DEXA vs 3D Body Scan vs InBody: Which Body Composition Test Is Accurate?

None of them give you your true body fat percentage, and the one that gets closest is not the one most people should be using.

That sounds like a dodge until you see how these machines work. Every body composition method you can book in the Chicago suburbs estimates your body fat from something else it can measure directly, and each one lands a few percentage points off in its own predictable direction. What separates a useful scan from a frustrating one is whether it makes the same mistake every single time, because a consistent error cancels out when you're comparing April to July.

What Each Method Actually Measures

No scanner can see fat. Each method measures something else and runs it through an equation.

DEXA passes two X-ray beams at different energies through you and measures how much of each gets absorbed. Fat, lean soft tissue, and bone mineral each absorb differently, so the machine can separate them across your whole body and by region. The radiation dose is small, roughly on par with a few hours of ordinary background exposure, but it is still ionizing radiation, which is why you'll be asked whether you might be pregnant.

A 3D optical scanner like Styku takes a different route. You stand on a rotating platform while infrared sensors build a high-resolution model of your body in about 30 seconds. From that model the software pulls volumes, circumferences, and shape ratios, then estimates fat and lean mass from those measurements. It never touches you and there's no radiation involved at all.

Bioelectrical impedance sends a small alternating current through your body and measures the resistance it meets. Lean tissue holds a lot of water and conducts well. Fat conducts poorly. The device converts that resistance into a body fat estimate using an equation built on an assumption about how hydrated you are, and that assumption is where most of the trouble starts.

The Comparison, Side by Side

Precision and Accuracy Are Not the Same Thing

Accuracy means how close a reading lands to the truth. Precision means how close two readings of the same thing land to each other. A bathroom scale that reads four pounds heavy every time is inaccurate and perfectly precise, and it will still tell you exactly how much weight you lost this month. A scale that swings randomly by four pounds is useless for that, even if it happens to average out correctly.

Most people shopping for a body composition scan are chasing accuracy. Almost everyone would be better served by precision, because the number that changes behavior is the difference between two scans rather than either scan on its own.

What the Research Says About 3D Scans

Two studies do most of the heavy lifting here, and they land in slightly different places, which is useful.

The larger one came out of the Shape Up! Adults study and looked specifically at the Styku S100 against DEXA in 188 adults, deliberately recruited across five ethnicities, three age bands, and the full BMI range.¹ Fat-free mass measured 54.1 kg by DEXA and 55.3 kg by the scanner, a mean difference of 1.2 kg, with a concordance correlation of 0.97. Fat mass concordance came in at 0.95. The limits of agreement were wider than that mean suggests, running from roughly 7 kg under to 5.6 kg over, so any individual reading can sit a fair distance from where DEXA would put it.

Repeatability is where it performed best. Scanning the same person twice moved fat mass by 0.41 kg and body fat percentage by 0.60 points, and waist and hip circumferences by less than 0.60 cm.¹ Results held steady across BMI, sex, and race or ethnicity, which is not something every body composition method can claim. The authors concluded the estimates are precise and accurate against criterion methods if the offsets are taken into account.¹

The second study is less flattering and we'd rather you hear it from us. Researchers at Texas Tech put four commercial 3D scanners, Styku among them, up against a four-component model, which combines DEXA, air displacement, and bioimpedance spectroscopy into a tougher reference than DEXA alone.⁴ All four scanners were reliable, with repeat-scan variation of 2.3% to 4.3% for body fat percentage and under 1.5% for fat-free mass. On absolute validity, three of the four scanners passed a 5% equivalence test against the reference model and Styku was the one that did not, and all four showed proportional bias, meaning the size of the error shifted depending on how lean or heavy the person was.⁴

A 3D scan is very good at telling you whether you changed, and less good at telling you your true body fat percentage on any given Tuesday. That's exactly how we use it.

Where Impedance Goes Sideways

Impedance is the most available option and the least stable one, because the current it measures is really measuring water.

In a study of 591 healthy adults, multifrequency impedance correlated with DEXA at 0.88 across the whole group and produced a mean body fat of 32.89% against DEXA's 34.72%.² Averages hide the problem. Broken out by body fat level, the device overestimated body fat by 3.03 points in men under 15% and by 4.40 points in women under 25%, then flipped and underestimated by 4.32 points in men above 25% and 2.71 points in women above 33%.² The bias runs in opposite directions at the two ends of the range, which means it can flatten real change in the middle of a weight loss phase.

Then there's day-to-day noise. A salty dinner, a hard training session, a few cups of coffee, or where you are in your cycle all move total body water, and the equation reads that shift as a change in lean mass. If you've ever stepped on a gym impedance scale two weeks apart and watched it claim you gained four pounds of muscle and lost four pounds of fat, this is why.

None of that makes impedance worthless. It's fast, free at most gyms, and fine for a rough screen. Just hold the numbers loosely, and take every reading fasted, unexercised, and at the same time of day if you're going to compare them at all.

The DEXA Problem

DEXA is the reference everything else gets measured against, and it is the strongest of the three. It also carries a spread of its own, which matters if you're treating a DEXA number as ground truth.

In July 2026, researchers scanned 99 adults on three DEXA systems on the same day: a Hologic Horizon A, a Hologic Discovery A, and a GE Lunar iDXA.³ Average body fat came back as 36.4%, 34.4%, and 35.2% depending on the machine, and the differences were statistically significant and not uniform across sexes or BMI categories.³ The authors concluded that whether another device looks accurate depends on which DEXA you call the truth.³

So a two-point gap between your DEXA result and your 3D scan might be a real difference in method, or it might be the same size as the gap between two DEXA machines in the same city. Either way, mixing methods mid-protocol will hand you a number you can't interpret.

Why This Matters More on a GLP-1

Weight loss without a body composition baseline is a guess about the most important question.

On a GLP-1, the scale moves reliably, and it tells you nothing about whether the pounds leaving are fat or lean tissue. Two people can drop the same 15 pounds over the same 12 weeks and end up in completely different metabolic positions. A scan every 4 to 8 weeks catches a lean-mass problem early, while there's still time to fix it with protein intake and resistance training rather than finding out at the end. We go deeper on that in our guide to protecting muscle mass on a GLP-1, and on what the trial data shows about the fat-to-lean split in body composition on a GLP-1.

This is also why we build the scan into the protocol instead of selling it as an add-on. Every medical weight loss program at Defiant starts with a baseline scan before week one, then re-scans on a schedule, so the dose conversation and the training conversation are both anchored to something real.

How to Scan So the Numbers Mean Something

Repeatability is mostly a function of how you show up. Five habits do most of the work:

• Same method, same machine, every time. Switching from an InBody to a DEXA halfway through a protocol produces a change you cannot attribute to anything.
• Same time of day. Morning is easiest to reproduce. Body water shifts over a day even if nothing else does.
• Similar hydration and food. Fasted or lightly fed, consistently. This matters enormously for impedance and modestly for the other two.
• Form-fitting clothing. Bulky layers change a 3D scan's volume estimate. Compression gear or a sports bra and shorts keeps it clean.
• Don't scan weekly. At a four-week interval, real change outruns measurement noise. At a one-week interval it usually doesn't, and you end up reacting to nothing.

Getting Scanned Near Naperville

Defiant is at 5100 Lincoln Ave in Lisle, about 10 minutes from downtown Naperville, and also convenient to Downers Grove, Wheaton, Bolingbrook, and Lombard. A single Styku 3D body scan is $50 and takes about 30 seconds on the platform, with 10 to 15 minutes total for check-in and a walk through your results. Most weight loss and peptide protocols build the scan in at a 4-to-8-week cadence, and scans are included in the Core Wellness credits on every membership tier.

For a research-grade snapshot, a DEXA at an imaging center or a university exercise lab is the right call, and we'll happily read the report alongside your scan history. Styku is a tracking tool rather than a clinical diagnostic, and we'd rather say that plainly than let anyone assume it replaces medical imaging.

Frequently Asked Questions

Last updated September 20, 2026.