Identification of important risk factors for all-cause mortality of acquired long QT syndrome patients using random survival forests and non-negative matrix factorization

Cheng, Chen; Zhou, Jiandong; Yu, Haozhe; Zhang, Qingpeng; Gao, Lianjun; Yin, Xiaomeng; Dong, Yan; Lin, Yajuan; Li, Daobo; Yang, Yiheng; Wang, Yunsong; Tse, Gary; Xia, Yunlong

doi:10.1016/j.hrthm.2020.10.022

Cited by 25 publications

(15 citation statements)

References 32 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This non-negativity makes the resulting matrices easier to inspect and makes the interpretation easier for real-world applications, such as identification of hidden stages in embryonic stem cell differentiation, 38 DNA methylation profiling of human cardiac tissue 39 and unsupervised cf-mRNA transcriptome decomposition. 40 NMF was recently used by our teams for mortality risk prediction in acquired long QT syndrome patients 41 and arrhythmic risk stratification in BrS patients. 42 In this study, these latent factors were then used as inputs an RSF model.…”

Section: Discussionmentioning

confidence: 99%

Territory-wide cohort study of Brugada syndrome in Hong Kong: predictors of long-term outcomes using random survival forests and non-negative matrix factorisation

Lee¹,

Zhou

et al. 2021

Open Heart

Self Cite

View full text Add to dashboard Cite

ObjectivesBrugada syndrome (BrS) is an ion channelopathy that predisposes affected patients to spontaneous ventricular tachycardia/fibrillation (VT/VF) and sudden cardiac death. The aim of this study is to examine the predictive factors of spontaneous VT/VF.MethodsThis was a territory-wide retrospective cohort study of patients diagnosed with BrS between 1997 and 2019. The primary outcome was spontaneous VT/VF. Cox regression was used to identify significant risk predictors. Non-linear interactions between variables (latent patterns) were extracted using non-negative matrix factorisation (NMF) and used as inputs into the random survival forest (RSF) model.ResultsThis study included 516 consecutive BrS patients (mean age of initial presentation=50±16 years, male=92%) with a median follow-up of 86 (IQR: 45–118) months. The cohort was divided into subgroups based on initial disease manifestation: asymptomatic (n=314), syncope (n=159) or VT/VF (n=41). Annualised event rates per person-year were 1.70%, 0.05% and 0.01% for the VT/VF, syncope and asymptomatic subgroups, respectively. Multivariate Cox regression analysis revealed initial presentation of VT/VF (HR=24.0, 95% CI=1.21 to 479, p=0.037) and SD of P-wave duration (HR=1.07, 95% CI=1.00 to 1.13, p=0.044) were significant predictors. The NMF-RSF showed the best predictive performance compared with RSF and Cox regression models (precision: 0.87 vs 0.83 vs. 0.76, recall: 0.89 vs. 0.85 vs 0.73, F1-score: 0.88 vs 0.84 vs 0.74).ConclusionsClinical history, electrocardiographic markers and investigation results provide important information for risk stratification. Machine learning techniques using NMF and RSF significantly improves overall risk stratification performance.

show abstract

Section: Discussionmentioning

confidence: 99%

Territory-wide cohort study of Brugada syndrome in Hong Kong: predictors of long-term outcomes using random survival forests and non-negative matrix factorisation

Lee¹,

Zhou

et al. 2021

Open Heart

Self Cite

View full text Add to dashboard Cite

show abstract

“…Both ECG indices have been reported to provide predictive value for arrhythmic risk stratification in the clinical context of acquired long QT syndrome for humans (115). Indeed, in a Chinese cohort of patients with acquired long QT syndrome, random survival forest analysis identified hypokalaemia as the second most important variable after cancer for predicting all-cause mortality (116).…”

Section: Hypokalaemia In the Clinical Contextmentioning

confidence: 99%

Arrhythmogenic Mechanisms in Hypokalaemia: Insights From Pre-clinical Models

Tse

Cheung

et al. 2021

Front. Cardiovasc. Med.

Self Cite

View full text Add to dashboard Cite

Potassium is the predominant intracellular cation, with its extracellular concentrations maintained between 3. 5 and 5 mM. Among the different potassium disorders, hypokalaemia is a common clinical condition that increases the risk of life-threatening ventricular arrhythmias. This review aims to consolidate pre-clinical findings on the electrophysiological mechanisms underlying hypokalaemia-induced arrhythmogenicity. Both triggers and substrates are required for the induction and maintenance of ventricular arrhythmias. Triggered activity can arise from either early afterdepolarizations (EADs) or delayed afterdepolarizations (DADs). Action potential duration (APD) prolongation can predispose to EADs, whereas intracellular Ca2+ overload can cause both EADs and DADs. Substrates on the other hand can either be static or dynamic. Static substrates include action potential triangulation, non-uniform APD prolongation, abnormal transmural repolarization gradients, reduced conduction velocity (CV), shortened effective refractory period (ERP), reduced excitation wavelength (CV × ERP) and increased critical intervals for re-excitation (APD–ERP). In contrast, dynamic substrates comprise increased amplitude of APD alternans, steeper APD restitution gradients, transient reversal of transmural repolarization gradients and impaired depolarization-repolarization coupling. The following review article will summarize the molecular mechanisms that generate these electrophysiological abnormalities and subsequent arrhythmogenesis.

show abstract

“…43 RSF-based models have been applied to enhance risk stratification in different clinical settings, including diabetes. [44][45][46][47][48] However, RSF model has been criticized for the bias due to favouring covariates with many split-points. 49 In our study, the CISF model was used for time-toevent survival data analysis in predicting AMI and non-AMI SCD, 11,12 which were shown to shown superior predictive performance compared to RSF and multivariate Cox models.…”

Section: Discussionmentioning

confidence: 99%

“…RSF models, as extensions of classification and regression trees and random forests, have been identified as alternative survival data analysis methods when the proportional hazard assumption is violated 43 . RSF‐based models have been applied to enhance risk stratification in different clinical settings, including diabetes 44‐48 . However, RSF model has been criticized for the bias due to favouring covariates with many split‐points 49 .…”

Section: Discussionmentioning

confidence: 99%

Predictive scores for identifying patients with type 2 diabetes mellitus at risk of acute myocardial infarction and sudden cardiac death

Lee¹,

Zhou

Guo

et al. 2021

Endocrino Diabet & Metabol

Self Cite

View full text Add to dashboard Cite

Introduction:The present study evaluated the application of incorporating non-linear J/U-shaped relationships between mean HbA1c and cholesterol levels into risk scores for predicting acute myocardial infarction (AMI) and non-AMI-related sudden cardiac death (SCD) respectively, amongst patients with type 2 diabetes mellitus.Methods: This was a territory-wide cohort study of patients with type 2 diabetes mellitus above the age 40 and free from prior AMI and SCD, with or without prescriptions of anti-diabetic agents between January 1st, 2009 to December 31st, 2009 at government-funded hospitals and clinics in Hong Kong. Patients recruited were followed up until 31 December 2019 or their date of death. Risk scores were developed for predicting incident AMI and non-AMI-related SCD. The performance of conditional inference survival forest (CISF) model compared to that of random survival forests (RSF) model and multivariate Cox model. Results:This study included 261 308 patients (age = 66.0 ± 11.8 years old, male = 47.6%, follow-up duration = 3552 ± 1201 days, diabetes duration = 4.77 ± 2.29 years). Mean HbA1c and low high-density lipoprotein-cholesterol (HDL-C) were significant predictors of AMI on multivariate Cox regression. Mean HbA1c was linearly associated with AMI, whilst HDL-C was inversely associated with AMI. Mean HbA1c and total cholesterol were significant multivariate predictors with a J-shaped relationship with non-AMI-related SCD. The AMI and SCD risk scores had an area under the curve (AUC) of 0.666 (95% confidence interval (CI) = [0.662, 0.669]) and 0.677 (95% CI = [0.673, 0.682]), respectively. CISF significantly improves prediction performance of both outcomes compared to RSF and multivariate Cox models. Conclusion:A holistic combination of demographic, clinical and laboratory indices can be used for the risk stratification of patients with type 2 diabetes mellitus for AMI and SCD.This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.

show abstract

Identification of important risk factors for all-cause mortality of acquired long QT syndrome patients using random survival forests and non-negative matrix factorization

Cited by 25 publications

References 32 publications

Territory-wide cohort study of Brugada syndrome in Hong Kong: predictors of long-term outcomes using random survival forests and non-negative matrix factorisation

Territory-wide cohort study of Brugada syndrome in Hong Kong: predictors of long-term outcomes using random survival forests and non-negative matrix factorisation

Arrhythmogenic Mechanisms in Hypokalaemia: Insights From Pre-clinical Models

Predictive scores for identifying patients with type 2 diabetes mellitus at risk of acute myocardial infarction and sudden cardiac death

Contact Info

Product

Resources

About