3-Dimensional optical scanning for body composition assessment: A 4-component model comparison of four commercially available scanners. Issue 10 (October 2020)

• Record Type: Journal Article
• Authors: Tinsley, Grant M.; Moore, M. Lane; Benavides, Marqui L.; Dellinger, Jacob R.; Adamson, Brian T.
• Is Part Of: Clinical nutrition. Volume 39:Issue 10(2020)
• Page Start: 3160
• Page End: 3167
• Publication Date: 2020-10
• Subjects: Digital anthropometry -- Body composition assessment -- 4-Compartment model -- 3D scanning -- Anthropometry -- Obesity; Critically ill -- Nutrition -- Periodicals; Diet therapy -- Periodicals; Parenteral feeding -- Periodicals; Enteral feeding -- Periodicals; Enteral Nutrition -- Periodicals; Parenteral Nutrition -- Periodicals; Metabolism -- Periodicals; Diétothérapie -- Périodiques; Alimentation parentérale -- Périodiques; Alimentation entérale -- Périodiques; Nutrition -- Périodiques; Diet therapy; Enteral feeding; Nutrition; Parenteral feeding; Electronic journals; Periodicals; Electronic journals; 615.854
• DOI: 10.1016/j.clnu.2020.02.008
• Languages: English
• ISSNs: 0261-5614

Abstract

Summary: Background & aims: Body composition assessment via 3-dimensional optical (3DO) scanning has emerged as a rapid and simple evaluation method. The aim of this study was to establish the precision of body composition estimates from four commercially available 3DO scanners and evaluate their validity as compared to a reference 4-component (4C) model. Methods: The body composition of 171 participants was assessed using four commercially-available 3DO scanners (FIT3D®, Naked Labs®, Size Stream®, and Styku®) and a 4C model utilizing data from dual-energy x-ray absorptiometry, air displacement plethysmography, and bioimpedance spectroscopy. Body composition estimates were compared via equivalence testing, Deming regression, Bland–Altman analysis, concordance correlation coefficients (CCC), root mean square error (RMSE), and related metrics. Precision metrics, including the root mean square coefficient of variation (RMS-%CV), precision error, and intraclass correlation coefficient, were generated for duplicate scans in 139 participants. Results: All scanners produced reasonably reliable estimates, with RMS-%CV of 2.3–4.3% for body fat percentage (BF%), 2.5–4.3% for fat mass (FM), and 0.7–1.4% for fat-free mass (FFM). ICC values ranged from 0.975 to 0.996 for BF% and 0.990 to 0.999 for FM and FFM. All scanners except Styku® demonstrated equivalence with 4C, using 5% equivalence regions, and constant errors of <1% for BF% and ≤0.5 kg for FM and FFM. However, the slopes of regression lines differed from the line of identity for most scanners and variables. CCC values ranged from 0.74 to 0.90 for BF%, 0.85 to 0.95 for FM, and 0.93 to 0.97 for FFM. RMSE values ranged from 3.7 to 6.1% for BF% and 2.8–4.6 kg for FM and FFM. Bland–Altman analysis indicated proportional bias of varying magnitudes was present for all scanners. Conclusions: Commercially available 3DO scanners produce relatively reliable body composition estimates. Three out of four scanners demonstrated equivalence with a 4C model for assessments of BF%, FM, and FFM, although other metrics of validity varied among scanners, and proportional bias was present for all scanners.