Table 1.
Studies of MCG in patients with stable CAD.
| Study | Diagnostic criteria of MCG | Indication/Test population (n)/Control (n) | Testing conditions | Specificity/Sensitivity (ROC AUC) | PPV/NPV (ROC AUC) | Reference |
|---|---|---|---|---|---|---|
| Park et al. (5) | Change in ST-segment fluctuation score between rest and stress with a cut-off of −39.0% Bulls-eye mapping of current between beginning of T wave and Tmax at rest vs. stress |
Anatomic CAD/Patients with suspected CAD with subsequent angiographically proven ≥50% stenosis of a vessel without acute MI in previous 3 months (42) and patients with angiographically proven non-obstructive CAD (5)/- | Shielded 64-channel Rest and exercise (bicycle ergometry test)/dobutamine stress |
74%/87% (0.84) (ST fluctuation score) 92%/91% (0.91) (mapping) |
– | Fractional flow reserve |
| Fenici et al. (6) | Angle (A), distance (D), and ratio (R) dynamics of the dipoles during the T wave interval and ST angle as prespecified criteria | Anatomic CAD Patients with IHD and angiographically proven >70% coronary stenosis and positive stress/SPECT (19) Healthy volunteers (20) |
Unshielded, 36-channel Rest |
20 Hz low pass filtering: 100%/32% (A) 90%/42% (D) 80%/42% (R) 70%/79% (STα) 50 Hz adaptive filtering: 100%/47% (A) 65%/74% (D) 50%/63% (R) 75%/79% (STα) |
20 Hz low pass filtering: 100%/60% (A) 80%/62% (D) 67%/59% (R) 71%/79% (STα) 50 Hz adaptive filtering: 100%/66% (A) 67%/72% (D) 55%/59% (R) 75%/79% (STα) |
EKG |
| Park et al. (9) | Reduction of epicardial current density and strength at QRSmax between rest and stress used as diagnostic for ischemia | Functional ischemia/ Patients with intermediate pre-test probability of CAD with subsequent angiographically proven ≥70% stenosis of a vessel (42) or with angiographically proven non-obstructive CAD (58)/- |
Shielded 55-channel Rest and pharmacologic (dobutamine) stress |
83%/98% | 80%/98% | EKG |
| Hänninen et al. (13) | ST slope increase and peak gradient orientation of the ST segment at cessation of stress, T-wave amplitude increase at two minutes recovery | Functional ischemia/ Patients with CAD with anginal pain, and a positive EKG stress test and either single-vessel disease (>50% luminal diameter stenosis in one of the main coronary arteries) with no history of MI (27) or triple-vessel disease (stenosis ≥70% luminal diameter) and ≥1 previous MI (17)/Healthy volunteers (26) |
Shielded 67-channel Exercise (supine bicycle ergometry test) |
– (0.83) (ST slope) (0.83) (ST peak gradient) (0.86) (T-wave increase) |
– | EKG |
| Shin et al. (18) | Quantitative and qualitative analysis of the change in ST-segment fluctuation score (–51% cut-off selected as best cut-off) and the non-dipole phenomenon during the interval from the beginning of the T wave to the Tmax |
Anatomic CAD and functional ischemia/ Patients with suspected CAD without acute MI in previous 3 months, with subsequent angiographically confirmed CAD (≥70% stenosis in ≥1 proximal epicardial coronary artery) and objective evidence of myocardial ischemia or ≥1 coronary stenosis of ≥80% and classic angina without provocative testing (71)/Asymptomatic patients without angiographically proven CAD (25) |
Shielded 64-channel Rest and exercise (bicycle ergometry test) |
82%/74% (0.79) (rST segment-fluctuation score) 88%/85% (0.86) (non-dipole) ROC AUC for combination 0.93 |
79%/77% (rST segment-fluctuation score) 87%/86% (non-dipole) |
EKG |
| Shin et al. (20) | Scoring system based on five MCG parameters (T wave score at stress; T wave dispersion at stress; T wave vector MCG at rest; % change in half RT interval vector MCG; and % change in T wave vector MCG) with cut-off of –0.27 shown as best discriminant of significant stenosis |
Anatomic CAD/ Training set: patients with indication for angiography due to chest pain or suspected CAD with ≥1 vessel with 70% stenosis, and without ACS or history of MI within 3 months (35) Internal cross-validation set: patients with indication for angiography due to chest pain or suspected CAD [45; Park et al. (5)]/Training set: patients with indication for angiography due to chest pain or suspected CAD without significant stenosis (73) |
Shielded 64-channel Rest and exercise (bicycle ergometry test) |
77%/89% (0.91) |
74%/91% | EKG |
| Huang et al. (24) | Machine learning approach to analysis of multilayer perceptron neural network as best model | Anatomic CAD/ Patients with chest pain and suspected CAD and underwent coronary angiography (209)/- |
Unshielded 4-channel Rest |
89%/90% for M10 92%/88% for M11 |
93%/85% for M10 92%/87% for M11 |
EKG |
| Tao et al. (25) | Machine learning classification (SVM-XGBoost model) of 164 MCG features measured during segments of the T wave and categorized as time domain, frequency domain, or information theory features | Anatomic CAD/Patients with IHD with clinically identified stenosis (227), including NSTEMI (16)/Healthy subjects (347) | Unshielded 4-channel Rest |
NR/97.8% (0.98) | 86.6%/NR | — |
| Kangwanariyakul et al. (26) | Machine-learning approach to analysis of the JT interval using algorithms of neural network, with BNN identified as best model | IHD/Patients with IHD (29)/Healthy subjects with no evidence of cardiac abnormal symptoms (22) | Not stated 9-channel Rest |
55%/97% (0.85) |
— | — |
| Steinberg et al. (28) | Algorithm-generated score of a scale of 0–100 based on four MCG parameters during Tmax/3 and Tmax: (1) Direction of the main vector from the plus to minus pole (α) between –20° and +110°; (2) Change in the angle of the main vector ≥45° in a time interval of 30 msec; (3) Change in the distance separating the plus and minus poles ≥20 mm in a time interval of 30 msec; (4) Change in the ratio of the pole strengths ≥0.3 in a time interval of 30 msec. Score cut-off of >49 applied based on a previous cohort |
Anatomic CAD Patients with suspected CAD and angiographically proven >50% stenosis (36) Patients with angiographically proven non-obstructive CAD (10) |
Unshielded 9-channel Rest |
40%/84% | 73%/57% | EKG |
| Ramesh et al. (29) | The presence of an abnormal MFM and an abnormal magnetic field angle | Anatomic CAD/Patients with chest pain with normal EKG, positive TMT (12) and negative TMT (17)/- | Shielded 37-channel |
94%/91% | - | Treadmill test |
| Huang et al. (30) | Pearson’s correlation coefficient by comparing each two T-waves by bivariate correlation analysis >0.55 | Anatomic CAD/Patients with an indication for coronary angiography due to angina-like symptoms and without a prior history of CAD; not requiring PCI (85) or requiring PCI (118)/- | Unshielded 4-channel |
66%/73% (0.75) |
75%/64% | EKG |
| Brisinda et al. (31) | STα and Tα, or one of the following: (1) Pattern with ≥2 dipoles in the time interval between 100 msec at the end of S wave (S100) and Tmax; (2) Direction of the current vector between –20° and +110° for the same time interval; (3) If the current vector direction lies between +110° and –20°, one of three parameters had to be satisfactory: (a) Change in the angle of the current vector >60 in 30 msec of the change of angle of S100–Tmax; (b) Change in the pole distance >20 mm (in 30 msec of S100–Tmax); c) Ratio magnetic field poles strength > ± 0.3 (in 30 msec of S100–Tmax) |
Anatomic CAD and functional ischemia Patients with documented CAD by angiography (four by SPECT and exercise bicycle ergometry test) (21) Healthy subjects (13) |
Unshielded, 36-channel Rest and exercise (bicycle ergometry test) |
92%/93% | 92%/NR | Stress EKG SPECT |
| Fenici et al. (32) | Machine learning classification based on scores for the dipoles (>0) and T wave extrema (angle [>45°], distance [>20 mm], ratio [>0.3]) of the MFM in 30 msec intervals during the Tmax/3 to Tmax, and STα and Tα (0–90° normal) as prespecified discriminatory criteria | Anatomic CAD Subgroup of patients classified as ischemic on the basis of clinical criteria and diagnostic tests, and who did not receive PCI (32) Healthy subjects with no evidence of CAD at clinical history, normal physical examination, and echocardiography (33) |
Unshielded, 36-channel Rest |
85%/75% | 83%/78% | EKG |
| Hänninen et al. (33) | Abnormalities in the orientation of the peak gradient of the precordial ST-segment and T-wave magnetic field | Functional ischemia/Patients with single-vessel CAD with angiographically proven stenosis (>50% luminal diameter) in one of the main coronary branches, anginal pain, and a positive EKG stress test, with no prior MI (27)/Healthy volunteers (17) | Shielded 67-channel Exercise (bicycle ergometry test) |
– | – | EKG |
| Van Leeuwen et al. (34) | Spatial distribution of the QT interval with SI cut-off of 3.18 selected as best discriminator | Anatomic CAD/Patients with CAD and angiographically proven ≥75% stenosis with prior MI (31) or without prior MI (23) Healthy subjects proven angiographically or volunteers with no history of CAD (20) |
Shielded 37-channel |
80%/74% | — | EKG |
| Van Leeuwen et al. (35) | >10% deviation from the normal course of the MFM orientation during QT interval selected as a discriminator | Anatomic CAD/ Patients with CAD with angiographically proven ≥75% stenosis of a vessel without evidence of MI (43) or with previous MI (36)/Patients with angiographically proven non-obstructive CAD and healthy volunteers (50) |
Shielded 37 or 61-channel Rest |
90%/68% (in patients without prior MI) 90%/85% (in patients with prior MI) |
– | EKG TTE Angiography |
| On et al. (36) | Sum of the integral values of the QRS (QRSi) or JT (JTi) intervals with JTi/QRSi <1.0 prespecified as discriminant | Anatomic CAD/Patients with angina pectoris and angiographically proven >75% stenosis of a vessel (14) with no (11) or previous (3) MI/Healthy volunteers (30) | Shielded 64-channel Rest |
80%/71% | – | EKG |
| Goernig et al. (37) | Spatiotemporal correlation analysis of 11 MCG parameters. Analysis combining three parameters (mean value correlation QRS at T, STDEV correlation T at QRS and QRS form) was identified as best discriminant | Anatomic CAD/ Patients who experienced MI 6–64 (mean 28) days earlier with angiographically proven >70% stenosis (108)/Subjects without known CAD and with echocardiographic proven normal LVEF (70) |
Shielded 31-channel Rest |
64%/73% | 86%/73% | EKG |
| Gapelyuk et al. (38) | Combination of Kullback-Leibler entropy at ST-T and normalized residual magnetic field strength at QRS selected as best discriminant index | Anatomic CAD/ Patients with symptomatic stable CAD and angiographically proven >50% stenosis in main coronary arteries without previous MI (101)/Healthy subjects with normal findings in EKG, echocardiography, and bicycle ergometry, and no history of cardiac symptoms (59) |
Shielded 7-channel Rest |
88%/88% (0.94) |
– | EKG |
| Wu et al. (39) | QTc dispersion (from the difference between the longest and shortest QTC interval on the QTc contour map) ≥ 79 ms or spatial smoothness index of QTc (SI-QTc) ≥ 9.1 ms | Anatomic CAD/Patients with stable angina and CAD (55)/- | Shielded 64-channel Rest |
68%/86% (0.77) |
– | Stress SPECT Treadmill test |
| Gapelyuk et al. (40) | Three-parameter index (based on ST slope at measurement positions A4 and A6, and the deviation in the MFM orientation) identified by LDA as best discriminant index | Anatomic CAD/ Patients with stable CAD and angiographically proven >50% stenosis without previous MI (101)/Healthy subjects with normal findings in EKG, echocardiography, and bicycle ergometry test, and no history of cardiac symptoms (59) |
Shielded 7-channel Rest |
83%/84% (0.91) |
– | EKG |
| Fenici et al. (41) | Automated analysis of the dynamic motion of the effective magnetic vector during the T wave identified as best discriminator |
Anatomic CAD/ Patients with stable angina and CAD (51), of whom 35 had prior MI/Healthy subjects (52) |
Unshielded 36-channel Rest |
96%/56% | 94%/69% | EKG |
α = average angle of direction for the abnormal current vector during ventricle repolarization period.
MCG, magnetocardiography; CAD, coronary artery disease; ROC, receiver operating curve; AUC, area under the curve; PPV, positive predictive value; NPV, negative predictive value; CAD, coronary artery disease; EKG, electrocardiography; MI, myocardial infarction; SI, smoothness index; MFM, magnetic field map; TTE, transthoracic echocardiography; LDA, linear discriminant analysis; STDEV, standard deviation; LVEF, left ventricular ejection fraction; QTc, corrected QT; Tmax, peak intensity of the T wave; ACS, acute coronary syndrome; STα, magnetic field map angle α for the ST segment; Tα, magnetic field map angle α for the T wave apex; SPECT, single-photon emission computed tomography; IHD, ischemic heart disease; Tmax/3, one-third of peak intensity; PCI, percutaneous coronary intervention; NR, not reported; BNN, Bayesian neural network; NSTEMI, non-ST segment elevation myocardial infarction.