Try a new search

Format these results:

Searched for:

person:huy07

in-biosketch:yes

Total Results:

13


Development and external validation of a dynamic risk score for early prediction of cardiogenic shock in cardiac intensive care units using machine learning

Hu, Yuxuan; Lui, Albert; Goldstein, Mark; Sudarshan, Mukund; Tinsay, Andrea; Tsui, Cindy; Maidman, Samuel D; Medamana, John; Jethani, Neil; Puli, Aahlad; Nguy, Vuthy; Aphinyanaphongs, Yindalon; Kiefer, Nicholas; Smilowitz, Nathaniel R; Horowitz, James; Ahuja, Tania; Fishman, Glenn I; Hochman, Judith; Katz, Stuart; Bernard, Samuel; Ranganath, Rajesh
BACKGROUND:Myocardial infarction and heart failure are major cardiovascular diseases that affect millions of people in the US with the morbidity and mortality being highest among patients who develop cardiogenic shock. Early recognition of cardiogenic shock allows prompt implementation of treatment measures. Our objective is to develop a new dynamic risk score, called CShock, to improve early detection of cardiogenic shock in cardiac intensive care unit (ICU). METHODS:We developed and externally validated a deep learning-based risk stratification tool, called CShock, for patients admitted into the cardiac ICU with acute decompensated heart failure and/or myocardial infarction to predict onset of cardiogenic shock. We prepared a cardiac ICU dataset using MIMIC-III database by annotating with physician adjudicated outcomes. This dataset that consisted of 1500 patients with 204 having cardiogenic/mixed shock was then used to train CShock. The features used to train the model for CShock included patient demographics, cardiac ICU admission diagnoses, routinely measured laboratory values and vital signs, and relevant features manually extracted from echocardiogram and left heart catheterization reports. We externally validated the risk model on the New York University (NYU) Langone Health cardiac ICU database that was also annotated with physician adjudicated outcomes. The external validation cohort consisted of 131 patients with 25 patients experiencing cardiogenic/mixed shock. RESULTS:CShock achieved an area under the receiver operator characteristic curve (AUROC) of 0.821 (95% CI 0.792-0.850). CShock was externally validated in the more contemporary NYU cohort and achieved an AUROC of 0.800 (95% CI 0.717-0.884), demonstrating its generalizability in other cardiac ICUs. Having an elevated heart rate is most predictive of cardiogenic shock development based on Shapley values. The other top ten predictors are having an admission diagnosis of myocardial infarction with ST-segment elevation, having an admission diagnosis of acute decompensated heart failure, Braden Scale, Glasgow Coma Scale, Blood urea nitrogen, Systolic blood pressure, Serum chloride, Serum sodium, and Arterial blood pH. CONCLUSIONS:The novel CShock score has the potential to provide automated detection and early warning for cardiogenic shock and improve the outcomes for the millions of patients who suffer from myocardial infarction and heart failure.
PMID: 38518758
ISSN: 2048-8734
CID: 5640892

Deep learning models for electrocardiograms are susceptible to adversarial attack

Han, Xintian; Hu, Yuxuan; Foschini, Luca; Chinitz, Larry; Jankelson, Lior; Ranganath, Rajesh
Electrocardiogram (ECG) acquisition is increasingly widespread in medical and commercial devices, necessitating the development of automated interpretation strategies. Recently, deep neural networks have been used to automatically analyze ECG tracings and outperform physicians in detecting certain rhythm irregularities1. However, deep learning classifiers are susceptible to adversarial examples, which are created from raw data to fool the classifier such that it assigns the example to the wrong class, but which are undetectable to the human eye2,3. Adversarial examples have also been created for medical-related tasks4,5. However, traditional attack methods to create adversarial examples do not extend directly to ECG signals, as such methods introduce square-wave artefacts that are not physiologically plausible. Here we develop a method to construct smoothed adversarial examples for ECG tracings that are invisible to human expert evaluation and show that a deep learning model for arrhythmia detection from single-lead ECG6 is vulnerable to this type of attack. Moreover, we provide a general technique for collating and perturbing known adversarial examples to create multiple new ones. The susceptibility of deep learning ECG algorithms to adversarial misclassification implies that care should be taken when evaluating these models on ECGs that may have been altered, particularly when incentives for causing misclassification exist.
PMID: 32152582
ISSN: 1546-170x
CID: 4349692

Optimizing cardiac resynchronization therapy to minimize ATP consumption heterogeneity throughout the left ventricle: a simulation analysis using a canine heart failure model

Hu, Yuxuan; Gurev, Viatcheslav; Constantino, Jason; Trayanova, Natalia
BACKGROUND:Cardiac resynchronization therapy (CRT) has been demonstrated to lead to restoration of oxygen consumption homogeneity throughout the left ventricle (LV), which is important for long-term reverse remodeling of the ventricles. However, research has focused exclusively on identifying the LV pacing sites that led to acute hemodynamic improvements. It remains unclear whether there exist LV pacing sites that could both improve the hemodynamics and result in ATP consumption homogeneity throughout the LV, thus maximizing both CRT short-term and long-term benefits. OBJECTIVE:The purpose of this study was to demonstrate the feasibility of optimizing CRT pacing locations to achieve maximal improvement in both ATPCTHI (an ATP consumption heterogeneity index) and stroke work. METHODS:We used an magnetic resonance image-based electromechanical model of the dyssynchronous heart failure (DHF) canine ventricles. ATPCTHI and stroke work improvement were determined for each of 34 CRT pacing sites evenly spaced over the LV epicardium. RESULTS:Results demonstrated the feasibility of determining the optimal LV pacing site that achieves simultaneous maximum improvements in ATPCTHI and stroke work. The optimal LV CRT pacing sites in the DHF canine ventricles were located midway between apex and base. The improvement in ATPCTHI decreased more rapidly with the distance from the optimal sites compared to stroke work improvement. CRT from the optimal sites homogenized ATP consumption by increasing septal ATP consumption and decreasing that of the lateral wall. CONCLUSION/CONCLUSIONS:Simulation results using a canine heart failure model demonstrated that CRT can be optimized to achieve improvements in both ATPCTHI and stroke work.
PMCID:4049272
PMID: 24657430
ISSN: 1556-3871
CID: 5272832

Efficient preloading of the ventricles by a properly timed atrial contraction underlies stroke work improvement in the acute response to cardiac resynchronization therapy

Hu, Yuxuan; Gurev, Viatcheslav; Constantino, Jason; Trayanova, Natalia
BACKGROUND:The acute response to cardiac resynchronization therapy (CRT) has been shown to be due to 3 mechanisms: resynchronization of ventricular contraction, efficient preloading of the ventricles by a properly timed atrial contraction, and mitral regurgitation reduction. However, the contribution of each of the 3 mechanisms to the acute response to CRT, specifically stroke work improvement, has not been quantified. OBJECTIVE:To use a magnetic resonance image-based anatomically accurate 3-dimensional model of failing canine ventricular electromechanics to quantify the contribution of each of the 3 mechanisms to stroke work improvement and identify the predominant mechanisms. METHODS:An MRI-based electromechanical model of the failing canine ventricles assembled previously by our group was further developed and modified. Three different protocols were used to dissect the contribution of each of the 3 mechanisms to stroke work improvement. RESULTS:Resynchronization of ventricular contraction did not lead to a significant stroke work improvement. Efficient preloading of the ventricles by a properly timed atrial contraction was the predominant mechanism underlying stroke work improvement. Stroke work improvement peaked at an intermediate atrioventricular delay, as it allowed ventricular filling by atrial contraction to occur at a low diastolic left ventricular pressure but also provided adequate time for ventricular filling before ventricular contraction. Reduction of mitral regurgitation by CRT led to stroke work worsening instead of improvement. CONCLUSION/CONCLUSIONS:Efficient preloading of the ventricles by a properly timed atrial contraction is responsible for a significant stroke work improvement in the acute CRT response.
PMCID:3852188
PMID: 23928177
ISSN: 1556-3871
CID: 5272822

Mechanistic insight into prolonged electromechanical delay in dyssynchronous heart failure: a computational study

Constantino, Jason; Hu, Yuxuan; Lardo, Albert C; Trayanova, Natalia A
In addition to the left bundle branch block type of electrical activation, there are further remodeling aspects associated with dyssynchronous heart failure (HF) that affect the electromechanical behavior of the heart. Among the most important are altered ventricular structure (both geometry and fiber/sheet orientation), abnormal Ca(2+) handling, slowed conduction, and reduced wall stiffness. In dyssynchronous HF, the electromechanical delay (EMD), the time interval between local myocyte depolarization and myofiber shortening onset, is prolonged. However, the contributions of the four major HF remodeling aspects in extending EMD in the dyssynchronous failing heart remain unknown. The goal of this study was to determine the individual and combined contributions of HF-induced remodeling aspects to EMD prolongation. We used MRI-based models of dyssynchronous nonfailing and HF canine electromechanics and constructed additional models in which varying combinations of the four remodeling aspects were represented. A left bundle branch block electrical activation sequence was simulated in all models. The simulation results revealed that deranged Ca(2+) handling is the primary culprit in extending EMD in dyssynchronous HF, with the other aspects of remodeling contributing insignificantly. Mechanistically, we found that abnormal Ca(2+) handling in dyssynchronous HF slows myofiber shortening velocity at the early-activated septum and depresses both myofiber shortening and stretch rate at the late-activated lateral wall. These changes in myofiber dynamics delay the onset of myofiber shortening, thus giving rise to prolonged EMD in dyssynchronous HF.
PMCID:3798786
PMID: 23934857
ISSN: 1522-1539
CID: 5263252

Effects of mechano-electric feedback on scroll wave stability in human ventricular fibrillation

Hu, Yuxuan; Gurev, Viatcheslav; Constantino, Jason; Bayer, Jason D; Trayanova, Natalia A
Recruitment of stretch-activated channels, one of the mechanisms of mechano-electric feedback, has been shown to influence the stability of scroll waves, the waves that underlie reentrant arrhythmias. However, a comprehensive study to examine the effects of recruitment of stretch-activated channels with different reversal potentials and conductances on scroll wave stability has not been undertaken; the mechanisms by which stretch-activated channel opening alters scroll wave stability are also not well understood. The goals of this study were to test the hypothesis that recruitment of stretch-activated channels affects scroll wave stability differently depending on stretch-activated channel reversal potential and channel conductance, and to uncover the relevant mechanisms underlying the observed behaviors. We developed a strongly-coupled model of human ventricular electromechanics that incorporated human ventricular geometry and fiber and sheet orientation reconstructed from MR and diffusion tensor MR images. Since a wide variety of reversal potentials and channel conductances have been reported for stretch-activated channels, two reversal potentials, -60 mV and -10 mV, and a range of channel conductances (0 to 0.07 mS/µF) were implemented. Opening of stretch-activated channels with a reversal potential of -60 mV diminished scroll wave breakup for all values of conductances by flattening heterogeneously the action potential duration restitution curve. Opening of stretch-activated channels with a reversal potential of -10 mV inhibited partially scroll wave breakup at low conductance values (from 0.02 to 0.04 mS/µF) by flattening heterogeneously the conduction velocity restitution relation. For large conductance values (>0.05 mS/µF), recruitment of stretch-activated channels with a reversal potential of -10 mV did not reduce the likelihood of scroll wave breakup because Na channel inactivation in regions of large stretch led to conduction block, which counteracted the increased scroll wave stability due to an overall flatter conduction velocity restitution.
PMCID:3616032
PMID: 23573245
ISSN: 1932-6203
CID: 5263242

Cardiac electromechanical models

Chapter by: Trayanova, N; Constantio, J; Hu, Yuxuan
in: Cardiac Electrophysiology: From Cell to Bedside by
pp. 361-370
ISBN: 9781455728565
CID: 5272882

Computational cardiology: how computer simulations could be used to develop new therapies and advance existing ones

Trayanova, Natalia A; O'Hara, Thomas; Bayer, Jason D; Boyle, Patrick M; McDowell, Kathleen S; Constantino, Jason; Arevalo, Hermenegild J; Hu, Yuxuan; Vadakkumpadan, Fijoy
This article reviews the latest developments in computational cardiology. It focuses on the contribution of cardiac modelling to the development of new therapies as well as the advancement of existing ones for cardiac arrhythmias and pump dysfunction. Reviewed are cardiac modelling efforts aimed at advancing and optimizing existent therapies for cardiac disease (defibrillation, ablation of ventricular tachycardia, and cardiac resynchronization therapy) and at suggesting novel treatments, including novel molecular targets, as well as efforts to use cardiac models in stratification of patients likely to benefit from a given therapy, and the use of models in diagnostic procedures.
PMCID:3482619
PMID: 23104919
ISSN: 1532-2092
CID: 5272852

A computational approach to understanding the cardiac electromechanical activation sequence in the normal and failing heart, with translation to the clinical practice of CRT

Constantino, Jason; Hu, Yuxuan; Trayanova, Natalia A
Cardiac resynchronization therapy (CRT) is an established clinical treatment modality that aims to recoordinate contraction of the heart in dyssynchrous heart failure (DHF) patients. Although CRT reduces morbidity and mortality, a significant percentage of CRT patients fail to respond to the therapy, reflecting an insufficient understanding of the electromechanical activity of the DHF heart. Computational models of ventricular electromechanics are now poised to fill this knowledge gap and provide a comprehensive characterization of the spatiotemporal electromechanical interactions in the normal and DHF heart. The objective of this paper is to demonstrate the powerful utility of computational models of ventricular electromechanics in characterizing the relationship between the electrical and mechanical activation in the DHF heart, and how this understanding can be utilized to devise better CRT strategies. The computational research presented here exploits knowledge regarding the three dimensional distribution of the electromechanical delay, defined as the time interval between myocyte depolarization and onset of myofiber shortening, in determining the optimal location of the LV pacing electrode for CRT. The simulation results shown here also suggest utilizing myocardial efficiency and regional energy consumption as a guide to optimize CRT.
PMCID:3484227
PMID: 22884712
ISSN: 1873-1732
CID: 5263232

Unstable QT interval dynamics precedes ventricular tachycardia onset in patients with acute myocardial infarction: a novel approach to detect instability in QT interval dynamics from clinical ECG

Chen, Xiaozhong; Hu, Yuxuan; Fetics, Barry J; Berger, Ronald D; Trayanova, Natalia A
BACKGROUND:Instability in ventricular repolarization in the presence of premature activations (PA) plays an important role in arrhythmogenesis. However, such instability cannot be detected clinically. This study developed a methodology for detecting QT interval (QTI) dynamics instability from the ECG and explored the contribution of PA and QTI instability to ventricular tachycardia (VT) onset. METHODS AND RESULTS/RESULTS:To examine the contribution of PAs and QTI instability to VT onset, ECGs of 24 patients with acute myocardial infarction, 12 of whom had sustained VT (VT) and 12 nonsustained VT (NSVT), were used. From each patient ECG, 2 10-minute-long ECG recordings were extracted, 1 right before VT onset (onset epoch) and 1 at least 1 hour before it (control epoch). To ascertain how PA affects QTI dynamics stability, pseudo-ECGs were calculated from an MRI-based human ventricular model. Clinical and pseudo-ECGs were subdivided into 1-minute recordings (minECGs). QTI dynamics stability of each minECG was assessed with a novel approach. Frequency of PAs (f(PA)) and the number of minECGs with unstable QTI dynamics (N(us)) were determined for each patient. In the VT group, f(PA) and N(us) of the onset epoch were larger than in control. Positive regression relationships between f(PA) and N(us) were identified in both groups. The simulations showed that both f(PA) and the PA degree of prematurity contribute to QTI dynamics instability. CONCLUSIONS:Increased PA frequency and QTI dynamics instability precede VT onset in patients with acute myocardial infarction, as determined by novel methodology for detecting instability in QTI dynamics from clinical ECGs.
PMCID:3247646
PMID: 21841208
ISSN: 1941-3084
CID: 5272842