Abstract

Human body composition measurements such as fat and fat-free mass play a critical role in qualifying health and nutritional status, the impact of disease, and change due to nutritional, therapeutic, or behavioral intervention. However, these measures are difficult to obtain, and require expensive equipment as well as specialized facilities and staff. Given the rise of low-cost mobile ultrasound systems, and ultrasound’s proven ability to assess measures of adipose tissue and skeletal muscle, our team will develop a novel ultrasound tool that can guide a user to collect high quality data and automatically determine body composition measures for nutritional evaluation. Through support from the Schiller Institute, we are launching a pilot study with clinical collaborators at the Brigham and Women’s Hospital (Boston, MA) and Jimma University (Jimma, Ethiopia). The pilot study will consist of: (1) clinical data collection, (2) developing machine learning models for image analysis and prediction, and (3) human-centered design for frontline healthcare workers. This project involves research collaboration across engineering, nursing, and computer science.