Summary: Researchers have created an AI-driven tool that detects chronic stress by measuring adrenal gland volume on routine chest CT scans. This imaging biomarker aligns with cortisol levels, validated stress questionnaires, and long-term cardiovascular outcomes, offering the first noninvasive, image-based way to quantify the body’s cumulative stress load.
The study found that larger adrenal gland volume is associated with higher physiological stress, greater allostatic load, and increased risk of heart failure and mortality. Because tens of millions of chest CT scans are already performed annually, extracting this biomarker from existing images could enable large-scale, low-cost screening and early prevention without additional tests or radiation exposure.
Key Facts
- AI Stress Biomarker: Adrenal gland volume measured on CT scans serves as an imaging biomarker that better reflects long-term chronic stress than single-point cortisol tests.
- Risk Prediction: Increased adrenal volume correlates with higher cortisol levels, elevated allostatic load, and a greater risk of future cardiovascular events, including heart failure.
- Clinical Potential: Because adrenal volume can be measured on routine chest CTs, this biomarker offers a scalable tool for identifying individuals at elevated stress-related disease risk and guiding preventive care.
Source: RSNA
Using a deep learning model, investigators have identified the first imaging biomarker of chronic stress that can be measured on routine CT scans, according to research to be presented at the annual meeting of the Radiological Society of North America (RSNA).
Chronic stress affects both physical and mental health and is linked to conditions such as anxiety, insomnia, muscle pain, high blood pressure and immune dysfunction. Over time, persistent stress contributes to major illnesses including heart disease, depression and obesity.
The study’s lead author, Elena Ghotbi, M.D., a postdoctoral research fellow at Johns Hopkins University School of Medicine, developed and trained a deep learning algorithm to automatically measure adrenal gland volume from existing chest CT scans.
Because tens of millions of chest CTs are performed yearly in the United States alone, the researchers emphasize that their approach uses widely available imaging data and enables population-level assessment of how chronic stress affects the body.
“This AI-driven biomarker could improve cardiovascular risk stratification and support preventive care without requiring additional testing or added radiation,” Dr. Ghotbi said.
Senior author Shadpour Demehri, M.D., professor of radiology at Johns Hopkins, noted that chronic stress is a common complaint among adults and that clinicians have lacked an objective, scalable way to quantify its cumulative biological effects.
“Until now, measuring the long-term burden of stress required questionnaires, occasional serum markers of inflammation, or cumbersome repeated cortisol collections,” Dr. Demehri said. “Adrenal volume gives us a physiologic marker that reflects the cumulative stress burden visible on scans patients already receive.”
Unlike point-in-time cortisol tests, which capture transient hormone levels, adrenal volume functions more like a biological barometer of chronic stress over time.
The researchers analyzed data from 2,842 participants (mean age 69.3; 51% women) enrolled in the Multi-Ethnic Study of Atherosclerosis. This dataset uniquely combined chest CT imaging with validated stress questionnaires, repeated salivary cortisol sampling and markers of allostatic load—an integrated measure of the physiological toll of long-term stress.
Applying the deep learning model to the CT scans, the team segmented the adrenal glands and calculated an Adrenal Volume Index (AVI), defined as adrenal volume in cm³ divided by height² in m². Saliva was sampled eight times per day over two days to capture cortisol patterns. Allostatic load was calculated from measures including body mass index, creatinine, hemoglobin, albumin, glucose, white blood cell count, heart rate and blood pressure.
Statistical analyses showed that AI-derived AVI correlated with validated stress questionnaires, circulating cortisol levels, and allostatic load. Higher AVI was associated with greater cortisol exposure and peak cortisol measures. Participants reporting higher perceived stress had larger AVI values than those reporting lower stress. AVI also correlated with higher left ventricular mass index, and each 1 cm³/m² increase in AVI was associated with increased risk of heart failure and mortality during follow-up.
“With up to a decade of follow-up on participants, we could link this imaging biomarker to clinically meaningful outcomes,” Dr. Ghotbi said. “This is the first validated imaging marker of chronic stress shown to have an independent effect on a cardiovascular outcome—heart failure.”
Teresa E. Seeman, Ph.D., a co-author and professor of epidemiology at UCLA, highlighted the study’s significance in operationalizing stress. “The work connects a routinely obtained imaging feature—adrenal volume—with biological and psychological measures of stress and shows independent prediction of a major clinical outcome,” she said.
The authors suggest that the adrenal volume biomarker could be applied across many conditions linked to chronic stress among middle-aged and older adults, enabling earlier intervention and targeted prevention strategies.
Other co-authors include Roham Hadidchi, Seyedhouman Seyedekrami, Quincy A. Hathaway, M.D., Ph.D., Michael Bancks, Nikhil Subhas, Matthew J. Budoff, M.D., David A. Bluemke, M.D., Ph.D., R. Graham Barr and Joao A.C. Lima, M.D.
Key Questions Answered:
A: They identified an imaging-based biomarker of chronic stress by measuring adrenal gland volume on chest CT scans using an AI model.
A: Adrenal volume reflects long-term physiological stress and correlates with cortisol measures, allostatic load, and future cardiovascular risk.
A: Cortisol levels vary throughout the day and require multiple samples to assess exposure; adrenal volume provides a more stable indicator of cumulative stress burden.
Editorial Notes:
- This article was edited by a Neuroscience News editor.
- Journal paper reviewed in full.
- Additional context added by editorial staff.
About this AI and stress research news
Author: Linda Brooks, RSNA
Source: RSNA
Contact: Linda Brooks – RSNA
Image: The image is credited to Neuroscience News
Original Research: Findings will be presented at the 111th Scientific Assembly and Annual Meeting of the Radiological Society of North America (RSNA).