Performance of miniature scalar atomic magnetometers for magnetocardiography in an unshielded hospital laboratory for clinical electrophysiology

Abstract

Background: Magnetocardiography (MCG) is under clinical evaluation as a novel, sensitive, radiation-free method for first-level non-invasive diagnosis of ischemic heart disease (IHD) and for rule-out of acute coronary syndromes (ACS) in chest pain patients with still normal ECG and serum markers. So far, multichannel MCG has been mainly performed with cryogenic sensors (SQUIDs). Alternatively, to reduce costs and/or to bring MCG to the patient bedside, novel non-cryogenic technologies have been developed, based on small induction coils or optically pumped magnetometers (OPMs). The magnetic field (MF) resolution of OPMs is nowadays comparable to that of the SQUIDs. However, OPMs operate in zero-field mode and consequently need magnetic shielding. Thus, presently used OPMs technology is not suitable for unshielded MCG. On the contrary, Miniature Scalar Atomic Magnetometers (MFAM™) operating in the total-field mode, with a sensitivity better than 2pT/√Hz to about 400Hz, were reported to detect the cardiac MF in an unshielded environment. This study aimed to investigate whether the MFAM, originally developed for geophysics and magnetic anomaly detection, could be suitable for MCG in an unshielded hospital laboratory for clinical electrophysiology (UHEPL). Method: Recordings were performed in the north-south direction, with sensors parallel to sources and tilted 30°. A dual-sensor MFAM module was used: one sensor, placed as close as possible to the source of interest, to detect its MF, and the other, more distant, to detect and subtract in real-time the background electromagnetic noise. The magnetic signal was acquired at 1-kHz sampling frequency (bandwidth DC-250 Hz). Signal averaging (SA) and digital filtering of RF noise (notch 50 Hz, LP 100 Hz) were used to improve the signal to noise ratio (S/Nr). First, the recordings were performed with a simplified phantom, filled with saline, using our amagnetic catheter to simulate current dipoles (CDs) of different geometry. Then MCG was attempted in 15 heathy volunteers (HV) (age 22–73 years) and compared with their SQUID MCG (Figure 1C, D). Results: After SA of 200 beats and filtering average peak-to-peak noise was 2–4 pT. S/N ratio (≥10 in the phantom (with 1 mA CDs) was twice lower in MFAM MCG. However ventricular MCG was feasible in all HV. On the contrary atrial MCG was adequate in one case only. Figure 1 summarizes the performance of the MFAM in UHEPL: (A) typical MFAM noise spectra. (B) 36-point SQUID MCG at the R wave peak, (C) MFAM-MCG signal taken at the fifth right parasternal intercostal space, D) SQUID MCG at same position for comparison. Conclusions: MFAM sensors were sufficiently stable to record MCG of several minutes in our UHEPL. The present sensitivity of MFAM is adequate for clinical MCG of ventricular de/repolarization (i.e. for detection/rule out of IHD/ACS), but not of atrial electrophysiology. A better performance is foreseen with development of more efficient gradiometer technology. Figure 1. Comparison between MFAM and SQUID Funding Acknowledgement: Type of funding source: None