3D body shape analysis for predicting sarcopenia and obesity in older adults

Project Summary If successful, the proposed study will provide accessible and scalable methods with measurable and validated accuracy to assess key indicators of an older adult’s health related to body composition phenotypes and associated physical function. The approach uses inexpensive and widely available commodity optical scanners

as a novel imaging modality with the potential to practically reduce healthcare costs and disparities. The study meets NIA strategic priorities by developing improved approaches for the early detection and diagnosis of disabling illnesses and age-related debilitating conditions (C1) and in providing a foundation for developing

interventions for treating, preventing, or mitigating the impact of age-related conditions (C3). Early screening and diagnosis of body composition and physical function combined with timely, dietary, exercise, and/or pharmacological interventions can mitigate the risk of functional decline and negative health outcomes in

individuals with sarcopenic obesity. Simple anthropometric measures are easy to perform but have poor diagnostic accuracy and are inconsistently associated with morbidity and long-term physical function. While other modalities (e.g., dual-energy X-ray absorptiometry (DEXA), CT, MRI) may have higher diagnostic accuracy, they

are impractical for widescale integration into clinical practice. There are relatively inexpensive systems and mobile apps that use 3D body shape from optical scanners for predicting body composition. While their validity is sufficient for consumer-oriented applications, these prediction algorithms may not be applicable for clinical use

on older adults ─ they were not trained on data from this population. These systems also do not predict both muscle mass and physical function which are critical in the diagnosis of sarcopenia and obesity. To address these limitations, our team has previously developed highly accurate prediction algorithms using optical body

scanning technology (R21HL124443, R01DK129809). These promising results merit us to further test and validate our system to translate such technologies into routine clinical and home-based care. We propose collecting data in an observational cross-sectional study of participants recruited from community-

dwelling older adults. Participants will undergo: (i) 3D optical body scans to determine body shape; (ii) DEXA to assess body composition; (iii) D3-creatine dilution tests to determine total muscle mass; and (iv) validated physical function assessments. This data will be used to train artificial intelligence algorithms to predict body

composition and physical function. We will investigate the usability of the approach for clinicians and for older adults using mobile platforms. We anticipate that the next step in this line of research is to conduct a cohort study that demonstrates the predictive nature on adverse outcomes in participants with sarcopenic obesity, or in using this system as part of

a clinical trial. Additionally, we envision transitioning this technology to real-world settings by developing a prototype system as part of a Small Business Technology Transfer program.