Inter-Patient ECG Heartbeat Classification with Temporal VCG Optimized by PSO.

Classifying arrhythmias can be a tough task for a human being and automating this task is highly desirable. Nevertheless fully automatic arrhythmia classification through Electrocardiogram (ECG) signals is a challenging task when the inter-patient paradigm is considered. For the inter-patient paradigm, classifiers are evaluated on signals of unknown subjects, resembling the real world scenario. In this work, we explore a novel ECG representation based on vectorcardiogram (VCG), called temporal vectorcardiogram (TVCG), along with a complex network for feature extraction. We also fine-tune the SVM classifier and perform feature selection with a particle swarm optimization (PSO) algorithm. Results for the inter-patient paradigm show that the proposed method achieves the results comparable to state-of-the-art in MIT-BIH database (53% of Positive predictive (+P) for the Supraventricular ectopic beat (S) class and 87.3% of Sensitivity (Se) for the Ventricular ectopic beat (V) class) that TVCG is a richer representation of the heartbeat and that it could be useful for problems involving the cardiac signal and pattern recognition.

Value of the QRS-T area angle in improving the prediction of sudden cardiac death after acute coronary syndromes.

BACKGROUND: Prediction of sudden cardiac death (SCD) after acute coronary syndromes (ACS) remains a challenge. Although electrophysiology measures obtained by 3-D vectorcardiography (VCG) shortly after ACS may be useful predictors of SCD, they have not been adopted into clinical practice. The main objective of our study was to assess whether the VCG-derived QRS-T area angle (between area vectors) and the QRS-T angle (between maximum vectors) have additional value beyond standard risk factors in predicting SCD after ACS. METHODS AND RESULTS: We studied 643 consecutive ACS patients for whom data on VCG and echocardiography during the index hospitalization were available. Seventy-seven patients (12%) died, 37 (6%) from SCD and 21 (3%) from other cardiac causes during the 30-month follow-up. After adjusting for 9 standard risk factors (age, sex, diabetes, previous stroke, left ventricular ejection fraction; and estimated glomerular filtration rate, heart rate, systolic blood pressure<100mmHg, and Killip class>1 on admission), QRS-T area angle and QRS-T angle were shown to have independent predictive value for both SCD and all cardiac deaths. Reclassification analysis showed that both measures had additional predictive value beyond the 9 standard risk factors. For SCD, net reclassification improvements for QRS-T area angle and QRS-T angle were 46% and 45% and relative integrated discriminative improvements were 16% and 13% (vs the average~11% of the 9 standard risk factors). CONCLUSIONS: The VCG-derived QRS-T area angle and QRS-T angle improved prediction of SCD after ACS beyond standard risk factors. Further evaluation of their clinical utility and cost-effectiveness is therefore warranted.

On designing and testing transformations for derivation of standard 12-lead/18-lead electrocardiograms and vectorcardiograms from reduced sets of predictor leads.

The aim of this study was to develop and evaluate transformation coefficients for deriving the standard 12-lead electrocardiogram (ECG), 18-lead ECG (with additional leads V7, V8, V9, V3R, V4R, V5R), and Frank vectorcardiogram (VCG) from reduced lead sets using 3 "limb" electrodes at Mason-Likar torso sites combined with 2 chest electrodes at precordial sites V1 to V6; 15 such lead sets exist and each can be recorded with 6-wire cable. As a study population, we used Dalhousie Superset (n = 892) that includes healthy subjects, postinfarction patients, and patients with a history of ventricular tachycardia. For each subject, 120-lead ECG recordings of 15-second duration were averaged, and all samples of the QRST complex for leads of interest were extracted; these data were used to derive--by regression analysis--general and patient-specific coefficients for lead transformations. These coefficients were then used to predict 12-lead/18-lead ECG sets and 3-lead VCG from 15 reduced lead sets, and the success of these predictions was assessed by 3 goodness-of-fit measures applied to the entire QRST waveform and to the ST deviation at J point; these 3 measures were similarity coefficient (SC in percentage), relative error (in percentage), and RMS error (in microvolts). Our results show that the best pair for predicting the standard 12-lead ECG by either general coefficients (mean SC = 95.56) or patient-specific coefficients (mean SC = 99.11) is V2 and V4; the best pair for deriving the 18-lead set by general coefficients (mean SC = 93.74) or by patient-specific coefficients (mean SC = 98.71) is V1 and V4; the best pair for deriving the Frank X, Y, Z leads is V1 and V3 for general coefficients (mean SC = 95.76) and V3 and V6 for patient-specific coefficients (mean SC = 99.05). The differences in mean SC among the first 8 to 10 predictor sets in each ranking table are within 1% of the highest SC value. Thus, in conclusion, there are several near-equivalent choices of reduced lead set using 6-wire cable that offer a good prediction of 12-lead/18-lead ECG and VCG; a pair most appropriate for the clinical application can be selected.

Standardization of the PQRST waveform and analysis of arrhythmias in the fetus using vector magnetocardiography.

Fetal magnetocardiography (fMCG) is useful for analysis of fetal cardiac events. However, fetal presentation and movement affect the fMCG waveform, making it difficult to standardize the waveform. The aim of this study was to investigate whether the use of vector magnetometers can compensate for these limitations. We studied 59 fetuses (gestational age, 22-40 wk, median, 32), including 41 with uncomplicated pregnancies and 18 with fetal cardiac disease. fMCG was recorded twice in each case, and the two waveforms were compared with each other in uncomplicated subjects to investigate the effects of fetal presentation. The superconducting quantum interference device (SQUID) system used in this study was a 12-channel vector magnetometer, by which the three components of the magnetic field (Bx, By, Bz) could be detected simultaneously at four recording points. By constructing the three components, a composite waveform (Bxyz) was obtained. The configuration of the composite waveforms was similar among normal fetuses always with positive polarity, independent of fetal presentation and movement. The difference in the time intervals (PR, QRS, and ventricular activation time [VAT]) between the first and second measurements was minimal in the composite waveforms (Bxyz) compared with that in each channel (Bx, By, Bz). Even before signal averaging, waveforms with high time resolution were recorded in at least one of the three components, making it possible to analyze fetal arrhythmias precisely. Our results indicate that vector magnetocardiography is potentially useful for standardization of the fMCG waveforms and to provide a more complete and accurate analysis of fetal arrhythmias.

Continuous vectorcardiography is superior to standard electrocardiography in the prediction of long-term outcome after thrombolysis in patients with acute myocardial infarction.

BACKGROUND: Thrombolytic therapy results in reperfusion of the occluded coronary vessel in approximately 75% of treated patients with acute myocardial infarction (AMI). Unsuccessful thrombolysis results in impaired outcome. This study was undertaken to evaluate reperfusion assessments with 12-lead standard static electrocardiography (ECG) and continuous vectorcardiography (VCG) in AMI patients treated with thrombolytic therapy, with particular emphasis on the value of these assessments in relation to long-term outcome. METHODS: ST-recovery analysis 90 and 180 min after the start of thrombolytic therapy was performed by repeated ECG and by VCG in 63 AMI patients. Median follow-up was 255 days. RESULTS: No significant differences in long-term outcome were found between patients with or without obtained reperfusion, as assessed by ECG. For VCG, we found significant elevated relative risks for experiencing death (relative risk = 11.00, confidence interval = 2.70-44.90); P = 0.0008 for the group with ST-vector magnitude recovery of less than 50% at 90 min from start of thrombolytic therapy. CONCLUSION: We demonstrated that early reperfusion assessment with VCG enables the prediction of long-term outcome and is superior to reperfusion assessment with standard static ECG in this regard. We therefore recommend continuous ischemia monitoring of AMI patients treated with thrombolytic therapy as a routine procedure.

Vectorelectrocardiography in coronary artery disease.

Continuous ST-segment monitoring by either continuous 12-lead ECG or continuous vectorcardiography provides reliable information regarding ST changes in patients with ongoing myocardial ischemia with or without concurrent chest pain. ST-segment monitoring enables the clinician to continuously follow the dynamic changes that characterize unstable angina and acute myocardial infarction syndromes. It provides important information for risk stratification in unstable coronary syndromes and helps in differentiating between extra-cardiac chest pain and acute coronary disease. The use of VECG for detection of perioperative myocardial damage is promising but much work still needs to be done to clarify the prognostic and clinical value of VECG in this setting.

Positional changes of spatial QRS- and ST-segment variables in normal subjects: implications for continuous vectorcardiography monitoring during myocardial ischemia.

Electrocardiographic QRS- and ST-segment changes are to be expected during changes in body posture. We prospectively analyzed the influence of changes in body position on continuous vectorcardiography monitoring of QRS-vector difference (QRS-VD) and ST change-vector magnitude (STC-VM) according to the currently used criteria of myocardial ischemia in 21 normal subjects. Fifteen (71%) and 6 (29%) subjects had significant positional QRS-VD and STC-VM changes, respectively. Vectorcardiography changes were most frequent and pronounced in the left lateral position. An alternative to the existing criterion of ischemia is proposed to improve the specificity of STC-VM. Subjects with positional QRS-VD changes had higher mean STC-VM values as compared with those without such changes. Otherwise no characteristics among those with positional vectorcardiography changes could be identified. There was no statistically significant association between positional QRS-VD and STC-VM changes (R = .13, P = .57). We conclude that the clinical use of QRS-VD in its present form for continuous vectorcardiography monitoring of myocardial ischemia seems to be of limited practical value, because of the presence of frequent "pseudo-ischemic" changes. STC-VM seems to have a significant potential of continuous vectorcardiography monitoring. However, an indicator of body position change or even an algorithm enabling on-line correction for positional vectorcardiography changes seems to be essential to improve the accuracy of this technique in identifying myocardial ischemia.

Comparison of scalar with vectorial electrocardiogram in axis determination.

Axis deviation is one of the variables most commonly sought for in Electrocardiography (ECG). Although no one doubts the superiority of vector cardiography (VCG) as the most accurate in axis determination, most clinicians adopt the Hexaxial Reference System (HRS) of the 12-Lead ECG (12LS) as the most accessible for routine use. The question therefore arises: How accurate is the HRS? The 12LS and Orthogonal (Frank Lead) ECG (OLS) were recorded in 664 adult Nigerians without heart or metabolic diseases. Their VCG were constructed manually for the QRS complexes. On each subject, QRS axis was determined by three methods: the HRS for the 12LS, the trigonometric method for the OLS and the direction of the maximum deflection vector for the VCG. Axes by the three modalities were analysed and compared statistically as applicable to paired samples. The frontal plane (FP) QRS axis ranged between O degree and +90 degrees in 98.2% of cases by VCG, 96% by the OLS and 93.6% by the 12LS. There was excellent correlation between axes obtained by VCG and OLS (r = 0.85; P < 0.0001). It was lower but highly significant between VCG and the 12LS (r = 0.70; P < 0.0001). In the horizontal plane (HP), the 97 per centile distribution ranged from 240 degrees through 0 degree to 30 degrees; that is, posteriorly and to the left. In the left sagittal plane (LSP), the 95 per centile distribution ranged from 60 degrees counter-clockwise to 210 degrees; that is inferiorly and posteriorly. In a sample of healthy adult Nigerians, the QRS axis by VCG was located posteriorly, inferiorly and to the left. Axis determination by the 12LS is limited to the FP only, and it bears a good correlation with VCG. This commends the H RS as a condonable tool for estimating wave axis routinely and for epidemiologic studies.

Normal limits of the derived vectorcardiogram in Caucasians.

The derived vectorcardiogram (dVCG) can be obtained from three orthogonal leads X, Y, Z, which are synthesized from the conventional 12-lead ECG by one of a number of mathematical techniques. The application of the dVCG in routine cardiological practice has the advantage of not requiring extra recording electrode other than those used for the conventional 12 electrocardiographic leads, and offers the simultaneous availability of additional dVCG diagnostic information. In this study, the dVCGs were obtained using the inverse Dower method from 1555 apparently healthy Caucasian individuals (884 men and 671 women) and were analysed to determine the age and sex dependent normal ranges of vector measurements in a Caucasian population. The vectorcardiographic parameters measured included the directions of inscription of the QRS vector loops, the magnitude of initial 20 and 30 ms QRS vectors and the direction of the initial 20 ms QRS vector. The maximal spatial QRS vector magnitude, as well as the maximal QRS and T vector magnitudes in the frontal, horizontal, and right sagittal planes, were observed to decrease significantly with advancing age in both sexes (P < 0.001) and were significantly larger in men in all age groups (P < 0.001). The observations from the current study illustrate the significant age and sex dependent differences in the normal ranges of dVCG parameters. These are of potential importance for diagnostic applications.

Assessment of myocardial ischemia by 12-lead electrocardiography and Frank vector system during coronary angioplasty: value of a new orthogonal lead system for quantitative ST segment monitoring.

The optimal number and placement of electrocardiographic (ECG) leads to detect myocardial ischemia induced by coronary balloon inflation was assessed by analyzing ST segment changes in the standard 12-lead ECG and Frank X, Y, Z leads at 90-s intervals during 34 consecutive coronary angioplasty procedures. Mean occlusion time during angioplasty was 218 +/- 65 s. Myocardial ischemia, defined as transient angina or ST segment deviation greater than or equal to 1 mm in at least one lead, occurred in 33 (97%) of the 34 procedures. The most sensitive single leads (V2 or V3) detected 17 (51%) of 33 ischemic episodes. The best dual-lead combinations (leads V2 and V5, leads a VF and V3 and leads V3 and Y) increased the sensitivity of 69% (23 of 33). The three-lead combination V2, V5, Y had the highest detecting power (78% [26 of 33]). The X, Y, Z leads by themselves had a sensitivity of only 60% (20 of 33). From this proposed orthogonal lead system (V2, V5, Y), which combines anteroposterior (V2), left to right (V5) and inferosuperior (Y) forces, the spatial ST vector magnitude was calculated and monitored during balloon inflations. A good correlation was observed between this ST vector magnitude and the sum of ST deviations on the standard ECG (r = 0.940, p less than 0.00001), and these data were reproducible over sequential balloon inflations. The results of the study suggest that this orthogonal lead system is of considerable value in the detection and quantification of acute myocardial ischemia and, in this respect, is more useful than the Frank orthogonal vector system.

Effect of pulmonary circulation on vector impedance cardiogram.

Vector impedance cardiograms in horizontal lead configuration [VICG(H)] were recorded in 34 normal subjects, 18 patients with mitral stenosis, 9 patients with mitral regurgitation, 14 patients with aortic regurgitation and 13 patients with non- cyanotic septal disorders. Data in normal subjects revealed that most of the phase reversal points in VICG(H) waveform did not coincide with those of conventional impedance cardiogram. The shape of VICG(H) waveform and values of VICG indices were observed to be markedly affected in patients having significant alteration in the pulmonary circulation and the changes observed were specific of the type of the disorder. The results of the study suggested that right side of the heart had dominant contribution in generation of VICG(H) waveform.

Lack of superiority of the vectorcardiogram over the electrocardiogram in detecting inferior wall myocardial infarction regardless of time since infarction.

Using more recent VCG and ECG criteria, the relative accuracy of these two tests in detecting inferior wall MI over time was evaluated in 38 of 236 patients undergoing elective left ventriculographic and coronary angiographic studies who had clinical plus angiographic evidence of inferior wall myocardial infarction. The overall sensitivity and specificity of the ECG criteria of the New York Heart Association and Warner did not differ from that of the VCG criteria of Starr and Takatsu. There was a trend toward decreased sensitivity in both VCG criteria and the ECG criterion of Warner in detecting inferior wall myocardial infarction greater than or equal to 18 months, although the difference did not reach statistical significance. Changing the age of infarction to greater than or equal to 3, greater than or equal to 6, greater than or equal to 12, or greater than or equal to 24 months did not yield a different result. It is concluded that VCG is not superior to ECG in the diagnosis of inferior wall MI regardless of time since occurrence of infarction.

Comparison of the classification ability of the electrocardiogram and vectorcardiogram.

Controversy exists over the classification ability of the standard 12-lead electrocardiogram (EGG) and the vectorcardiogram (VCG). In this study the diagnostic information content and classification performance of the ECG and VCG were examined using multivariate statistical techniques and a large validated data base of 3,266 cases. Logistic classification models were developed to differentiate between 7 diagnostic entities: normal (n = 538), left (n = 557), right (n = 323) and biventricular (n = 437) hypertrophy, and anterior (n = 390), inferior (n = 657) and combined (n = 364) myocardial infarction. The models were obtained from a learning sample (n = 2,446) using an optimal set of computer derived ECG and VCG measurements. They were subsequently applied to a test sample (n = 820). In the learning sample, the discrimination models resulted in a total correct classification rate of 69.6% for the ECG and 69.4% for the VCG. The total accuracy rate was slightly lower in the test set: 66.3% for the ECG and 67.1% for the VCG. The combined use of the best ECG and VCG variables did not increase total diagnostic accuracy. When cases with biventricular hypertrophy and combined infarction were deleted, accuracy rates of more than 80% were achieved for both lead systems. Differences in the classification rates for the subgroups were not statistically significant. Thus, the conventional 12-lead ECG is as good as the VCG for the differential diagnosis of 7 main entities, provided identical procedures are used in the design of the classifiers.(ABSTRACT TRUNCATED AT 250 WORDS)

Assessment of the performance of electrocardiographic computer programs with the use of a reference data base.

To allow an exchange of measurements and criteria between different electrocardiographic (ECG) computer programs, an international cooperative project has been initiated aimed at standardization of computer-derived ECG measurements. To this end an ECG reference library of 250 ECGs with selected abnormalities was established and a comprehensive reviewing scheme was devised for the visual determination of the onsets and offsets of P, QRS, and T waves. This task was performed by a group of cardiologists on highly amplified, selected complexes from the library of ECGs. With use of a modified Delphi approach, individual outlying point estimates were eliminated in four successive rounds. In this way final referee estimates were obtained that proved to be highly reproducible and precise. This reference data base was used to study measurement results obtained with nine vectorcardiographic and 10 standard 12-lead ECG analysis programs. The medians of program determinations of P, QRS, and T wave onsets and offsets were close to the final referee estimates. However, an important variability could be demonstrated between measurements from individual programs and mean differences from the referee estimates amounted to 10 msec for QRS for certain programs. In addition, the variances of all programs with respect to the referee point estimates were variable. Some programs proved to be more accurate and stable when the data from high- vs low-noise recordings were analyzed. Average Q wave durations calculated from ECGs for which programs agreed on the presence of a Q or QS wave differed by more than 8 msec in several program-to-program comparisons. Such differences may have important consequences with respect to diagnostic performance. Various factors that might explain these differences have been determined. The present study demonstrates that to allow an exchange of results and diagnostic criteria between different ECG computer programs, definitions, minimum wave requirements, and measurement procedures urgently need to be standardized.

Delayed initial QRS forces of intraventricular pacing: unreliability of the scalar ECG in their diagnosis.

Delta waves, similar to those observed in the Wolff-Parkinson-White syndrome, are currently observed in the ECG tracings of patients with endocardial pacing. With the aid of standard ECG and vectorcardiography (VCG), we studied the initial QRS forces during right endocardial pacing in 20 patients. More details were obtained by VCG concerning the amplitude duration, electrical axis and shape of the delta waves of paced ventricular beats, than by ECG. It is remarkable that delayed initial QRS forces, diagnosed by VCG analysis, were present in only 9 of 20 patients with delta waves in their ECG. In other patients, the slurring or thickening at the onset of the R waves in the ECG resulted from the perpendicular direction of the initial QRS vectors to the lead axis, or from multiple changes in the direction of the early, nondelayed QRS vectors. The superiority of VCG in the study of the initial QRS forces, particularly in the diagnosis of delayed initial QRS forces, is discussed.

Time-domain remarks on the cycle-to-cycle variability in some quantities measured on the vectorcardiogram magnitude.

Traces from five patients with different pathophysiological conditions were considered. For each of them more than ten successive cardiac cycles were observed. In each cycle, fourteen temporal intervals were measured. Diagrams were prepared with all values from many measurements made on the electrocardiogram of a patient. Means, variances and coefficients of variation were also computed. A consistent variability was detected in all measurements on every patient. The causes of such a situation have to be studied in detail. However, the coefficient of variation appears make it possible by itself, to distinguish some pathological cases (infarcts and conduction defects) from all others considered here.

Studies on the normal polarcardiogram using right to left axis as the polar axis.

Polarcardiograms using right to left axis as the polar axis were recorded in 112 cases of the clinically healthy subjects. The wave forms of the polarcardiogram of the normal subjects were classified into the several difinite patterns. The quantitative values obtained by the measurement of the polarcardiograms agreed well with those obtained by the vectorcardiograms reported by the various authors. Polarcardiography has advantages not only to let us know the spatial polar coordinates of the cardiac vectors at any instant very easily and quickly, but also to give us the informations of the time-related continuous pattern of the polar components of the cardiac vectors very easily which are difficult to obtain by the conventional ECG and VCG.