MR Imaging Body composition as a non-invasive tool to detect endogenous hypercortisolism: an exploratory study

Background Cushing's syndrome (CS) is a complex endocrine disorder resulting from prolonged exposure to endogenous hypercortisolism. Up to 73% of CS patients also meet the criteria for metabolic syndrome, further blurring the clinical picture and delaying diagnosis. We aim to compare the body composition of two groups—healthy volunteers and patients with endogenous hypercortisolism—using magnetic resonance imaging (MRI). Our objective was to identify a specific anthropometric phenotypic signature in patients with endogenous hypercortisolism that could facilitate earlier clinical suspicion. Methods 30 patients with a recent diagnosis of endogenous hypercortisolism (77% adrenal origin, 10% pituitary origin, 13% other origin), were matched with 15 healthy controls using a propensity score analysis, considering age, sex, and body mass index. We acquired the abdominal and thigh images using a 3T Philips Ingenia MRI scanner and a quantitative Dixon sequence. Results Robust Principal Component Analysis (RPCA) and cluster analysis showed a sensitivity of 93% and specificity of 73% for classifying control and endogenous hypercortisolism patients using only body composition information. Logistic regression analysis showed that belonging to the cluster found by RPCA has an OR of 38.5, p < 0.001, for presenting endogenous hypercortisolism. Conclusions MRI-based body composition may offer a novel approach for the early suspicion of endogenous hypercortisolism in overweight and obese patients, and the muscle-to-liver fat fraction ratio is a specific anthropometric phenotypic signature.