Contactless three-dimensional electro anatomical imaging based on magnetocardiography is reliable in unshielded hospital environments: a retrospective study of 635 patients

Magnetocardiographic mapping (MCG) is a contactless method for body surface recording of cardiac magnetic fields (CMF) generated by currents flowing within myocardial fibers. MCG is increasingly reported as a sensitive, radiation-free, method for different kinds of clinical applications, including emergency rule-out of acute ischemic heart disease, non-invasive three-dimensional cardiac electro-anatomical imaging (3D-EAI), and the prenatal diagnosis of fetal arrhythmias, wellbeing, neural maturation, and autonomic function. So far, MCG has been mostly performed with cryogenic instrumentations or optical magnetometers (OPM) needing electromagnetic shielding (EMS); however, there is a growing interest for recently developed novel, budget-priced, MCG-systems working in unshielded hospital environments. Since clinical experience with unshielded MCG is still relatively limited, this single-center study aimed at retrospectively quantifying the feasibility, sensitivity, and reproducibility of ambulatory MCG recorded in an unshielded laboratory fully equipped for interventional electrophysiology (UEFL). Method: MCG data of 635 patients were retrospectively analyzed. MCG was typically recorded in the supine position (sampling 1 KHz in the bandwidth DC-200) with the CMI-2409 9-channel DC-SQUID coupled to second-order axial gradiometers, pick-up coil 19 mm and 55-70 mm baselines), or with its custom 36-channel equivalent (A3619a). Proprietary software provided automatic analysis of CMF dynamics and source localization, after solution of the "inverse problem" with the effective magnetic dipole (EMD) model and pseudo-current reconstruction. The reproducibility of repeated measurements was estimated by calculating the intraclass correlation coefficient (ICC) with values in the range of 0-0.2 indicating slight, 0.2-0.4 fair, 0.4-0.6 moderate, 0.6-0.8 substantial, and 0.8-1.0 almost perfect agreement. Results: In our UEFL, peak-to-peak background noise was 7-30 picotesla (pT) (raw signals) and 1-2 pT, after adaptive filtering of 50 Hz. After signal averaging the sensitivity was 20-40 fT/√Hz above 1 Hz. Inverse solution of CMF and analysis of the EMD dynamics have have been proven reliable and reproducible (ICC: >0.7) in localizing cardiac sources with and average 3D uncertainty better than 10 mm, and to detect abnormal atrial and ventricular electrophysiological events (EPE). By merging MCG results within a 3D model of cardiac anatomy, reconstructed from orthogonal fluoroscopic images, or from 3D rendering of cardiac MRI, provided accurate pre-interventional localization of focal arrhythmogenic substrates, useful to guide catheter ablation. Conclusions: Unshielded SQUID-based MCG provides S/NR, high signal reproducibility, and 3D source localization accuracy adequate for clinical use in UEFLs. The availability of new OPM technology, stably operational without EMS, is foreseen to favor the development of novel contactless MCG-based 3D-EAI compatible with UEFL.