Electrocardiographic mapping. Noninvasive electrophysiological cardiac imaging

Assessment of regularization techniques for electrocardiographic imaging

A widely used approach to solving the inverse problem in electrocardiography involves computing potentials on the epicardium from measured electrocardiograms (ECGs) on the torso surface. The main challenge of solving this electrocardiographic imaging (ECGI) problem lies in its intrinsic ill-posedness. While many regularization techniques have been developed to control wild oscillations of the solution, the choice of proper regularization methods for obtaining clinically acceptable solutions is still a subject of ongoing research. However there has been little rigorous comparison across methods proposed by different groups. This study systematically compared various regularization techniques for solving the ECGI problem under a unified simulation framework, consisting of both 1) progressively more complex idealized source models (from single dipole to triplet of dipoles), and 2) an electrolytic human torso tank containing a live canine heart, with the cardiac source being modeled by potentials measured on a cylindrical cage placed around the heart. We tested 13 different regularization techniques to solve the inverse problem of recovering epicardial potentials, and found that non-quadratic methods (total variation algorithms) and first-order and second-order Tikhonov regularizations outperformed other methodologies and resulted in similar average reconstruction errors.

Comparison of an automated algorithm to expert physician interpretation of 80-lead body surface mapping in the evaluation of acute myocardial ischemia and infarction in patients presenting to the emergency department with chest pain: results from the Optimal Cardiovascular Diagnostic Evaluation Enabling Faster Treatment of Myocardial Infarction trial

INTRODUCTION/BACKGROUND

Eighty-lead (80 L) body surface map (BSM) technology provides electrocardiogram data for the clinician to interpret. A BSM device also offers an automated interpretation. Little information is available about the performance of automated algorithm interpretation in comparison to human interpretation of the 80 L BSM.

METHODS

Interpretations of BSMs by automated algorithm and a core laboratory of physician readers from The Optimal Cardiovascular Diagnostic Evaluation Enabling Faster Treatment of Myocardial Infarction trial were compared. The κ statistic and its 95% confidence interval for concordance were calculated. The effect of BSM quality on concordance was also analyzed.

RESULTS

3405 maps for 1601 subjects were reviewed by the core laboratory and automated algorithm. There was a combined concordance rate of 87.3% (κ = 0.46; 95% confidence interval, 0.40-0.52). A decrease in signal quality was associated with a decrease in concordance between human and automated algorithm interpretation (κ = 0.52 for good quality vs κ = 0.30 for poor quality).

CONCLUSION

A moderate degree of concordance was noted between physician and automated algorithm interpretation of 80 L BSMs. Signal quality of 80 L electrocardiographic BSM directly affected concordance.

Body surface mapping: potential role in a chest pain critical care pathway

Recent advances in biomarkers have improved the evaluation of patients with acute chest pain, but current critical care pathways may still lead to important delays in early diagnosis and, hence, treatment of acute myocardial infarction (AMI). Electrocardiographic changes may occur within seconds of an ischemic insult, but the conventional 12-lead electrocardiogram (ECG) typically has only 50% to 60% sensitivity for diagnosis of AMI. Recording of multiple ECGs over a larger thoracic surface area, including the right ventricular, high left lateral, and posterior regions, by body surface mapping (BSM) has been made feasible in the setting of acute coronary syndromes by novel developments in electrode technology and simultaneous multichannel ECG data acquisition. Clinical studies of an Food and Drug Adminstration-approved BSM system (PRIME-ECG) have demonstrated improved early diagnosis of AMI in patients without 12-lead ST elevation and improved detection of right ventricular or posterior involvement in ST elevation MI. The improved diagnostic sensitivity compared with the conventional 12-lead ECG coupled with the potential reduction of delay to diagnosis compared with biomarkers suggest that BSM may have an important role as part of a chest pain critical care pathway for evaluation of patients with ischemic type chest pain.

Acute detection of ST-elevation myocardial infarction missed on standard 12-Lead ECG with a novel 80-lead real-time digital body surface map: primary results from the multicenter OCCULT MI trial

STUDY OBJECTIVE

Although 80-lead ECG body surface mapping is more sensitive for ST-elevation myocardial infarction (STEMI) than the 12-lead ECG, its clinical utility in chest pain in the emergency department (ED) has not been studied. We sought to determine the prevalence, clinical care patterns, and clinical outcomes of patients with STEMI identified on 80-lead but not on 12-lead (80-lead-only STEMI).

METHODS

The Optimal Cardiovascular Diagnostic Evaluation Enabling Faster Treatment of Myocardial Infarction trial was a multicenter prospective observational study of moderate- to high-risk chest pain patients presenting to the ED. Patients received simultaneous 12-lead and 80-lead ECGs as part of their initial evaluation and were treated according to the standard of care, with clinicians blinded to the 80-lead results. The primary outcome of the trial was door-to-sheath time in patients with 80-lead-only STEMI versus patients with STEMI identified by 12-lead alone (12-lead STEMI). Secondary outcomes included angiographic and clinical outcomes at 30 days.

RESULTS

One thousand eight hundred thirty patients were evaluated, 91 had a discharge diagnosis of 12-lead STEMI, and 25 patients met criteria for 80-lead-only STEMI. Eighty-four of the 91 12-lead STEMI patients underwent cardiac catheterization, with a median door-to-sheath time of 54 minutes, versus 14 of the 25 80-lead-only STEMI patients, with a door-to-sheath time of 1,002 minutes (estimated treatment difference in median=881; 95% confidence interval 181 to 1,079 minutes). Clinical outcomes and revascularization rates, however, were similar between 80-lead-only STEMI and 12-lead STEMI patients.

CONCLUSION

The 80-lead ECG provides an incremental 27.5% increase in STEMI detection versus the 12-lead. Patients with 80-lead-only STEMI have adverse outcomes similar to those of 12-lead STEMI patients but are treated with delayed or conservative invasive strategies.

Diagnosing Old MI by Searching for a Linear Boundary in the Space of Principal Components

Body surface potential mapping (BSPM) is a technique employing multiple electrodes to capture, via noninvasive means, an indication of the heart's condition. An inherent problem with this technique is the resulting high-dimensional recordings and the subsequent problems for diagnostic classifiers. A data set, recorded from a 192-lead BSPM system, containing 74 records is investigated. QRS isointegral maps, offering a summary of the information obtained during ventricular depolarization, were derived from 30 old inferior myocardial infarction and 44 normal recordings. Principal component analysis was applied to reduce the dimensionality of the recordings and a linear classifier was employed for classification. This perceptron-based classifier has been adapted so that the final weight and bias values are estimated prior to the learning process. This estimation process, referred to as the linear hyperplane approach (LHA), derives the estimated weights from a bisector hyperplane, placed orthogonal to the means of two class distributions in an n-dimensional Euclidean space. Estimating weights encourages a network to exhibit better generalization ability. Utilizing a number of different principal components as input features, the LHA achieved an average sensitivity and specificity of 79.58% and 76.45%, respectively, across all experiments. The average accuracy of 76.73% achieved with this approach was significantly better than the other benchmark classifiers evaluated against it.

Selection of optimal recording sites for limited lead body surface potential mapping: a sequential selection based approach

Background

In this study we propose the development of a new algorithm for selecting optimal recording sites for limited lead body surface potential mapping. The proposed algorithm differs from previously reported methods in that it is based upon a simple and intuitive data driven technique that does not make any presumptions about deterministic characteristics of the data. It uses a forward selection based search technique to find the best combination of electrocardiographic leads.

Methods

The study was conducted using a dataset consisting of body surface potential maps (BSPM) recorded from 116 subjects which included 59 normals and 57 subjects exhibiting evidence of old Myocardial Infarction (MI). The performance of the algorithm was evaluated using spatial RMS voltage error and correlation coefficient to compare original and reconstructed map frames.

Results

In all, three configurations of the algorithm were evaluated and it was concluded that there was little difference in the performance of the various configurations. In addition to observing the performance of the selection algorithm, several lead subsets of 32 electrodes as chosen by the various configurations of the algorithm were evaluated. The rationale for choosing this number of recording sites was to allow comparison with a previous study that used a different algorithm, where 32 leads were deemed to provide an acceptable level of reconstruction performance.

Conclusion

It was observed that although the lead configurations suggested in this study were not identical to that suggested in the previous work, the systems did bear similar characteristics in that recording sites were chosen with greatest density in the precordial region.

Temporal analysis of the depolarization wave of healed myocardial infarction in body surface potential mapping

BACKGROUND

We studied the ability of different time segments of the depolarization wave recorded with body surface potential mapping (BSPM) to detect and localize myocardial infarction (MI).

METHODS

BSPM was recorded in 24 patients with remote MI and in 24 healthy controls. Cine and contrast-enhanced magnetic resonance imaging (MRI) was used as a reference method. Patients were grouped according to anatomical location of their MI. The QRS complex was divided into six temporally equal segments, for which time integrals were calculated.

RESULTS

The time segments of the QRS complex showed different MI detection capability depending on MI location. For anterior infarction the second segment of the QRS complex was the best in MI detection and the optimal area was on the right inferior quadrant of the thorax (time integral average -1.5 +/- 1.8 mVms patients, 1.0 +/- 1.6 mVms controls, P = 0.002). For lateral infarction the first segment of the QRS complex performed best and the optimal area for MI detection was the left fourth intercostal area (time integral average 1.8 +/- 1.0 mVms patients, 0.7 +/- 0.5 mVms controls, P = 0.024). For inferior and posterior MI the mid-phases of the QRS complex were the best and the optimal area was the mid-inferior area of the thorax (time integral average -6.2 +/- 8.3 mVms patients, 3.3 +/- 4.3 mVms controls, P = 0.002; -9.1 +/- 6.1 mVms patients, 0.6 +/- 7.1 mVms controls, P = 0.001, respectively).

CONCLUSIONS

Time segment analysis of the depolarization wave offers potential for improving the detection and localization of healed MI.

Quantitative assessment of myocardial ischemia by electrocardiographic and scintigraphic imaging

We calculated distributions of epicardial potentials from body-surface electrocardiograms (ECGs) recorded during controlled myocardial ischemia and compared them with scintigraphic estimates of ischemia's extent/severity. The study population consisted of patients suffering from single-vessel coronary artery disease, referred for elective percutaneous transluminal coronary angioplasty of either the left anterior descending (n=7), the right coronary (n=9), or the left circumflex (n=2) artery. After the target vessel had been dilated, a 1960s "study" inflation was performed with a non-perfusion-type balloon catheter; at its commencement, technetium-99m sestamibi was injected via a femoral-vein catheter, and ECGs were recorded throughout the inflation from 120 leads. Single photon emission computed tomographic imaging was performed one hour after the injection of radionuclide to obtain an "occlusion image", and again one hour after a repeat injection 24 hours later to obtain a "control image"; the latter image was subtracted from the former, to derive a scintigraphic difference map (Delta map). The ECGs were signal-averaged over a 10-s window at preinflation and peak-inflation states, the preinflation averaged complexes were subtracted from the peak-inflation ones to produce body-surface Delta maps, and the corresponding Delta maps of epicardial potentials were calculated by applying the electrocardiographic inverse solution; this procedure is referred to as electrocardiographic imaging. The ECG-derived epicardial Delta maps related spatially to the scintigraphic Delta maps in all patients. The percent areas and surface integrals of positive values in ECG-derived Delta maps were found to be very good single-variable predictors of the extent (r=0.73; p=0.0006) and severity (r=0.72; p=0.0008) of the scintigraphically-estimated perfusion defect; a regression equation using two ECG-derived predictors further improved the agreement with scintigraphic estimates (r=0.81; p=0.0004 for estimates of severity). These findings suggest that noninvasive electrocardiographic imaging might provide quantitative estimates of the extent/severity of myocardial ischemia that agree closely with those provided by scintigraphic techniques.

Noninvasive assessment of reperfusion after fibrinolytic therapy for acute myocardial infarction

Assessment of reperfusion by the 12-lead electrocardiogram (ECG) or biochemical markers is limited by suboptimal sensitivity and/or specificity. Body surface mapping (BSM) improves the spatial sampling of the 12-lead ECG. Serial 12-lead ECGs and 64-lead anterior BSMs were recorded from 67 patients with acute myocardial infarction undergoing coronary angiography 90 minutes after fibrinolytic therapy. ECG-1 and BSM-1 were recorded before/shortly after therapy (median 18 minutes). ECG-2 and BSM-2 were recorded after the 90-minute angiogram (median 30 minutes). The maximum ST elevation on ECG-1 was noted and > or = 30% ST resolution on ECG-2 was taken to represent partial/complete reperfusion. Patients were randomly divided into a training set and validation set. Isointegral and isopotential ST-T variables from BSMs of training-set patients were compared with Thrombolysis In Myocardial Infarction (TIMI) trial flow using discriminant analysis to identify which variables best classified reperfusion. Reperfusion (TIMI 2/3 flow) occurred in 32 of 34 training-set patients and in 29 of 33 validation-set patients. In the training set, > or = 30% ST resolution correctly classified reperfusion with 72% sensitivity (23 of 32) and 50% specificity (1 of 2). In the validation set, > or = 30% ST resolution classified reperfusion with 59% sensitivity (17 of 29) and 50% specificity (2 of 4). In comparison, a model containing 24 BSM variables correctly classified all training-set patients, and when prospectively tested in the validation-set, correctly classified 28 of 29 patients who achieved reperfusion (97% sensitivity) and all 4 patients who failed to reperfuse (p = 0.035). In conclusion, BSM is more useful than the 12-lead ECG for noninvasive assessment of reperfusion after fibrinolytic therapy for acute myocardial infarction.

Value of magnetocardiographic QRST integral maps in the identification of patients at risk of ventricular arrhythmias

It has been shown that regional ventricular repolarization properties can be reflected in body surface distributions of electrocardiographic QRST deflection areas (integrals). We hypothesize that these properties can be reflected also in the magnetocardiographic QRST areas and that this may be useful for predicting vulnerability to ventricular tachyarrhythmias. Magnetic field maps were obtained during sinus rhythm from 49 leads above the anterior chest in 22 healthy (asymptomatic) control subjects (group A) and in 29 patients with ventricular arrhythmias (group B). In each subject, the QRST deflection area was calculated for each lead and displayed as an integral map. The mean value of maximum was significantly larger in the control group A than in the patient group B (1,626+/-694 pTms vs. 582+/-547 pTms, P<0.0001). To quantitatively assess intragroup variability in the control group A and intergroup variability of the control and patient groups, we used the correlation coefficient r and covariance sigma. These indices showed significantly less intragroup than intergroup variation (e.g., in terms of sigma, 28.0x10(-6)+/-12.3x10(-6) vs. 3.4x10(-6)+/-12.5x10(-6), P<0.0001). Each QRST integral map was also represented as a weighted sum of 24 basis functions (eigenvectors) by means of Karhunen-Loeve transformation to calculate the contribution of the nondipolar eigenvectors (all eigenvectors beyond the third). This percentage nondipolar content of magnetocardiographic QRST integral maps was significantly higher in the patient group B than in the control group A (13.0%+/-9.1 % vs. 2.6%+/-2.0%, P<0.0001). Discriminations between control subjects and patients with ventricular arrhythmias based on magnitude of the maximum, covariance sigma, and nondipolar content were 90.2%, 90.2%, and 86.3% accurate, with a sensitivity of 89.7%, 93.1%, and 75.9%, and a specificity of 90.9%, 86.4%, and 100%. We have shown that magnitude of the maximum and indices of variability and nondipolarity of the magnetocardiographic QRST integral maps may predict arrhythmia vulnerability. This finding is in agreement with earlier studies that used body surface potential mapping and suggests that magneticfield mapping may also be a useful diagnostic tool for risk analysis.

Assessment of spatial resolution of pace mapping when using body surface potentials

Using computer simulations and statistical methods, the resolution of pace mapping when used in combination with body surface potentials was systematically investigated. In an anatomical model of the human ventricular myocardium, pre-excitation sequences were initiated at 69 sites positioned along the atrioventricular (AV) ring and corresponding body surface potential maps (BSPMs) were calculated at 32 leads placed on the anterior torso. For each time after the onset of pre-excitation (every 4 ms to 40 ms) and each root-mean-square (RMS) noise level (5, 10, 20 and 50 microV), BSPMs were cros-correlated and the spatial resolution defined as the largest pacing site separation at which the differences in correlation coefficients were not statistically significant (level p > or = 0.05). The findings indicate that when random RMS noise of 5 microV was added to the simulated BSPMs, average spatial resolution over all 60 sites was at 20 ms after the onset of pre-excitation within 3.5 +/- 0.9 mm. The results provide theoretical evidence that statistical analysis of BSPMs obtained during pace mapping can offer improved means for subcentimetre identification of accessory pathways located along the AV ring.