Related Experiment Video
Updated: Aug 26, 2025
Estimating Bilateral Atrial Function by Cardiovascular Magnetic Resonance Feature Tracking in Patients with Paroxysmal Atrial Fibrillation
Published on: July 20, 2022
Performance of multilabel machine learning models and risk stratification schemas for predicting stroke and bleeding
Juan Lu1, Rebecca Hutchens2, Joseph Hung3
1Department of Computer Science and Software Engineering, The University of Western Australia, Perth, Australia; Medical School, The University of Western Australia, Perth, Australia; Harry Perkins Institute of Medical Research, Perth, Australia.
Insights
Machine learning models show improved prediction of major bleeding and death in atrial fibrillation (AF) patients compared to traditional risk scores. These advanced models offer better risk stratification for anticoagulant therapy decisions.
Area of Science:
- Cardiology
- Medical Informatics
- Machine Learning
Background:
- Assessing stroke and bleeding risks is crucial for anticoagulant therapy in atrial fibrillation (AF) patients.
- Current risk stratification schemas like CHA 2 DS 2 -VASc and HAS-BLED have limited predictive accuracy for AF patients.
- Multilabel machine learning (ML) offers potential for enhanced predictive performance in AF risk assessment.
Purpose of the Study:
- To compare the predictive performance of multilabel ML models against established clinical risk scores for outcomes in AF patients.
- To evaluate the ability of ML models to improve risk stratification for anticoagulant therapy in non-valvular AF.
Main Methods:
- Retrospective cohort study of 9670 non-valvular AF patients with 1-year follow-up.
- Outcomes included ischemic stroke, major bleeding, and all-cause death.
- Compared discrimination and calibration of ML models (gradient boosting, neural networks, SVM) with CHA 2 DS 2 -VASc and HAS-BLED using AUC and NRI.
Main Results:
- A multilabel gradient boosting classifier chain achieved superior AUCs for stroke (0.685), major bleeding (0.709), and death (0.765).
- ML models significantly improved major bleeding prediction (NRI=22.8%, p<0.05) and death prediction (p<0.05) compared to HAS-BLED and CHA 2 DS 2 -VASc.
- ML models identified additional risk factors, including hemoglobin level and renal function.
Conclusions:
- Multilabel ML models demonstrate superior predictive capabilities for major bleeding and death in non-valvular AF patients.
- ML models offer a promising advancement over traditional risk scores for personalized anticoagulant therapy decisions.
Background:
Appropriate anticoagulant therapy for patients with atrial fibrillation (AF) requires assessment of stroke and bleeding risks. However, risk stratification schemas such as CHA2DS2-VASc and HAS-BLED have modest predictive capacity for patients with AF. Multilabel machine learning (ML) techniques may improve predictive performance and support decision-making for anticoagulant therapy. We compared the performance of multilabel ML models with the currently used risk scores for predicting outcomes in AF patients.
Methods:
This was a retrospective cohort study of 9670 patients, mean age 76.9 years, 46% women, who were hospitalized with non-valvular AF, and had 1-year follow-up. The outcomes were ischemic stroke (167), major bleeding (430) admissions, all-cause death (1912) and event-free survival (7387). Discrimination and calibration of ML models were compared with clinical risk scores by area under the curve (AUC). Risk stratification was assessed using net reclassification index (NRI).
Results:
Multilabel gradient boosting classifier chain provided the best AUCs for stroke (0.685 95% CI 0.676, 0.694), major bleeding (0.709 95% CI 0.703, 0.716) and death (0.765 95% CI 0.763, 0.768) compared to multi-layer neural networks and classifier chain using support vector machine. It provided modest performance improvement for stroke compared to AUC of CHA2DS2-VASc (0.652, NRI = 3.2%, p-value = 0.1), but significantly improved major bleeding prediction compared to AUC of HAS-BLED (0.522, NRI = 22.8%, p-value < 0.05). It also achieved greater discriminant power for death compared with AUC of CHA2DS2-VASc (0.606, p-value < 0.05). ML models identified additional risk features such as hemoglobin level, renal function.
Conclusions:
Multilabel ML models can outperform clinical risk stratification scores for predicting the risk of major bleeding and death in non-valvular AF patients.
