High Reliability, Limited Validity: Assessment of a 3D Body Scanner to Estimate Anthropometric and Body Composition Measurements.

INTRODUCTION

Body composition and anthropometric measurements are valuable for predicting health and performance outcomes. Three-dimensional body scanners provide a portable, cost-effective, and rapid method for collecting anthropometric and body composition metrics; however, their accuracy must be validated against clinical reference methods to ensure that the outputs are reliable and valid.

METHODS

This study evaluated the reliability and validity of a portable 3D body scanner (Styku S100X) compared to manual anthropometry and dual-energy X-ray absorptiometry (DXA). Fifty participants (25 male, 25 female) underwent manual anthropometric assessments, two 3D full-body scans, and a full-body DXA scan. Reliability was assessed using intra-class correlations (ICC; k = 3, absolute agreement, two-way mixed-effects model), while validity was examined using Pearson's r correlations with manual anthropometry and DXA measurements. Data were stratified by sex, body mass index, body mass, stature, and age to provide exploratory insight into the potential influence of these factors on the 3D scanner's accuracy. For each group, mean difference with 95% CI, root mean square error (RMSE), and mean absolute percentage error (MAPE) were calculated.

RESULTS

The 3D body scanner demonstrated excellent reliability for segment circumferences and body composition (ICC > 0.95, p < 0.001) and good to excellent reliability for segment lengths and widths (ICC = 0.80-0.97, p < 0.001). Segment circumference estimates showed very large to nearly perfect correlations with manual anthropometry (r = 0.78-0.99, MAPE < 8%), but length and width estimates did not accurately reflect ISAK-standard measurements due to differences in landmark identification (r = 0.09-0.81, MAPE: 6-42%). Lean mass estimates demonstrated the strongest agreement with DXA (r = 0.61-0.95, MAPE < 8%), whereas body fat percentage (MAPE: 11-29%), fat mass (MAPE: 11-29%), and android and gynoid fat mass (MAPE: 59-74%) showed poor agreement.

CONCLUSIONS

These findings suggest that while the 3D scanner is suitable for tracking anthropometric and body composition measurements over time, body composition estimates (aside from lean mass) should not be considered accurate representations of an individual's body composition. Caution is advised when using 3D body scanner estimates for body composition assessment in clinical, sport, or fitness settings.