Sex-Specific Patterns of Mortality Predictors Among Patients Undergoing Cardiac Resynchronization Therapy: A Machine Learning Approach

Abstract

Background: The relative importance of variables explaining sex-related differences in outcomes is scarcely explored in patients undergoing cardiac resynchronization therapy (CRT). We sought to implement and evaluate machine learning (ML) algorithms for the prediction of 1- and 3-year all-cause mortality in CRT patients. We also aimed to assess the sex-specific differences in predictors of mortality utilizing ML.Methods: Using a retrospective registry of 2,191 CRT patients, ML models were implemented in 6 partially overlapping patient subsets (all patients, females, or males with 1- or 3-year follow-up). Each cohort was randomly split into training (80%) and test sets (20%). After hyperparameter tuning in the training sets, the best performing algorithm was evaluated in the test sets. Model discrimination was quantified using the area under the receiver-operating characteristic curves (AUC). The most important predictors were identified using the permutation feature importances method.Results: Conditional inference random forest exhibited the best performance with AUCs of 0.728 (0.645–0.802) and 0.732 (0.681–0.784) for the prediction of 1- and 3-year mortality, respectively. Etiology of heart failure, NYHA class, left ventricular ejection fraction, and QRS morphology had higher predictive power, whereas hemoglobin was less important in females compared to males. The importance of atrial fibrillation and age increased, while the importance of serum creatinine decreased from 1- to 3-year follow-up in both sexes.Conclusions: Using ML techniques in combination with easily obtainable clinical features, our models effectively predicted 1- and 3-year all-cause mortality in CRT patients. Sex-specific patterns of predictors were identified, showing a dynamic variation over time.

Machine Learning Prediction of Cardiac Resynchronisation Therapy Response From Combination of Clinical and Model-Driven Data

“…Despite of intensive research performed in the field, the fraction of patients with low response to the therapy remains as high as 30–50% depending on which criteria are used for assessing CRT outcome. New artificial intellegence and ML based approaches to data analysis have been extensively used in attempts to increase the accuracy of patient differentiation (Kalscheur et al, 2018 ; Feeny et al, 2019 , 2020 ; Tokodi et al, 2021 ). Computational models based on clinical data are also employed to identify mechanisms responsible for the poor efficacy and develop approaches improving CRT outcomes (Lumens et al, 2015 ; Huntjens et al, 2018 ; Lee et al, 2018 ; Isotani et al, 2020 ).…”

“…Recent studies using ML techniques have achieved impressive results in preoperative clinical data analysis for selecting patients for CRT. Predictive models have been developed to estimate mortality or hospitalization risks from the baseline clinical parameters (Kalscheur et al, 2018 ; Tokodi et al, 2020 , 2021 ), to assess improvements in EF based on baseline indices and analysis of medical records (Hu et al, 2019 ) and to stratify patients by an unsupervised learning approach implementing ECG traces (Cikes et al, 2019 ) and electrocardiography (Feeny et al, 2020 ). In a recent study (Feeny et al, 2019 ), Feeny and co-authors using supervised ML approaches selected 9 clinical features (QRS morphology, QRS duration, New York Heart Association CHF classification, LV EF and end-diastolic diameter (EDD), sex, ischemic cardiomyopathy, atrial fibrillation, and epicardial LV lead) that were sufficient to predict patient improvement with fairly high accuracy.…”

Machine Learning prediction of cardiac resynchronisation therapy response from combination of clinical and model-driven data

“…Despite of intensive research performed in the field, the fraction of patients with low response to the therapy remains as high as 30-50% depending on which criteria are used for assessing CRT outcome. New artificial intellegence and ML based approaches to data analysis have been extensively used in attempts to increase the accuracy of patient differentiation [7,9,13,12]. Computational models based on clinical data are also employed to identify mechanisms responsible for the poor efficacy and develop approaches improving CRT outcomes [14,34,35,36].…”

“…We also compared the accuracy of ML classifiers built on the hybrid dataset for CRT response defined by 5, 10, 15% LV EF improvement and by coupled EF10 and ESV15 criteria (see Tables 8,9 in the Supplementary Materials). For every response definition, our best classifiers demonstrate improved performance as compared with all clinical and ML predictors reported in [13].…”

“…Recent studies using ML techniques have achieved impressive results in preoperative clinical data analysis for selecting patients for CRT. Predictive models have been developed to estimate mortality or hospitalization risks from the baseline clinical parameters [7,8,9], to assess improvements in EF based on baseline indices and analysis of medical records [10] and to stratify patients by an unsupervised learning approach implementing ECG traces [11] and electrocardiography [12]. In a recent study [13], Feeny and co-authors showed 9 features (QRS morphology, QRS duration, New York Heart Association CHF classification, LV EF and end-diastolic diameter (EDD), sex, ischemic cardiomyopathy, atrial fibrillation, and epicardial LV lead) were sufficient to predict patient improvement with fairly high accuracy.…”