of the patient charts was performed and tabulated in an ongoing, serial fashion within 7 days after MRI to assess changes in management and patient outcomes. This served to remove bias that did not rely on recollection. Clinical records, ancillary studies, and notes of all physicians involved in the care of the patients were studied in detail, and all changes in the plan of care for the patients were classified and codified. The review was performed by a team that included an MRI technologist and 2 cardiologists (H.S. and R.W.W.B.). In the rare event of discord, blinded review by another cardiologist facilitated consensus.
A checklist of 4 essential questions was answered for each patient after scan interpretation and all pertinent clinical records review:
1. Did the patients’ suspected diagnosis change post-MRI?
2. Did MRI provide additional information to the existing diagnosis?
3. Was the pre-MRI (tentative) diagnosis confirmed?
4. Was patient management altered as a result of the scan data?
If “Yes” was answered to any of the preceding questions, the MRI was considered to have added value to patient diagnosis and/or therapy. A separate query was recorded if there was an equivocal or “could not rule out” interpretation.
Continuous variables are summarized as mean and SEM, and discrete variables are summarized as absolute numbers and percentages. Immediate and long-term lead parameters were compared with the unpaired Student t test. Statistical analyses were performed using IBM-SPSS Statistics (version 20.0, IBM-SPSS Inc., Armonk, New York). All tests were 2-tailed, and p < 0.05 was considered significant.
Image quality was not affected when the PM or ICD was located outside the field of view. In 2 patients where the ICD (St. Jude Medical, St. Paul, Minnesota) was inside the thoracic field of view, the device caused severe image distortion and field degradation. In 1 patient, only partial field-of-view data were obtained, but the data were of diagnostic value (excluding inferior/inferoseptum and posterolateral cardiac amyloidosis). In all but 1 of the remaining patients, images were interpretable and were generally of good to excellent quality, and sufficient to answer the diagnostic question with neither the generator nor the leads causing distortion. In 1 cardiac patient, the study was uninterpretable for viability, and a positron emission tomography scan was recommended.
MRI scanning was performed in 136 consecutive patients with implantable devices (representing 170 distinct MRI examinations): 42 patients with ICDs, 6 patients with a retained lead, and 88 patients with dual-chamber conventional (non-MR conditional) PMs. No MR-approved or MR conditional devices were incorporated into this cohort. There were no safety issues before or after the scan. There were no deaths, arrhythmias, or power-on-resets. No clinically significant change in lead parameters, including sensitivity, capture threshold, lead impedance, or shock impedance were noted. No significant change in battery voltage and/or elective replacement indicator was noted. We used the same parameters of clinically significant change as in the MagnaSafe Registry (5). No post-procedure revisions to the generator and/or lead were required. No reimplantations were required. The average total time that the patient was in the magnet environment for noncardiac scans was 29 ± 8 min, whereas for cardiac scans, it was 23 ± 10 min (p = NS).
No symptoms consistent with device movement, torque, or heating were reported or observed during MRI examinations. Transient asynchronous pacing at the device-specific magnet rate (generally 99 to 100 beats/min) was rarely observed while in the bore of the magnet (<5%), which typically and appropriately ceased on full patient positioning within the bore. No unexpected or rapid activation of pacing, power-on-reset, or therapeutic delivery was observed during any MRI. Also, all devices functioned appropriately after MRI, and no changes in device programming were observed. Comparison of device interrogation results obtained before and immediately after MRI revealed no significant individual or mean changes in battery voltage, lead thresholds, lead impedances, or sensing signal amplitudes for all (100%) of the patients, independent of number of scans. In PM-dependent patients (with or without ICDs; 4%), no specific differences in symptoms, pre-PM metrics, or post-interrogation metrics were present.
Of the scanned regions, 100 (73%) patients underwent MRI of the brain or spinal cord, 31 (22%) patients underwent a cardiovascular examination, and 5 (3%) patients underwent MRI for orthopedic indications (Table 1).
In the patients who underwent a central nervous system MRI, patient diagnosis and subsequent management were significantly altered by the results of the scan in 28 (28%) cases. In an additional 47 patients (47%), MRI provided important clinically relevant information that aided in patient management even if the primary diagnosis was unchanged. Thus, 75% of the neurological MRI studies added substantial value to patient care.
Of the 31 cardiovascular cases, the patient diagnosis and management were significantly altered in 12 (38%) patients. In 16 (52%) patients, new information was obtained from the MRI scan that contributed to patient management. In 90% of the cardiovascular studies, CMR affected the diagnostic value. In 2 (6%) patients, MRI only confirmed the pre-scan diagnosis (Tables 1 and 3, Figures 3 and 4).
In 1 (20%) of the 5 studies performed for orthopedic pathologies, the post-scan diagnosis was different from the pre-scan diagnosis, and changed the surgical course of the patients, whereas in the remaining 4 patients (80%), additional valuable information was obtained from the study, which altered their medical care. Therefore, all the MRIs in this group added value to patient management (Tables 1 and 4).
Of note, individual examination of immediate and long-term device parameters in the 2 patients who underwent higher SAR sequences revealed no significant changes compared with baseline values.
Repeat MRI examination was performed in 11 patients (9 neurologic/neurosurgical and 2 cardiac [evaluation of rejection]) without adverse effects. One patient underwent 9 scans over 3 years in 1 of the index cases to define a roadmap for neurosurgical evacuation of a life-threatening impinging herniating glioblastoma multiforme. At the time of this report, he remains alive and functional.
The contribution in certain disease states of MRI as a valuable diagnostic modality, especially in central nervous system pathologies, is widely recognized (10). Recently, CMR has become indispensable in the diagnosis and management of a variety of unique cardiac conditions that are not well interrogated by standard modalities, particularly infiltrative and inflammatory myocardial diseases. A number of studies have shown that the diagnostic yield of the scar and/or fibrosis in these suspected disease states is high (11–15). Using this noninvasive imaging tool permits some invasive procedures to be appropriately avoided, which represents a substantial cost savings with improved resource use and translates into earlier therapeutic interventions. Similarly, more effective targeted therapeutics can be appropriately and confidently used when MRI-directed care, despite risk, can be implemented. However, currently, patients with implanted devices are still not routinely imaged via MRI or CMR due to overwhelming safety concerns. Although MRI in patients with conventional (non-MR conditional) devices is no longer an absolute contraindication, it requires a multidisciplinary approach and detailed assessment of risk-to-benefit ratios before scanning, together with careful monitoring during the scan and device reprogramming before and after the scan by a well-disciplined, experienced team. It follows that the decision to perform MRI in patients with implantable devices is frequently made by considering the potential benefit of the MRI relative to the attendant risk.
In 2007, the American Heart Association/American College of Cardiology determined that “MR examination of non-pacemaker dependent patients is discouraged and should only be considered in cases where there is a strong clinical indication and in which the benefits clearly outweigh the risk” (16). With regard to PM-dependent patients, they recommended that MRI examinations “should not be performed unless there are highly compelling circumstances and when the benefits clearly outweigh the risks” (16). The more recently published German Roentgen Society guidelines state that although a nonconditional implanted device no longer represents an absolute contraindication, but rather a relative contraindication, for an MR examination, it should be performed taking into consideration the individual risk-to-benefit profile and should still be regarded as “off-label” (17). The position statement from the Canadian Heart Rhythm Society recommends that MRI of a non-MR conditional device should only be performed at centers with a high level of expertise in MRI and device management, and should be restricted to clinical scenarios in which MR scanning might provide crucial information for the management of the patient’s care (18). Taken together, our findings underscored these 2 position papers’ anecdotal and expertise-driven consensus positions, and defined the first in humans confirmation of these recommendations.
The risks associated with MRI scans in this patient population are well known (16). The recent plethora of safety data in this regard, especially the recent large MagnaSafe registry data, repeatedly showed that with careful selection and monitoring of these patients and their devices by trained personnel before, during, and after the scan, serious complications can be easily avoided, and the study can be performed safely in a time efficient manner (6). This then, importantly, was therefore not the focus of our study. Regardless, there were no safety concerns or device-related issues in any patient during the entire study. However, imaging a patient with an implantable device clearly carries a higher risk than in patients without such devices; therefore, to be clinically justified, the benefit-to-risk ratio has to be high. Although the risks were well studied in this patient population, the benefit of MRI in this high-risk population was traditionally extrapolated from data on low-risk patients without implanted devices.
It should be made clear that routine use of MRI would have naturally been indicated if the patients did not have a device; thus, our approach provided contemporary insight into the value to an otherwise currently disadvantaged population. In our study, we did not eliminate any patients based on device characteristics other than restricting devices that were implanted before 2004; that is, we included patients with PMs and ICDs from all commercial manufacturers, including PM-dependent patients, patients with a history of ventricular tachycardia, and those with retained leads. Thus, our study represented a real-world patient population, and one that was avoided in studies and registries to date. Not only did we extend the findings of others, but for the first time, we documented our cardinal finding, namely, the value of performing MRI in these high-risk patients. Unstated but understood were the obligatory efforts to ensure safety as manifested in our zero-complication rate to patient or device.
Our seminal finding was that, in this patient population, MRI retained its high diagnostic yield; the risk-to-benefit ratio clearly justified the performance of the study. With current Food and Drug Administration−approved MR-conditional PMs (Biotronik, Berlin, Germany; Medtronic, Minneapolis, Minnesota; Abbott [previously St. Jude’s], Abbott Park, Illinois; and Boston Scientific, Marlborough, Massachusetts) and now approval for ICDs (Biotronik and Medtronic), assuming parallel cost-considerations naturally follow, the apprehensions about MRI scanning will inevitably decline; however, a generation of non-MRI conditional devices have been implanted while the aging population exerts additional pressure to perform MRIs.
This was a small study with 136 patients performed in a single institution, which represented nearly 200 MRI scans with only 2 MRI and/or cardiology-trained physicians who evaluated the clinical indications before authorizing the scan. However, it was the first and only study, to our knowledge, that placed MRI efficacy in this population as the primary focus. We did not control for other imaging studies that were performed before the MRI. Also, the finding of a confirmation of a tentative pre-MRI scan diagnosis, which did not add tangible value with regard to altering either diagnosis or therapy, we propose that it did improve diagnostic certainty, and thus indelibly improved downstream confidences of initiated therapies. However, the inclusion of this metric (representing ≈18% of the population) might not be universally accepted. Most of our studies were for neurological, cardiac, and orthopedic indications. The diagnostic yield for other organ systems was not evaluated. However, our patient population and their clinical indications represented a real-world scenario in which the safety and clinical usefulness of MR is often debated. As such, the diagnostic yield of MRI in patients in the present study was not compared with the yield of alternative modalities in a control group. In <3% of patients, the scan was performed in the presence of conventional contraindications (<6 weeks post-implantation, epicardial leads, abandoned leads). Similarly, patients with ICDs who were PM-dependent were included in our study; these patients were traditionally excluded from all studies, including the MagnaSafe registry. Although no problems were encountered, this was a small number of patients and no valid conclusions can be drawn. Finally, it was obvious this study could not be performed in a randomized controlled manner, yet it represents a real-world scenario in which patients with devices are rarely referred for MRI, even when other imaging modalities have been inconclusive; predominantly due to safety concerns. Thus, our goal was to determine the diagnostic usefulness of MRI in this real-world population in whom it had never been studied before.
The presumption and attainment of safety as shown by other investigators and reinforced by this observation was not a trivial encumbrance. This was a clinical investigation approved by our institutional review board, and certain precautions well beyond standard clinical MRI, both described and not described, were undertaken. The reader is reminded that other than the PMs that were approved by the Food and Drug Administration for MRI compatibility, the undertaking of imaging in patients with such devices is not suggested without suitable protocol, training, and expertise, and that catastrophic complications, including death, have occasionally been reported.
Our study added valuable new data that demonstrated a high diagnostic yield of MRI in patients with implanted cardiac devices, allowing the risk-to-benefit ratio to be better evaluated in the future. This study complimented the widely accepted robust safety data to justify MRI in this high-risk population. It further added to the notion that under careful protocol-driven imaging, performance of life-changing and often life-saving MRI can be performed in settings in which they are typically forbidden without loss of life, patient complications, or damage to the implantable devices. Extrapolation to the general public conservatively predicts innumerable daily changes in patient management and subsequent appropriate interventions worldwide, all while avoiding needless misdiagnosis and its ramifications. Until such time these devices become universally MRI compatible, there will be a great public need for centers with expertise and ability to perform MRI, which this study uniquely reinforced.
COMPETENCY IN MEDICAL KNOWLEDGE: Under the notion that multiple studies have demonstrated the safety for performance of PMs and/or ICDs in the MRI environment, an additive value and frequent life-saving or life-changing diagnoses are demonstrated when a policy of early consideration for MRI is endorsed.
TRANSLATIONAL OUTLOOK: Consideration for performance of MRI in patients with PMs and/or ICDs should be strongly considered when it is expected that a change in patient management or life-changing care might be anticipated. If such a policy requires transfer of a patient to a center with such expertise, it is expected that patient care will benefit. Because of the capabilities to ensure patient safety have sufficiently matured over the last decade, it is anticipated that this strategy will prove to be efficacious, resulting in improved resource use, and most importantly, substantial saving of lives.
The authors have reported that they have no relationships relevant to the contents of this paper to disclose. Presented in part at SCMR, 2012, 2013, and 2014; AHA 2014, 2015, and 2016; and ACC 2016.
All authors attest they are in compliance with human studies committees and animal welfare regulations of the authors’ institutions and Food and Drug Administration guidelines, including patient consent where appropriate. For more information, visit the JACC: Clinical Electrophysiology author instructions page.
- Abbreviations and Acronyms
- cardiac magnetic resonance
- implantable cardioverter-defibrillator
- magnetic resonance imaging
- specific absorption rate
- Received December 5, 2016.
- Revision received March 7, 2017.
- Accepted March 13, 2017.
- Taruya A.,
- Tanaka A.,
- Nishiguchi T.,
- et al.
- Faris O.P.,
- Shein M.
- Nazarian S.,
- Roguin A.,
- Zviman M.M.,
- et al.
- Raphael C.E.,
- Alpendurada F.,
- Prasad S.K.,
- Pennell D.J.,
- Mohiaddin R.
- ↵ACR appropriateness criteria. Available at: https://www.acr.org/Quality-Safety/Appropriateness-Criteria. Accessed June 2017.
- Friedrich M.G.,
- Sechtem U.,
- Schulz-Menger J.,
- et al.
- Syed I.S.,
- Glockner J.F.,
- Feng D.,
- et al.
- Vogelsberg H.,
- Mahrholdt H.,
- Deluigi C.C.,
- et al.
- Smedema J.P.,
- Snoep G.,
- van Kroonenburgh M.P.,
- et al.
- Levine G.N.,
- Gomes A.S.,
- Arai A.E.,
- et al.
- Sommer T.,
- Luechinger R.,
- Barkhausen J.,
- et al.