Related Experiment Video
Updated: Aug 5, 2026
07:40
Multi-Modal Home Sleep Monitoring in Older Adults
Published on: January 26, 2019
A Multimodal Machine Learning Model for Framingham Risk Score-Based Cardiovascular Risk Stratification in Patients
Feng Zhao1,2, Yin Li3, Chenyang Li1,2
1Department of Otolaryngology Head and Neck Surgery, Shanghai Key Laboratory of Sleep Disordered Breathing, Shanghai Sixth People's Hospital Affiliated to Shanghai Jiao Tong University School of Medicine, Shanghai, China.
Journal of Sleep Research
|July 28, 2026
Summary
Machine learning models can improve cardiovascular risk stratification in patients with obstructive sleep apnea (OSA). Integrating sleep-specific markers with clinical data enhances risk prediction beyond traditional factors.
Area of Science:
- Biomedical data analysis
- Cardiovascular disease research
- Sleep medicine
Background:
- Cardiovascular disease (CVD) is a leading cause of mortality.
- Obstructive sleep apnea (OSA) is increasingly recognized as a CVD risk factor.
- Current cardiovascular risk stratification tools may not fully capture risks in OSA patients.
Purpose of the Study:
- To apply a machine learning framework to classify cardiovascular risk categories in patients with OSA.
- To investigate the utility of integrating polysomnographic, clinical, and haematological data for risk stratification.
- To determine if OSA-related phenotypes enhance cardiovascular risk assessment.
Main Methods:
- Utilized a multilayer perceptron (MLP) machine learning framework.
- Integrated diverse data types including polysomnography, clinical, and haematological parameters.
- Classified patients into Framingham Risk Score-based cardiovascular risk categories.
Main Results:
- The MLP model achieved robust discriminative performance with an AUROC of 0.822.
- OSA-related phenotypes provided additional information for cardiovascular risk stratification.
- Demonstrated the potential of sleep-specific markers in cardiovascular risk assessment.
Conclusions:
- Machine learning offers a powerful approach for precision cardiovascular risk stratification in OSA populations.
- Incorporating sleep-specific markers alongside conventional factors improves risk assessment.
- Further research with larger cohorts and explainable AI is warranted for clinical application.