Machine learning-based detection of sleep-disordered breathing in hypertrophic cardiomyopathy

Area of Science:

• Cardiology
• Sleep Medicine
• Machine Learning in Healthcare

Background:

• Hypertrophic cardiomyopathy (HCM) frequently co-occurs with sleep-disordered breathing (SDB).
• Untreated SDB can lead to adverse cardiovascular events in HCM patients.
• Current SDB detection rates in HCM populations are suboptimal.

Purpose of the Study:

• To develop a machine learning-based discriminant model for identifying SDB in patients with HCM.
• To improve the detection of concomitant SDB in HCM patients.

Main Methods:

• A multicentre study involving 369 HCM patients who underwent nocturnal oximetry.
• Development of a random forest discriminant model using clinical parameters to predict high Oxygen Desaturation Index (ODI >10).
• Comparison of the model's performance against a previous logistic regression model.

Main Results:

• The machine learning model achieved an area under the receiver operating characteristic curve of 0.86 in the test set.
• High sensitivity (0.91) and specificity (0.68) were observed for the model.
• The model significantly outperformed a previous logistic regression model (p=0.03).

Conclusions:

• This is the first study to develop a machine learning model for SDB detection in HCM patients.
• The model shows potential for cost-effective screening and timely treatment of SDB in this population.
• Improved SDB management can help prevent cardiac remodeling and adverse events in HCM.