Coronary artery disease (CAD) remains the leading cause of death worldwide and is associated with substantial individual, economic, and societal burdens. Only a minority of patients present mild symptoms in stable coronary artery disease (SCAD) (1). To lower the coronary deaths in asymptomatic adults, numerous measurements of risk factors and markers, as well as stress tests, are employed as screening investigations.

Fractional flow reserve (FFR) assesses physiological stenosis severity. However, coronary angiography in low-risk patients imposes unnecessary burden and risk, while, FFR, which requires specialized pressure guidewires, is both technically complex and costly. Routine electrocardiogram and echocardiography demonstrate relatively high diagnostic performance for acute myocardial infarction, but are insufficient for SCAD. Although the exercise treadmill testing, stress echocardiography, cardiac magnetic resonance imaging (CMR) and single-photon emission computed tomography (SPECT) can also evaluate myocardial ischemia, they cannot assess coronary artery stenosis, moreover, these methods present several limitations, including complex procedures, high technical demands, substantial costs, and potential radiation exposure (2, 3). Therefore, a non-invasive, safe and simple method is required to facilitate the early diagnosis of stable coronary artery disease.

Computed tomography angiography-derived fractional flow reserve (CTFFR) is a novel noninvasive approach for precisely localizing ischemia-causing coronary stenosis. It utilizes computational fluid dynamics (CFD) to calculate “3-vessel” FFR from standard coronary computed tomography angiography (CCTA) images without requiring additional imaging or vasodilators (4, 5). Furthermore, the PLATFORM study demonstrates that CTFFR enhances diagnostic certainty by reducing invasive coronary angiography cancellations by 61% and significantly lowering the rate of detecting non-obstructive CAD (6).

Magnetocardiography (MCG) measures changes in the heart's electromagnetic field, which are altered in the diseased hearts, likely due to modifications in the trajectory or number of moving ions or electrons. An objective scoring system facilitates interpretation and comparison of results. Given the absence of radiation, MCG is ideal for screening and frequent monitoring of heart conditions, especially in CAD (7, 8). MCG detects minute electrophysiological changes that lead to abnormal currents, especially occurring in early ischemia, enabling early CAD diagnosis (9).

Nowadays, MCG has shown certain potential in the diagnosis of CAD (10), arrhythmia and inflammatory cardiomyopathy (11), and has become a research hotspot worldwide. However, most of the relevant studies are exploratory studies, with a small sample size and lacking in-depth research. Also, the performance of MCG in patients with relatively low risk of myocardial ischemia during routine check-ups is still inconclusive. This study aims to use the latest MCG equipment to assess myocardial ischemic function in patients with stable coronary artery disease.

To the best of our knowledge, this was the first study to evaluate the diagnostic performance of the MCG against CTFFR. The main finding of our analysis was that, compared to CTFFR, MCG showed high sensitive and specific for predicting myocardial ischemia in patients with cardiovascular risk factors or suspected symptoms who had 30%–90% diameter stenosis on CCTA.

MCG is a non-contact, non-invasive, non-radiation method to detect myocardial ischemia by recording magnetic fields generated by cardiac electrical activity (17). Previous studies confirmed that MCG demonstrates high sensitivity and specificity, compared to other diagnostic modalities, such as electrocardiography (ECG), SPECT, and FFR (18–20). An evidence gap exists regarding the diagnostic value of CAD between MCG and CTFFR. Additionally, nearly all the studies enrolled high-risk patients with CTA- or angiographically-confirmed CAD, or classic with angina pectoris, while evidence for MCG in low-risk patients remains limited (21, 22). Lastly, when interpreting previous results, small sample size limitations cannot be overlooked (23). Thus, conducting this study to evaluate MCG's diagnostic performance against CTFFR in patients with 30%–90% diameter stenosis on CCTA was meaningful.

In the current study, we achieved a sensitivity of 69.6%, specificity of 87.9%, PPV of 72.7%, NPV of 86.2%, and accuracy of 82. 1%. Our finding was different from many previous studies considering MCG as a highly sensitive but less specific method (24–26). First, low-risk suspected CAD patients comprised most enrolled participants. Patients in our study had low rates of typical symptoms, hypertension, hyperlipidemia, and diabetes. However, when reviewing previous studies, the high risks mentioned above were considerably prevalent (24–26). Second, the coronary artery stenosis was less severe than other studies (26, 27). There were 132 patients with coronary stenosis ≥50% but only 92 patients with CTFFR ≤0.8. Concurrently, the scatter plot of CTFFR values and MCG predictions showed us most CTFFR values >0.60. Thus, our cohort represented an unselected population, better reflecting real-world research. Based on these results, we believe that MCG could be a promising screening tool for individuals with low risk of CAD.

In addition to our study, other studies have also confirmed that MCG exhibits higher specificity and lower sensitivity. Van Leeuwen et al. found that MCG had a sensitivity of 85% in CAD patients with prior myocardial infarction vs. 68% in those without, and a specificity of 90% (21). Fenici et al. reported that two machine-learning techniques combined to identify abnormal MCG ventricular repolarization (VR) in 25 IHD patients and exclude VR abnormalities in 28 controls, yielding 75% sensitivity and 85% specificity (28). The results of these studies are consistent with our findings. In clinical practice, high specificity can guide decisions for further angiography, reducing the frequency of invasive procedures.

According to the subgroup result of CTFFR and MCG, for patients whose CTFFR was >0.8 and MCG was positive, the discrepancy may be explained by the incomplete correspondence between hemodynamic alterations in major coronary branches and the state of myocardial injury caused by ischemia. In SCAD, hemodynamic alterations may exhibit a temporal delay in inducing myocardial damage and electrophysiological alterations. During this subclinical phase, CTFFR may detect hemodynamic abnormalities while MCG remains negative due to the absence of established tissue injury. Furthermore, hemodynamic abnormalities in major coronary vessels are not the sole cause of myocardial ischemia. Coronary microvascular dysfunction (CMVD)—a structural and functional disorder affecting small arteries, arterioles, and capillaries—frequently induces myocardial ischemia despite normal hemodynamics in major coronary branches (29, 30), in the study by Quesada et al., MCG demonstrated the capability to detect CMVD through a 90-second non-invasive scan, eliminating the need for intravenous vasodilators or ionizing radiation (31). In such scenarios, CTFFR may yield negative results, whereas MCG could demonstrate positive findings.

In the present study, an approximately 21.3% rate of mismatch between CCTA and CTFFR was found. Among these mismatched data, MCG showed a positivity rate of 54.5% (6/11) with positive mismatch (CTFFR ≤0.8, CCTA <50%), while demonstrating a positivity rate of 5.9% (3/51) with negative mismatch (FFR>0.8, CCTA ≥50%), which indicated a high correlation between MCG and hemodynamically significant lesions. This is like the previous work by Park that MCG could accurately detect CAD vs. invasively determined FFR with a sensitivity of 86.7% and specificity of 73.9% (20). Considering the increasing need for early detection of ischemia in people at high risk of cardiovascular disease, MCG has been proposed as a noninvasive and contactless technique for functional diagnosis of the heart (32, 33). ECG relies on horizontal body surface potentials, making it vulnerable to signal cancellation that masks tangential/vortex currents. MCG directly measures perpendicular magnetic fields from cardiac activity. Tangential currents' magnetic fields are easily captured by sensors. Though electrically undetectable due to cancellation, vortex currents still generate measurable percardial magnetic fields via the right-hand rule (34, 35), MCG has the potential to benefit the assessment of suspected patients without CAD. Based on the application of artificial intelligence in the evaluation of disease, techniques like machine learning methods were implemented to automate diagnosis (36). In the previous study, back-propagation neural network (BNN) and direct kernel self-organizing map (DK-SOM) were applied to explore the IHD pattern recorded by MCG and exhibited accuracy of 74.5% and 80.4% respectively (37).

In terms of convenience and resource utilization, the implementation of an MCG pathway could significantly reduce patients' length of stay, and it just need 10 min to finish the assessment of the myocardial ischemia. Besides, compared to other measures, MCG yielded significantly higher patients' satisfaction (38, 39). Furthermore, given the substantial number of patients presenting with chest pain in outpatient and emergency departments, magnetocardiography (MCG) holds significant potential as a non-invasive diagnostic tool. When used complementarily to established modalities such as CCTA and SPECT (40, 41), MCG may effectively reduce both misdiagnosis rates and unnecessary exposure to surgical or other invasive procedures (42).

Several limitations should be acknowledged. Firstly, it was a single-center study conducted in China, the results might not be applied to whole population. Secondly, it is noted that the vast majority of patients in the current study were male. Though such a sex bias could not be ignored, the principle of MCG was to record magnetic fields generated by the electrical activity and there was no evidence revealing a difference in sex subgroup in the field of MCG. Thus, we recognized that sex would not change the final results. Thirdly, we were lack of the results of coronary angiography and FFR regarded as the golden standard of functional evaluation. Lastly, there are only 35 patients whose stenosis ≥70 confirmed by CCTA, the small sample of this subgroup might contribute to the low sensitivity.

Compared to CTFFR, MCG yielded a good specificity and acceptable sensitivity for the detection of CAD in patients whose diameter stenosis CTA was confirmed from 30% to 90%. It may provide an alternative to functional evaluation before other invasive or radiation exposure methods. Further studies are warranted to determine the role of MCG in this context.