Myosteatosis Emerges as a Potential COPD Biomarker
AI-quantified thoracic muscle fat may outperform emphysema-like measurements in predicting COPD
AI may be able to turn coronary artery calcium (CAC) CT scans performed for other reasons into an early warning tool for chronic obstructive pulmonary disease (COPD). A large prospective study published in Radiology: Cardiothoracic Imaging found that AI-derived myosteatosis measurements could identify likely COPD cases before symptoms appear.
COPD is the third leading cause of death worldwide, according to the World Health Organization. Risk assessment has traditionally focused on smoking history, symptoms, spirometry and lung findings such as emphysema. However, the disease is typically detected only after symptoms have appeared or worsened.
“A scalable biomarker is needed that can be extracted opportunistically and identify systemic vulnerability before COPD is diagnosed,” said lead author Morteza Naghavi, MD, founder and president of HeartLung.AI, in Houston. “Our study suggests that a muscle quality biomarker outside the lung may outperform an emphysema-like lung measurement in predicting future COPD. This supports a broader view of COPD, not only as a lung disease, but as part of a systemic cardiopulmonary and metabolic phenotype that may be detectable years earlier.”
AI-based Quantification of Muscle Fat
Myosteatosis is a sign of poor muscle quality in addition to low muscle mass. Because it has also been linked to worse cardiovascular outcomes, the researchers investigated whether it may signal the kind of systemic metabolic and inflammatory dysfunction that also predisposes people to future COPD.
“This is exactly where AI can change medicine,” Dr. Naghavi said. “Radiologists have been looking at CT scans for decades and they can see emphysema, lung nodules and coronary calcium. But subtle quantitative muscle fat infiltration is not something the human eye can reliably measure in daily practice. AI makes it measurable, reproducible and monitorable over time.”
To test the idea, Dr. Naghavi and colleagues analyzed 5,535 participants ages 45 to 84 from the Multi-Ethnic Study of Atherosclerosis (MESA), a diverse U.S. cohort with detailed imaging and longitudinal clinical data. Forty-eight percent were male, 51% were never smokers and none had clinical cardiovascular disease at baseline, making the cohort well suited for studying links between body composition and COPD.
Using AI-CVD, HeartLung.AI’s FDA-cleared opportunistic CT platform, the researchers extracted myosteatosis and emphysema measurements from MESA participants’ baseline CAC CT scans.
Looking across a 20-year follow-up period, the researchers noted that 7.1% of subjects (396) were diagnosed with COPD, according to hospital discharge records.
Myosteatosis, defined as the lowest quartile of thoracic skeletal muscle mean attenuation, was independently associated with an approximately threefold increased risk of clinically diagnosed COPD compared with the healthiest muscle density after adjusting for smoking, body mass index, inflammation and other risk factors.
Myosteatosis also remained a powerful predictor even after adjusting for emphysema-like lung percentage. However, the researchers noted that CAC CT scans may not fully capture upper lung regions where emphysema typically begins, which limits the head-to-head comparison with the emphysema-like biomarker.
The association held across subgroups including age, sex, obesity, smoking history and physical activity, and remained robust after excluding early COPD cases and participants with baseline asthma.
“Even participants who did not smoke, or who did not have obvious emphysema-like findings at baseline, could still develop COPD years later if they had substantial myosteatosis,” Dr. Naghavi said. “This research points toward a prevention model in which imaging can identify risk years earlier and direct patients toward spirometry, smoking cessation, exercise, nutrition, pulmonary rehabilitation and metabolic intervention before irreversible damage occurs.”
The findings underscore how closely COPD prevention is tied to cardiovascular and metabolic prevention, Dr. Naghavi noted. “More COPD patients die with cardiovascular disease, particularly heart failure, than with respiratory dysfunction,” he said. “The same patient may have increased coronary calcium, excess epicardial fat, fatty liver, low bone density and myosteatosis. Our AI-CVD platform was built with a goal to help clinicians see the whole risk phenotype from one scan and intervene earlier.”
Validating Myosteatosis as a Cardiopulmonary Biomarker
According to Dr. Naghavi, the next step is large-scale validation of the emerging predictive biomarker in additional cohorts. He also called for longitudinal studies to determine whether worsening muscle quality precedes lung function decline and whether improving muscle quality reduces COPD risk. On the technology side, the goal is to integrate reproducible AI-based myosteatosis measurements into actionable clinical workflows.
The researchers also plan to publish an additional paper examining whether AI-quantified myosteatosis could independently predict long-term atrial fibrillation and heart failure risk.
“Validation in other cohorts and in predicting additional life-threatening disease would suggest that we should teach radiologists and clinicians to think about muscle quality the same way we now think about visceral fat, coronary calcium index, liver fat and bone density,” Dr. Naghavi said. “Myosteatosis should not be dismissed as an incidental finding. If we can measure it, monitor it and eventually modify it, it could become an potential biomarker for both cardiovascular and pulmonary prevention.”
For More Information
Access the Radiology: Cardiothoracic Imaging article, “Artificial Intelligence-derived Measurements of Myosteatosis from Coronary Artery Calcium CT Scans to Predict COPD: The Multi-Ethnic Study of Atherosclerosis.”
Read previous RSNA News stories on chest imaging: