Diagnostic quality and clinical utility of cardiovascular magnetic resonance in patients with subcutaneous ICDs

Funding Acknowledgements Type of funding sources: None. Background Cardiovascular magnetic resonance (CMR) is increasingly requested in patients with cardiac implantable electronic devices (CIEDs) – particularly for scar identification in patients with ventricular arrhythmias and implantable cardiac defibrillators (ICDs). Image artefact can be challenging in patients with larger devices, and patients with subcutaneous ICDs (S-ICDs) may not be referred due to concerns about non-diagnostic imaging. Modified sequences including wide band MOCO with PSIR (WB-MOCO) have been used to reduce image artefact in patients with CIEDs. We evaluated the image quality and clinical impact of gradient echo cine imaging and LGE CMR imaging in patients with S-ICDs. Methods Twenty CMR scans were performed adhering to manufacturer conditions in 19 patients (mean age 42± 13 years, 68% male) with MR conditional S-ICDs over four years.All patients were scanned at 1.5T at a single UK tertiary centre, with sequence modification using a pre-defined sequence protocol including gradient echo cine imaging, and WB-MOCO where conventional LGE showed artefact. LGE images were reviewed and graded for artefact by two independent observers, both by AHA segment and for overall quality on a scale of 0–4 (Grade 0 = no artefact, Grade 2 = 2–3 segments obscured by artefact, Grade 4 = entire myocardium obscured). Segments were viewed in both long and short axis orientations and deemed diagnostic if visible in at least one orientation. Clinical utility was assessed using pre-defined criteria, based on review of electronic medical records. Ethical approval was obtained from the Health Research Authority(REC21/EE/0037). Results CMR scans were performed in all patients without complication, excepting loss of ICD beeper function in 3 (16%) of patients (manufacturer recognised complication). S-ICDs were implanted for secondary prevention in 50% of cases with 63% patients having non-ischaemic cardiomyopathies. Scan indications included assessment of aetiology of heart failure (25%), ischaemic heart disease (20%), suspected left ventricular thrombus (20%), myocarditis (20%), scar assessment for viability (5%) miscellaneous (10%). Volumetric analysis from cine imaging was possible in 80%. LGE was present in 19 (95%) scans. Conventional LGE was only successful in 10% (2; 1 grade 0, 1 grade 1). Where WB-MOCO was used (n=18) this improved to 83% (grade 0 in 72%, grade 1 in 11%) with 95%being diagnostic (just 5% grade 3, 0% grade 4). Artefact was most commonly seen in the mid inferolateral segment (33% scans). Where follow-up data was available (19/20 scans), CMR led to a new diagnosis in 2 (10%) of scans and changes in management in 10 (53%) of scans. Conclusion Scar imaging using CMR can be safely performed in patients with S-ICDs, with diagnostic LGE imaging feasible in the majority (95% scans grades 0–2) and diagnostic cine imaging (in 80%) with the results providing high clinical yield.