CT-derived sarcopenia and myosteatosis predict treatment escalation in hospitalized patients with inflammatory Bowel disease.

Authors

Abstract

Sarcopenia and myosteatosis reflect muscle quantity and quality and are linked to adverse outcomes in chronic diseases. Their role in predicting treatment escalation in inflammatory bowel disease (IBD) remains unclear. We retrospectively analyzed 308 IBD patients (251 ulcerative colitis, 57 Crohn's disease) who underwent abdominal CT scans at the level of the third lumbar vertebra. Patients were randomly assigned to a training set (n = 217) and a validation set (n = 91). Sarcopenia and myosteatosis were quantified using skeletal muscle index (SMI) and skeletal muscle density (SMD). Treatment escalation was defined as initiation of biologics, cyclosporine, or surgery following relapse. Independent predictors were identified via multivariate logistic regression. Five machine learning models-logistic regression, random forests, extreme gradient boosting, support vector machine, and light gradient boosting machine (LightGBM)-were constructed and evaluated using receiver operating characteristic, calibration, and decision curve analysis. Age, sarcopenia, and myosteatosis were independent risk factors for treatment escalation. The LightGBM model achieved the highest predictive performance (The area under the curve: 0.839 training set, 0.763 validation set), demonstrated good calibration, and provided superior clinical net benefit. The corresponding Nomogram allowed intuitive individualized risk assessment. CT-derived sarcopenia and myosteatosis independently predict treatment escalation in IBD. Machine learning models integrating these parameters with clinical features can effectively identify high-risk patients, supporting early intervention and personalized therapy. Incorporating additional imaging markers, biomarkers, and functional assessments may further refine predictive accuracy and guide strategies to improve muscle health and clinical outcomes.

Tags

Topics