Company Profile / Digital Health & AI Diagnostics
Edina, Minnesota
DexaFit
The body-scan company that taught an AI to read the same seven-minute scan for signs of chronic disease.
Body Composition • Metabolism • Disease Risk
The body-scan company that taught an AI to read the same seven-minute scan for signs of chronic disease.
Body Composition • Metabolism • Disease Risk
Most people meet heart disease the hard way - in an emergency room, after it has already announced itself. DexaFit's wager is that the warning was there earlier, hiding inside a routine scan most people take for a very different reason.
DexaFit started in 2011 as something far more familiar to the fitness world: a place to get a DXA scan. Dual-energy X-ray absorptiometry - the same technology clinics use to measure bone density - can, in a few minutes, break a body down into fat, muscle, visceral fat and bone. For athletes, dieters and the biohacker crowd, it replaced the bathroom-scale guess with hard numbers. Co-founders Adam Kadela and former NFL player Matt Ulrich, teammates at Northwestern, opened the first studio in Chicago and turned that measurement into a business.
Over the following decade the brand spread through a network of independently operated partner centers, adding VO2 max cardio-fitness tests and resting metabolic rate measurements to the menu, and wrapping the results in an app that scores longevity and tracks trends over time. The premise stayed consistent: give people objective, medical-grade numbers instead of generic advice.
Then came the harder idea. The DXA scan already captures an enormous amount of data - far more than a body-fat percentage. What if a machine-learning model, trained on hundreds of thousands of those scans, could spot the fingerprints of disease that no technician would see by eye? That question became DexaFit Dx, an AI system aimed at screening for coronary artery disease, type 2 diabetes and musculoskeletal conditions from a scan that costs little and exposes the patient to almost no radiation.
It is a neat piece of leverage. Rather than invent a new machine or a new patient behavior, DexaFit is asking a trusted, decades-old scanner to do a second job. The company reports its model was trained on more than 250,000 scans and 5 billion data points, and claims over 95% accuracy with an under-2% false-positive rate. Those are the company's own figures; the tool is not yet commercially available and is pursuing FDA clearance as a Software as a Medical Device.
Chronic diseases like heart disease and type 2 diabetes account for a large share of deaths, and they are often silent until late. Screening exists, but much of it is invasive, expensive, or reserved for people already flagged as high-risk. DexaFit's answer is to attach detection to a test people are willing to take anyway - and to make the abstract idea of "prevention" into something concrete: a number that moves.
Illustrative view of the signals DexaFit reads from one scan. Bars reflect relative emphasis in the company's offering, not clinical performance.
Two layers: the in-center tests that built the brand, and the software that is trying to extend it into clinical territory.
A full-body scan measuring body fat, muscle mass (ALMI), visceral fat, regional breakdown and bone density - with minimal radiation.
Cardiorespiratory fitness testing that measures maximal oxygen uptake to gauge aerobic capacity and endurance.
Resting metabolic rate testing - how many calories the body burns at rest - to guide nutrition and weight plans.
Turns testing data into a longevity score, diabetes and heart-disease risk estimates, trend tracking and goal-setting.
Machine learning that analyzes standard DXA scans to screen for coronary artery disease, type 2 diabetes and musculoskeletal conditions. FDA clearance pending.
On the body-composition side, DexaFit competes with consumer-scanning brands like BodySpec, Fit3D and InBody, plus hospital and clinic DXA services. Its differentiator there is breadth - bundling DXA, VO2 max and RMR into one data-driven experience with an app on top.
On the AI-screening side, the comparison shifts to coronary calcium CT scoring and a growing field of cardiovascular-risk and medical-imaging AI companies. DexaFit's angle is cost and convenience: piggybacking detection on a low-radiation scan people already take, rather than sending them for a dedicated, pricier test.
The strategy has two reinforcing layers. Local partner centers supply reach and the human touch; a shared AI model supplies leverage. Reach without leverage is just a franchise; leverage without reach has nowhere to run. DexaFit is betting the combination is the point.
COO / CPO. Opened the first DexaFit in Chicago in 2011; earlier led foreign exchange at a Chicago trading and technology firm.
Former Northwestern teammate and NFL alum who co-founded DexaFit alongside Kadela.
Leads technology for DexaFit Dx. Studied theoretical mathematics at Stanford; worked on computer vision and machine learning, including a stint at Dropbox.
Chief Executive Officer, steering the company across its testing network and AI ambitions.
Sajad Zalzala and Sandra Bender serve as medical directors, anchoring the clinical side of the business.
Adam Kadela and Matt Ulrich open the first DexaFit body-composition studio.
Independently operated DexaFit centers open across the U.S., including Kansas City.
An app translates DXA, VO2 max and RMR data into longevity scores and risk insights.
Development starts on AI to detect coronary artery disease and other conditions from DXA scans.
Partner locations bundle RMR and VO2 max testing alongside DXA scans.
DexaFit Dx works toward Software as a Medical Device clearance; early-access sign-ups open.
Your health is a narrative, a story unfolding with every heartbeat, and we're here to help you script every chapter with clarity and purpose.
A simple, 7-minute full-body DXA scan becomes a radically simple and easy-to-do screening test.
A world where there are no sudden deaths from heart attacks, and cancer is detected when it can still be cured.
Sources: DexaFit Dx, DexaFit.com, LinkedIn, The Org, Crunchbase, PitchBook, published interviews. Figures such as accuracy, scan counts and radiation are company-reported and approximate.