AI Study Links Brown Fat to Lower Cardiometabolic and Psychiatric Disease Risk

A new study published in Nature on 14 October 2026 details how artificial intelligence was used to analyze medical imaging data from more than 100,000 patients. The research identifies a link between the presence of brown adipose tissue (BAT) and a lower prevalence of various cardiometabolic and psychiatric conditions. The study also highlights a potential drug repurposing opportunity involving losartan, a common blood pressure medication.

AI-Driven Analysis of Patient Cohorts

The research team, led by Paul Cohen and colleagues, expanded a previous cohort to include 101,243 patients from Memorial Sloan Kettering Cancer Center. The data was derived from 18F-fluorodeoxyglucose (18F-FDG) positron emission tomography (PET)–computed tomography (CT) scans taken between 2018 and 2024. Of these patients, 8,897 were identified as BAT positive (BAT+), while 92,346 were BAT negative (BAT−).

To move beyond simple presence or absence, the researchers employed a machine learning approach called TriDFusion (3DF). This tool allowed for the semi-automatic quantification of BAT metabolic activity in approximately 8,700 patients. The AI analyzed total and depot-specific BAT activity, defined as the product of the mean standard uptake value (SUV) and BAT volume in millilitres.

Associations with Cardiometabolic and Psychiatric Health

Using propensity score matching to balance for age, sex, body mass index (BMI), and ambient temperature, the study found significant associations between BAT status and health outcomes. The analysis revealed that cardiovascular and psychiatric diseases were less prevalent in individuals with BAT.

Key findings included:

• Type 2 Diabetes (T2D): Patients with BAT had an odds ratio (OR) of 0.42 for T2D, indicating a substantially lower likelihood of diagnosis.
• Heart Disease: Chronic ischaemic heart disease showed an OR of 0.53 in the BAT+ group.
• Psychiatric Conditions: The study noted decreased odds of anxiety disorders (OR 0.78), depressive episodes (OR 0.68), and sleep disorders (OR 0.77) among those with BAT.

The researchers also examined specific BAT depots. The axillary depot showed the strongest association with reduced cardiometabolic disease risk. Higher axillary BAT activity was linked to lower odds of hypertension, type 2 diabetes, and congestive heart failure. Additionally, patients with BAT exhibited healthier white fat distribution, characterized by a higher ratio of subcutaneous adipose tissue (SAT) to visceral adipose tissue (VAT).

Medication Links and Losartan

The study analyzed medication usage to identify pharmacologic regulators of BAT. The use of angiotensin receptor blockers (ARBs) was positively associated with BAT presence. As a proof of concept, the researchers demonstrated that losartan, an ARB, increased brown adipocyte activity and whole-body energy expenditure through a peroxisome proliferator-activated receptor-γ (PPARγ)-dependent mechanism.

Laboratory values further supported the metabolic benefits of BAT. Patients with BAT had lower levels of glucose, haemoglobin A1C (HbA1c), triglycerides, and brain natriuretic peptide (BNP). They also had higher levels of high-density lipoprotein (HDL) cholesterol. While total cholesterol and low-density lipoprotein cholesterol (LDL-C) were higher in the BAT+ group, remnant cholesterol—a risk factor for atherosclerotic cardiovascular disease—was significantly lower.

Why it matters

This study demonstrates the utility of AI in extracting clinical phenotypes from existing medical imaging data. By quantifying BAT activity across a large cohort, the researchers identified new health associations beyond cardiometabolic disease, including psychiatric outcomes. The identification of losartan as a potential BAT activator offers a pathway for drug repurposing, potentially providing a therapeutic method to harness the metabolic benefits of brown fat in humans.

---

Sources

• nature.com: Artificial intelligence-enabled clinical phenotyping of human brown fat | Nature

This article was drafted with AI assistance and checked against the sources above. Company claims are reported as claims. Cover image is AI-generated.