Researchers have identified the first imaging-based biomarker of chronic stress using a deep learning model applied to routine CT scans. The findings, presented at the Radiological Society of North America (RSNA) annual meeting, highlight how artificial intelligence can detect long-term stress by measuring adrenal gland volume.
Chronic Stress and Its Health Impact
Chronic stress affects both mental and physical well-being, contributing to anxiety, insomnia, hypertension, immune dysfunction and major diseases such as heart disease, depression and obesity. Until now, clinicians have largely relied on questionnaires, inflammatory markers and difficult-to-obtain cortisol samples to assess chronic stress.
How the AI Model Measures Stress
Lead author Dr. Elena Ghotbi from Johns Hopkins University developed a deep learning algorithm. It automatically measures adrenal gland volume on existing chest CT scans. Since clinicians perform tens of millions of such scans annually in the U.S., this approach enables stress evaluation without additional imaging or radiation.
“Our method uses widely available imaging data and allows large-scale assessment of chronic stress using routine CT scans,” Dr. Ghotbi said. She added that this biomarker may strengthen cardiovascular risk prediction and guide preventive care.
Senior author Dr. Shadpour Demehri explained that adrenal volume offers a more reliable and long-term indicator of stress. Single cortisol measurements, by contrast, only reflect momentary stress levels.
Study Design and Participant Insights
As reported by MedicalXpress, the research team analyzed data from 2,842 adults in the Multi-Ethnic Study of Atherosclerosis. This study uniquely integrates imaging, cortisol measures, stress questionnaires, and allostatic load—a metric capturing cumulative stress effects. They retrospectively applied the AI model to calculate an Adrenal Volume Index (AVI), defined as adrenal volume divided by height squared.
Key Findings and Clinical Implications
The AI-derived AVI correlated strongly with cortisol levels, allostatic load, depression scores and perceived stress. Higher AVI values were associated with elevated cortisol, increased allostatic load and greater left ventricular mass. Each 1 cm³/m² rise in AVI corresponded to increased risks of heart failure and mortality.
“With up to 10 years of follow-up, AVI independently predicted heart failure—the first validated imaging marker of chronic stress,” Dr. Ghotbi noted.
Co-author Dr. Teresa Seeman emphasized the importance of linking a routine imaging feature with biological and psychological stress indicators. She called it a major step forward in quantifying stress-related health impacts.
Future Applications
According to Dr. Demehri, this biomarker offers a practical, physiologically sound measure of chronic stress accessible through widely performed CT scans. Researchers believe it could be applied across numerous stress-associated diseases in middle-aged and older adults. This could help advance early detection and prevention strategies.