Answers are grounded in this paper and cite their sources.
Objective To develop a CT-based differential elasticity mapping-derived habitat radiomics model for preoperative prediction of right recurrent laryngeal nerve lymph node (RRLNLN) metastasis in esophageal squamous cell carcinoma (ESCC). Methods In total, 415 ESCC patients from two hospitals were retrospectively enrolled. Differential elasticity maps were generated from CT images, and the local Moran’s index was used to partition each lymph node into five habitat subregions. Radiomic features were extracted and selected via t-test, Pearson correlation, and least absolute shrinkage and selection operator (LASSO) regression. Multiple machine learning classifiers were constructed and evaluated. Results Thirteen key radiomic features were retained. Ensemble learning methods outperformed conventional models, with light gradient boosting machine (LightGBM) achieving the best performance. In the independent test set, LightGBM yielded an area under the curve (AUC) of 0.865, sensitivity of 0.677, specificity of 0.915, and F1 score of 0.724. The model showed good calibration (Hosmer–Lemeshow p = 0.830) and provided significant clinical net benefit across a wide threshold range. The DeLong test confirmed superiority over logistic regression and support vector machine (SVM) (p < 0.05). Conclusion This first integration of CT elastography with local Moran’s index-based habitat analysis enables accurate preoperative prediction of RRLNLN metastasis in ESCC, offering a promising tool to guide individualized surgical planning.
No takes yet. Share an insight, caveat, or question.
Huang et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: