Aims:

(1)          First, to determine image quality using different magnetic resonance (MR) perfusion protocols in patients with all available device types in a real-world setting, including non-conditional devices.

(2)          Second, to demonstrate feasibility of high-quality perfusion imaging using spoiled gradient echo (sGE) protocols for non-invasive stress testing.

Methods:

From 08/2020 to 11/2021, N=168 patients with active cardiac implantable electronic devices (CIED) were scanned on a 1.5-T MR scanner (Philips Ingenia and Ambition). Patients were scheduled in consideration of current recommendations. Our MR scanning protocol was tailored to the clinical indication, and whenever myocardial perfusion imaging was possible (stress and/or rest), both a conventional steady-state-free-precession (SSFP)-based and a modified sGE-perfusion protocol were applied. Such a tailored perfusion protocol was performed in N=81 patients (72% men) with exclusively left-sided devices (pacemaker, n=31; implantable cardioverter-defibrillator (ICD), n=34; subcutaneous ICD (S-ICD), n=9 and cardiac resynchronization therapy device (CRT-D or -P), n=7) with a percentage of 10 % non-conditional devices. For assessment of image quality, a semi-quantitative 4-point grading scale was used based on a standard 16-segment model.

Results:

A total of N=26 stress tests with either regadenoson, adenosine or dobutamine and N=55 rest perfusion protocols were performed. Asynchronous pacing was required in approx. 30% of the patients due to a missing sufficient rhythm (heart rate < 40 bpm). Device interrogation before and after MR scanning showed no significant changes in battery status, pacing threshold, lead impedance and sensing. One patient suffered from a re-occurrence of atrial fibrillation when combining asynchronous pacing with regadenoson stress testing. Image quality in myocardial perfusion imaging was substantially better in sGE-based perfusion protocols compared to conventional SSFP-based perfusion in PM, ICD, CRT and s-ICD-Patients (p < 0.001), whereas there was no relevant subjective difference in patients with pacemakers since image quality was neither impaired in SSFP- nor in sGE-based protocols. Most device artefacts - precluding a meaningful assessment of the respective segments during perfusion imaging – were located primarily in the segments 13,1 and 7 in PM and in 7 and 8 in ICD-patients. A significant relationship between the extent of device artefacts PM-patients were found for weight, p = 0.04 and  BSA, p = 0.03 and in CRT-patients for BMI, p = 0.04.

Conclusion:

Myocardial (stress)-perfusion imaging by CMR is safe and feasible with high image-quality in patients with all kinds of CIEDs – including MR-conditional as well as non-conditional devices. When performing CMR-based myocardial perfusion imaging in patients with left-sided ICD/CRT/s-ICD, a sGE-based perfusion protocol should be preferred compared to conventional SSFP-based perfusion protocols in order to achieve artefact-free and best interpretable images.