Related Experiment Video
Updated: May 21, 2026
07:13
Comparison of Predictive Performance of Three Lymph Node Staging Systems in Colorectal Signet Ring Cell Carcinoma Based on Machine Learning Model
Published on: April 18, 2025
Radiomics-based prediction model for lymph node metastasis in thyroid microcarcinoma.
Yifeng Yang1, Anlong Yuan2, Cuicui Huang3
1Department of Thyroid and breast Surgery, Qingdao Chengyang District People's Hospital, Qingdao, 266000, China.
BMC Surgery
|May 20, 2026
Summary
This study developed a predictive model for lymph node metastasis in thyroid microcarcinoma using clinical, ultrasound, and radiomic data. The Random Forest model accurately identifies key predictors like TSH level, tumor size, and sphericity for risk assessment.
Area of Science:
- Oncology
- Radiology
- Medical Informatics
Background:
- Thyroid microcarcinoma (TMC) poses challenges in predicting lymph node metastasis (LNM).
- Accurate preoperative risk stratification is crucial for individualized surgical planning in TMC patients.
Purpose of the Study:
- To develop and validate a predictive model for LNM in TMC.
- Integrate clinical, ultrasonographic, and radiomic features for enhanced prediction.
- Provide a tool for preoperative risk assessment and surgical planning.
Main Methods:
- Retrospective analysis of 426 TMC patients, divided into training (n=300) and validation (n=126) sets.
- Collected demographic, clinical, ultrasonographic, laboratory (TSH), and radiomic data.
- Employed univariate analysis, LASSO regression, and multivariate logistic regression to identify predictors. Constructed Random Forest (RF), K-nearest neighbors (KNN), and gradient boosting (GB) models.
Main Results:
- Multivariate analysis identified tumor size, lymph node size, TSH level, central lymph node metastasis, and 3D tumor volume as independent risk factors for LNM.
- Sphericity emerged as an independent protective factor.
- The RF model demonstrated superior predictive performance (training AUC: 0.838, validation AUC: 0.815) compared to KNN and GB models.
Conclusions:
- The developed Random Forest model effectively predicts lymph node metastasis in thyroid microcarcinoma.
- TSH level, tumor size, and sphericity are identified as key predictors.
- The model shows high clinical utility for preoperative risk assessment and surgical planning in TMC.
