ROC Analysis of PVC Detection Algorithm using ECG and Vector-ECG Charateristics

ECG analysis method was developed using ROC analysis of PVC detecting algorithm. ECG signal of MIT-BIH arrhythmia database was analyzed by MATLAB. First of all, the baseline was removed by median filter to preprocess the ECG signal. R peaks were detected for ECG analysis method, and normal VCG was extracted for VCG analysis method. Four PVC detecting algorithm was analyzed by ROC curve, which parameters are maximum amplitude of QRS complex, width of QRS complex, r-r interval and geometric mean of VCG. To set cut-off value of parameters, ROC curve was estimated by true-positive rate (sensitivity) and false-positive rate. sensitivity and false negative rate (specificity) of ROC curve calculated, and ECG was analyzed using cut-off value which was estimated from ROC curve. As a result, PVC detecting algorithm of VCG geometric mean have high availability, and PVC could be detected more accurately with amplitude and width of QRS complex.

Authors:

• J. S. Nah
• A. Y. Jeon
• J. H. Ro
• G. R. Jeon

Keywords:

• Vectorcardiogram (VCG)
• Premature Ventricular contraction (PVC)
• ROC (receiver operating characteristic) curve
• ECG

References:

[1] J. Chee, R. Acharya U, K. Er, W. Tan and C. Kuang Chua, "Visualization
of cardiac health using vector cardiogram", IRBM, vol.29, no.4,
pp.245-254, 2008.
[2] A. Redz, "Presentation and Analysis of Vector Electrocardiograms",
Department of Numerical Analysis and Computer Science, Royal Institute
of Technology, Sweden, 1998.
[3] A. Shvilkin, B. Bojovic, B. Vajdic, I. Gussak, K.K. Ho, P. Zimetbaum and
M.E. Josephson, "Vectorcardiographic and electrocardiographic criteria
to distinguish new and old left bundle branch block.", Heart Rhythm, vol.
7, no. 8, pp. 1085-1092, 2010.
[4] Jekova and V. Krasteva, "Fast Algorithm for Vectorcardiogram and
Interbeat Intervals Analysis: Application for Premature Ventricular
Contractions Cloassification.", Bioautomation, vol.3, pp.82-93, 2005.
[5] R. Lazzara, "Spatial vectorcardiogram to predict risk for sudden
arrhythmic death: phoenix risen from the ashes.", Heart Rhythm., vol.7,
no.11, pp.1606-1613, 2010.
[6] M. Ghasemi, A. Jalali, H. SadAbadi, M. Atarod, H. Golbayani, P.
Ghorbanian and A. Ghaffari, "Electrocardiographic imaging of
myocardial infarction using heart vector analysis.", Computers in
Cardiology, vol.34, pp.625-628, 2007.
[7] G. T. Kang, K. T. Park, G. R. Kim, B. C. Choi and D. K. Jung, "Real time
gait analysis using acceleration signal.", J. of the Korean Sensors Society,
vol.18, no.6, pp. 449-455, 2009.
[8] A. R. Pérez Riera, A. H. Uchida, C. F. Filho, A. Meneghini, C. Ferreira ,
E. Schapacknik, S. Dubner and P. Moffa, "Significance of
vectorcardiogram in the cardiological diagnosis of the 21st century.",
Clin. Cardiology, vol.30, no.7, pp.319-323, 2007.
[9] J. K. Kim, D. H. Kang and M. H. Lee, "An Adaptive Classification
Algorithm of Premature Ventricular Beat With Optimization of Wavelet
Parameterization", J. Biomed. Eng., vol.30, pp.294-305, 2009.
[10] H. K. Jeon, I. S. Cho and, H. S. Kwon, "The Detection of PVC based
Rhythm Analysis and Beat Matching", KIMICS, vol.13, no.11,
pp.2391-2398, 2009.
[11] I. S. Cho and H. S. Kwon, "R Wave Detection Algorithm Based Adaptive
Variable Threshold and Window for PVC Classification", KICS, vol.34,
no.11, pp.1289-1295, 2009.