Development and validation of an ultrasound-based predictive model for central lymph node metastasis in papillary...

Patients with peripherally calcified papillary thyroid carcinoma (PTC) represent a unique subtype with potentially distinct metastatic behavior. However, a specific tool for predicting central lymph node metastasis (CLNM) in this subgroup is lacking, creating a knowledge gap in personalized preoperative assessment. This study aimed to develop and validate a preoperative prediction model for this patient population, addressing the absence of a specific risk assessment tool.

We retrospectively enrolled 210 consecutive patients with pathology-confirmed, solitary PTC exhibiting peripheral calcification on preoperative ultrasound, who underwent initial surgery (lobectomy or total thyroidectomy) with central lymph node dissection (CLND) between January 2017 and July 2025. Patients with incomplete data, prior neck treatment, or distant metastases were excluded. The primary outcome, CLNM, was definitively diagnosed by postoperative pathology. The data were subsequently divided into training and validation datasets at a 7:3 ratio using 1,000 bootstrap resamples. In addition to a therapeutic or preventive CLND, each patient underwent an ultrasonography examination and either a thyroid lobectomy or a total thyroidectomy. The most significant risk factors were identified using the least absolute shrinkage and selection operator (LASSO) regression approach, and a Clinical-ultrasound (Clin-US) nomogram was created. The area under the receiver operating characteristic (ROC) curve was used to evaluate the model's performance. Accuracy and clinical utility were evaluated using calibration and decision curve analysis (DCA) curves.

The study cohort comprised 210 patients, with 131 (62.4%) females and 79 (37.6%) males. The median age was 39.0 years in the training set and 36.0 years in the validation set. Postoperative pathology confirmed CLNM in 120 cases (57%). Age, abutment-to-lesion perimeter ratio (A/P), tumor location, US-reported central lymph node (CLN) status, halo sign, extrusion beyond calcification, and type of peripheral calcification were identified as independent risk factors. The developed Clin-US model demonstrated high discriminative performance, yielding an area under the ROC curve (AUC) of 0.942 [95% confidence interval (CI): 0.904-0.980] in the training cohort and 0.870 (95% CI: 0.777-0.962) in the validation cohort. Corresponding sensitivities and specificities were 91.9% and 89.1% for the training set, and 73.5% and 84.6% for the validation set, respectively. Calibration curve indicated good agreement between predicted and observed probabilities, and DCA curve suggested clinical utility across a wide probability threshold range (0.12-0.89).

In this study, we developed and preliminarily validated a nomogram for predicting CLNM in patients with peripherally calcified PTC. The model showed promising performance in our cohort and may serve as a reference tool to aid in individualized preoperative decision-making for this specific subtype.