Lack of a balanced diet can have a significant impact on most organs of the body. Traditionally, evaluation of these conditions relied heavily upon body mass index "BMI" measurements, which are limited and open to inaccurate interpretation or omission of critical data. Advances in imaging allow better recognition of these conditions using accurate qualitative and quantitative data and correlation with any morphological changes in organs. Body composition evaluations include the assessment of the bone mineral density (BMD), visceral fat, subcutaneous fat, liver fat and iron overload and muscle fat (including the lean muscle ratio), with differential evaluation of specific muscle groups when required. Such measurements are important as a baseline and for monitoring the effect of therapies and various interventions. In addition, they may predict and help alleviate any potential complications, allowing counselling of patients in a relatable manner. This positively influences patient compliance and outcomes during early counselling, monitoring and modulation of therapy. This encourages patients suffering from obesity and eating disorders to better understand their often chronic but reversible condition. We present a review of current literature with reflection on our own practices. We discuss the importance of monitoring the reversibility of certain parameters in specific cohorts of patients. We consider the role of artificial intelligence and deep learning in developing software algorithms that can help the reading radiologist evaluate large volumes of data and present the results in a format that is easier to interpret, thereby reducing interobserver and intraobserver variabilities.
Keywords: Artificial intelligence; Body composition; Diet; Dual energy X-ray absorptiometry; Eating disorders; Magnetic resonance imaging; Obesity.
Copyright © 2020 Elsevier B.V. All rights reserved.