Ischemic heart disease (IHD) remains a significant global public health issue, and its prevalence has been increasing over the years. According to the 2023 report from the National Center for Biotechnology Information (NCBI), IHD is responsible for 17.8 million deaths annually, positioning it as the third most common cause of mortality worldwide (1). However, identifying myocardial ischemia in patients with suspected coronary artery disease (CAD) remains a challenging aspect of routine cardiological diagnostics with its diverse manifestation and the complexities involved in distinguishing non-IHD. Functional or stress testing, which aims to detect inducible myocardial ischemia, has traditionally been considered the “gold standard” and is the most commonly used as a non-invasive method for diagnosing CAD (2). However, a non-invasive evaluation is performed on less than half of the patients before percutaneous coronary intervention (PCI) (3, 4). This is primarily due to limitations in testing, which include low diagnostic accuracy and the potential radiation risks associated with coronary computed tomography (CT) or single-photon emission computed tomography (SPECT) (5).

Magnetocardiography (MCG) is a non-contact, non-invasive, radiation and contrast-free method that enables the recording of magnetic fields generated by the electrical activity of the heart (6–9). Although electrocardiography (EKG) and MCG provide information about the same electrical activities of the heart, MCG presents several advantages. Cardiac magnetic fields remain unaffected by variations in the conductivity of body tissues or fluids, without attenuation or distortion (10). Additionally, its high sensitivity and non-invasive, contactless procedure make it a valuable tool for early diagnosis of myocardial ischemia that may otherwise go undetected by EKG (11). Several clinical studies have already demonstrated the superior sensitivity of MCG compared to EKG in detecting ischemic myocardium both at rest and during stress (11–17). The remarkable ability of MCG to identify patients with CAD has been widely recognized (5, 17–20). Various MCG investigations have employed a variety of devices, including cryogenic superconducting quantum interference devices (SQUIDs) (21, 22). These devices have primarily been utilized in magnetically shielded rooms (MSR) to eliminate background environmental noise, for instance, noise emanating from nearby instruments. However, they can also yield reliable outcomes in unshielded environments by incorporating a second (or higher order) gradiometer configuration of the pick-up coils and/or utilizing real-time electronic noise subtraction (10). Recently, advancements have been made in non-cryogenic MCG devices, offering alternative options (23). Furthermore, a variety of quantitative methods and computer algorithms have been devised to facilitate the interpretation of diverse magnetic field patterns (24–27).

This review will provide an overview of the evidence supporting the utility of MCG, a valuable tool for diagnosing myocardial ischemia that is currently available, and discuss the potential impact of these findings on the future integration of MCG into clinical practice.

Previous studies have explored the application of MCG for the diagnosis or ruling out of stable CAD in Table 1. Other studies have investigated its use for the detection or ruling out of acute coronary syndrome (ACS) in Table 2. These studies have utilized a range of techniques to qualitatively and quantitatively analyze the magnetic field throughout the cardiac cycle. In most of the studies, the quantitative analysis has been focused on evaluating changes in the magnetic field during ventricular repolarization, typically occurring at the end of the ST segment (prior to the T wave) and/or the T wave. These methods encompass the analysis of various aspects, such as the extrema and dynamics of the magnetic field angle, as well as the dynamics of distance and ratio involving the minimum and maximum poles. These measurements are typically taken during the ascending T wave, specifically from one-third of the peak intensity (Tmax/3) to the peak intensity (Tmax) (6, 28–32, 42, 49). Additionally, other studies have also investigated different parameters related to the ST segment and T wave, particularly during or after exercise (13, 33). Due to the typically higher magnetic field and signal-to-noise (S/N) ratio during rest, many subsequent studies have focused on utilizing variations of parameters measured during the T wave, initially described by Park et al. (42). Additionally, other MCG parameters have been investigated during the QT and QRS intervals (23, 34–39), and there have been reports on the application of machine-learning approaches for interpreting MCG signals (24–26).

Numerous studies have provided evidence that MCG, whether conducted in a shielded or unshielded environment, at rest, or under conditions of exercise or pharmacologic stress, can effectively differentiate between patients with angiographically confirmed stable CAD and healthy individuals (13, 25, 26, 34, 36–38, 40, 41). Additionally, MCG has shown potential in distinguishing patients with chest pain but without evidence of CAD on angiography or other diagnostic tests (5, 9, 18, 20, 28, 35, 39). However, it is important to proceed with caution when interpreting these results, as many of the studies enrolled small populations and included highly selected patient cohorts with or without the disease, which may not fully represent the broader population encountered in clinical practice.

Several studies have subsequently examined the patterns of resting magnetic fields in individuals with CAD. These studies have evaluated different parameters of MCG and have endeavored to enhance diagnostic accuracy and minimize background noise by employing various analytical approaches and algorithms. The earlier study revealed significant differences in multiple MCG parameters such as ST slope, ST shift, T peak amplitude, ST-T integral, and magnetic field map (MFM) orientation between patients with CAD (n = 101) and a control group of healthy subjects (n = 59) (40). They yielded a specificity and sensitivity of 83% and 84% respectively [with an area under the curve (AUC) of 91.2% for the receiver operating curve (ROC)], and the accuracy of CAD classification at 84% remained consistent regardless of the number of affected vessels or the severity of stenosis. In addition, various quantitative methods have been employed to differentiate CAD. These methods include binary classification approaches utilizing threshold values for MCG indices (5, 28, 35, 36), integrated indices derived from MCG parameter values (20, 50–52), the assessment of the number of abnormal MCG parameters (31), spatial distribution analysis of the QT interval (34), and the utilization of automated machine learning algorithms (25–27). In a recent study, a combination of quantitative (change in ST-segment fluctuation score) and qualitative (non-dipole phenomenon) parameters was utilized to improve the diagnostic accuracy of shielded MCG in distinguishing patients with stable angina from asymptomatic individuals without CAD (18). The inclusion of the non-dipole phenomenon resulted in an increased AUC of the ROC curve, elevating it from 0.79 to 0.93.

Initial investigations on MCG in patients with CAD demonstrated its capacity to identify alterations in multiple MCG parameters during stress induced by exercise or drugs. The analysis indicated that ST segment MCG parameters exhibited greater sensitivity to exercise-induced ischemia in patients without a history of MI (n = 27), whereas T wave MCG parameters were most sensitive to changes in patients with prior MI (n = 17) (13). For the assessment of 42 patients with CAD following a dobutamine-stress test, an analytical approach centered on the epicardial current distribution at the point of maximum amplitude of the QRS complex (QRSmax) was employed (9). MCG demonstrated a sensitivity of over 90% for detecting CAD, irrespective of the location of stenosis or the number of affected vessels.

Several studies have directly compared the diagnostic efficacy of MCG with other tests. In a study by Park et al., MCG exhibited superior sensitivity compared to 12-lead EKG in detecting CAD using a conventional dobutamine stress protocol (9). Another study demonstrated higher sensitivity, along with comparable specificity, and similar positive predictive value (PPV) and negative predictive value (NPV) for MCG compared to EKG in the diagnosis of stable angina (41). In another study, MCG showed higher specificity and comparable sensitivity, PPV, and NPV when compared to single photon emission computed tomography (SPECT) for discriminating patients with angina (39).

In studies involving patients experiencing acute chest pain and suspected ACS, the analysis of MCG data, measured either at rest or after exercise, in shielded or unshielded environments, has revealed qualitative and quantitative distinctions that facilitate differentiation between patients with ACS and healthy individuals (15, 16, 23, 43, 44, 53, 54). Moreover, MCG has been successful in distinguishing patients without definitive evidence of ACS or CAD in diagnostic examinations (7, 8, 14, 15, 42, 44–46, 55, 56). A previous study utilizing a shielded, 64-channel MCG system showed the capability of 15 MCG parameters to discriminate between patients diagnosed with non-ST segment elevation myocardial infarction (NSTEMI) (n = 83) and age-matched individuals presenting with chest pain but without clinical indications of CAD (15). Among these parameters, the field map angle of the T wave peak exhibited the highest diagnostic accuracy, with a sensitivity of 86% and a specificity of 75%. In a prospective study involving 402 patients experiencing acute chest pain without ST-segment elevation in the EKG, it was observed that abnormalities in the MFM between the onset and peak of the T wave at admission were predictive of an elevated risk of mortality over a 3-year period. The relative risk for MCG abnormalities was 4.58, compared to 1.69 for EKG, and 2.58 for elevated troponin levels (8). Another study found that MCG has the potential to differentiate patients with ACS and bundle branch block, a condition that can complicate the diagnosis of ACS when using EKG (47, 48). MCG has also shown promise in discriminating patients with reduced left ventricular ejection fraction (37) and those with a history of previous MI (57). However, further studies with larger patient populations are necessary to explore the full potential of MCG in these particular conditions. Additionally, a direct comparison between MCG, utilizing either visual or automated analysis, and other diagnostic tests such as EKG, cardiac troponin I, and echocardiography, revealed that MCG showed higher sensitivity, comparable specificity, comparable positive predictive value (PPV), and higher negative predictive value (NPV) in distinguishing patients with CAD and acute chest pain from patients with chest pain but normal results on diagnostic tests (42).

Previous studies evaluated various MCG parameters to improve the detection of stable CAD or ACS in patients with different clinical presentations. MCG proved effective in identifying ischemia, even in patients with normal EKG and cardiac biomarker results. Initial evidence suggests acceptable sensitivity and specificity for detecting IHD in selected cohorts with stable CAD or ACS, with MCG outperforming EKG, echocardiography, and cardiac troponin assays. MCG could be a valuable initial test for suspected CAD or ACS, but more research is needed to determine the best parameters and validate its diagnostic performance across diverse patient populations. Further studies should focus on integrating MCG into clinical practice and assessing its incremental value in existing diagnostic pathways, potentially leading to the development of MCG criteria for early exclusion of non-ischemic or non-CAD patients, reducing unnecessary testing and hospital resource utilization. In addition, to address the challenges posed by the evolving nature of MCG technology and diagnostic criteria in CAD studies conducted over several decades, a meta-analysis of current data or the following approaches are needed. Although significant progress has been made in MCG device technology and machine-learning analysis techniques, further validation of potential diagnostic parameters is necessary, particularly in large patient cohorts that represent a diverse range of cases.

The use of MCG has the potential to benefit the assessment of patients with suspected ACS, particularly in the field of emergency medicine. Chest pain is a common reason for emergency department visits, but a significant portion of patients (60%–90%) do not have an acute cardiac cause for their pain. Current diagnosis of ACS in patients with acute undifferentiated chest pain involves a resting 12-lead EKG, multiple measurements of cardiac troponin levels over several hours, and clinical judgment. Integrating MCG into the diagnostic pathway could help reduce the time to diagnosis and the costs associated with serial troponin testing. Another challenge in emergency medicine is the risk of missed diagnoses of patients with NSTEMI or unstable angina, which can lead to adverse outcomes after discharge. MCG has the potential to decrease the likelihood of missed diagnoses and improve clinical outcomes. The benefits of early identification of patients with non-cardiac chest pain have been demonstrated through accelerated risk algorithms that incorporate high-sensitivity cardiac troponin assays, resulting in significant improvements in time to discharge, cardiac outcomes, and hospital resource utilization. Further evaluation through prospective observational studies involving unselected cohorts of patients presenting to the emergency department with acute chest pain will provide insights into whether MCG could be used prior to cardiac troponin testing to expedite patient assessment. Most of the original multichannel MCG devices have specific operational requirements and high running costs, primarily due to the need for external electromagnetic shielding (EMS) or liquid helium cooling. However, the recent development of portable MCG devices holds the potential for bedside assessment of patients with acute chest pain upon their initial presentation to the emergency department (23, 58). Enhancements in the practical aspects of MCG devices such as device footprint, ease of use, operator training requirements, and the need for a shielded operating environment will play a crucial role in determining their ease of implementation in clinical practice.

Finally, validation studies are necessary to determine the diagnostic accuracy of MCG parameters compared to current diagnostic pathways in undifferentiated patient populations. Validated MCG diagnostic criteria should be evaluated in well-defined cohorts including patients with stable CAD, ACS, inducible ischemia, and non-ischemic chest pain. Furthermore, there are indications in the literature that MCG may have broader clinical applications in CAD beyond diagnosis. For instance, its use in stress testing to detect functional ischemia could provide valuable prognostic information for risk stratification. Future clinical studies should explore other endpoints such as infarction location and severity, as well as the prediction of major adverse cardiac events and post-MI arrhythmias.

MCG presents a non-invasive and non-contact imaging modality that is free from emissions, offering potential improvements in the management of patients with CAD. It has demonstrated the ability to detect myocardial ischemia in patients with stable CAD and ACS. However, further clinical studies are necessary to evaluate the use of MCG in undifferentiated patient cohorts. It is also important to validate and standardize MCG analytical techniques and parameters. Prospective, multicenter observational studies are currently needed to investigate the effectiveness of MCG in ruling out ACS in emergency settings. These studies will help determine the utility of newer MCG devices and their potential integration into routine clinical practice as complementary diagnostic tools.