Your phone can now measure body fat as well as hospital scanners — and predict heart risk better than BMI

Researchers built a smartphone app that estimates body composition from photos with accuracy matching hospital-grade scanners, then showed it predicts insulin resistance better than BMI alone. This means doctors could screen for cardiometabolic risk using a tool patients already carry, without expensive equipment or clinic visits.

BMI has been the standard risk measure for decades because it's cheap and scalable — but it's a blunt instrument that misses the actual problem: where fat sits on your body matters more than how much you have. Visceral fat (around organs) drives heart disease and diabetes; subcutaneous fat (under skin) doesn't. This paper shows a smartphone measurement captures that distinction well enough to change clinical decisions about who needs intervention. The catch is deployment: the model was trained on UK Biobank data and tested on small cohorts. Whether hospitals and insurers actually adopt this instead of continuing to use BMI — which requires nothing but a scale and a height measurement — depends on whether the accuracy advantage translates into measurable health outcomes at scale, not just in research settings.