2018 Computing in Cardiology Conference (CinC)最新文献_第5页

Automatic Sleep Arousal Identification From Physiological Waveforms Using Deep Learning 利用深度学习从生理波形中自动识别睡眠唤醒

2018 Computing in Cardiology Conference (CinC) Pub Date : 2018-12-30 DOI: 10.22489/CinC.2018.242

Daniel Miller, Andrew Ward, N. Bambos

{"title":"Automatic Sleep Arousal Identification From Physiological Waveforms Using Deep Learning","authors":"Daniel Miller, Andrew Ward, N. Bambos","doi":"10.22489/CinC.2018.242","DOIUrl":"https://doi.org/10.22489/CinC.2018.242","url":null,"abstract":"The 2018 PhysioNet Computing in Cardiology Challenge focused on diagnosing sleep disorders, motivated by enabling treatment to alleviate the associated mental and physical health consequences. The dataset consists of 1,985 patients monitored at an MGH sleep laboratory where vital signs were recorded, and arousal regions were annotated by experts. This work presents a deep-learning method to identify sleep arousals. In traditional machine learning, feature extraction is one of the most time-intensive considerations, requiring a great deal of domain expertise and experimentation. In contrast, deep learning techniques automatically learn variable interactions between pairs or groups of signals, and any relevant temporal dependencies. This allows such algorithms to automatically extract sleep patterns from rich physiological time series. The model presented here integrates ideas from several successful deep learning models to construct a multi-channel time-series convolutional-deconvolutional neural network. This network was trained using crossentropy loss, and evaluated on a 20% held-out validation set. Hyper-parameters were selected on the AUPRC metric, and training utilized early stopping to prevent overfitting. The resultant model achieved an AUPRC of 0.369 and an AUROC of 0.855 on the final competition test set.","PeriodicalId":215521,"journal":{"name":"2018 Computing in Cardiology Conference (CinC)","volume":"150 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-12-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"117313402","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 13

Predictive Models for Risk Assessment of Worsening Events in Chronic Heart Failure Patients 慢性心力衰竭患者恶化事件风险评估的预测模型

2018 Computing in Cardiology Conference (CinC) Pub Date : 2018-12-30 DOI: 10.22489/CinC.2018.249

Maria Carmela Groccia, D. Lofaro, R. Guido, D. Conforti, A. Sciacqua

{"title":"Predictive Models for Risk Assessment of Worsening Events in Chronic Heart Failure Patients","authors":"Maria Carmela Groccia, D. Lofaro, R. Guido, D. Conforti, A. Sciacqua","doi":"10.22489/CinC.2018.249","DOIUrl":"https://doi.org/10.22489/CinC.2018.249","url":null,"abstract":"This work aims at developing and assessing a machine learning based Knowledge Discovery task for risk prediction of major cardiovascular worsening events in chronic heart failure patients. Clinical data from 50patients with chronic heartfailure was analyzed. For each patient, personal data, different vital and clinical parameters and the presence of cardiovascular worsening events have been stored every three months per two years. We defined the Knowledge Discovery analysis as a predictive task stated as supervised binary classification problem. The class label was defined based on the occurrence or not of cardiovascular worsening events between two consecutive visits. To take into account the temporality of the worsening events, six different temporal weighting strategies, applied to the vital parameters, were tested. Several machine learning algorithms were applied for each strategy obtaining different predictive models. Models performance have been evaluated mainly in term of area under the ROC curve (AUC), and Linear Support Vector Machine got the best performing predictive model. The implemented Knowledge Discovery task have shown to be a reliable tool for support cardiologists for riskpredictions of major cardiovascular worsening events.","PeriodicalId":215521,"journal":{"name":"2018 Computing in Cardiology Conference (CinC)","volume":"89 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-12-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116163678","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 4

Machine Learning Based Classification of Myocardial Infarction Conditions Using Smartphone-Derived Seismo- and Gyrocardiography 基于机器学习的心肌梗死分类使用智能手机衍生的地震和陀螺仪心动图

2018 Computing in Cardiology Conference (CinC) Pub Date : 2018-12-30 DOI: 10.22489/CinC.2018.110

Saeed Mehrang, Mojtaba Jafari Tadi, O. Lahdenoja, M. Kaisti, T. Vasankari, T. Kiviniemi, J. Airaksinen, Mikko Pänkäälä, T. Koivisto

{"title":"Machine Learning Based Classification of Myocardial Infarction Conditions Using Smartphone-Derived Seismo- and Gyrocardiography","authors":"Saeed Mehrang, Mojtaba Jafari Tadi, O. Lahdenoja, M. Kaisti, T. Vasankari, T. Kiviniemi, J. Airaksinen, Mikko Pänkäälä, T. Koivisto","doi":"10.22489/CinC.2018.110","DOIUrl":"https://doi.org/10.22489/CinC.2018.110","url":null,"abstract":"In this paper, we attempt to classify the pre- and post-operation cardiac conditions of ST-elevation myocardial infarction (STEMI) utilizing seismocardiography (SCG) and gyrocardiography (GCG) signals recorded solely by a smartphone. SCG and GCG signals were recorded from 20 MI patients who were admitted to Emergency Department of Turku Hospital. Two measurements were recorded from each subject, one before they proceeded to percutaneous coronary intervention (pre-operation) and one afterwards (post-operation) with an average time interval of 2 days. Noise and artefact removal were applied to the signals and subsequently 25 features were extracted. Two classification algorithms, random forest (RF) and support vector machines (SVM), were deployed to discriminate the two cardiac conditions. Accuracy rates of 74% and 78% were obtained for RF and SVM, respectively. The results indicate that smartphone SCG-GCG based ischaemia analysis has clinical implications that warrants further investigations.","PeriodicalId":215521,"journal":{"name":"2018 Computing in Cardiology Conference (CinC)","volume":"12 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-12-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124686971","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 10

Noninvasive Electrocardiographic Imaging of Scar-Related Ventricular Tachycardia: Association With Magnetic Resonance Scar Imaging 疤痕相关室性心动过速的无创心电图成像:与磁共振疤痕成像的关联

2018 Computing in Cardiology Conference (CinC) Pub Date : 2018-12-30 DOI: 10.22489/CinC.2018.303

Omar A. Gharbia, Susumu Tao, A. Lardo, H. Halperin, Linwei Wang

引用次数: 0

Feature-Based Cardiac Cycle Segmentation in Phonocardiogram Recordings 心音图记录中基于特征的心动周期分割

2018 Computing in Cardiology Conference (CinC) Pub Date : 2018-12-30 DOI: 10.22489/CinC.2018.222

Jussi Taipalmaa, M. Zabihi, S. Kiranyaz, M. Gabbouj

{"title":"Feature-Based Cardiac Cycle Segmentation in Phonocardiogram Recordings","authors":"Jussi Taipalmaa, M. Zabihi, S. Kiranyaz, M. Gabbouj","doi":"10.22489/CinC.2018.222","DOIUrl":"https://doi.org/10.22489/CinC.2018.222","url":null,"abstract":"Phonocardiogram (PCG) conveys crucial information for cardiac health evaluation in ambulatory care and is an essential diagnostic test for heart assessment. Thus, identification and positioning of the first and second heart sound within PCG is a vital step in automatic heart sound analysis. This study proposes a solution for individual cardiac cycle segmentation of PCG recordings. It extracts a rich set of features that are used for the segmentation of each cardiac cycle in a PCG recording by localizing the PCG peaks, S1 and S2. To accomplish this objective, a rich set of 66 discriminative features are selected and extracted from each frame in a PCG recording and several classifiers are evaluated to find out the one that achieves the highest segmentation accuracy. Finally, a post-processing method is proposed to reduce the classification noise and hence improve the segmentation performance Contrary to the earlier methods proposed in the literature, this method is evaluated on one of the largest datasets available consisting of 48 877s PCG recordings. The proposed method has achieved F1-score of 93.45%, and Sensitivity and Specificity values of 94.23% and 98.16% respectively. Moreover, it has been tested on the Pascal benchmark dataset, and has achieved Sensitivity and Specificity values of 96.42% and 98.12%, respectively.","PeriodicalId":215521,"journal":{"name":"2018 Computing in Cardiology Conference (CinC)","volume":"45 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-12-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129693118","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

High Frequency QRS Analysis From Orthogonal Leads 正交导联的高频QRS分析

2018 Computing in Cardiology Conference (CinC) Pub Date : 2018-12-30 DOI: 10.22489/CinC.2018.051

J. Halámek, P. Leinveber, M. Malik, G. Schmidt, F. Plesinger, M. Matejkova, J. Lipoldova, P. Jurák

{"title":"High Frequency QRS Analysis From Orthogonal Leads","authors":"J. Halámek, P. Leinveber, M. Malik, G. Schmidt, F. Plesinger, M. Matejkova, J. Lipoldova, P. Jurák","doi":"10.22489/CinC.2018.051","DOIUrl":"https://doi.org/10.22489/CinC.2018.051","url":null,"abstract":"We analysed high frequency averaged QRS (HFQRS) in orthogonal leads and different passbands. Three groups of subjects were compared: healthy subjects, ischemic heart disease (IHD) and dilated cardiomyopathy (DCM) patients. Among the IHD group, those with heart failure (HF) symptoms were identified. Investigated parameters included HFQRS maximal amplitude, HFQRS power, and HFQRS fragmentation based on normalized length of the HFQRS line. The study aimed at assessing (1) group differences in relation to the passband, lead, and parameter, and (2) the reproducibility of parameters. Results: Significant differences were found between healthy subjects and IHD or DCM in all parameters and passbands. Some singularities of significance existed between IHD and DCM. Significant differences were also found between IHD sub-groups with and without HF symptoms, and these existed over more frequency bands. Conclusion: HFQRS parameters are frequency dependent and this dependency should be tested to eliminate singularities in statistical significances. Differences between groups with or without HF symptoms were found mainly at higher passbands, regardless of deterioration of reproducibility. Lead X appeared to be the lead with maximal differences between groups.","PeriodicalId":215521,"journal":{"name":"2018 Computing in Cardiology Conference (CinC)","volume":"22 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-12-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128040166","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Cardiac Fibrosis Detection Applying Machine Learning Techniques to Standard 12-Lead ECG 应用机器学习技术检测标准12导联心电图的心脏纤维化

2018 Computing in Cardiology Conference (CinC) Pub Date : 2018-12-30 DOI: 10.22489/CinC.2018.174

F. Melgarejo-Meseguer, F. Gimeno-Blanes, J. Rojo-álvarez, M. Salar-Alcaraz, J. Gimeno-Blanes, A. García-Alberola

{"title":"Cardiac Fibrosis Detection Applying Machine Learning Techniques to Standard 12-Lead ECG","authors":"F. Melgarejo-Meseguer, F. Gimeno-Blanes, J. Rojo-álvarez, M. Salar-Alcaraz, J. Gimeno-Blanes, A. García-Alberola","doi":"10.22489/CinC.2018.174","DOIUrl":"https://doi.org/10.22489/CinC.2018.174","url":null,"abstract":"Hypertrophic cardiomyopathy (HCM) is a myocardial disorder that affects 0.2% of the population and it is genetically transmitted. Several ECG findings have been related to the presence of fibrosis in other cardiac diseases, but data for HCM in this setting are lacking. Our hypothesis is that fibrosis affects the electrical cardiac propagation in patients with HCM in a relatively specific way and that this effect may be detected with suitable postprocessing applied to the ECG signals. We used 43 standard 12-lead ECGs from patients with previous clinical diagnosis of HCM. Principal Component Analysis (PCA) was applied by combining the ECG-leads oriented to different anatomic regions, hence assessing the potential fibrosis effects in the resulting leads for postprocessing convenience. Linear classifier of Support Vector Machine type were used with several statistics extracted from the resulting PCA-components, including normalized power, standard deviation, kurtosis, skewness, and local maxima. Results reached 75.0% sensitivity, 80.0% specificity, 85.7% positive predictive value, 66.7% negative predictive value, and 76.9% accuracy in our database. There is evidence that myocardial fibrosis can be detected in patients with HCM by postprocessing their ECG signals.","PeriodicalId":215521,"journal":{"name":"2018 Computing in Cardiology Conference (CinC)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-12-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131097153","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 1

A Computationally Efficient Method to Quantify the Biometric Properties of Ventricular Repolarization Irregularities in Healthy and Diseased Human Subjects 一种计算有效的方法来量化健康和患病人类受试者心室复极不规则性的生物特征

2018 Computing in Cardiology Conference (CinC) Pub Date : 2018-12-30 DOI: 10.22489/CinC.2018.080

J. Palhalmi

引用次数: 0

Automatic Generation of Bi-Ventricular Models of Cardiac Electrophysiology for Patient Specific Personalization Using Non-Invasive Recordings 自动生成的双心室模型的心脏电生理的病人具体个性化使用无创记录

2018 Computing in Cardiology Conference (CinC) Pub Date : 2018-12-30 DOI: 10.22489/CinC.2018.265

K. Gillette, A. Prassl, J. Bayer, E. Vigmond, A. Neic, G. Plank

{"title":"Automatic Generation of Bi-Ventricular Models of Cardiac Electrophysiology for Patient Specific Personalization Using Non-Invasive Recordings","authors":"K. Gillette, A. Prassl, J. Bayer, E. Vigmond, A. Neic, G. Plank","doi":"10.22489/CinC.2018.265","DOIUrl":"https://doi.org/10.22489/CinC.2018.265","url":null,"abstract":"Introduction: Personalized in silico models of cardiac electrophysiology based on non-invasive recordings, such as body surface potential maps, are considered of pivotal importance in clinical modeling applications. Efficient, automated workflows are desired to construct patientspecific models for clinical use. Objective: We aimed to develop an automated workflow for the generation of a parameterizable cardiac EP model capable of simulating body surface potential maps independent of user interaction. Methods: A cardiac bi-ventricular model with torso was segmented and meshed from clinical MRI scans. Universal ventricular coordinates were computed for userindependent definition of fibers, a fast conducting endocardial layer, and earliest activation on the endocardium. The extracellular epicardial potential distribution was simulated and projected to the torso surface to acquire a body surface potential map. Results: Total model generation from segmentation required approximately 2 hours. Automatized simulation of a single depolarization sequence required approximately 30 minutes using a forward element method implementation. Discussion: The proposed workflow integrated recentlydeveloped technologies to generate a parameterizable cardiac EP model within clinical time scales.","PeriodicalId":215521,"journal":{"name":"2018 Computing in Cardiology Conference (CinC)","volume":"9 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-12-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121099312","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 7

Validation of Activation Recovery Interval in Structurally Normal Human Ventricles by Optical Mapping 结构正常人脑室激活恢复间隔的光学映射验证

2018 Computing in Cardiology Conference (CinC) Pub Date : 2018-12-30 DOI: 10.22489/CinC.2018.132

Peter Langfield, J. Duchâteau, R. Walton, F. Sacher, J. Rogier, L. Labrousse, F. Brette, M. Hocini, M. Haïssaguerre, O. Bernus, E. Vigmond

{"title":"Validation of Activation Recovery Interval in Structurally Normal Human Ventricles by Optical Mapping","authors":"Peter Langfield, J. Duchâteau, R. Walton, F. Sacher, J. Rogier, L. Labrousse, F. Brette, M. Hocini, M. Haïssaguerre, O. Bernus, E. Vigmond","doi":"10.22489/CinC.2018.132","DOIUrl":"https://doi.org/10.22489/CinC.2018.132","url":null,"abstract":"Background: A large Dispersion of Repolarization (DoR) is associated with an increased arrhythmogenic risk. This can be measured clinically by calculating the Activation Recovery Interval (ARI) to estimate Action Potential Duration (APD). However, the ability of ARI to accurately predict APD dispersion in patients with repolarization abnormality has not been determined. Objective: Compare ARI calculated from patients with optical mapping of human hearts to establish the validity of ARI as a surrogate for APD. Methods: Optical mapping (OM) was performed on the left ventricles of 4 explanted human hearts. APD and repolarization times were measured endo- and epicardially on the anterior of the LV. Electroanatomic mapping was performed with CARTO over the entire endo- and epicardial surfaces of 3 patients. Activation and repolarization were calculated, dispersion of ARI was measured. Results: APD and ARI were consistent between mapping methods over most of the sub-regions studied. Epicardium ARI dispersion was consistently higher than that of the endocardium in both OM and CARTO datasets. Conclusion: APD distribution, and consequently DoR, agree between mapping methods. Measuring DoR by ARI accurately assesses the underlying repolarization abnormalities in patients.","PeriodicalId":215521,"journal":{"name":"2018 Computing in Cardiology Conference (CinC)","volume":"18 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2018-12-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122310778","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0