Mount Sinai to develop sleep apnea outcome risk prediction models

Researchers from Mount Sinai have been awarded a four-year, $3 million grant from the National Heart, Lung, and Blood Institute of the National Institutes of Health (NIH) to develop artificial intelligence (AI)-driven prediction models to flag risk of cardiovascular disease events in patients with obstructive sleep apnea.

The American Heart Association (AHA) indicates that obstructive sleep apnea increases patients’ risk of cardiovascular disease, including coronary artery disease, hypertension and stroke. The use of continuous positive airway pressure (CPAP) machines is often prescribed to treat sleep apnea, but evidence to suggest the benefits of CPAP use in relation to cardiovascular event rates is limited.

To bridge this gap, the research team will build machine learning (ML) tools to identify obstructive sleep apnea patients at high risk for atherosclerosis progression and cardiovascular events like stroke and heart attack.

The researchers underscored that this approach could help forecast cardiovascular treatment effectiveness of CPAP use in patients considered “non-sleepy” based on their responses to a clinical test. Doing so may shed light on which patients would benefit most from CPAP therapy and which should avoid the treatment.

This work is based on earlier Mount Sinai research highlighting the potential harm of CPAP use in non-sleepy patients, necessitating an increased focus on personalized sleep apnea treatment.

“We are inspired by the transformative potential of machine learning techniques in health care, particularly in analyzing vast amounts of complex data to personalize treatment strategies,” stated principal investigator Girish Nadkarni, MD, MPH, the Irene and Dr. Arthur M. Fishberg Professor of Medicine, director of The Charles Bronfman Institute of Personalized Medicine, and system chief of Data-Driven and Digital Medicine at Icahn Mount Sinai, in the press release. “Our study has the potential to revolutionize the management of obstructive sleep apnea by offering decision support tools that optimize treatment plans, improve patient outcomes, and reduce the burden of sleep apnea-related cardiovascular disease events on both individuals and health care systems.”

To build their predictive models, the researchers will utilize data from two cohorts: the Sleep Apnea Cardiovascular Endpoints (SAVE) randomized clinical trial – containing information from over 2,500 non-sleepy participants with moderate to severe obstructive sleep apnea and established cardiovascular disease – and the Multi-Ethnic Study of Atherosclerosis (MESA) cohort – made up of data from over 6,000 non-sleepy participants who are generally healthy and come from ethnically diverse backgrounds.

These data will be used to flag key predictors of atherosclerosis progression and cardiovascular events alongside identifying risk-based patient subgroups with differential CPAP treatment outcomes for cardiovascular events.

The predictive models will be validated using electronic health record (EHR) data from Mount Sinai.

"Through precision medicine, we are prioritizing rigorous intervention to enhance cardiovascular disease risk reduction,” said principal investigator Mayte Suarez-Farinas, PhD, associate director of the Center for Biostatistics, and professor of Population Health Science and Policy, and Genetics and Genomic Sciences at Icahn Mount Sinai. “Health care providers will be equipped with innovative tools to identify patients at heightened risk for heart attack or stroke and be able to predict treatment outcomes of CPAP therapy in sleep apnea patients. This personalized approach will enable clinicians to tailor treatment strategies to individual patient needs, optimizing CPAP adherence and efficacy.”

This research furthers Mount Sinai’s efforts to improve sleep apnea care through advanced analytics technologies.

“Supported by a transformative grant, I’m thrilled to lead a project that stands at the intersection of cutting-edge artificial intelligence and sleep medicine,” stated primary principal investigator Neomi Shah, MD, MPH, MSc, associate dean for Faculty Career Advancement, vice chair for Faculty Affairs in the Mount Sinai Health System Department of Medicine, and professor of Medicine (Pulmonary, Critical Care and Sleep Medicine) at the Icahn School of Medicine at Mount Sinai.

“Our work will epitomize the wealth of expertise and collaborative effort across the Mount Sinai Health System to both enrich our understanding of the condition and improve patient care, impacting millions in the United States. We are committed to validating our AI tools within Mount Sinai’s clinical dataset to translate our research into real-world practice, thereby, effectively bridging the research to practice gap,” Shah continued.

Last month, a Mount Sinai team was awarded a $4.1 million NIH grant to develop AI-driven models to predict adverse outcomes in sleep apnea patients.

The researchers noted that current diagnostic tools for sleep apnea – which typically rely on counting the number of breathing disturbances a patient experiences during sleep – are not accurate predictors of short- and long-term adverse outcomes like excessive daytime sleepiness, cardio-cerebrovascular morbidity and neurocognitive impairment.

The AI models are designed to address the shortcomings of these tools by analyzing hypoxic, arousal and ventilatory variables to generate a personalized patient risk score.

Update 2024/03/14: This article has been updated with a quote from Dr. Neomi Shah.