The procedures for body size, shape, and composition assessment through commercially available solutions for three-dimensional optical imaging enable the rapid collection of accurate and reproducible data. Clinicians could implement the acquisition of innovative and useful biomarkers ("e-tape" measurements) in routine evaluations of patients to assist in characterizing their health status.
The body size and composition assessment is commonly included in the routine management of healthy athletes as well as of different types of patients to personalize the training or rehabilitation strategy. The digital anthropometric analyses described in the following protocol can be performed with recently introduced systems. These new tools and approaches have the potential to be widely used in clinical settings because they are very simple to operate and enable the rapid collection of accurate and reproducible data. One system consists of a rotating platform with a weight measurement plate, three infrared cameras, and a tablet built into a tower, while the other system consists of a tablet mounted on a holder. After image capture, the software of both systems generates a de-identified three-dimensional humanoid avatar with associated anthropometric and body composition variables. The measurement procedures are simple: a subject can be tested in a few minutes and a comprehensive report (including the three-dimensional scan and body size, shape, and composition measurements) is automatically generated.
Anthropometry is the study of the physical measures of the human body. Height, weight, lengths, skinfold thicknesses, and circumferences are commonly used anthropometric measures that proved to be useful for investigating patients with endocrine and metabolic disorders and for monitoring growth, aging, and body size and composition adaptations elicited by diet and training in athletes1,2. For example, the assessment of waist and hip circumferences proved to be useful for the management of persons with obesity: both circumferences assess the distribution of adiposity that can be considered a predictor of all-cause mortality3.
Limb circumferences are frequently assessed in rehabilitative and sports medicine because of their usefulness for detecting and/or monitoring the decrease in appendicular lean mass (e.g., calf circumference is used as a simple and practical skeletal muscle marker for diagnosing low skeletal muscle and sarcopenia)1,2 and the inter-limb asymmetry that impacts both physical performance and risk of injuries in athletes and quality of life in patients (e.g., cancer patients with unilateral extremity swelling)1,2. Further, a large number of anthropometrics-based body composition prediction models have been proposed over the last several decades to estimate the amount of fat mass or fat-free mass from a combination of different anthropometric measures such as body circumferences or skinfold thicknesses1,2,4,5,6,7.
Because conventional anthropometric (i.e., tape-based and caliper-based) measurements may not be culturally or socially acceptable and also exhibit poor reliability8, there was the need for the development and validation of non-invasive, reproducible, and valid approaches. Recently developed three-dimensional (3D) optical imaging systems enabled to provide non-invasive, precise, and accurate measurements8,9,10,11, as well as digital consumer cameras and smartphones offer easy-to-use and widely available tools suitable to be used in clinical and non-clinical settings to evaluate both patients and healthy subjects8,9,10,11,12,13,14,15,16,17,18,19,20. The aim of the protocol reported in the following section is to describe the procedures for evaluating body size, shape, and composition through two commercially available solutions for 3D optical imaging that became pervasive during the last years both in the healthcare setting (to evaluate patients) and in non-clinical settings (to evaluate athletes).
The protocol follows the research integrity guidelines of the Polytechnic of Turin21. The acquisition of optical images was performed within research studies approved by the local ethics committees (data are reported in the legends of Figure 1 and Figure 2) and the investigated subjects gave their written consent (for study participation and publication of anonymized images).
1. Subject preparation
NOTE: All prescanning preparations described in this section are similar between testing procedures #1 and #2.
2. Subject registration for the testing procedure #1
3. Testing procedure #1
4. Subject registration for the testing procedure #2
5. Testing procedure #2
After the image capture, the software of system #1 generates a de-identified 3D humanoid avatar (Figure 1C: point clouds are converted to a mesh connected by triangles with approximately 25,000 vertices and 50,000 faces) and automated anthropometry, which includes lengths, circumferences, volumes, surface areas, and body composition estimates. The dashboard of system #1 enables for each subject to visualize (and download a report including) the 3D scan (Figure 1C), measurements of body weight, size, and shape (i.e., body shape rating, waist circumference, waist-to-hip ratio, trunk-to-leg volume ratio), estimates of basal metabolic rate and body composition (i.e., BF%, fat mass, lean mass), and standard circumference measurements (neck, bust, waist, hips, left and right biceps, left and right forearm, left and right thigh, left and right calf).
Moreover, the results of the posture and balance assessments can also be visualized and are included in the report. The posture assessment results include the front, side, and back views of the 3D scan with associated shift (defined as a part of the body that moves into a sloping position and the rate at which it slopes in any one direction) and tilt (defined as the "sliding" movement forward, backward, left or right, a slight change or variation in position from the center point) measurements: i) front and back views: shift measurements toward right or left with respect to the sagittal plane (represented as a vertical line between right and left hemisomes) and tilt percentages with respect to the transverse (horizontal plane) for head, shoulder, underbust, hip, knee; ii) side view: shift measurements forward or backward with respect to the frontal (coronal) plane (represented as a vertical line up from the ankle joint) for head, shoulder, hip, knee.
The balance assessment result includes the weight distribution during standing posture for the anterior and posterior regions of the right and left feet. The dashboard of system #1 also enables for each subject to download .OBJ and .GIF image files and a .CSV file with the anthropometric measurements and body composition estimates listed in Table 1.
All body composition estimates are obtained by using proprietary algorithms, with the exceptions of the basal metabolic rate estimation and of the body shape index calculation that are obtained, respectively, according to the Mifflin-St. Jeor equation23 and the Krakauer equation24 reported in Table 2.
After the image capture, the software of system #2 generates a de-identified 3D humanoid avatar (Figure 2D: point clouds are converted to a mesh connected by triangles with approximately 50,000 vertices and 100,000 faces) and automated anthropometry, which includes lengths, circumferences, volumes, surface areas, and body composition estimates. The dashboard of system #2 (data download dashboard in the Table of Materials) enables for each subject to download .OBJ and .PNG image files and the following three .CSV files:
The "App Measures.csv" file reports the following anthropometric and body composition measurements: weight, body surface area, BF%, visceral adipose tissue, fitness index, arms lean mass, legs lean mass, lean body mass, total bone mineral content, shoulder width, back shoulder width (through back neck), circumferences of neck, overarm, biceps (right/left), forearm (right/left), wrist (right/left), chest, underbust, bust (with drop), stomach, waist, paint waist, hips (taken 8 inches down from small of back), seat, thigh (right/left), calf (right/left), back-neck-to-waist length, sleeve length (right/left), crotch length, inseam, outseam (right/left).
The "Body Composition.csv" file reports the following anthropometric and body composition measurements: body fat, body mass index, body surface area, bone mineral content, fat mass index, fitness index, height, lean body index, lean body mass, arms lean mass, legs lean mass, resting metabolic rate, stomach circumference, visceral adipose tissue, waist-to-height ratio, waist-to-hip ratio, weight.
The "Core Measures.csv" file reports the anthropometric measurements listed in Table 1.
All body composition estimates are obtained by using proprietary algorithms, with the exceptions of the basal metabolic rate estimation and the BF% estimation that are obtained, respectively, according to the Katch-McArdle equation25 and to the two equations previously developed and validated by Harty et al.26. These two equations (BF% Equation 1 and Equation 2 de Table 2) are adopted, respectively, for individuals with lower abdominal circumference <103.5 cm (<40.75 inches) and ≥103.5 cm (≥40.75 inches). Although previous studies demonstrated the accuracy of Equation 1 for BF% estimation in healthy adults15,26, we recently found that it overestimated (with respect to dual-energy X-ray absorptiometry) the BF% in young athletes18. Therefore, we proposed the reparameterization (Equation 3) reported in Table 2 to provide an accurate estimation of BF% in young soccer players of both sexes18.
In addition to the above-listed body composition variables, the appendicular lean mass (ALM) can also be estimated for the scans performed with system #2 through the device-specific equation recently proposed by McCarthy et al.27 for sedentary subjects (ALM Equation 1 for males and ALM Equation 2 for females in Table 2) that we adapted for ALM estimation in young athletes (ALM Equation 3 for males and ALM Equation 4 for females in Table 2)18.
Figure 3 shows representative avatars obtained in a male athlete (body mass index: 26.0 kg/m2: Figure 3A,B) and a person with obesity (body mass index: 44.0 kg/m2: Figure 3C,D) with the system #1 (Figure 3A,C) and with the system #2 (Figure 3B,D). The anthropometric measurements and body composition estimates obtained by system #1 and system #2 for two scans of both subjects are reported in Table 3.
Circumference measurements obtained in the athlete differed between the scans obtained with the two systems (especially for the neck, arms, hips, and legs: data are reported in Table 3). BF% values (averages of the two scans) were 19.1% and 16.1% for the scans performed with systems #1 and #2, respectively, and lean mass values (averages of the two scans) were 69.7 kg and 72.2 kg. The ALM average value for the scans performed with system #2 was 38.4 kg.
Circumference measurements obtained in the person with obesity differed between the scans obtained with the two systems (especially for waist and legs: data are reported in Table 3). BF% values (averages of the two scans) were 44.1% and 46.3% for the scans performed with systems #1 and #2, respectively, and lean mass values (averages of the two scans) were 67.9 kg and 64.9 kg. The ALM average value for the scans performed with system #2 was 25.1 kg.
Figure 1: Images taken with system #1. (A,B) Standardized A-pose assumed and maintained by a representative male subject during the rotation of the platform and (C) the relative 3D avatar. Acquisition of optical images was performed within a research study approved by the ethics committee of the University of Turin (protocol n. 0115311). Please click here to view a larger version of this figure.
Figure 2: Images taken with system #2. (A) Image of a representative male subject standing over a carpet with the feet guide symbol (black oval symbol over the carpet) aligned with the green feet marker shown on the screen of the tablet of system #2. Acquisition of the (B) frontal and (C) lateral images in the representative subject and (D) the relative 3D avatar. Acquisition of optical images was performed within a research study approved by the ethics committee of the University of Turin (protocol n. 0115311). Please click here to view a larger version of this figure.
Figure 3: Representative avatars obtained with systems #1 and #2. (A,B) Male athlete and (C,D) a person with obesity investigated with (A,C) system #1 and (B,D) system #2. Each subject underwent two scans, with repositioning: the avatar obtained from the first scan was shown for both subjects, while the anthropometric and body composition estimates obtained by system #1 and system #2 for two scans of both subjects are reported in Table 3. Acquisition of optical images was performed within research studies approved by the ethics committee of the University of Turin (protocol n. 0115311) and by the Territorial Ethics Committee (CET – protocol n. 0065654). Please click here to view a larger version of this figure.
Table 1: Extended measurement sets downloadable from the dashboard of systems #1 and #2. Please click here to download this Table.
Table 2: Equations for estimation of basal metabolic rate, body fat percentage, and appendicular lean mass. Basal metabolic rate estimations: the units of measurement are kg for weight and lean mass, cm for height, and years for age. A body shape index estimation: the units of measurement are m for waist circumference and height are measured, kg/m2 for body mass index. Body fat percentage estimation: sex is coded as male = 1 and female = 0, body surface area is measured in cm2, and all remaining variables indicated as circ. are measured in cm. Muscle to stomach index is obtained as (right biceps circ. + left biceps circ. + right thigh circ. + left thigh circ. + right calf circ. + left calf circ.) / maximum stomach circ. Appendicular lean mass estimation: the units of measurement are cm for all circumferences and lengths; cm2 for surface areas; cm3 for volumes; kg for weight; years for age. Three equations of this table are from Minetto et al.18. Abbreviations: BMR = basal metabolic rate; ABSI = a body shape index; BF% = body fat percentage; circ. = circumferences; ALM = appendicular lean mass; NHOPI = Native Hawaiian and other Pacific Islander. Please click here to download this Table.
Table 3. Circumference measurements and body composition estimates obtained by the two systems in each of the two representative subjects (one athlete and one person with obesity). Please click here to download this Table.
Supplemental Figure S1: Subject registration form for testing procedure #1. Please click here to download this File.
Supplemental Figure S2: Subject registration form for testing procedure #2. Please click here to download this File.
The procedures presented in this article can be used to evaluate body size, shape, and composition through two commercially available solutions for 3D optical imaging that have been previously developed and validated9,10,11,12,13,14,15,16,17,18,19,20. These solutions are simple to operate, and valid data can be quickly collected and automatically organized into a report. Moreover, the presented systems enable the collection of reproducible data (as suggested by the comparison of the results from the two scans performed with both systems in our two representative cases and documented by previous studies)9,10,11,12,13,14,15,16,17,18,19,20 and can therefore be used to monitor the training- or diet-induced changes.
As system #2 has a limited weight (~4 kg in total for tablet and holder), it is easily portable. However, a limitation of system #2 is that the generation of a 3D avatar from 2D images can produce 3D reconstructions that are less accurate than those obtained with system #1, especially in persons with obesity (as shown in the representative example of Figure 3 C,D) or in patients presenting localized abnormalities of the body shape (e.g., patients after bariatric surgery presenting troublesome skin excess or cancer patients with unilateral upper or lower limb lymphedema).
The availability of adequate space is critical for the scan acquisition with both systems: a clear area of 157 x 198 cm for system #1 and of 86 x 166 cm for system #2 is required. Moreover, system #2 requires the subject to be placed close to a blank wall without mirrors, glossy posters, or windows. Both systems require that no natural sunlight and no reflective surfaces should be in view of the cameras. Both systems also require a constant and consistent wi-fi internet connection to process scans effectively.
The main limitation of the above-described procedures is that they require the investigated subject to be able to assume the standing position. Therefore, these approaches cannot be used in severely ill patients (such as seriously impaired neurological patients or critically ill patients) who are unable to get out of bed. Moreover, the investigated subjects must be able to maintain the standing position (i.e., A-pose and side pose) without movements that can change the shape of the avatar10,22 and bias the estimation of body circumferences.
A limitation of the above-described parameters is that they are obtained using proprietary device-specific algorithms: this implies that the body size, shape, and composition measurements are unique to the particular scanning system. Therefore, comparing or pooling data acquired with different systems is precluded by analytical (i.e., between scanners) variability. Consistently, circumference measurements obtained in our representative two subjects shown in Figure 3 differed between the two systems. However, device-agnostic solutions have already been developed to overcome this limitation: these solutions reformat and edit the 3D mesh, then automatically detect different landmarks (such as armpits, crotch, and feet) and then calculate body size measurements28,29,30,31,32,33,34,35. Another limitation of the above-described body composition parameters is that they are obtained through conventional anthropometrics-based prediction models. However, recent studies showed that body shape-based models could be required to capture information about body composition beyond conventional anthropometric measurements36,37.
Despite some limitations, the digital anthropometric approach must be considered ready to be used in the clinical setting. 3D imaging systems provide non-invasive measurements that can be more acceptable compared to manual (tape-based and/or caliper-based) measurements that are based on the identification of anatomical landmarks through observation and palpation. Moreover, 3D optical scanning is also faster compared to other investigations (e.g., magnetic resonance imaging and dual-energy X-ray absorptiometry) commonly adopted in research and clinical settings for body size and composition assessment. In addition, as it is relatively inexpensive and radiation-free, it is safe to be used for subsequent scans (e.g., the image acquisition can easily and quickly be repeated if the experimenter notices body movements or an improper limb placement that can produce changes in the shape of the avatar) and for repeated investigations38 as well as safe to be used in special populations (such as children, adolescents, and pregnant women)35,39.
Clinicians could therefore implement the acquisition of innovative and useful biomarkers ("e-tape" measurements and derived body composition estimations) in routine evaluations of healthy subjects (e.g., athletes) to assist in predicting and characterizing their physical performance and injury risk40,41,42,43 as well as to monitor injury recovery. For example, leg strength and lean mass symmetry influence physical performance and (re-)injury risk44. Therefore, the recovery of a normal symmetry of the thigh/calf circumferences can be included among the general goals to consider for returning to play45. The routine evaluation of patients also could be improved by the integration of digital anthropometry into healthcare. The assessment of body circumferences and shape (that is driven by the internal distribution of soft and fat tissues) can be useful to detect the low mass muscle (e.g., in patients suspected to be sarcopenic), to predict the metabolic disease risk46, to assess the outcome of a surgical procedure, as well as to monitor the patient progress following an intervention38. Patients with diseases that have nutritional components as key contributors to their pathophysiology can specifically benefit from longitudinal monitoring of body size and composition to reduce symptoms and co-existing conditions47. For example, in the case of diet- and/or drug-based management of obesity, it may not be appropriate to only monitor weight because the well-known "25/75 rule of thumb" (i.e., the general assumption that weight loss is typically 25% fat-free mass loss and 75% fat loss) may not accurately describe intervention efficacy38 that could be unraveled by anthropometry-based assessment of the relative amount of muscle and fat loss. Furthermore, digital anthropometry, integrated into healthcare, has the potential to expand healthcare services to remote locations, thereby improving patient assistance and adherence and reducing healthcare costs.
The authors have nothing to disclose.
The Authors are grateful to Dr. Federico Della Vecchia and Dr. Alessandro Cairo (University of Turin) for their valuable support with the manuscript preparation. This work was supported by grants from Fondazione CRT (Turin, Italy), the University of Turin (Fondo per la Ricerca Locale – ex-60%), and the National Institutes of Health (grant R01DK109008, Shape UP! Adults).
System #1 | |||
Proscanner | Fit3D Inc., San Mateo, CA, USA | Version 5 | "System #1" in the manuscript |
Fit3D Proscanner app | Fit3D Inc., San Mateo, CA, USA | Version 5 | "App #1" in the manuscript |
CHUWI tablet PC | Chuwi Technology Co., Ltd., Shenzhen, CHINA | Hi10X | "Tablet #1" in the manuscript |
Fit3D dashboard | Fit3D Inc., San Mateo, CA, USA | https://dashboard.fit3d.com | |
System #2 | |||
Mobile Scanner 1 (MS-1) app | Size Stream LLC, Cary, NC, USA | Version 2 | "System #2" in the manuscript |
iPad | Apple Inc., Cupertino, CA, USA | 9th generation | "Tablet #2" in the manuscript |
iPad Floor Stand | Displays2go LLC, Fall River, MA, USA | SKU: TABFLATBBK | www.displays2go.com/P-29987/Universal-Tablet-Floor-Stand-Anti-Theft-Locking-Kit |
Size Stream registration dashboard | Size Stream LLC, Cary, NC, USA | https://measure.mobilefit.sizestream.com | |
Size Stream data download dashboard | Size Stream LLC, Cary, NC, USA | https://data.mobilefit.sizestream.com |
.