A Nomogram Model Based on Radiomic and Clinical Features of Erector Spinae Muscle CT Images for Diagnosing Uremic Sarcopenia

Objective To develop a diagnostic model for uremic sarcopenia by combining radiomic features of the erector spinae muscle from chest CT images with clinical data. Methods A total of 195 patients with uremia were collected, including 84 diagnosed with uremic sarcopenia based on the gold standard, and 111 without sarcopenia. A nomogram model was developed by integrating radiomic and clinical features. The diagnostic performance of the nomogram model, radiomics signature, and clinical signature was individually compared to validate their effectiveness. Receiver operating characteristic (ROC) curves were generated to assess diagnostic accuracy. Results A nomogram combining radiomic and clinical features was developed using a logistic regression machine learning algorithm. The predictive performance of the nomogram was evaluated using the area under the ROC curve (AUC) and calibration curves. In the test set, the nomogram demonstrated superior predictive performance, with an AUC of 0.951 (95% CI: 0.886–1.000). Decision curve analysis (DCA) was used to assess the clinical benefits of the model, and the nomogram showed a higher clinical benefit in diagnosing uremic sarcopenia. Conclusion The nomogram developed by combining CT radiomic features of the erector spinae muscle with clinical features demonstrates excellent diagnostic performance for uremic sarcopenia, achieving perfect specificity and eliminating false positives compared to individual models.