Mount Sinai researchers have created an analytic tool using machine learning that can predict cardiovascular disease risk in millions of patients with obstructive sleep apnoea, according to findings recently published in Nature Communications Medicine.
The team said their study is the first to provide estimates of whether continuous positive airway pressure (CPAP), a widely used therapy for obstructive sleep apnoea, will increase or decrease an individual’s cardiovascular risk. It highlights the potential for precision medicine and varied approaches to tailor clinical care and reduce cardiovascular disease risk in vulnerable patients.
Obstructive sleep apnoea is a common, serious condition in which breathing repeatedly stops and starts during sleep. It is associated with elevated risks for cardiovascular disease, including stroke and heart disease. CPAP, which provides a continuous stream of pressurised air through a mask and helps eliminate breathing disturbances during sleep, remains the most effective treatment for sleep apnoea. However, prior large studies have not shown that CPAP lowers risks for cardiovascular disease in patients with this disease.
The Mount Sinai researchers used a machine learning algorithm to create an analysis model that predicts how CPAP could affect an individual’s cardiovascular health – estimating each patient’s likeliness of benefit or harm from the therapy, based on their sleep and health information.
“Our findings represent a significant advancement in personalised medicine, moving away from a one-size-fits-all strategy in the treatment of obstructive sleep apnoea,” said co-corresponding author Neomi A. Shah, MD, MPH, MSc, ATSF, Professor of Medicine (Pulmonary, Critical Care and Sleep Medicine) and Artificial Intelligence and Human Health, and System Vice Chair of Faculty Affairs for the Department of Medicine at Icahn School of Medicine at Mount Sinai. “This underscores the value of new data-driven approaches like our model to assist clinicians in making informed decisions about CPAP treatment recommendations, enhancing personalised care to meet the individual needs of every patient.”
The Mount Sinai team analysed data from the Sleep Apnea Cardiovascular Endpoints (SAVE) trial, the largest clinical cohort evaluating CPAP for cardiovascular disease prevention with more than 2,600 participants from 89 sites in seven countries, to estimate individualised treatment effect scores. They considered more than 100 predictors from sleep and health information to establish 23 key baseline features, such as prior medical conditions and smoking status, in their analysis model.
The researchers found that treatment response significantly varied across the cohort. The model identified a subgroup who were expected to have improved cardiovascular risk with CPAP treatment; participants in this subgroup who were randomly assigned to receive the therapy experienced a 100-fold improvement in future cardiac risk compared with patients from this subgroup on usual care. Conversely, those in a subgroup predicted to be harmed by the therapy experienced a greater than 100-fold increase in cardiovascular disease outcomes, including recurrent strokes and heart attacks, when receiving CPAP compared with usual care.
“These results demonstrate the power of machine learning for prediction of treatment effects in an era of precision medicine; however, such models require careful validation to prove their utility in clinical practice,” said co-primary author Oren Cohen, MD, Assistant Professor of Medicine (Pulmonary, Critical Care and Sleep Medicine) at the Icahn School of Medicine.
“Artificial intelligence in medicine must move beyond pattern recognition to causal reasoning,” said co-corresponding author Mayte Suarez-Farinas, PhD, Co-Director for the Division of Biostatistics and Data Science, and Professor of Population Health Science and Policy, and Artificial Intelligence and Human Health, at the Icahn School of Medicine. “By estimating individualised treatment effects over time using randomised clinical trial data, we move predictive AI toward decision-support tools grounded in causality and capable of informing real-world treatment decisions and improving outcomes.”
Source: Mount Sinai
