Residual risk prediction in anticoagulated patients with atrial fibrillation using machine learning: A report from the GLORIA-AF registry phase II/III.
Yang Liu, Yang Chen, Ivan Olier, Sandra Ortega-Martorell, Bi Huang, Hironori Ishiguchi, Ho Man Lam, Kui Hong, Menno V Huisman, Gregory Y H Lip
{"title":"Residual risk prediction in anticoagulated patients with atrial fibrillation using machine learning: A report from the GLORIA-AF registry phase II/III.","authors":"Yang Liu, Yang Chen, Ivan Olier, Sandra Ortega-Martorell, Bi Huang, Hironori Ishiguchi, Ho Man Lam, Kui Hong, Menno V Huisman, Gregory Y H Lip","doi":"10.1111/eci.14371","DOIUrl":null,"url":null,"abstract":"<p><strong>Background: </strong>Although oral anticoagulation decreases the risk of thromboembolism in patients with atrial fibrillation (AF), a residual risk of thrombotic events still exists. This study aimed to construct machine learning (ML) models to predict the residual risk in these patients.</p><p><strong>Methods: </strong>Patients with newly diagnosed non-valvular AF were collected from the Global Registry on Long-Term Oral Anti-Thrombotic Treatment in Patients with Atrial Fibrillation (GLORIA-AF) registry. To predict the residual risk of the composite outcome of thrombotic events (defined as ischemic stroke, systemic embolism, transient ischemic attack and myocardial infarction), we constructed four prediction models using the logistic regression (LR), random forest, light gradient boosting machine and extreme gradient boosting machine ML algorithms. Performance was mainly evaluated by area under the receiver-operating characteristic curve (AUC), g-means and F1 scores. Feature importance was evaluated by SHapley Additive exPlanations.</p><p><strong>Results: </strong>15,829 AF patients (70.33 ± 9.94 years old, 55% male) taking oral anticoagulation were included in our study, and 641 (4.0%) had residual risk, sustaining thrombotic events. In the test set, LR had the best performance with higher AUC trend of 0.712. RF has highest g-means of 0.295 and F1 score of 0.249. This was superior when compared with the CHA<sub>2</sub>DS<sub>2</sub>-VA score (AUC 0.698) and 2MACE score (AUC 0.696). Age, history of TE or MI, OAC discontinuation, eGFR and sex were identified as the top five factors associated with residual risk.</p><p><strong>Conclusion: </strong>ML algorithms can improve the prediction of residual risk of anticoagulated AF patients compared to clinical risk factor-based scores.</p>","PeriodicalId":12013,"journal":{"name":"European Journal of Clinical Investigation","volume":" ","pages":"e14371"},"PeriodicalIF":4.4000,"publicationDate":"2024-12-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"European Journal of Clinical Investigation","FirstCategoryId":"3","ListUrlMain":"https://doi.org/10.1111/eci.14371","RegionNum":3,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"MEDICINE, GENERAL & INTERNAL","Score":null,"Total":0}
引用次数: 0
Abstract
Background: Although oral anticoagulation decreases the risk of thromboembolism in patients with atrial fibrillation (AF), a residual risk of thrombotic events still exists. This study aimed to construct machine learning (ML) models to predict the residual risk in these patients.
Methods: Patients with newly diagnosed non-valvular AF were collected from the Global Registry on Long-Term Oral Anti-Thrombotic Treatment in Patients with Atrial Fibrillation (GLORIA-AF) registry. To predict the residual risk of the composite outcome of thrombotic events (defined as ischemic stroke, systemic embolism, transient ischemic attack and myocardial infarction), we constructed four prediction models using the logistic regression (LR), random forest, light gradient boosting machine and extreme gradient boosting machine ML algorithms. Performance was mainly evaluated by area under the receiver-operating characteristic curve (AUC), g-means and F1 scores. Feature importance was evaluated by SHapley Additive exPlanations.
Results: 15,829 AF patients (70.33 ± 9.94 years old, 55% male) taking oral anticoagulation were included in our study, and 641 (4.0%) had residual risk, sustaining thrombotic events. In the test set, LR had the best performance with higher AUC trend of 0.712. RF has highest g-means of 0.295 and F1 score of 0.249. This was superior when compared with the CHA2DS2-VA score (AUC 0.698) and 2MACE score (AUC 0.696). Age, history of TE or MI, OAC discontinuation, eGFR and sex were identified as the top five factors associated with residual risk.
Conclusion: ML algorithms can improve the prediction of residual risk of anticoagulated AF patients compared to clinical risk factor-based scores.
期刊介绍:
EJCI considers any original contribution from the most sophisticated basic molecular sciences to applied clinical and translational research and evidence-based medicine across a broad range of subspecialties. The EJCI publishes reports of high-quality research that pertain to the genetic, molecular, cellular, or physiological basis of human biology and disease, as well as research that addresses prevalence, diagnosis, course, treatment, and prevention of disease. We are primarily interested in studies directly pertinent to humans, but submission of robust in vitro and animal work is also encouraged. Interdisciplinary work and research using innovative methods and combinations of laboratory, clinical, and epidemiological methodologies and techniques is of great interest to the journal. Several categories of manuscripts (for detailed description see below) are considered: editorials, original articles (also including randomized clinical trials, systematic reviews and meta-analyses), reviews (narrative reviews), opinion articles (including debates, perspectives and commentaries); and letters to the Editor.