Introduction

ECG is the most commonly used triggering method for cardiac MRI. However, the patient preparation required to apply and adjust ECG electrodes can be time consuming. At higher field strengths, the ECG signal is prone to distortion due to gradient interference and the magnetohydrodynamic effect. BSC triggering is a Pilot Tone based method that does not require extra patient preparation. It utilizes an external radiofrequency (RF) signal that is modulated by the patient’s physiological motion^1-3. In this study, we evaluated the performance of BSC vs. ECG triggering in a comprehensive cardiac MRI exam by comparing the quality of images and quantitative measurements.

Methods

Both volunteer and patient data were collected with prototype sequences on a clinical 3T system (MAGETOM Vida, Siemens Healthcare, Erlangen, Germany). A calibration procedure was performed once at the beginning of the exam to train and calibrate BSC triggering signals and correct for signal interference caused by RF pulses. A comprehensive, non-contrast cardiac exam including localizer, morphology, cine, 2D flow, and parametric mapping was acquired in 14 volunteers (38 ± 14.5 years, 6 females) with both BSC and ECG triggering. Biventricular cardiac function and flow in aorta and main pulmonary artery (MPA, collected in 9 volunteers) were evaluated in suiteHEART (NeoSoft, Pewaukee, WI, USA). Myocardial T1, T2, and T2* values were measured in the interventricular septum. Pairwise student t-tests were performed to compare quantitative measurements from images acquired with BSC vs. ECG triggering.

To evaluate the effectiveness of BSC triggering post contrast agent administration, BSC triggered pre- and post-contrast T1 maps and LGE were added to a standard ECG triggered clinical exam in 6 patients (27 ± 10.4 years, 4 females). Calibration was performed prior to contrast injection to evaluate the robustness of BSC triggering post-contrast. Scan parameters of BSC triggered sequences were set to match the ECG triggered clinical scans performed in the same patient. Inversion time (TI) in BSC triggered LGE scans were increased slightly relative to the ECG scans to account for the longer time after contrast injection. Pre- and post-contrast T1 maps were used to calculate extracellular volume (ECV). LGE image quality and sharpness were evaluated pairwise by a CMR expert (30 years of experience) blinded to the triggering method. Sequence types and corresponding imaging parameters for both volunteer and patient scans are shown in Table 1.

Results

BSC triggered exams were successfully performed in all healthy volunteers and patients. All data from one volunteer were discarded due to gross patient motion observed between scans, resulting in poor correspondence of slice positions between BSC and ECG triggered images. Example images from healthy volunteers using both triggering methods are shown in Figure 1. All LGE images acquired in patients were reviewed and no significant differences in image quality or sharpness were detected. Figure 2 shows the LGE and T1 maps from one patient, showing mild enhancement in the basal to mid anterior and inferior wall in the long axis view, and at the RV attachment site in the mid short axis view. A summary of quantitative results measured in both volunteers and patients is shown in Table 2. Due to the limited size of the patient cohort, statistical analysis was performed in only healthy volunteers.

Discussion

The feasibility of BSC triggering was evaluated by comparing with ECG triggering. The BSC utilizes an external RF signal to detect physiological motion. Depending on Pilot Tone transducer placement relative to the patient’s anatomy, the received signal characteristics can vary. Thus, a patient-specific training scan is required to calibrate the BSC trigger. Overall, nearly all comparisons demonstrated equivalent image quality and quantitative results. However, we did observe a systematic reduction in peak aortic velocity with BSC triggering. This could have been due to either a misalignment of image frames between ECG and BSC, or slight variations in BSC triggering between heartbeats. The triggering constancy relative to ECG needs further investigation. In Figure 2, signal to noise ratio in the blood pool appears slightly lower in the BSC triggered LGE and the post contrast T1 map. This could be due to the contrast agent washout from the blood pool because BSC triggered images were acquired later after contrast injection.

Conclusion

Beat Sensor Cardiac triggered MR exam was successfully evaluated and validated in healthy volunteers and patients. Its performance was compared with standard ECG triggering across multiple imaging applications. The overall image quality was equivalent for BSC and ECG. Quantitative measurements in function, flow, and parametric maps in healthy volunteers showed no significant difference in all measurements except peak aortic velocity.

Acknowledgements

Debbie Scandling, The Ohio State University, Columbus, OH, United States, for helping with volunteer recruitment