Artificial intelligence is turning a familiar medical workhorse into a window on one of the most pervasive health threats of modern life: chronic stress. By training deep learning systems on routine CT scans, researchers have uncovered a subtle anatomical pattern that appears to encode the body’s long term response to pressure, fatigue and worry. The finding hints at a future in which stress is not just something patients describe, but something clinicians can measure and track with the same precision they bring to blood pressure or cholesterol.
Instead of ordering new tests or exotic imaging, the new approach mines scans that patients are already getting for other reasons, from chest pain to cancer screening. I see that shift as quietly radical, because it reframes stress from a vague psychological complaint into a quantifiable risk factor that can be spotted early, monitored over time and, crucially, addressed before it hardens into disease.
How a deep learning model spotted stress hiding in plain sight
The core breakthrough rests on a simple but powerful idea: if chronic stress reshapes the body over time, those changes should be visible somewhere on medical images, even if the human eye has never been trained to see them. Researchers built a deep learning model and fed it thousands of routine CT scans, asking the system to learn patterns that correlate with markers of long term strain on the body. Instead of focusing on obvious disease, the algorithm was tuned to pick up more diffuse signals that might reflect the cumulative toll of stress on organs and tissues.
According to the reporting, the team used this deep learning approach to identify what they describe as the first imaging based biomarker of chronic stress, derived directly from everyday CT images that patients already undergo for other clinical reasons. By letting the model search for subtle, multivariate patterns that radiologists would not normally flag, the researchers were able to connect specific anatomical features to the long term effects of stress on health, a link that is detailed in their description of how Researchers used a deep learning system to uncover this hidden signal.
The Adrenal Volume Index, explained in plain language
At the center of this work is a new metric with a technical name that masks a straightforward concept: the Adrenal Volume Index. The adrenal glands sit atop the kidneys and help regulate the body’s response to stress through hormones like cortisol. The researchers found that by carefully measuring the size and characteristics of these glands on CT scans, and then standardizing those measurements into a single score, they could capture how chronically activated a person’s stress system appears to be.
This Adrenal Volume Index, often shortened to AVI, effectively turns a fuzzy notion of “being under pressure” into a number that can be compared across patients and tracked over time. The reporting notes that Adrenal Volume Index tracks stress, hormones and heart risk, tying the physical size and structure of the adrenal glands to downstream problems, namely heart failure. In other words, AVI is not just a descriptive label, it is a bridge between what the scan shows and what might happen to a patient years down the line.
From hormones to heart risk: what AVI actually measures
To understand why AVI matters, it helps to follow the chain from gland to hormone to disease. Chronic stress keeps the adrenal glands working overtime, which can alter their volume and internal architecture. Those changes, in turn, reflect shifts in cortisol and other hormones that influence blood pressure, metabolism and inflammation. By quantifying adrenal volume in a standardized way, AVI captures how revved up the stress response has been over a long period, rather than just offering a snapshot of how someone feels on a particular day.
The team did not stop at measuring glands, they examined how AVI related to cortisol levels, broader concepts like allostatic load (the cumulative impact of stress on the body) and concrete cardiovascular outcomes. Their analysis showed that higher AVI scores tracked with hormonal markers of stress and with increased risk of heart problems, including heart failure, which underscores why this metric is more than an anatomical curiosity. The reporting describes how The team then examined how AVI related to cortisol, allostatic load and the cumulative impact of stress, tying the imaging signal to both biochemical and clinical endpoints.
Why chronic stress finally has a visible biomarker
For decades, clinicians have known that chronic stress is linked to conditions like heart disease, depression and obesity, but they have had to rely on questionnaires, intermittent lab tests and patient self reports to gauge its severity. That has made stress feel intangible, more like a background factor than a vital sign. By contrast, an imaging based biomarker like AVI gives stress a physical footprint that can be seen, measured and compared, much like a coronary calcium score or a liver fat percentage.
The reporting emphasizes that this is the first time a deep learning artificial intelligence model has been used to identify a biomarker of chronic stress directly from CT scans, and that this marker is tied to serious outcomes such as heart disease, depression and obesity. In practical terms, that means a patient who comes in for a scan of the chest or abdomen could walk away not only with answers about their immediate complaint, but also with an objective measure of how much chronic tension their body has been carrying. The work is described as Using a deep-learning artificial intelligence model to discover a first-of-its-kind biomarker of chronic stress that is associated with heart disease, depression and obesity.
Turning routine CT scans into stress early-warning systems
One of the most striking aspects of this research is that it does not require new machines or special appointments. Hospitals already perform vast numbers of CT scans every day, from emergency evaluations after a car crash to follow up imaging for cancer survivors. By layering AI analysis on top of those existing scans, clinicians can extract additional information about a patient’s stress burden without exposing them to extra radiation or asking them to come back for another test.
In practice, that could turn CT scanners into early warning systems for chronic stress, flagging patients whose AVI scores suggest a heightened risk of long term tension and its downstream complications. Reporting on this work notes that using artificial intelligence and routine CT scans, researchers say they can help experts “see” stress and better understand the risks of long term tension that might otherwise go unnoticed in a busy clinic. That idea is captured in the description that Using artificial intelligence and routine CT scans can help experts see stress on images that were originally ordered for entirely different reasons.
What this means for radiologists and everyday clinical practice
For radiologists, the emergence of AVI and similar AI derived metrics signals a shift in how imaging is interpreted and reported. Instead of focusing solely on discrete abnormalities like tumors, fractures or blockages, radiology reports could begin to include standardized scores that reflect systemic conditions such as chronic stress. I see that as part of a broader evolution in which imaging becomes a tool not just for diagnosing acute problems, but for profiling long term risk and guiding preventive care.
In a typical workflow, an AI system could automatically calculate AVI whenever a CT scan includes the adrenal glands, then present that score alongside traditional findings. Radiologists could highlight elevated AVI in their reports, prompting primary care physicians or cardiologists to discuss stress management, mental health support or lifestyle changes with the patient. The reporting that describes how experts can now “see” stress on routine scans suggests that these tools are meant to augment, not replace, human judgment, giving clinicians another lens through which to interpret the images they already read every day.
Patients, prevention and the promise of personalized stress care
For patients, the idea that a CT scan can reveal chronic stress may feel unsettling at first, but it also opens the door to more personalized care. Instead of being told in general terms that stress is bad, a person could see how their own body has responded over time, quantified in an AVI score that can be tracked across years. That kind of concrete feedback can be a powerful motivator, much like seeing a rising A1C level pushes some people to take prediabetes more seriously.
Because AVI is tied to outcomes like heart disease, depression and obesity, an elevated score could trigger a more proactive approach to prevention. Clinicians might refer patients to cognitive behavioral therapy, prescribe stress reduction apps such as Headspace or Calm, adjust medications that affect cortisol, or recommend structured exercise programs that are known to blunt the physiological impact of chronic tension. The key is that the conversation would be anchored in a measurable risk, not just a vague admonition to “relax,” and that shift could help move stress management from the margins of care into the mainstream of cardiovascular and mental health prevention.
The limits of what AI can see, and what remains unverified
As promising as AVI and related tools appear, it is important to recognize what they do not yet tell us. A high AVI score indicates that the adrenal glands have changed in ways associated with chronic stress and heart risk, but it does not pinpoint the source of that stress, nor does it guarantee that a particular individual will develop heart failure or depression. The relationship is probabilistic, not deterministic, and it will take longer term follow up studies to refine how AVI should be used in risk calculators and treatment guidelines. Any attempt to claim that AVI alone can diagnose specific psychiatric conditions or predict exact timelines for disease would be unverified based on available sources.
There are also questions about how AVI behaves across different populations, including people with adrenal disorders, those taking steroids, or patients whose scans are affected by technical factors like motion or contrast timing. The current reporting focuses on the association between adrenal volume, hormones and heart risk, but it does not yet spell out how the metric performs in every subgroup or how it should be interpreted alongside other imaging biomarkers. Until those details are clarified, clinicians will need to treat AVI as a powerful new clue rather than a standalone verdict, integrating it with clinical history, lab results and patient preferences.
Where stress imaging could go next
The discovery of a stress related signal in CT scans is likely just the beginning of a broader wave of AI driven imaging biomarkers. If deep learning can tease out patterns in adrenal glands that correlate with chronic tension, similar models might find signatures of stress in the brain, the heart or the immune system on MRI, echocardiography or even retinal photographs. I expect that future work will explore how AVI interacts with other markers, such as coronary calcium scores or measures of visceral fat, to build a more holistic picture of how stress weaves into the fabric of chronic disease.
At the same time, the use of routine scans as a platform for new insights raises important questions about consent, privacy and communication. Patients may need clearer explanations of what additional information is being extracted from their images and how it will be used. Health systems will have to decide when and how to report incidental findings like elevated AVI, especially if there is not yet a consensus on the best intervention. Those debates are a sign that imaging is moving into a new era, one in which the quiet hum of a CT scanner can reveal not only broken bones or hidden tumors, but also the long shadow that chronic stress casts over the body.
More from MorningOverview