Agreement Between A 2-Dimensional Digital Image-Based 3-Compartment Body Composition Model and Dual Energy X-Ray Absorptiometry for The Estimation of Relative Adiposity

Abstract

The purpose of this study was to compare relative adiposity (%Fat) derived from a 2-dimensional image-based 3-component (3C) model (%Fat_3C-IMAGE) and dual-energy X-ray absorptiometry (DXA) (%Fat_DXA) against a 5-component (5C) laboratory criterion (%Fat_5C). 57 participants were included (63.2% male, 84.2% White/Caucasian, 22.5±4.7 yrs., 23.9±2.8 kg/m²). For each participant, body mass and standing height were measured to the nearest 0.1 kg and 0.1 cm, respectively. A digital image of each participant was taken using a 9.7 inch, 16g iPad Air 2 and analyzed using a commercially available application (version 1.1.2, made Health and Fitness, USA) for the estimation of body volume (BV) and inclusion in %Fat_3C-IMAGE . %Fat_3C-IMAGE and %Fat_5C included measures of total body water derived from bioimpedance spectroscopy. The criterion %Fat_5C included BV estimates derived from underwater weighing and bone mineral content measures via DXA. %Fat_DXA estimates were calculated from a whole-body DXA scan. A standardized mean effect size (ES) assessed the magnitude of differences between models with values of 0.2, 0.5, and 0.8 for small, moderate, and large differences, respectively. Data are presented as mean ± standard deviation. A strong correlation (r = 0.94, p <.001) and small mean difference (ES = 0.24, p <.001) was observed between %Fat_3C-IMAGE (19.20±5.80) and %Fat_5C (17.69±6.20) whereas a strong correlation (r = 0.87, p <.001) and moderate-large mean difference (ES = 0.70, p <.001) was observed between %Fat_DXA (22.01±6.81) and %Fat_5C. Furthermore, %Fat_3C-IMAGE (SEE = 2.20 %Fat, TE= 2.6) exhibited smaller SEE and TE than %Fat_DXA (SEE = 3.14 %Fat, TE = 5.5). The 3C image-based model performed slightly better in our sample of young adults than the DXA 3C model. Thus, the 2D image analysis program provides an accurate and non-invasive estimate of %Fat within a 3C model in young adults. Compared to DXA, the 3C image-based model allows for a more cost-effective and portable method of body composition assessment, potentially increasing accessibility to multi-component methods.

Introduction

A 3 –component (3C) model of body composition assessment is considered superior to simpler 2 –component (2C) models, as it reduces the proportional assumptions of 2C models by accounting for components of fat –free mass (FFM). Typically, a 3C model measures fat mass (FM) and divides FFM into total body water (TBW) and fat-free dry mass, thus, accounting for potential variations in FFM hydration (1). A laboratory-based 3C model utilizes underwater weighing (UWW) or air displacement plethysmography (ADP) to measure body volume (BV), and isotope dilution technique to estimate TBW. An alternative laboratory-based 3C model that is commonly used in body composition research is dual energy X-ray absorptiometry (DXA); this method quantifies FM and partitions FFM into bone mineral content (BMC) and lean soft tissue (1-3). However, a limitation of DXA, and similar to 2C models, is it assumes FFM is of constant hydration across all individuals (1,4).

The necessity to utilize expensive, time-consuming, and complex laboratory-based methods restricts the accessibility of a 3C model outside of a research setting. Nevertheless, advancements in technology and research have mitigated some of the burden and increased the potential portability and practicality of a 3C body composition assessment. For instance, research has commonly used bioimpedance spectroscopy (BIS) for TBW measurements in a 3C model due to the ease of administration (5-7). However, an alternative approach for the measurement of BV, that can easily be obtained in various settings, has been more challenging and remains an area of interest in body composition research.

Alternative 3C models using ultrasound or skinfold measures to estimate BV are strongly correlated with laboratory-based 3C and 4-component (4C) models (6,8-10). Skinfold measures of BV provide practitioners with an inexpensive, relatively quick, and portable method of 3C body composition assessment (11). However, research highlighting inter-rater reliability concerns of skinfold measures infers that the accuracy of skinfold measurements heavily relies on the experience and skill of the evaluator (12-14). Recent advancements in technology has increased the accessibility of ultrasound. However, ultrasound devices are expensive and, similar to skinfold measures, the accuracy of ultrasound measurements is technician-dependent (15). Additionally, there has yet to be a consensus on ultrasound protocol standards for body composition assessment, making comparisons across studies difficult (15,16). Collectively, the development of a simpler and more objective method for BV has been sought.

Simple, portable, and accurate methods for BV measurement would be a valuable resource to practitioners. A 2-dimensional (2D) image analysis program has recently been developed to estimate FM, FFM, body fat percentage (%Fat), and BV from a single digital image, taken on a smartphone or tablet (17). The 2D image analysis program automatically identifies and measures the horizontal linear diameter of various anatomical landmarks and processes this information via an undisclosed proprietary algorithm (17). Recently published research has demonstrated that the novel 2D image analysis program accurately estimates BV when compared to an UWW laboratory criterion (18,19). Body volume estimated from the novel 2D image analysis program does not require specialized equipment, other than a smartphone or tablet, and can be completed by most individuals with minimal training when compared to alternative BV measurements such as skinfold and ultrasound.

The novel 2D image analysis program has also demonstrated the capacity to accurately estimate relative adiposity in a 3C model (%Fat_3C-IMAGE) (18,19). However, while DXA has been used as a measure of BMC and BV in previous validation studies, none included DXA as a stand-alone measure of %Fat (%Fat_DXA). Although DXA is commonly used as a criterion method in body composition research, its inability to account for TBW may increase measurement error, especially in populations whose FFM hydration commonly deviates from assumed values (20-23). Furthermore, DXA requires expensive equipment, exposes participants to low levels of radiation, and subsequently, may require additional state or county approval due to radiation exposure (24), further limiting its accessibility and application.

Currently, no research has assessed the relative agreement between the 3C image-based model that includes body mass, BV, and TBW to a 3C DXA estimate which does not account for TBW. Furthermore, the validity of the 3C image-based model has yet to be examined against a 5-component (5C) lab criterion measure of body composition (%Fat_5C). Therefore, the purpose of this study was to compare %Fat estimated from a 3C model using BV derived from the 2D image analysis program (%Fat_3C-IMAGE), dual-energy X-ray absorptiometry (%Fat_DXA) and a %Fat_5C laboratory criterion using BV derived from UWW. The authors hypothesized that the %Fat_3C-IMAGE and %Fat_DXA would provide comparable and acceptable agreement when compared to the %Fat_5C criterion.

Methods

Participants

This study includes data collected between 2019 and 2021. A sub-sample of participants included in the current study were included in a recently published study by Fedewa et al. and Sullivan et al. (18, 19). All participants met the following pre-established eligibility criteria: 18 years or older, willing to comply with the study procedures, and ability to provide their own transportation to the testing laboratory. Fifty-seven adults (36 males, 21 females) with complete body composition data (including 2D image, UWW, BIS, and DXA) were included in this study. Participant characteristics are displayed in Table 1. Written informed consent was obtained from each participant prior to data collection. Additionally, all participants completed a medical history form and physical activity readiness questionnaire before involvement in the current study. This study was approved by the University of Alabama’s Institutional Review Board.

Table 1.

Descriptive Characteristics Of Study Participants (n = 57).

Age (years)	22.5±4.7
Height (cm)	174.7±8.7
Weight (kg)	73.4 ±12.6
Body Mass Index (kg/m2)	23.9±2.8
Bone Mineral Content (kg)	3.0±0.6
Body Volume (UWW)	69.0±11.7
Body Volume (IMAGE)	69.1±11.8
Total Body Water (L)	40.9±9.7

Notes: cm, centimeters. kg, kilograms. kg/m², kilograms per meter squared. L, liters. UWW, underwater weighing.

Procedures

All testing procedures were completed during a single visit to the Exercise Physiology laboratory at the University of Alabama. Prior to arriving to the laboratory for their testing visit, participants were instructed to abstain from exercise and the ingestion of food and drink (except water) for a minimum of 12 hours.

Anthropometrics

For each participant, body mass was measured, to the nearest 0.1 kg, with a calibrated digital scale (Tanita BWB −800, Tanita, Arlington Heights, IL) and standing height (without shoes) was measured to the nearest 0.1 cm with a manual stadiometer (SECA 213, Seca Ltd., Hamburg, Germany). For descriptive purposes, body mass index (BMI) was calculated as BMI = weight (kg) ÷ [height (m)]² and reported in kg/m².

Digital Image Analysis

For image procurement, participants wore snug-fitting athletic clothing that allowed for the automated 2D image analysis program to identify the necessary anatomical points of interest. Participants with long hair were instructed to pull their hair “back” and “up” to allow the digital image to show the diameter of the neck. Participants stood with their feet flat in front of a white photography background facing away from the digital camera, with weight evenly distributed on both feet. The heels were placed together with the feet pointed slightly outward at a 60-degree angle. Participants were required to remain motionless with arms abducted at a 45-degree angle away from the torso and aligned within the coronal plane, with palms facing away from the camera. Once correctly positioned, a single digital image that included the head, feet, and arms of the individual was obtained from the rear/posterior view using a 9.7 inch, 16g iPad Air 2 and analyzed using a commercially available application (version 1.1.2, made Health and Fitness, USA. www.mymadeapp.com). Body volume was derived from the 2D digital image, using a proprietary algorithm which automatically identifies and measures the horizontal linear diameter of various anatomical landmarks (United States Utility Patent 16/841,944), and was used in the 3C model calculation of %Fat_3C-IMAGE (17). The development of the proprietary algorithm was completed using participants not included in the current study.

Bioimpedance Spectroscopy

Hand-to-foot BIS (Imp SFB7, ImpediMed Limited, Queensland, Australia) was used to determine TBW for the %Fat_3C-IMAGE and %Fat_5C models. Multi-component models of body composition assessment commonly use BIS for the estimation of TBW as it is generally accepted as a more convenient alternative to the traditional isotopic dilution criterion, when applied properly in healthy adults (6,7). Prior to electrode placement, sites were cleaned with alcohol pads and excess hair was removed with a razor. In accordance with manufacturer’s specifications, electrodes were placed on the right hand and right foot with participants in a supine position with the arms ≥30° away from the body with legs separated.

Dual Energy X-Ray Absorptiometry

Bone mineral content and %Fat_DXA were estimated with a whole-body DXA scan (Lunar Prodigy, version 14.10.022, GE Healthcare, Madison, WI). Prior to each scan, the DXA was calibrated according to the manufacturer’s instructions using a standard calibration block. Participants were instructed to remove shoes, all jewelry, and bulky clothing prior to the scan. Whole body scans were performed with participants lying supine with their arms at their side and palms against their legs. Velcro straps were placed around the ankles and knees of each participant in order to prevent lower limb movement. Total body %Fat_DXA was estimated using the manufacturers proprietary algorithm. For use in the 5C criterion model, the BMC measures from DXA were converted to total body bone mineral (Mo) and total body soft tissue mineral (M_S) using the following equations (25,26) where BMC, M_O, and M_S are measured in kilograms and TBW is in liters:

$$\begin{gathered} M_O=BMC\times 1.0436 \\ {M}_S=[0.882\times (12.9\ast TBW)+37.9]∕1000 \end{gathered}$$

Underwater Weighing

For each participant, residual lung volume was measured on land, before UWW, using Parvo software (TrueOne 2400, Parvo Medics, Sandy, UT, UAS). All UWW occurred in a heated, custom-built tank with participants wearing form-fitted clothing or a bathing suit. Participants entered the UWW tank and were positioned into a sling seat suspended from a calibrated Chatillon 15-kg scale (Model #1315DD-H, Largo, FL). Participants were then instructed to fully submerge underwater and maximally exhale; with all body parts submerged, the participant’s underwater weight was recorded (to the nearest 0.025 kg). 6 to 10 trials were completed per participant with the average of the three highest underwater weight values used to determine BV_UWW for inclusion in the criterion %Fat_5C model. BV_UWW was calculated using the following equation, where body mass is in kilograms, underwater weight is in pounds, and residual lung volume is in liters.

$$\text{BV}=(\text{body mass}-\text{underwater weight})÷(0.9937))-(0.1+\text{RV})$$

Three- and Five-Component Model Calculations

Body volume estimates from the 2D image analysis system were combined with TBW and body mass for the %Fat_3C-IMAGE model calculation as described by Siri et al. (27).

$$\begin{gathered} FM=(2.118\times BV)-(0.78\times TBW)-(1.351\times BM) \\ \%Fa{t}_{3C-IMAGE}=(FM∕BM)\times 100 \end{gathered}$$

The criterion estimate of %fat was calculated using the %Fat_5C model described by Wang et al. (26), and was determined using the following equations, where FM, BM, M_O, M_S, are in kilograms and BV and TBW are in liters.

$$\begin{gathered} FM=(2.748\times BV)-(0.715\times TBW)+(1.129\times M_O) \\ +(1.222\times M_S)-(2.051\times BM) \\ \%Fa{t}_{5C}=(FM∕BM)\times 100 \end{gathered}$$

Statistical Analysis

Statistical analyses were performed using SPSS for Windows (SPSS 25.0, Chicago, IL). The number of participants included in the current study, exceeded the minimum number of participants (n = 34) required to detect a small mean difference (ES = 0.2) between measures with 0.8 power, assuming a correlation between measures of 0.9 and an alpha level of 0.05. Values of %Fat were compared using repeated measures analysis of variance (ANOVA), with a priori planned contrasts between each alternative measure (%Fat_3C-IMAGE and %Fat_DXA) and the %Fat_5C criterion measure. The constant error (CE) was calculated as the difference between the criterion and alternative measure, such that a positive CE would indicate the alternative measure overestimated %Fat and a negative CE would indicate the alternative measure underestimated %Fat. In addition, the absolute value of the CE was used to calculate the mean absolute error. This approach is commonly used when assessing the validity agreement between different clinical measures to enhance the identification of poor agreement, which can be masked when overestimated and underestimated values are averaged together resulting in no observed mean difference (28-30). The magnitude of the differences between the %Fat_3C-IMAGE, %Fat_DXA, and %Fat_5C models were assessed using a standardized mean effect size (ES), by dividing the difference between the criterion and alternative measure by the standard deviation of the criterion (31). Threshold values for the standardized ES were 0.2, 0.5, and 0.8 for small, moderate, and large differences, respectively (32). Equivalence testing was used to determine if the %Fat3C-IMAGE and %FatDXA measures could be considered equal to the %Fat5C criterion measure, even in the presence or absence of statistical differences, based on a ±5% equivalence region and 90% confidence limits for the individual comparisons (33). Bivariate correlations and univariate linear regression determined the strength of the association between %Fat derived from the alternative methods and 5C criterion method. Regression procedures were used to determine the Pearson’s product moment correlation coefficients (r), standard error of the estimate (SEE), and total error (TE) (24, 34). Furthermore, standards outlined by Lohman and Heyward were used to qualitatively describe the level of agreement and accuracy observed from each %Fat SEE as follows: 2.0 as ideal, 2.5 as excellent, 3.0 as very good, 3.5 as good, 4.0 as fairly good, 4.5 as fair, and 5.0 as poor (24,35). The strength of each r value was described as follows: 0.2, 0.5, and 0.8, categorized as small, moderate, and large, respectively (32). Fisher’s r to z transformation was used to determine if the strength of the observed correlations were statistically different from each other (36). The Bland-Altman method was used to identify the 95% limits of agreement for all metrics of body composition (37). Data were screened for outliers and normal distribution with skewness or kurtosis >2 indicating non-normal distribution. Statistical significance was determined using an alpha p <.05. All data are expressed as mean ± standard deviation (M±SD), unless otherwise indicated.

Results

Participant age ranged from 18 to 39, with the majority of participants (82.5%) between the ages of 18 to 24. Caucasian/White participants accounted for a large portion of the sample (n = 48, 84.2%), with a small number of Hispanic/Latino (n = 4, 7.0%), Asian (n = 2, 3.5%), Black/African American (n = 2, 3.5%) participants. One participant self-identified as multi-racial. In addition, female participants accounted for over one-third of the sample (n = 21, 36.8%). Body mass index ranged from 15.3 to 31.0 kg/m², with the majority of participants (70.1%) categorized as “normal weight” (>18.5 kg/m² and <25 kg/m²). See Table 1 for additional descriptive characteristics of the study sample.

Data was normally distributed as skewness and kurtosis values were all between −2 and +2. Although Mauchly’s Test of Sphericity was statistically significant, the omnibus test for potential differences between measures was statistically significant using both the Greenhouse-Geisser (p <.001) and Huynh-Feldt (p <.001) correction. The ±5% equivalence region for %Fat_5C required that the confidence interval for the difference between criterion and alternative measures fell between −0.88 %Fat and 0.88 %Fat. The corresponding 90% confidence interval for the difference between %Fat_3C-IMAGE and the %Fat_5C criterion was 1.02 to 1.98 %Fat, which was not inside the equivalence region. Similarly, the corresponding 90% confidence interval for the difference between %Fat_3C-IMAGE and the %Fat_5C criterion was 3.55 to 5.07 %Fat, which was also not inside the equivalence region. Thus, the means of the %Fat_3C-IMAGE and %Fat_DXA measures were not considered equivalent to the %Fat_5C criterion measure. Comparisons between %Fat values derived from the three models are shown in Table 2.

Table 2.

Agreement Between Alternative Measures of a Field 3–Compartnent Model and Dual-Energy X-Ray Absorptiometry When Compared to a 5–Component Criterion Model in Healthy Adults.

	Mean ± SD	p	ES	r	SEE	TE	Limits of Agreement
							CE ± 1.96 SD	Upper	Lower	Trend
%Fat5C	17.69 ± 6.20	-	-	-	-	-	-	-	-	-
%Fat3C-IMAGE	19.20 ± 5.80	<.001	0.24	0.94	2.20	2.63	1.50 ± 4.27	5.77	−2.77	−0.19
%FatDXA	22.01 ± 6.81	<.001	0.70	0.87	3.14	5.49	4.31 ± 6.73	11.04	−2.41	0.18

Note: %Fat, percent body fat. 3C, 3-component; 5C, 5-component; CE, constant error; DXA, dual energy X-ray absorptiometry; ES, effect size; FFM, fat-free mass; FM, fat mass; r, Pearson Product Moment correlation coefficient; SD, standard deviation; SEE, standard error of the estimate; TE, total error.

When comparing the mean differences between measures, a small difference was observed when comparing %Fat_3C-IMAGE to the %Fat_5C criterion (ES = 0.24, p <.001), whereas a moderate-large difference was observed when comparing %Fat_DXA to the %Fat_5C criterion (ES = 0.70, p <.001), such that both methods overestimated %Fat. Better agreement between the %Fat_3C-IMAGE and %Fat_5C criterion was also indicated by the smaller mean absolute error (2.22±1.41 %Fat), which was significantly less than the mean absolute error between the %Fat_DXA and %Fat_5C criterion (4.64±2.96 %Fat) (p <.001). In addition, %Fat_3C-IMAGE (r = 0.94, p <.001) and %Fat_DXA (r = 0.87, p <.001) were strongly correlated with %Fat_5C, however the %Fat_3C-IMAGE model yielded a significantly stronger relationship (p =.006). The observed agreement between %Fat_3C-IMAGE and %Fat_5C (SEE = 2.20 %Fat) was categorized as “ideal-excellent,” whereas the agreement between %Fat_DXA and %Fat_5C (SEE = 3.14 %Fat) was categorized as “good-very good.” The Bland–Altman Plots comparing the agreement between %Fat_3C-IMAGE, %Fat_DXA, and %Fat_5C are shown in( Figs. 1 and 2). No significant trends in regression lines of the Bland-Altman plots were observed for %Fat in either model. The 95% limits of agreement and bias trend values displayed in Table 2.

Fig. 1.

Bland–Altman plot of the difference between relative adiposity (%Fat) measured by the 3–component image-based model (%Fat_3C-IMAGE) and the 5–component criterion method (%Fat_5C). The solid line indicates the line of best fit. The fine dotted line in the middle indicates the mean difference, and the dotted lines at the top and bottom of the graph indicate the upper and lower 95% limits of agreement.

Fig. 2.

Bland–Altman plot of the difference between relative adiposity (%Fat) measured using dual energy X-ray absorptiometry (%Fat_DXA) and the 5–component criterion (%Fat_5C). The solid line indicates the line of best fit. The fine dotted line in the middle indicates the mean difference, and the dotted lines at the top and bottom of the graph indicate the upper and lower 95% limits of agreement.

Discussion

The purpose of this study was to determine the relative agreement of a field 3C body composition model and DXA with a 5C laboratory criterion when estimating %Fat in a sample of young adults. These data indicate that both techniques yielded comparable agreement, however, the field 3C exhibited a smaller mean difference, as well as lower SEE and LOA than DXA. Thus, the hypotheses were accepted as these findings indicate that the novel 2D image analysis program provides a valid estimate of %Fat when incorporated in a 3C model.

The results of the current study are consistent with previous research in NCAA Division I female athletes conducted by Moon et al. (38). In their sample of 29 participants, the laboratory %Fat_3C (r >0.99, SEE = 0.38 %Fat) and %Fat_DXA (r = 0.93, SEE = 1.78 %Fat) were strongly correlated with %Fat_5C. While both techniques yielded stronger correlations than were observed in the current study, the %Fat_3C yielded superior agreement than %Fat_DXA(38). Similar results were also observed in a sample of 27 male and female bodybuilders, in which a laboratory %Fat_3C model including total body water estimates from bioimpedance spectroscopy (r = 0.99, SEE = 0.69 %Fat), multi-frequency bioimpedance assessment (r = 0.92, SEE = 1.84 %Fat), and single-frequency bioimpedance assessment (r = 0.93, SEE = 1.82 %Fat) all yielded better agreement than %Fat_DXA (r = 0.80, SEE = 1.99 %Fat) (39).

Comparable results were also reported by Andreoli et al., where %Fat_3C estimates including TBW and BV yielded greater agreement than %Fat_DXA values in a small sample of professional water polo players (4). Additionally, Tinsley et al., recently compared the validity of %Fat estimates from several body composition methods against a %Fat_5C model (40). Tinsley et al. reported that %Fat estimates from a 3C (R² >0.99, SEE = 0.39 %Fat) and 4C model (R² >0.99, SEE = 0.02 %Fat) were not different from the %Fat_5C criterion. Furthermore, Tinsley et al. reported that although %Fat_DXA was strongly correlated with a %Fat_5C criterion, (R² = 0.95, SEE = 2.06 %Fat), DXA produced %Fat values that were significantly different than the 5C model, such that DXA tended to overestimate %Fat (40), further demonstrating that a %Fat_3C model with TBW tends to perform better than %Fat_DXA for the estimation of %Fat. These cumulative results, including those of the current study, highlight the possible limitations of DXA derived %Fat values, due to its inability to account for the variability in TBW.

The use of DXA method reduces the amount of equipment needed and some of the physical burden of a multi-component body composition assessment. Furthermore, DXA not only provides estimates of %Fat but also leans mass, and regional composition estimates. Although convenient, DXA requires expensive equipment, exposes participants to low levels of radiation, and may require additional state or county approval due to radiation exposure (24). Furthermore, DXA’s failure to estimate TBW in populations whose FFM hydration substantially differs from assumed values, may lead towards increased measurement error. For instance, older adults, athletes/individuals engaging in consistent exercise training, and certain ethnicities have all shown to exhibit variations in the composition of FFM that violate the proportional assumption that FFM is of constant hydration across all individuals (20-23).

In comparison to an alternative field-based 3C model which used skinfold measures to estimate BV and BIS for TBW (%Fat_3C-SKF), the %Fat_3C-IMAGE model in the current study demonstrated slightly better agreement with a 5C criterion than did the %Fat_3C-SKF model when compared against a 4C criterion (6,10). When compared against a 4C model, the field-based %Fat_3C-SKF model produced %Fat estimates that were strongly correlated to the 4C criterion (r = 0.94, SEE = 2.87 %Fat). However, as mentioned previously, skinfold measurements are prone to measurement inconsistencies, potentially limiting their utility despite their widespread use among practitioners (13,14). It is important to highlight that the field-based %Fat_3C-SKF model was compared to a %Fat_4C lab criterion method, while the %Fat_3C-IMAGE model in the current study was compared against a more robust %Fat_5C criterion model.

This study is not without its limitations. The primary being the lack of diversity of our sample. Future research should confirm these results in a larger, more diverse sample to ensure generalizability across varying races/ethnicities, body types, and ages. Another limitation of the current study is the inability to analyze the agreement between methods in males and females separately, due to a lack of statistical power. Additionally, comparisons across previous research should be interpreted with caution as the r values are not directly comparable. Despite this, it should be noted that the patterns in previous studies are similar to those reported in the current study. Lastly, it may be prudent of future research to compare the relative agreement between the %Fat_3C-IMAGE model and other proposed field-based %Fat3C models, using various techniques to estimate BV. Furthermore, it is encouraged that future research examine the ability of the %Fat_3C-IMAGE model to track changes in body composition over a period of time.

In conclusion, The results of the current study indicate that the novel 2D image analysis program may be used successfully in a %Fat_3C model with TBW, to accurately estimate relative adiposity. Furthermore, the %Fat_3C-IMAGE model performed slightly better in our sample young adults than the %Fat_DXA 3C model. As such, the %Fat_3C-IMAGE model tested in the current study provides a practical, non-invasive, and portable alternative to complex laboratory-based methods.