Abstract 4140246: AI and Quantum Sensors: Realization of a Safe and Effective Unshielded Bedside Magnetocardiogram to Detect Ischemia in the Emergency Room

Introduction: Magnetocardiography (MCG) has potential to significantly improve cardiac disease diagnosis, but adoption in clinical practice is hindered by high cost and large footprints of traditional systems. Hence, we developed a bedside, non-contact MCG system that utilizes quantum magnetometers and AI methods to extract validated cardiac biomarkers, while rejecting ambient magnetic interference. Goal: To investigate the safety and efficacy of the bedside MCG device in identifying non-ST-segment elevation myocardial infarction (N-STEMI) among patients from an urban emergency department. Methods: The MCG system is a custom prototype with 26 optically pumped magnetometers - quantum sensors that measure cardiac magnetic fields with picoTesla sensitivity. We recruited a diverse sample of 78 adult patients (72 with analyzable data) presenting to the ED with chest pain symptoms suggestive of ACS who were not experiencing STEMI (mean age=56 years; 64% male; 11% Asian, 28% Black, 7% American Indian/Alaska Native, 44% White, 1% Native Hawaiian/Pacific Islander, 9% other/unknown; 6 technically unusable recordings, 8%). A 5-minute MCG was recorded within 2 hours of an initial 12-lead ECG. Signal processing and AI algorithms were used to remove ambient interference and recover cardiac signal. Literature-based MCG markers were inputted to a classification and regression tree (CART) model to classify patients as NSTEMI ACS rule-in or rule-out. Results: The CART model successfully discriminated clinically determined ACS rule-in patients (n=7) from rule-outs (n=65) with sensitivity=86% and specificity=92%. The most predictive MCG markers were the field pattern angle difference between R- and T-wave maximum, the power ratio between the R- and T-waves, and the field maximum to minimum ratio. No adverse events were reported. Conclusions: MCG is a promising, non-contact diagnostic tool for emergency physicians diagnosing suspected ACS cases. This pilot study shows that a bedside, quantum-sensing MCG that utilizes AI has the potential to provide safe and accurate diagnostic results, and it is feasible to integrate the device into the ED workflow.