Most of the high-frequency electrocardiogram (HFECG) classification algorithms at this stage are either two-class classification of myocardial infarction (MI) and non-infarct or infarct class classification algorithms, which cannot detect cases in the early myocardial ischemic stage of infarction. For this reason, an ischemic cardiac disease classification algorithm based on HFECG is proposed in this paper. The algorithm selects and improves six high-frequency component parameters as features and uses the XGBoost model to classify the samples. Compared with the traditional algorithm, this algorithm adds the classification of ischemic (ISC) cases, which can detect potential cases of MI early. In addition, this paper improves the solving process and morphological index of the reduced amplitude zone (RAZ) in the high-frequency component parameters, which improves the classification performance of the algorithm. The algorithm in this paper is used in the experiments on the PTB-XL data set, and the clinical data is used to verify it. The experimental results show that the HFECG features used in this paper have strong characterization ability for myocardial ischemic abnormalities, with accuracy rates of 83.9%, 81.7%, 88.2% and 93.9% for PTB-XL data set for the four classification categories: normal (NORM), abnormity (ABNORM), ISC and MI, in that order. The algorithm can effectively tap cases in the early myocardial ischemic stage of infarction.
定义3 对于、在处的幅值下降区域RAZ,若满足、幅值中较小值大于较大值的30%,且V处幅值大于、幅值较小值的30%,则称该幅值下降区域为一个满足百分比条件的幅值下降区域(reduced amplitude zone percentage,RAZP)。
美国国家航空航天局(national aeronautics and space administration,NASA)研发的ECG(electrocardiogram, 心电图)软件中,对幅值下降区域的判断标准为[22,23]:同时存在两个上包络幅值极大值点、和其间上包络幅值最小值点,两个下包络幅值极小值点P、P和其间下包络幅值最大值点,且每一个RAZ均需满足定义3,由此可得定义4。
定义4 对于一个HFQRS信号,若在信号的上包络存在两个幅值极大值点、,且下包络存在两个极小值P、P,必然有一个与间上包络的最小值点和P与P间下包络的最大值点,则称由这样的、及其间的幅值采样点构成的点集为、在处的幅值下降区域RAZ由P与P及其间的幅值采样点构成的点集为P与P在处的幅值下降区域RAZ。当RAZ与RAZ同时满足定义3时,则称该组幅值下降区域RAZ与RAZ为一个满足NASA条件的幅值下降区域(reduced amplitude zone NASA,RAZN)。
ZHOUM G, WANGH D, ZENGX Y, et al. Mortality, morbidity, and risk factors in China and its provinces, 1990-2017: A systematic analysis for the global burden of disease study 2017[J]. Lancet,2019, 394(10204): 1145-1158. DOI: 10.1016/S0140-6736(19)30427-1 .
YY 0782-2010 Medical electrical equipment Part 2-51:Particular requirements for safety, including essential performance,of recording and analysing single channel and multichannel of electrocardiographs [S]. Beijing: Standards Press of China,2010 (Ch). 10.1007/978-1-137-09339-4_25
LING Y, ZHAO C M. Prediction of acute vascular occlusion and cardiac death in patients with NSTEMI by serum NT-proBNP[J]. Journal of Guiyang Medical College,2015,40(11):1245-1248 (Ch).
[6]
LANGNERP H, GESELOWITZD B. Characteristics of the frequency spectrum in the normal electrocardiogram and in subjects following myocardial infarction [J]. Circulation Research,1960, 8: 577-584. DOI: 10.1161/01.RES.8.3.577 .
[7]
LANGNERP H, LAUERJ A. The relative significance of high-frequency and low-frequency notching in the electrocardiogram[J]. American Heart Journal,1966, 71 (1): 34-42. DOI: 10.1016/0002-8703(66)90654-5 .
[8]
LANGNERP H, GESELOWITZD B, BRILLERS A.Wide band recording of the electrocardiogram and coronary heart disease [J]. American Heart Journal,1973, 86(3): 308-317. DOI: 10.1016/0002-8703(73)90039-2 .
[9]
ABBOUDS, COHENR J, SELWYNA, et al. Detection of transient myocardial ischemia by computer analysis of standard and signal-averaged high-frequency electrocardiograms in patients undergoing percutaneous transluminal coronary angioplasty [J]. Circulation,1987, 76(3): 585-596. DOI: 10.1161/01.CIR.76.3.585 .
[10]
SCHLEGELT T, KULECZW B, DEPALMAJ L, et al. Real-time 12-lead high-frequency QRS electrocardiography for enhanced detection of myocardial ischemia and coronary artery disease[J]. Mayo Clinic Proceedings,2004, 79(3): 339-350. DOI: 10.4065/79.3.339 .
[11]
LEINVEBERP, HALAMEKJ, JURAKP, et al. Higher frequencies in QRS complex for the detection of myocardial ischemia [C]//2016 Computing in Cardiology Conference (CinC). New York: IEEE Press,2016:297-300. DOI: 10.22489/cinc.2016.087-361 .
[12]
LUCAS A, LIN Y, DONG S, et al. Applications of machine learning algorithms in predicting coronary artery disease and myocardial infarction [J].North American Academic Research,2018, 1(2): 81-93.
[13]
SUNL, LUY P, YANGK T, et al. ECG analysis using multiple instance learning for myocardial infarction detection [J]. IEEE Transactions on Biomedical Engineering,2012, 59(12): 3348-3356. DOI: 10.1109/TBME.2012.2213597 .
[14]
GED F. Classification of myocardial infraction stages using high-frequency ECG[J]. Communications in Information Science and Management Engineering, 2(11):77-80. DOI: 10.1109/icbbe.2010.5516250 .
[15]
KORAP, KALVAS R. Improved Bat algorithm for the detection of myocardial infarction [J]. SpringerPlus,2015, 4(1): 1-18. DOI: 10.1186/s40064-015-1379-7 .
[16]
ACHARYAU R, FUJITAH, SUDARSHANV K, et al. Automated detection and localization of myocardial infarction using electrocardiogram: A comparative study of different leads [J]. Knowledge⁃Based Systems,2016, 99: 146-156. DOI: 10.1016/j.knosys.2016.01.040 .
[17]
ACHARYAU R, FUJITAH, ADAMM, et al. Automated characterization and classification of coronary artery disease and myocardial infarction by decomposition of ECG signals: A comparative study [J]. Information Sciences,2017, 377: 17-29. DOI: 10.1016/j.ins.2016.10.013 .
[18]
ACHARYAU R, FUJITAH, OH S L, et al. Application of deep convolutional neural network for automated detection of myocardial infarction using ECG signals[J]. Information Sciences,2017, 415/416:190-198. DOI: 10.1016/j.ins.2017.06.027 .
[19]
LUI H W, CHOWK L. Multiclass classification of myocardial infarction with convolutional and recurrent neural networks for portable ECG devices [J].Informatics in Medicine Unlocked,2018, 13: 26-33. DOI:10.1016/j.imu.2018.08.002 .
[20]
FENGK, PIX T, LIUH Y, et al. Myocardial infarction classification based on convolutional neural network and recurrent neural network [J]. Applied Sciences,2019, 9 (9): 1879. DOI: 10.3390/app9091879 .
[21]
PANJ P, TOMPKINSW J. A real-time QRS detection algorithm [J]. IEEE Transactions on Biomedical Engineering,1985, BME⁃32 (3): 230-236. DOI: 10.1109/tbme.1985.325532 .
WANGL, SUNW, CHENY B, et al. ECG denoising method based on adaptive wavelet threshold selection [J].Computer Engineering Applications,2018,54(15):29-33. DOI: 10.3778/j.issn.1002-8331.1805-0352(Ch ).
[24]
AMITG, GALANTEO, DAVRATHL R, et al. High-frequency QRS analysis in patients with acute myocardial infarction: A preliminary study [J]. Annals of Noninvasive Electrocardiology,2013, 18(2): 149-156. DOI: 10.1111/anec.12023 .
[25]
SCHLEGEL T T, DEPALMA J L, MORADI S. Real⁃time, High Frequency QRS Electrocardiograph: US,09/906013[P]. 2003-01-16. 10.4065/79.3.339
[26]
SCHLEGEL T T, DEPALMA J L, MORADI S. Real⁃Time, High Frequency QRS Electrocardiograph with Reduced Amplitude Zone Detection: US,7539535[P]. 2009-05-26. 10.4065/79.3.339
[27]
CHENT Q, GUESTRINC. XGBoost: A scalable tree boosting system [C]//Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. New York:ACM,2016:785-794. DOI: 10.1145/2939672.2939785 .
[28]
WAGNERP, STRODTHOFFN, BOUSSELJOTR D, et al. PTB-XL, a large publicly available electrocardiography dataset[J]. Scientific Data,2020, 7 (1): 154. DOI: 10.1038/s41597-020-0495-6 .
[29]
BOUSSELJOT R, KREISELER D, SCHNABEL A.Nutzung der EKG-signaldatenbank CARDIODAT der PTB über das internet[J]. Biomedizinische Technik,1995, 40(s1): 317-318.