Introduction

The Beat Sensor Cardiac triggering technology, which is the application of Pilot Tone to cardiac triggering, is a relatively new technique that allows for electrode-free triggering of cardiac MRI sequences (1 - 3). In this work we evaluated this technique at five MRI centers and at two field strengths in both patients and in healthy volunteers. We were interested in assessing the robustness of the method in different subject types, field strengths, and sequence flavors, as well as the stability of the technique with respect to workflow and usability.

Methods

A prototype MR sequence and signal processing package was developed that allows for Beat Sensor Cardiac triggering of a wide range of MR sequence types typically used in standard cardiac MRI examinations. The pilot tone signal transmission and detection were performed using the commercially available BioMatrix Beat Sensor body array coil (Siemens Healthcare GmbH, Erlangen, Germany). The coil was positioned over the thorax, centered on the estimated heart position. The prototype software package 1) trains and calibrates the pilot tone signal, combining signals from all active local coil arrays and extracting the cardiac motion component, as well as correcting for signal corrupted e.g., during an RF pulse and 2) calculates a final filtered signal that is the inverted temporal derivative of the cardiac motion component. The trigger time point corresponds to the time of early systolic contraction and occurs ca. 200 ms after the R-wave. To facilitate the acquisition of static images of the heart in the end-diastolic phase of the cardiac cycle, the timepoint of the R-wave was estimated for each individual case from the cardiac motion component and used to adjust the timing parameters of the sequence protocol.

The prototype was tested in a total of 55 subjects, of which 20 were patients, at five different sites and at two different field strengths (1.5 T, 3 T) running the software version syngo XA31A. Relevant for this evaluation were the subject sizes (height and weight) as well as the range of heart rates and heart rate variations, the presence of devices in the scanner room that could potentially interfere with the pilot tone signal and the number of different operators. Of the patients examined, indications included sarcoidosis, hypertrophic cardiomyopathy, myocarditis, chest pain, atrial fibrillation, among others. For reference purposes an ECG or pulse triggering device was attached and additional comparison data were acquired with ECG triggering in 30 subjects. Patients were positioned supine, head first in the scanner. All scanning facts are summarized in Figure 1.

Results

In all 55 subjects a wide range of standard clinical cardiac MR sequences were successfully triggered with the novel Beat Sensor Cardiac trigger technique. The exact type of sequences tested varied from examination to examination and site to site. However, in all cases at least localization - with single shot TrueFISP and dark-blood HASTE - and/or cine or late-enhancement sequences (TrueFISP) were performed and were accompanied by additional sequences depending on the clinical query or test protocol. In total, the following sequences were acquired successfully:

• 47 cases cine
• 46 cases T1 mapping
• 27 cases T2 mapping
• 45 cases late-enhancement in free-breathing or breath-hold
• 21 cases dynamic sequences, of which 7 with contrast agent and one with adenosine
• 22 cases flow quantification in breath-hold or free-breathing
• 22 cases dark-blood prepared turbo spin echo both with and without STIR preparation.

At three sites in-room devices including monitors and contrast agent power injectors were switched on and the Beat Sensor signal monitored for disturbances: no noteworthy signal changes were found. On 8 of 55 occasions the pilot tone signal calibration and learning phase had to be repeated. This was due to incorrect operator timing of the two-step signal training and calibration process and associated with the prototype nature of these adjustment steps. In all cases, once the training and calibration were successfully completed, no further signal training and calibration was necessary, despite patient table movement for contrast administration. Thus, even for longer examinations (> 60 minutes) a stable signal was maintained. On 5 occasions an inverted trigger signal was documented, which usually could be resolved by repeating the calibration, although in all cases this did not compromise the trigger quality but had to be considered when optimizing the timing of static image acquisitions. In several individuals with highly variable heart rates, including a patient with atrial fibrillation, Beat Sensor Cardiac triggering functioned well (Figure 2). All examinations were successfully performed in breath-hold on expiration, combined with free-breathing scans as appropriate. In a subset of 5 subjects, comparison of quantitative parameters derived from Beat Sensor Cardiac triggered image data e.g., EF, EDV, ESV, T1, T2 and T2* values were comparable to those derived from ECG triggered image data (4). See figure 3 for example images.

Conclusions

We were able to demonstrate in a five-center evaluation of the Beat Sensor Cardiac triggering device that pilot tone-based triggering of standard cardiac examinations is feasible in a range of different patients and subject types and at different field strengths. This technology offers an alternative means to trigger clinical cardiac examinations without the need to attach electrodes, and, unlike ECG triggering, is robust to gradient interference.

Acknowledgements

No acknowledgement found.