AI-Powered Body Atlas Reveals Hidden Health Risks Beyond BMI

A new AI-driven atlas of human body composition has been developed using whole-body MRI scans from over 66,000 people. The research challenges traditional health assessments like BMI by providing detailed insights into fat and muscle distribution across the body. The study used advanced imaging and open-source AI to analyse scans with minimal human input. This framework can also be applied to routine chest or abdominal CTs and MRIs, improving patient monitoring in clinical settings.

Findings revealed that low skeletal muscle mass independently raises the risk of death from any cause by 1.44 times. High levels of fat within muscles were linked to a 1.54-fold increase in major cardiovascular events. Visceral fat, in particular, showed a strong connection to diabetes, with a 2.26-fold higher risk for those with elevated levels. Researchers also created reference curves showing how body composition changes with age, separated by sex and height. These trajectories could help personalise medical care, especially in oncology, where muscle and fat levels affect treatment side effects, survival rates, and cancer recurrence.

The atlas offers a more precise way to assess health risks than BMI alone. Its open-source AI tools allow wider use in hospitals, potentially improving early disease detection and treatment planning. The detailed body composition data could reshape how doctors evaluate and manage patient health.