BACKGROUND Type 2 diabetes mellitus (T2DM) is a prevalent metabolic disorder often accompanied by sarcopenia, both of which pose significant public health challenges. Irisin, a myokine involved in energy metabolism, has been increasingly recognized for its role in the pathophysiology of T2DM with sarcopenia. OBJECTIVE This study aims to explore the relationship between Irisin levels and T2DM combined with sarcopenia, and to develop predictive models for early identification of high-risk individuals using machine learning techniques. METHODS A total of 200 subjects from Zhoupu Hospital Affiliated to Shanghai University of Medicine & Health Sciences were enrolled and divided into four groups: healthy controls, sarcopenia only, T2DM only, and T2DM with sarcopenia. Five machine learning algorithms－decision tree (C4.5), logistic regression, random forest, naive Bayes, and multilayer perceptron－were applied to evaluate the influence of Irisin levels on clinical outcomes. Model performance was assessed using accuracy, Kappa coefficient, mean absolute error (MAE), and relative absolute error (RAE). RESULTS Among the five models, the optimized C4.5 decision tree demonstrated superior performance (accuracy = 99.0%, Kappa = 0.9840, MAE = 0.0050, RAE = 0.0159). Sixteen valid knowledge rules were extracted, revealing that individuals with Irisin levels >94 ng/ml exhibited normal muscle strength and gait speed, while those CONCLUSIONS Irisin serves as a promising biomarker for distinguishing metabolic states and assessing the risk of T2DM with sarcopenia. The decision tree-based knowledge rules enable early identification of high-risk individuals and offer valuable insights for clinical decision-making. Further large-scale, multicenter validation is warranted to enhance generalizability.
Xu et al. (Wed,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: