"Victor Gonçalves Marques, A. Gharaviri, Simone Pezzuto, A. Auricchio, P. Bonizzi, S. Zeemering, U. Schotten
In some persistent atrial fibrillation (AF) patients, localized drivers may sustain AF and thus could represent possible ablation targets. In this work, we test in silico the feasibility of locating AF drivers from high-density electrode grid catheter mapping. A volumetric 3D atrial model was used to simulate 8 AF episodes driven by a stable reentry around a region of scar tissue (5 left atrium [LA], 3 right atrium [RA]). Sequential mapping in 1s segments was performed with a high-density electrode grid, starting from 20 uniformly distributed regions (12 LA, 8 RA). Conduction velocities estimated for each AF cycle were used to obtain temporal and directional parameters of the propagation. Trajectories of connected activation times were used to detect reentries or radial spread of activations. If no pattern was detected, the electrode array was moved in 5mm steps upstream of the propagation direction. The algorithm obtained accuracy, sensitivity, and precision of 87.2%, 23.4%, and 56.3% for reentries and 87.0%, 8.5%, and 26.8% for radial spread of activations, respectively. Reentries were found in average within 1.52 steps15 mm from the initial position of the grid. The results indicate that propagation patterns may be sufficient to track localized AF drivers sequentially during high-density mapping.
{"title":"Tracking of Atrial Fibrillation Drivers Based on Propagation Patterns: An In-Silico Study","authors":"\"Victor Gonçalves Marques, A. Gharaviri, Simone Pezzuto, A. Auricchio, P. Bonizzi, S. Zeemering, U. Schotten","doi":"10.22489/CinC.2022.174","DOIUrl":"https://doi.org/10.22489/CinC.2022.174","url":null,"abstract":"In some persistent atrial fibrillation (AF) patients, localized drivers may sustain AF and thus could represent possible ablation targets. In this work, we test in silico the feasibility of locating AF drivers from high-density electrode grid catheter mapping. A volumetric 3D atrial model was used to simulate 8 AF episodes driven by a stable reentry around a region of scar tissue (5 left atrium [LA], 3 right atrium [RA]). Sequential mapping in 1s segments was performed with a high-density electrode grid, starting from 20 uniformly distributed regions (12 LA, 8 RA). Conduction velocities estimated for each AF cycle were used to obtain temporal and directional parameters of the propagation. Trajectories of connected activation times were used to detect reentries or radial spread of activations. If no pattern was detected, the electrode array was moved in 5mm steps upstream of the propagation direction. The algorithm obtained accuracy, sensitivity, and precision of 87.2%, 23.4%, and 56.3% for reentries and 87.0%, 8.5%, and 26.8% for radial spread of activations, respectively. Reentries were found in average within 1.52 steps15 mm from the initial position of the grid. The results indicate that propagation patterns may be sufficient to track localized AF drivers sequentially during high-density mapping.","PeriodicalId":117840,"journal":{"name":"2022 Computing in Cardiology (CinC)","volume":"296 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-09-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125754748","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}
Maximiliano Mollura, Cristian Drudi, Li-wei H. Lehman, Riccardo Barbieri
Introduction: Fluids and vasopressors represent the cornerstone for hemodynamic instability management in the intensive care unit (ICU). However, optimal personalized treatments strategies are still missing. Goal: To evaluate the ability of a reduced set of cardiovascular features in determining optimal actions with a reinforcement learning approach. Methods: Data were extracted from the MIMIC-III database Patients' trajectories were modeled as a Markov decision process with a target reward based on 90-day mortality. Performances with a reduced set of cardiovascular features (CARDIO), including heart rate, systolic and diastolic blood pressure, shock index, and oxygen saturation were compared with a random policy model (RANDOM) and a model with a full set of 48 clinical variables including physiologic, laboratory measurement, and ventilation parameters (FULL). Results: The CARDIa model achieved the highest results with a 95% lower bound (LB) of estimated policy value equal to 96.17 compared with the 86.00 obtained from the FULL model and 82.62 from the RANDOM policy model. Conclusions: Results show that cardiovascular features and ongoing treatments have the potential to determine the optimal dosage of fluids and vasopressors for septic patients when using reinforcement learning tools for the development of medical decision support systems.
{"title":"Optimal Fluid And Vasopressor Interventions In Septic ICU Patients Through Reinforcement Learning Model","authors":"Maximiliano Mollura, Cristian Drudi, Li-wei H. Lehman, Riccardo Barbieri","doi":"10.22489/CinC.2022.189","DOIUrl":"https://doi.org/10.22489/CinC.2022.189","url":null,"abstract":"Introduction: Fluids and vasopressors represent the cornerstone for hemodynamic instability management in the intensive care unit (ICU). However, optimal personalized treatments strategies are still missing. Goal: To evaluate the ability of a reduced set of cardiovascular features in determining optimal actions with a reinforcement learning approach. Methods: Data were extracted from the MIMIC-III database Patients' trajectories were modeled as a Markov decision process with a target reward based on 90-day mortality. Performances with a reduced set of cardiovascular features (CARDIO), including heart rate, systolic and diastolic blood pressure, shock index, and oxygen saturation were compared with a random policy model (RANDOM) and a model with a full set of 48 clinical variables including physiologic, laboratory measurement, and ventilation parameters (FULL). Results: The CARDIa model achieved the highest results with a 95% lower bound (LB) of estimated policy value equal to 96.17 compared with the 86.00 obtained from the FULL model and 82.62 from the RANDOM policy model. Conclusions: Results show that cardiovascular features and ongoing treatments have the potential to determine the optimal dosage of fluids and vasopressors for septic patients when using reinforcement learning tools for the development of medical decision support systems.","PeriodicalId":117840,"journal":{"name":"2022 Computing in Cardiology (CinC)","volume":"21 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-09-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126399124","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}
C. Nagel, Johannes Osypka, L. Unger, D. Nairn, A. Luik, R. Wakili, O. Doessel, A. Loewe
Fibrotic atrial cardiomyopathy is characterized by a replacement of healthy atrial tissue with diffuse patches exhibiting slow electrical conduction properties and altered myocardial tissue structure, which provides a substrate for the maintenance of reentrant activity during atrial fibrillation (AF). Therefore, an early detection of atrial fibrosis could be a valuable risk marker for new-onset AF episodes to select asymptomatic subjects for screening, allowing for timely intervention and optimizing therapy planning. We examined the potential of estimating the fibrotic tissue volume fraction in the atria based on P waves of the 12-lead ECG recorded in sinus rhythm in a quantitative and non-invasive way. Our dataset comprised 68,282 P waves from healthy subjects and 42,227 P waves from AF patients with low voltage areas in the atria, as well as 642,400 simulated P waves of a virtual cohort derived from statistical shape models with different extents of the left atrial myocardium replaced by fibrosis. The root mean squared error for estimating the left atrial fibrotic volume fraction on a clinical test set with a neural network trained on features extracted from simulated and clinical P waves was 16.57 %. Our study shows that the 12-lead ECG contains valuable information on atrial tissue structure. As such it could potentially be employed as an inexpensive and widely available tool to support AF risk stratification in clinical practice.
{"title":"Improving Clinical ECG-based Atrial Fibrosis Quantification With Neural Networks Through In Silico P waves From an Extensive Virtual Patient Cohort","authors":"C. Nagel, Johannes Osypka, L. Unger, D. Nairn, A. Luik, R. Wakili, O. Doessel, A. Loewe","doi":"10.22489/CinC.2022.124","DOIUrl":"https://doi.org/10.22489/CinC.2022.124","url":null,"abstract":"Fibrotic atrial cardiomyopathy is characterized by a replacement of healthy atrial tissue with diffuse patches exhibiting slow electrical conduction properties and altered myocardial tissue structure, which provides a substrate for the maintenance of reentrant activity during atrial fibrillation (AF). Therefore, an early detection of atrial fibrosis could be a valuable risk marker for new-onset AF episodes to select asymptomatic subjects for screening, allowing for timely intervention and optimizing therapy planning. We examined the potential of estimating the fibrotic tissue volume fraction in the atria based on P waves of the 12-lead ECG recorded in sinus rhythm in a quantitative and non-invasive way. Our dataset comprised 68,282 P waves from healthy subjects and 42,227 P waves from AF patients with low voltage areas in the atria, as well as 642,400 simulated P waves of a virtual cohort derived from statistical shape models with different extents of the left atrial myocardium replaced by fibrosis. The root mean squared error for estimating the left atrial fibrotic volume fraction on a clinical test set with a neural network trained on features extracted from simulated and clinical P waves was 16.57 %. Our study shows that the 12-lead ECG contains valuable information on atrial tissue structure. As such it could potentially be employed as an inexpensive and widely available tool to support AF risk stratification in clinical practice.","PeriodicalId":117840,"journal":{"name":"2022 Computing in Cardiology (CinC)","volume":"5 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-09-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126595661","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}
"Ricardo Maximiliano Rosales, Konstantinos A. Mountris, M. Doblaré, M. Mazo, Emilio L. Pueyo
Depolarization sequences triggering mechanical contraction of the heart are largely determined by the cardiac conduction system $(CS)$. Many biophysical models of cardiac electrophysiology still have poor representations of the $CS$. This work proposes a semiautomatic method for the generation of an anatomically-realistic porcine $CS$ that reproduces ventricular activation properties in swine computational models. Personalized swine biventricular models were built from magnetic resonance images. Electrical propagation was described by the monodomain model. The $CS$ was defined from manually-determined anatomic landmarks using geodesic paths and a fractal tree algorithm. Two $CS$ distributions were defined, one restricted to the subendocardium and another one by performing a subendo-to-intramyocardium projection based on histological porcine data. Depolarization patterns as well as left ventricular transmural and inter-ventricular delays were assessed to describe ventricular activation by the two $CS$ distributions. The electrical excitations calculated using the two $CS$ distributions were in good agreement with reported activation patterns. The pig-specific subendo-intramyocardial $CS$ led to improved reproduction of experimental activation delays in ventricular endocardium and epicardium.
{"title":"Ventricular Conduction System Modeling for Electrophysiological Simulation of the Porcine Heart","authors":"\"Ricardo Maximiliano Rosales, Konstantinos A. Mountris, M. Doblaré, M. Mazo, Emilio L. Pueyo","doi":"10.22489/CinC.2022.030","DOIUrl":"https://doi.org/10.22489/CinC.2022.030","url":null,"abstract":"Depolarization sequences triggering mechanical contraction of the heart are largely determined by the cardiac conduction system $(CS)$. Many biophysical models of cardiac electrophysiology still have poor representations of the $CS$. This work proposes a semiautomatic method for the generation of an anatomically-realistic porcine $CS$ that reproduces ventricular activation properties in swine computational models. Personalized swine biventricular models were built from magnetic resonance images. Electrical propagation was described by the monodomain model. The $CS$ was defined from manually-determined anatomic landmarks using geodesic paths and a fractal tree algorithm. Two $CS$ distributions were defined, one restricted to the subendocardium and another one by performing a subendo-to-intramyocardium projection based on histological porcine data. Depolarization patterns as well as left ventricular transmural and inter-ventricular delays were assessed to describe ventricular activation by the two $CS$ distributions. The electrical excitations calculated using the two $CS$ distributions were in good agreement with reported activation patterns. The pig-specific subendo-intramyocardial $CS$ led to improved reproduction of experimental activation delays in ventricular endocardium and epicardium.","PeriodicalId":117840,"journal":{"name":"2022 Computing in Cardiology (CinC)","volume":"595 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-09-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127517342","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}
Ismail Elnaggar, Tero Hurnanen, Jonas Sandelin, O. Lahdenoja, A. Airola, M. Kaisti, T. Koivisto
Bed based ballistocardiography (BCG) is a prime candidate for at home and nighttime monitoring especially in the growing elderly population because co-operation from the user is not required to be able to record signals. One issue with BCG is that the signal quality has intra-and inter-person variability based on factors such as age, gender, body position, and motion artifacts, making it challenging to accurately measure heart rate. A rule-based algorithm which considers all eight available BCG channels simultaneously from a given time epoch was developed using continuous wavelet transform (CWT) to extract the localized time-frequency representation of each epoch and then an averaging method was applied across the different scales of the CWT to produce a 1-dimensional array. Autocorrelation was then applied to this array to produce a heart rate estimate based on the lag between the autocorrelation maximum and the first side peak. This method does not require identification of individual heart beats to estimate heart rate and does not require annotated training data. This model produces an average mean absolute error (MAE) of 1.09 bpm across 40 subjects when compared to heart rate derived from ECG. This method produces competitive results without the need for annotated training data, which can be challenging to collect.
{"title":"Multichannel Bed Based Ballistocardiography Heart Rate Estimation Using Continuous Wavelet Transforms and Autocorrelation","authors":"Ismail Elnaggar, Tero Hurnanen, Jonas Sandelin, O. Lahdenoja, A. Airola, M. Kaisti, T. Koivisto","doi":"10.22489/CinC.2022.364","DOIUrl":"https://doi.org/10.22489/CinC.2022.364","url":null,"abstract":"Bed based ballistocardiography (BCG) is a prime candidate for at home and nighttime monitoring especially in the growing elderly population because co-operation from the user is not required to be able to record signals. One issue with BCG is that the signal quality has intra-and inter-person variability based on factors such as age, gender, body position, and motion artifacts, making it challenging to accurately measure heart rate. A rule-based algorithm which considers all eight available BCG channels simultaneously from a given time epoch was developed using continuous wavelet transform (CWT) to extract the localized time-frequency representation of each epoch and then an averaging method was applied across the different scales of the CWT to produce a 1-dimensional array. Autocorrelation was then applied to this array to produce a heart rate estimate based on the lag between the autocorrelation maximum and the first side peak. This method does not require identification of individual heart beats to estimate heart rate and does not require annotated training data. This model produces an average mean absolute error (MAE) of 1.09 bpm across 40 subjects when compared to heart rate derived from ECG. This method produces competitive results without the need for annotated training data, which can be challenging to collect.","PeriodicalId":117840,"journal":{"name":"2022 Computing in Cardiology (CinC)","volume":"70 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-09-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127260947","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}
The field of automated auscultation has been growing in popularity in the past decade due to manual auscultation being a challenging task requiring years of training. Many efforts in the field focus on achieving high accuracy, with confident, albeit sometimes wrong, classifiers. Such model over-confidence is especially dangerous in health-care setting. Leveraging the release of the new heart sound dataset as a part of PhysioNet 2022 challenge, we explored a novel murmur detection methodology using uncertainty-aware tandem learning. To separate unknown samples and detect heart sounds with murmur present, we developed two binary classifiers, under the assumption that training two models to solve simpler tasks could improve the overall sensitivity. First, we used a support vector machine for identification of unknown samples, followed by a Deep Neural Network (DNN) for prediction of murmur. In addition, we implemented uncertainty estimation in DNN using Monte Carlo dropouts for further eliminating any samples that should be labelled as unknown. Our team mobihealth achieved 63% and 69% sensitivity and specificity of murmur, scoring 0.467 (ranked 34th out of 40) and 11032 (ranked 25th out of 39) on the hidden validation set and 0.374 (ranked 40th out of 40) and 18754 (ranked 39th out of 39) on the hidden testing set during the challenge for murmur and outcome prediction tasks, respectively.
{"title":"Towards Uncertainty-Aware Murmur Detection in Heart Sounds via Tandem Learning","authors":"E. Bondareva, Tong Xia, Jing Han, Cecilia Mascolo","doi":"10.22489/CinC.2022.234","DOIUrl":"https://doi.org/10.22489/CinC.2022.234","url":null,"abstract":"The field of automated auscultation has been growing in popularity in the past decade due to manual auscultation being a challenging task requiring years of training. Many efforts in the field focus on achieving high accuracy, with confident, albeit sometimes wrong, classifiers. Such model over-confidence is especially dangerous in health-care setting. Leveraging the release of the new heart sound dataset as a part of PhysioNet 2022 challenge, we explored a novel murmur detection methodology using uncertainty-aware tandem learning. To separate unknown samples and detect heart sounds with murmur present, we developed two binary classifiers, under the assumption that training two models to solve simpler tasks could improve the overall sensitivity. First, we used a support vector machine for identification of unknown samples, followed by a Deep Neural Network (DNN) for prediction of murmur. In addition, we implemented uncertainty estimation in DNN using Monte Carlo dropouts for further eliminating any samples that should be labelled as unknown. Our team mobihealth achieved 63% and 69% sensitivity and specificity of murmur, scoring 0.467 (ranked 34th out of 40) and 11032 (ranked 25th out of 39) on the hidden validation set and 0.374 (ranked 40th out of 40) and 18754 (ranked 39th out of 39) on the hidden testing set during the challenge for murmur and outcome prediction tasks, respectively.","PeriodicalId":117840,"journal":{"name":"2022 Computing in Cardiology (CinC)","volume":"85 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-09-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122529925","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}
Sven Festag, Gideon Stein, Tim Büchner, M. Shadaydeh, J. Denzler, C. Spreckelsen
We, the team UKJ_FSU, propose a deep learning system for the prediction of congenital heart diseases. Our method is able to predict the clinical outcomes (normal, abnormal) of patients as well as to identify heart murmur (present, absent, unclear) based on phonocardiograms recorded at different auscultation locations. The system we propose is an ensemble of four temporal convolutional networks with identical topologies, each specialized in identifying murmurs and predicting patient outcome from a phonocardiogram taken at one specific auscultation location. Their intermediate outputs are augmented by the manually ascertained patient features such as age group, sex, height, and weight. The outputs of the four networks are combined to form a single final decision as demanded by the rules of the George B. Moody PhysioNet Challenge 2022. On the first task of this challenge, the murmur detection, our model reached a weighted accuracy of 0.567 with respect to the validation set. On the outcome prediction task (second task) the ensemble led to a mean outcome cost of 10679 on the same set. By focusing on the clinical outcome prediction and tuning some of the hyper-parameters only for this task, our model reached a cost score of 12373 on the official test set (rank 13 of 39). The same model scored a weighted accuracy of 0.458 regarding the murmur detection on the test set (rank 37 of 40).
我们，UKJ_FSU团队，提出了一个用于预测先天性心脏病的深度学习系统。我们的方法能够预测患者的临床结果(正常、异常)，并根据不同听诊位置记录的心音图识别心脏杂音(存在、不存在、不清楚)。我们提出的系统是四个具有相同拓扑结构的时间卷积网络的集合，每个网络都专门用于识别杂音并预测患者在特定听诊位置的心音图结果。他们的中间输出通过人工确定的患者特征(如年龄、性别、身高和体重)得到增强。根据George B. Moody PhysioNet挑战赛2022的规则要求，将四个网络的输出组合成一个最终决定。对于这个挑战的第一个任务，杂音检测，我们的模型相对于验证集达到了0.567的加权精度。在结果预测任务(第二个任务)上，集成导致同一集合上的平均结果成本为10679。通过专注于临床结果预测并仅针对该任务调整一些超参数，我们的模型在官方测试集中达到了12373的成本分数(39个中的第13位)。同样的模型在测试集上的杂音检测的加权精度为0.458(40个中的第37位)。
{"title":"Outcome Prediction and Murmur Detection in Sets of Phonocardiograms by a Deep Learning-Based Ensemble Approach","authors":"Sven Festag, Gideon Stein, Tim Büchner, M. Shadaydeh, J. Denzler, C. Spreckelsen","doi":"10.22489/CinC.2022.137","DOIUrl":"https://doi.org/10.22489/CinC.2022.137","url":null,"abstract":"We, the team UKJ_FSU, propose a deep learning system for the prediction of congenital heart diseases. Our method is able to predict the clinical outcomes (normal, abnormal) of patients as well as to identify heart murmur (present, absent, unclear) based on phonocardiograms recorded at different auscultation locations. The system we propose is an ensemble of four temporal convolutional networks with identical topologies, each specialized in identifying murmurs and predicting patient outcome from a phonocardiogram taken at one specific auscultation location. Their intermediate outputs are augmented by the manually ascertained patient features such as age group, sex, height, and weight. The outputs of the four networks are combined to form a single final decision as demanded by the rules of the George B. Moody PhysioNet Challenge 2022. On the first task of this challenge, the murmur detection, our model reached a weighted accuracy of 0.567 with respect to the validation set. On the outcome prediction task (second task) the ensemble led to a mean outcome cost of 10679 on the same set. By focusing on the clinical outcome prediction and tuning some of the hyper-parameters only for this task, our model reached a cost score of 12373 on the official test set (rank 13 of 39). The same model scored a weighted accuracy of 0.458 regarding the murmur detection on the test set (rank 37 of 40).","PeriodicalId":117840,"journal":{"name":"2022 Computing in Cardiology (CinC)","volume":"9 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-09-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127878702","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}
Jiayi Yan, Hanshuang Xie, Huaiyu Zhu, Yamin Liu, Fan Wu, Yun Pan
Accurate P wave detection is important for arrhythmia diagnosis, e.g. P wave absence or P duration for atrial fibrillation diagnosis and other atrial arrhythmias. Phasor transform is an effective method for ECG fiducial points delineation. It maps each ECG sample into a phasor to enhance slight variations and preserves morphology and magnitude characteristics. In this paper, we optimized the automatic P wave delineation method based on phasor transform in four aspects, i.e., signal denoising, wave localization, candidate points detection, and optimal points selection. In our experiments, the length of the search window and the degree of phasor transform were established through various trials. Especially, along with zero-crossing points of the phasor signal, intersections of the phasor signal and the original ECG signal are obtained as candidates, which make the most contribution to delineation results. For validation, the QT Database with 3194 P wave annotations from 105 records of two leads is adopted. As a result, we reached F1 scores of 94.67% and 93.56% with detection error rates (DERs) of 10.80% and 13.06% for P wave onset and offset points detection, respectively. The F 1 score and DER for P peak detection under a tolerance of 75 ms were 95.33% and 9.46%, respectively, which outperforms other reproducible works and their combinations.
{"title":"An Optimized Automatic P Wave Delineation Method Based on Phasor Transform","authors":"Jiayi Yan, Hanshuang Xie, Huaiyu Zhu, Yamin Liu, Fan Wu, Yun Pan","doi":"10.22489/CinC.2022.122","DOIUrl":"https://doi.org/10.22489/CinC.2022.122","url":null,"abstract":"Accurate P wave detection is important for arrhythmia diagnosis, e.g. P wave absence or P duration for atrial fibrillation diagnosis and other atrial arrhythmias. Phasor transform is an effective method for ECG fiducial points delineation. It maps each ECG sample into a phasor to enhance slight variations and preserves morphology and magnitude characteristics. In this paper, we optimized the automatic P wave delineation method based on phasor transform in four aspects, i.e., signal denoising, wave localization, candidate points detection, and optimal points selection. In our experiments, the length of the search window and the degree of phasor transform were established through various trials. Especially, along with zero-crossing points of the phasor signal, intersections of the phasor signal and the original ECG signal are obtained as candidates, which make the most contribution to delineation results. For validation, the QT Database with 3194 P wave annotations from 105 records of two leads is adopted. As a result, we reached F1 scores of 94.67% and 93.56% with detection error rates (DERs) of 10.80% and 13.06% for P wave onset and offset points detection, respectively. The F 1 score and DER for P peak detection under a tolerance of 75 ms were 95.33% and 9.46%, respectively, which outperforms other reproducible works and their combinations.","PeriodicalId":117840,"journal":{"name":"2022 Computing in Cardiology (CinC)","volume":"4 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-09-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126512012","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}
Phase contrast MRI can provide a comprehensive analysis for the hemodynamic changes in the aorta which is useful for the diagnosis of several aortic diseases. However, an initial step of accurate segmentation of the aorta is necessary, which is usually a time-consuming and subjective step. Several methods have been proposed to automate this step using classical segmentation methods and recently deep learning models. Most of the current models combine the magnitude and phase images equally across all time phases which hinder the potential advantage that the frames of higher velocity might have more useful information compared to the low velocity frames. In this work, we propose a novel adaptive combination model that combines the output probability maps of both the magnitude and phase models based on an initial velocity estimation as a surrogate for the confidence level in the velocity images. We applied our model on the 2D-PC images of 215 patients and our results shows an accuracy of 87% for the magnitude images, 68% for the velocity images, 87.1% for the combined images, and 89.1 % for our proposed combination model.
{"title":"Improving Aorta Segmentation from Phase Contrast MRI Using Adaptive Velocity-Dependent Weighting on the Deep Learning Output for Magnitude and Phase Images","authors":"Mohamed A Elbayumi, S. Saraya, T. Basha","doi":"10.22489/CinC.2022.244","DOIUrl":"https://doi.org/10.22489/CinC.2022.244","url":null,"abstract":"Phase contrast MRI can provide a comprehensive analysis for the hemodynamic changes in the aorta which is useful for the diagnosis of several aortic diseases. However, an initial step of accurate segmentation of the aorta is necessary, which is usually a time-consuming and subjective step. Several methods have been proposed to automate this step using classical segmentation methods and recently deep learning models. Most of the current models combine the magnitude and phase images equally across all time phases which hinder the potential advantage that the frames of higher velocity might have more useful information compared to the low velocity frames. In this work, we propose a novel adaptive combination model that combines the output probability maps of both the magnitude and phase models based on an initial velocity estimation as a surrogate for the confidence level in the velocity images. We applied our model on the 2D-PC images of 215 patients and our results shows an accuracy of 87% for the magnitude images, 68% for the velocity images, 87.1% for the combined images, and 89.1 % for our proposed combination model.","PeriodicalId":117840,"journal":{"name":"2022 Computing in Cardiology (CinC)","volume":"2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-09-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131243949","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}
Cardiovascular disease (CVD) patients were targeted from cardiology department in this study to segregate who had stenosis and also identify the principal diseased coronary artery using PPG and ECG signals. After pre-processing these signals, dicrotic notch region of PPG and S-T segment of ECG, within each cardiac cycle was extracted as templates. A new fused segment was generated from two templates by a proposed algorithm. Utilizing statistics on three templates we defined the term Coronary Health Index (CHI) to evaluate the status of coronary arteries. Setting CHI thresholding values, healthy and stenosed artery were differentiated. Using CHI values from patients with stenosis, the classification of arteries (LAD, RCA, and LCx) was performed using Graph Attentive Convolution Network. Among 408 CVD patients 256 had occlusion in either LAD or RCA or LCx. Binary classification among presence and absence of stenosis was carried out with 0.92 accuracy, 0.91 recall, 0.91 precision, 0.90 specificity, and 0.92 F-score. Identification of stenosed artery was measured with Kappa score (0.89) and Youden's J statistic value (0.84). AUC(0.93) and AP(0.92) values from ROC and PRC curves, respectively. This derived CHI could be able to study stenosis in non-invasive, easy and cost-effective manner.
{"title":"Coronary Health Index (CHI) as A Determinant for Arterial Stenosis, Derived Using PPG and ECG Signals","authors":"Poulomi Pal, M. Mahadevappa","doi":"10.22489/CinC.2022.316","DOIUrl":"https://doi.org/10.22489/CinC.2022.316","url":null,"abstract":"Cardiovascular disease (CVD) patients were targeted from cardiology department in this study to segregate who had stenosis and also identify the principal diseased coronary artery using PPG and ECG signals. After pre-processing these signals, dicrotic notch region of PPG and S-T segment of ECG, within each cardiac cycle was extracted as templates. A new fused segment was generated from two templates by a proposed algorithm. Utilizing statistics on three templates we defined the term Coronary Health Index (CHI) to evaluate the status of coronary arteries. Setting CHI thresholding values, healthy and stenosed artery were differentiated. Using CHI values from patients with stenosis, the classification of arteries (LAD, RCA, and LCx) was performed using Graph Attentive Convolution Network. Among 408 CVD patients 256 had occlusion in either LAD or RCA or LCx. Binary classification among presence and absence of stenosis was carried out with 0.92 accuracy, 0.91 recall, 0.91 precision, 0.90 specificity, and 0.92 F-score. Identification of stenosed artery was measured with Kappa score (0.89) and Youden's J statistic value (0.84). AUC(0.93) and AP(0.92) values from ROC and PRC curves, respectively. This derived CHI could be able to study stenosis in non-invasive, easy and cost-effective manner.","PeriodicalId":117840,"journal":{"name":"2022 Computing in Cardiology (CinC)","volume":"194 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-09-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131405815","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}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：