Dataset Information

A method using deep learning to discover new predictors from left-ventricular mechanical dyssynchrony for CRT response.

ABSTRACT:

Background

Studies have shown that the conventional parameters characterizing left ventricular mechanical dyssynchrony (LVMD) measured on gated SPECT myocardial perfusion imaging (MPI) have their own statistical limitations in predicting cardiac resynchronization therapy (CRT) response. The purpose of this study is to discover new predictors from the polarmaps of LVMD by deep learning to help select heart failure patients with a high likelihood of response to CRT.

Methods

One hundred and fifty-seven patients who underwent rest gated SPECT MPI were enrolled in this study. CRT response was defined as an increase in left ventricular ejection fraction (LVEF) > 5% at 6 [Formula: see text] 1 month follow up. The autoencoder (AE) technique, an unsupervised deep learning method, was applied to the polarmaps of LVMD to extract new predictors characterizing LVMD. Pearson correlation analysis was used to explain the relationships between new predictors and existing clinical parameters. Patients from the IAEA VISION-CRT trial were used for an external validation. Heatmaps were used to interpret the AE-extracted feature.

Results

Complete data were obtained in 130 patients, and 68.5% of them were classified as CRT responders. After variable selection by feature importance ranking and correlation analysis, one AE-extracted LVMD predictor was included in the statistical analysis. This new AE-extracted LVMD predictor showed statistical significance in the univariate (OR 2.00, P = .026) and multivariate (OR 1.11, P = .021) analyses, respectively. Moreover, the new AE-extracted LVMD predictor not only had incremental value over PBW and significant clinical variables, including QRS duration and left ventricular end-systolic volume (AUC 0.74 vs 0.72, LH 7.33, P = .007), but also showed encouraging predictive value in the 165 patients from the IAEA VISION-CRT trial (P < .1). The heatmaps for calculation of the AE-extracted predictor showed higher weights on the anterior, lateral, and inferior myocardial walls, which are recommended as LV pacing sites in clinical practice.

Conclusions

AE techniques have significant value in the discovery of new clinical predictors. The new AE-extracted LVMD predictor extracted from the baseline gated SPECT MPI has the potential to improve the prediction of CRT response.

SUBMITTER: He Z

PROVIDER: S-EPMC10961110 | biostudies-literature | 2023 Feb

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

A method using deep learning to discover new predictors from left-ventricular mechanical dyssynchrony for CRT response.

He Zhuo Z Zhang Xinwei X Zhao Chen C Ling Xing X Malhotra Saurabh S Qian Zhiyong Z Wang Yao Y Hou Xiaofeng X Zou Jiangang J Zhou Weihua W

Journal of nuclear cardiology : official publication of the American Society of Nuclear Cardiology 20220801 1

<h4>Background</h4>Studies have shown that the conventional parameters characterizing left ventricular mechanical dyssynchrony (LVMD) measured on gated SPECT myocardial perfusion imaging (MPI) have their own statistical limitations in predicting cardiac resynchronization therapy (CRT) response. The purpose of this study is to discover new predictors from the polarmaps of LVMD by deep learning to help select heart failure patients with a high likelihood of response to CRT.<h4>Methods</h4>One hund ...[more]

PMID: 35915327

Similar Datasets

Project description:BackgroundCardiac resynchronization therapy (CRT) patients with different pathophysiology may influence mechanical dyssynchrony and get different ventricular resynchronization and clinical outcomes.MethodsNinety-two dilated cardiomyopathy (DCM) and fifty ischemic cardiomyopathy (ICM) patients with gated single-photon emission computed tomography (SPECT) myocardial perfusion imaging (MPI) were included in this retrospective study. Patients were classified based on the concordance between the left ventricular (LV) lead and the latest contraction or relaxation position. If the LV lead was located on or adjacent to both the latest contraction and relaxation position, the patient was categorized into the both match group; if the LV lead was located on or adjacent to the latest contraction or relaxation position, the patient was classified into the one match group; if the LV lead was located on or adjacent to neither the latest contraction nor relaxation position, the patient was categorized to the neither group. CRT response was defined as [Formula: see text] improvement of LV ejection fraction at the 6-month follow-up. Variables with P < .05 in the univariate analysis were included in the stepwise multivariate model.ResultsDuring the follow-up period, 58.7% (54 of 92) for DCM patients and 54% (27 of 50) for ICM patients were CRT responders. The univariate analysis and stepwise multivariate analysis showed that QRS duration, systolic phase bandwidth (PBW), diastolic PBW, diastolic phase histogram standard deviation (PSD), and left ventricular mechanical dyssynchrony (LVMD) concordance were independent predictors of CRT response in DCM patients; diabetes mellitus and left ventricular end-systolic volume were significantly associated with CRT response in ICM patients. The intra-group comparison revealed that the CRT response rate was significantly different in the both match group of DCM (N = 18, 94%) and ICM (N = 24, 62%) patients (P = .016). However, there was no significant difference between DCM and ICM in the one match and neither group. For the inter-group comparison, Kruskal-Wallis H-test revealed that CRT response was significantly different in all the groups of DCM patients (P < .001), but not in ICM patients (P = .383).ConclusionsCompared with ICM patients, systolic PBW, diastolic PBW and PSD have better predictive and prognostic values for the CRT response in DCM patients. Placing the LV lead in or adjacent to the latest contraction and relaxation position can improve the clinical outcomes of DCM patients, but it does not apply to ICM patients.

Project description:AimsLeft ventricular mechanical dyssynchrony has been described in heart failure with preserved ejection fraction (HFpEF), but its prognostic significance is not known.Methods and resultsOf 3445 patients with HFpEF enrolled in the Treatment of Preserved Cardiac Function Heart Failure with an Aldosterone Antagonist (TOPCAT) trial, dyssynchrony analysis was performed on 424 patients (12%) by multiple speckle tracking echocardiography strain-based criteria. The primary dyssynchrony analysis was the standard deviation of the time to peak longitudinal strain (SD T2P LS). Cox proportional hazards models assessed the association of dyssynchrony with the composite outcome of cardiovascular death or heart failure hospitalization. Mean age was 70 ± 10 years, LVEF was 60 ± 8%, and QRS duration was 101 ± 27 ms. Worse dyssynchrony, reflected in SD T2P LS, was associated with wider QRS, prior myocardial infarction, larger LV volume and mass, and worse systolic (lower LVEF and global longitudinal strain) and diastolic (lower e' and higher E/e') function. During a median follow-up of 2.6 (interquartile range 1.5-3.8) years, 107 patients experienced the composite outcome. Worse dyssynchrony was associated with the composite outcome in unadjusted analysis [hazard ratio (HR) 1.04, 95% confidence interval (CI) 1.01-1.07; P = 0.021, per 10 ms increase], but not after adjusting for clinical characteristics, or after further adjustment for LVEF, AF, NYHA class, stroke, heart rate, creatinine, haematocrit, and QRS duration (HR 1.03, 95% CI 0.99-1.06; P = 0.16, per 10 ms increase).ConclusionWorse LV mechanical dyssynchrony, assessed by speckle tracking echocardiography, is not an independent predictor of adverse outcomes in HFpEF, suggesting that mechanical dyssynchrony is unlikely to be an important mechanism underlying this syndrome. These findings warrant validation in an independent study specifically designed to assess the prognostic utility of mechanical dyssynchrony in HFpEF.Trial registrationNCT00094302.

Project description:BackgroundComplete bundle branch block in individuals without structural heart disease is known as isolated complete bundle branch block. Isolated complete left bundle branch block (CLBBB) is correlated with ventricular dysfunction secondary to dyssynchrony; however, few studies have investigated isolated complete right bundle branch block (CRBBB), which was previously considered benign but was recently found to be associated with adverse cardiovascular outcomes. This study aimed to evaluate cardiac mechanical synchrony, and systolic and diastolic function in patients with isolated CRBBB and compare cardiac synchrony and function to patients with isolated CLBBB.MethodsThis cross-sectional study was conducted at The First Hospital of China Medical University in Shenyang, China, from 2020 to 2021. A total of 44 isolated CRBBB patients, 44 isolated CLBBB patients, and 42 healthy subjects were enrolled in the study. Transthoracic echocardiography was performed in all subjects. Synchrony parameters, including the mechanical dispersion of the right ventricle [the standard deviation of time to the peak longitudinal strain of six right ventricular (RV) segments] and atrioventricular dyssynchrony parameter [the ratio of left ventricular (LV) diastolic filling time to the time interval between two adjacent R waves (RR interval) measured by tissue Doppler imaging]. RV and LV function were assessed by the global longitudinal strain (GLS) of six RV segments and 18 LV segments, and the ratio of the peak early diastolic flow velocity to annular velocity (E/e') of the tricuspid valve and mitral valve. Statistical analyses were performed, including an analysis of variance, Pearson correlation analysis, and linear regression analysis.ResultsCompared with the healthy subjects, the mechanical dispersion of the right ventricle was significantly increased, and ventricular function was impaired as evidenced by the decreased RV GLS and LV GLS, and the increased E/e' of the tricuspid valve and mitral valve in the isolated CRBBB patients (all P<0.001). Moreover, compared with the isolated CLBBB patients, the mechanical dispersion of the right ventricle and E/e' of the tricuspid valve were increased, and RV GLS was significantly reduced in the isolated CRBBB patients (all P<0.001). Mechanical dispersion of the right ventricle was independently associated with RV GLS [coefficient, 0.13; 95% confidence interval (CI): 0.004-0.26; P=0.04] in the isolated CRBBB patients. RV GLS (coefficient, 0.10; 95% CI: 0.01-0.20; P=0.03) and the ratio of the LV diastolic filling time to the RR interval measured (coefficient, -0.30; 95% CI: -0.53 to -0.07; P=0.01) were independent factors of LV GLS.ConclusionsThe isolated CRBBB patients had impaired cardiac mechanical synchrony and ventricular function, and more decreased RV synchrony and function than the isolated CLBBB patients. Right intraventricular synchrony was independently associated with RV systolic dysfunction in patients with isolated CRBBB. Atrioventricular synchrony and RV systolic function were independently associated with the LV systolic function. Therefore, comprehensive evaluations of echocardiography results and close monitoring is required for isolated CRBBB patients.

Dataset Information

A method using deep learning to discover new predictors from left-ventricular mechanical dyssynchrony for CRT response.

Background

Methods

Results

Conclusions

Publications

A method using deep learning to discover new predictors from left-ventricular mechanical dyssynchrony for CRT response.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets