Feed

Cervical lymph node metastasis (CLNM) plays a crucial role in determining the surgical strategy for patients with papillary thyroid carcinoma (PTC). This study aimed to identify predictive factors for CLNM based on ultrasound features and gene mutation characteristics, and to develop a nomogram model for individualized risk prediction. A total of 171 patients with pathologically confirmed PTC who underwent surgery between January 2023 and October 2025 were retrospectively analyzed. Patients were randomly divided into a training set (n = 131) and a validation set (n = 40) at a ratio of 7:3. Clinical characteristics, ultrasound imaging features, and thyroid-related gene mutations were collected. Univariate and multivariate logistic regression analyses were performed to identify independent risk factors for CLNM, and a nomogram model was subsequently constructed. Model performance was evaluated using receiver operating characteristic curves, calibration analysis with the Hosmer-Lemeshow test, and decision curve analysis. Significant differences were observed between groups in terms of sex, tumor shape, boundary, calcification, and capsule contact (all P < .05). Multivariate logistic regression identified sex (odds ratio [OR] = 2.817), irregular shape (OR = 5.585), obscure boundary (OR = 2.074), calcification (OR = 3.515), and capsule contact (OR = 2.927) as independent predictors of CLNM in PTC (all P < .05). A nomogram was constructed based on these variables. The area under the curve was 0.800 (95% CI: 0.723-0.876) in the training set and 0.775 (95% CI: 0.628-0.922) in the validation set. The model demonstrated good calibration and clinical utility as indicated by the Hosmer-Lemeshow test (P > .05) and decision curve analysis. A nomogram integrating ultrasound features and clinical factors was developed to predict CLNM in patients with PTC, showing favorable predictive performance and potential clinical utility. Future prospective, multicenter studies with larger sample sizes and more comprehensive datasets are needed to further validate and optimize the model, thereby improving its generalizability and supporting individualized clinical decision-making.