A multivariable model of ultrasound and biochemical parameters for predicting high-volume lymph node metastases of papillary thyroid carcinoma with Hashimoto's thyroiditis.

Objectives: This study aims to develop a nomogram to predict high-volume (> 5) lymph node metastases (HVLNM) in papillary thyroid carcinoma concomitant with Hashimoto's thyroiditis by combining ultrasound with clinicopathologic data. Materials and methods: The study reviewed 187 patients diagnosed with papillary thyroid cancer (PTC) concomitant with Hashimoto's thyroiditis from the First People's Hospital of Kunshan between March 2018 and December 2022. These patients underwent preoperative ultrasound and postoperative examinations. They were divided into two groups based on the size of their lymph nodes (LNs). A predictive model was developed using LASSO regression and multifactor logistic regression analysis. The receiver operating characteristic (ROC) curve was used to validate the predictive model. Results: A total of 187 patients were randomized into 132 participants for training and 55 participants for external validation. Four predictors including tumor size, extrathyroidal extension, histological grade and vascularity, were selected from 13 variables based on LASSO regression analysis. In the training set, the model built from the above four predictor has a satisfactory predictive power, with an area under the ROC curve of 0.914, and validation set with the ROC curve of 0.889, which indicated that the nomogram can be used effectively in clinical settings. Conclusion: In summary, the nomogram constructed by tumor size, extrathyroidal extension, histological grade and vascularity, is useful for predicting the risk of HVLNMs in patients with papillary thyroid carcinoma associated with Hashimoto's thyroiditis, which is expected to provide the basis for adequate and accurate management before the primary surgery.