Abstract
OBJECTIVE: To develop a nomogram based on ultrasound features and preoperative serum thyroid function of patients with primary papillary thyroid carcinoma (PTC) to quantify the probability of atypical metastatic cervical lymph nodes. METHODS: A retrospective study involving 316 patients diagnosed with PTC at Chengdu Integrated TCM & Western Medicine Hospital from January 2023 to December 2024 was conducted. Patients with typical ultrasound features of metastatic cervical lymph nodes or incomplete data were excluded, and 158 PTC patients with atypical ultrasound features were included in the study. The patients were divided into two groups based on the presence of cervical lymph node metastasis in the postoperative pathologic findings. The thyroid function and ultrasound data of the two groups were analyzed to identify independent risk factors for metastatic cervical lymph nodes with atypical ultrasound features. A nomogram prediction model was established and evaluated for discrimination and calibration via receiver operating characteristic (ROC) curves, calibration curves, and 5-fold cross-validation. RESULTS: Of the 158 patients, 59 were assigned to the metastatic group, and 99 were assigned to the nonmetastatic group. Multivariate analysis revealed the following independent risk factors for metastatic cervical lymph nodes with atypical ultrasound features: age ≤ 45 years (OR=2.898, 95% CI=1.294-6.810), male sex (OR=3.224, 95% CI=1.468-7.333), contact with capsule (OR=7.346, 95% CI=2.448-27.049), internal blood flow (grade II-III, OR=4.915, 95% CI=1.626-15.882), and TGAb positivity (OR=5.173, 95% CI=2.026-14.355). Based on these factors, a nomogram model was developed, which demonstrated an AUC of 0.805, a sensitivity of 72.88%, a specificity of 76.77%, and an accuracy of 75.32%. CONCLUSION: The nomogram, which is based on age, sex, the distance between the nodule and the adjacent capsule, internal blood flow, and TGAb levels, has a strong ability to predict cervical lymph node metastasis in PTC patients with atypical ultrasound features. This model may assist in reducing the incidence of misdiagnoses of metastatic lymph nodes by providing imaging and laboratory data to facilitate clinical decision-making.