Researchers Uncover New Biomarker for Chronic Stress Using AI

Researchers have made significant strides in identifying a biomarker for chronic stress through advanced imaging techniques. This breakthrough involves using artificial intelligence (AI) to analyze routine CT scans, potentially providing a method to assess long-term health risks associated with stress. The findings were presented at the annual meeting of the Radiological Society of North America (RSNA) in early December.

The study employed a deep learning model, a type of AI designed to identify patterns within large datasets, to automatically analyze medical images. Specifically, the model focused on the adrenal glands, small organs crucial in the body’s stress response due to their role in cortisol production. According to Health Canada, chronic stress is linked to various health issues, including heart disease and mental illness, yet there has been a lack of objective measures to quantify its biological effects over time.

Lead author Elena Ghotbi, a postdoctoral research fellow at The Johns Hopkins University School of Medicine, explained the rationale behind the study. “Our initial hypothesis was based on the fact that there is no widely accessible and approved marker to measure chronic stress,” she stated. The researchers proposed that measuring adrenal gland size via chest CT scans could correlate with chronic stress markers.

Unlike traditional cortisol tests, which provide a snapshot of stress levels, the volume of the adrenal glands may indicate prolonged physiological strain. By using an AI model trained to segment the adrenal glands on CT scans, the team calculated an Adrenal Volume Index (AVI). This index is defined as the total adrenal gland volume in cubic centimetres, adjusted for a person’s height squared in metres. The study found that individuals reporting high stress levels had a significantly higher AVI than those with lower perceived stress.

The AI model was tested on imaging and health data from nearly 3,000 participants in the Multi-Ethnic Study of Atherosclerosis, a comprehensive cohort study that integrates chest CT scans with cortisol measurements. Ghotbi noted that the study demonstrated a clear relationship between adrenal volume, cortisol levels, and self-reported stress, as well as long-term cardiovascular outcomes.

Senior author Shadpour Demehri, a professor of radiology at Johns Hopkins, emphasized the potential of this approach. He remarked, “There is no quick or objective measure of chronic stress,” highlighting the variability in how individuals express their stress. The focus of the study is on the biological implications rather than the psychological aspects of stress.

Both Ghotbi and Demehri acknowledged the preliminary nature of their findings, which require validation in diverse populations and scanning technologies. They emphasized that external validation is crucial before this measure can be integrated into clinical practice. Nevertheless, Demehri expressed optimism about the implications of their research. “The ability to apply the algorithm to millions of existing CT scans illustrates how AI can reveal biological signals that were previously difficult to quantify,” he stated.

Recent data indicates that approximately 6.4 million publicly funded CT examinations were conducted in Canada during the 2022–2023 fiscal year, averaging 160 exams per 1,000 people across the nation. The Canadian Medical Imaging Inventory collected primary data from May 5, 2023, to October 31, 2023. Demehri highlighted the potential impact of applying their algorithm to all existing CT machines, envisioning a future where such analyses could become routine. He remarked, “Just imagine if this algorithm could run across all CT machines and extract the data we need. While there are no guarantees in medicine, we are hopeful.”

As the research progresses, the team aims to explore the clinical applications of their findings, potentially transforming the way chronic stress is monitored and addressed in healthcare settings.