4 Evaluation of accelerated cardiac magnetic resonance sequences for diagnosis of right heart pathologies

Introduction 2010 Task Force Criteria (TFC) for the diagnosis of Arrhythmogenic Right Ventricular Cardiomyopathy (ARVC) include a reduced right ventricular ejection fraction (RVEF) assessed using cardiac magnetic resonance (CMR).¹ ² Conventional CMR sequences to assess TFC require extended or repeated breath holds of 12–20 seconds which patients may find difficult to tolerate. Accelerated compressed sensing (CS) sequences reduce breath hold times but require a compromise in image quality. We aim to establish if CS impacts the quantitative diagnostic parameters derived from CMR images. Materials and Methods Short-axis stack cardiac cine images were acquired from 26 patients (16 male, 10 female, mean age 51, min. age 16, max. age 76) using 2-shot CS sequences in addition to standard TruFISP sequences on a 1.5T Siemens Magnetom Sola scanner. Measurements were made by two radiologists of the RVEF using Circle CVI software. Measured RVEF was compared between conventional and CS 2-shot sequences using a paired t-test (p=0.05). Radiologist agreement was assessed using Bland-Altman analysis. Results Ejection fraction measurement: A statistically significant difference was found in measurements of EF with the mean difference being a 3.1% underestimate of EF when measured on CS 2-shot for Radiologist 1 and a 1.7% underestimate for Radiologist 2. Radiologist agreement: Radiologist 1 typically measured the RVEF to be lower than Radiologist 2 by ~5% for both sequence types. Scan length: CS scan times were 30 seconds, compared to 2 minutes 15 seconds for TruFISP. Discussion The mean difference in EF between the radiologists was greater in magnitude than the mean difference in EF between the sequences. Therefore the difference between the sequences is not clinically significant. Despite one patient being a gene carrier for ARVC and two patients being found to have an EF <40%, no patients included in this study were diagnosed with ARVC. While we have tested CS for measuring RV volume and ejection fraction, this has not yet been applied to any symptomatic ARVC patients. Conclusion CS sequences significantly reduce the total breath hold time without degrading the diagnostic information that can be derived from clinical images. CS is therefore a feasible replacement for TruFISP in right heart imaging. Acknowledgements Thank you to Seb Atkinson for acting as second radiologist for the right ventricle contouring; to Sarah Branfoot, Andy Sawyer and other radiographers at the Freeman Hospital for carrying out the additional sequences, and to Mark Rickaby and Fiona Rose for their assistance with data anonymisation. References Marcus FI, et al. Circulation. 2010. doi:10.1161/circulationaha.108.840827. Bosman LP, et al. EP Europace. 2020. doi:10.1093/europace/euaa039.