Development of a Machine Learning-Based Predictive Model for Central Lymph Node Metastasis in Papillary Thyroid...

Area of Science:

• Oncology
• Medical Informatics
• Surgical Oncology

Background:

• Papillary thyroid microcarcinoma (PTMC) frequently exhibits central lymph node metastasis (CLNM).
• Prophylactic central lymph node dissection (CLND) is debated due to potential overtreatment.
• Accurate CLNM prediction is crucial for personalized PTMC management.

Purpose of the Study:

• To develop and validate a machine learning (ML) model for predicting CLNM in PTMC patients.
• To identify key clinical and ultrasound features associated with CLNM.
• To evaluate the performance of different ML algorithms for CLNM prediction.

Main Methods:

• A cohort of 228 PTMC patients was randomly divided into training and validation sets.
• Least Absolute Shrinkage and Selection Operator (LASSO) regression was used for feature selection.
• Eight ML models were developed and evaluated using cross-validation, AUC, calibration curves, and decision curves. SHAP values were used for interpretability.

Main Results:

• Key predictors of CLNM included age, gender, tumor diameter, T3, T4, TPOAb, and ultrasound findings.
• The Support Vector Machine (SVM) model showed the best performance with 0.783 accuracy and 0.805 specificity in the validation set.
• Age, gender, and tumor diameter were the most influential factors in the SVM model.

Conclusions:

• A machine learning framework, particularly the SVM model, effectively predicts CLNM in PTMC.
• This model offers a promising tool to guide decisions regarding prophylactic CLND.
• The findings may help minimize unnecessary surgeries and improve patient outcomes.