Ventricular geometry–regularized QRSd predicts cardiac resynchronization therapy response: machine learning from crosstalk between electrocardiography and echocardiography

Juan Lei, Yi Grace Wang, Luna Bhatta, Jamal Ahmed, Dali Fan, Jingfeng Wang, Kan Liu

    Research output: Contribution to journalArticle

    Abstract

    Up to one-third of patients selected by current guidelines do not respond to cardiac resynchronization therapy (CRT), the aim of this study was to find out novel analytical approaches to improve pre-implantation CRT response prediction. Among 31 pre-implantation features of clinical, laboratory, electrocardiography (ECG), and echocardiography variables in a consecutive cohort of patients receiving a first-time CRT device (CRT-pacemaker or CRT-defibrillator), we developed a machine learning (ML) model with three classification algorithms (support vector machines (SVM), K nearest neighbors, and random subspaces) with the best features combination to predict CRT response. Three categorical variables, left bundle branch block (LBBB), nonischemic cardiomyopathy, and female gender, were independently associated with CRT responses. Among continuous variables, including septal wall thickness, posterior wall thickness, and relative wall thickness (RWT), could regularize ECG QRS duration (QRSd) and significantly enhance the correlation between QRSd and CRT response. The 3 ML algorithms in a total of 38 features combinations constantly recognized that the features combined with QRSd/RWT outperformed the combinations without it. For each of three algorithms, the triplet feature combination of QRSd/RWT, LBBB, and nonischemic cardiomyopathy repeatedly increased the classification rate more than 8%. The best performance for CRT response prediction occurred with SVM model, which proposed actual QRSd/RWT values that favored CRT responses in patients both with and without LBBB. Lower QRSd/RWT values were required for CRT responses in patients with ischemic cardiomyopathy compared to those with non-ischemic cardiomyopathy. ML from ventricular remodeling characteristics–regularized QRSd improves CRT response prediction.

    Original languageEnglish (US)
    JournalInternational Journal of Cardiovascular Imaging
    DOIs
    StatePublished - Jan 1 2019

    Fingerprint

    Cardiac Resynchronization Therapy
    Echocardiography
    Electrocardiography
    Cardiomyopathies
    Bundle-Branch Block
    Machine Learning
    Cardiac Resynchronization Therapy Devices
    Ventricular Remodeling
    Defibrillators
    Guidelines

    Keywords

    • Cardiac resynchronization therapy
    • Classification
    • Machine learning
    • QRS duration
    • Ventricular geometric characteristics

    ASJC Scopus subject areas

    • Radiology Nuclear Medicine and imaging
    • Cardiology and Cardiovascular Medicine

    Cite this

    @article{95fc3a82f4f1483f97e0e950b119fe4a,
    title = "Ventricular geometry–regularized QRSd predicts cardiac resynchronization therapy response: machine learning from crosstalk between electrocardiography and echocardiography",
    abstract = "Up to one-third of patients selected by current guidelines do not respond to cardiac resynchronization therapy (CRT), the aim of this study was to find out novel analytical approaches to improve pre-implantation CRT response prediction. Among 31 pre-implantation features of clinical, laboratory, electrocardiography (ECG), and echocardiography variables in a consecutive cohort of patients receiving a first-time CRT device (CRT-pacemaker or CRT-defibrillator), we developed a machine learning (ML) model with three classification algorithms (support vector machines (SVM), K nearest neighbors, and random subspaces) with the best features combination to predict CRT response. Three categorical variables, left bundle branch block (LBBB), nonischemic cardiomyopathy, and female gender, were independently associated with CRT responses. Among continuous variables, including septal wall thickness, posterior wall thickness, and relative wall thickness (RWT), could regularize ECG QRS duration (QRSd) and significantly enhance the correlation between QRSd and CRT response. The 3 ML algorithms in a total of 38 features combinations constantly recognized that the features combined with QRSd/RWT outperformed the combinations without it. For each of three algorithms, the triplet feature combination of QRSd/RWT, LBBB, and nonischemic cardiomyopathy repeatedly increased the classification rate more than 8{\%}. The best performance for CRT response prediction occurred with SVM model, which proposed actual QRSd/RWT values that favored CRT responses in patients both with and without LBBB. Lower QRSd/RWT values were required for CRT responses in patients with ischemic cardiomyopathy compared to those with non-ischemic cardiomyopathy. ML from ventricular remodeling characteristics–regularized QRSd improves CRT response prediction.",
    keywords = "Cardiac resynchronization therapy, Classification, Machine learning, QRS duration, Ventricular geometric characteristics",
    author = "Juan Lei and Wang, {Yi Grace} and Luna Bhatta and Jamal Ahmed and Dali Fan and Jingfeng Wang and Kan Liu",
    year = "2019",
    month = "1",
    day = "1",
    doi = "10.1007/s10554-019-01545-5",
    language = "English (US)",
    journal = "International Journal of Cardiovascular Imaging",
    issn = "1569-5794",
    publisher = "Springer Netherlands",

    }

    TY - JOUR

    T1 - Ventricular geometry–regularized QRSd predicts cardiac resynchronization therapy response

    T2 - machine learning from crosstalk between electrocardiography and echocardiography

    AU - Lei, Juan

    AU - Wang, Yi Grace

    AU - Bhatta, Luna

    AU - Ahmed, Jamal

    AU - Fan, Dali

    AU - Wang, Jingfeng

    AU - Liu, Kan

    PY - 2019/1/1

    Y1 - 2019/1/1

    N2 - Up to one-third of patients selected by current guidelines do not respond to cardiac resynchronization therapy (CRT), the aim of this study was to find out novel analytical approaches to improve pre-implantation CRT response prediction. Among 31 pre-implantation features of clinical, laboratory, electrocardiography (ECG), and echocardiography variables in a consecutive cohort of patients receiving a first-time CRT device (CRT-pacemaker or CRT-defibrillator), we developed a machine learning (ML) model with three classification algorithms (support vector machines (SVM), K nearest neighbors, and random subspaces) with the best features combination to predict CRT response. Three categorical variables, left bundle branch block (LBBB), nonischemic cardiomyopathy, and female gender, were independently associated with CRT responses. Among continuous variables, including septal wall thickness, posterior wall thickness, and relative wall thickness (RWT), could regularize ECG QRS duration (QRSd) and significantly enhance the correlation between QRSd and CRT response. The 3 ML algorithms in a total of 38 features combinations constantly recognized that the features combined with QRSd/RWT outperformed the combinations without it. For each of three algorithms, the triplet feature combination of QRSd/RWT, LBBB, and nonischemic cardiomyopathy repeatedly increased the classification rate more than 8%. The best performance for CRT response prediction occurred with SVM model, which proposed actual QRSd/RWT values that favored CRT responses in patients both with and without LBBB. Lower QRSd/RWT values were required for CRT responses in patients with ischemic cardiomyopathy compared to those with non-ischemic cardiomyopathy. ML from ventricular remodeling characteristics–regularized QRSd improves CRT response prediction.

    AB - Up to one-third of patients selected by current guidelines do not respond to cardiac resynchronization therapy (CRT), the aim of this study was to find out novel analytical approaches to improve pre-implantation CRT response prediction. Among 31 pre-implantation features of clinical, laboratory, electrocardiography (ECG), and echocardiography variables in a consecutive cohort of patients receiving a first-time CRT device (CRT-pacemaker or CRT-defibrillator), we developed a machine learning (ML) model with three classification algorithms (support vector machines (SVM), K nearest neighbors, and random subspaces) with the best features combination to predict CRT response. Three categorical variables, left bundle branch block (LBBB), nonischemic cardiomyopathy, and female gender, were independently associated with CRT responses. Among continuous variables, including septal wall thickness, posterior wall thickness, and relative wall thickness (RWT), could regularize ECG QRS duration (QRSd) and significantly enhance the correlation between QRSd and CRT response. The 3 ML algorithms in a total of 38 features combinations constantly recognized that the features combined with QRSd/RWT outperformed the combinations without it. For each of three algorithms, the triplet feature combination of QRSd/RWT, LBBB, and nonischemic cardiomyopathy repeatedly increased the classification rate more than 8%. The best performance for CRT response prediction occurred with SVM model, which proposed actual QRSd/RWT values that favored CRT responses in patients both with and without LBBB. Lower QRSd/RWT values were required for CRT responses in patients with ischemic cardiomyopathy compared to those with non-ischemic cardiomyopathy. ML from ventricular remodeling characteristics–regularized QRSd improves CRT response prediction.

    KW - Cardiac resynchronization therapy

    KW - Classification

    KW - Machine learning

    KW - QRS duration

    KW - Ventricular geometric characteristics

    UR - http://www.scopus.com/inward/record.url?scp=85066074746&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=85066074746&partnerID=8YFLogxK

    U2 - 10.1007/s10554-019-01545-5

    DO - 10.1007/s10554-019-01545-5

    M3 - Article

    C2 - 31104177

    AN - SCOPUS:85066074746

    JO - International Journal of Cardiovascular Imaging

    JF - International Journal of Cardiovascular Imaging

    SN - 1569-5794

    ER -