Ultrasound-based nomogram for predicting lateral cervical lymph node metastasis in thyroid papillary carcinoma

Objective: To explore the risk factors of lateral cervical lymph node metastasis(LLNM) in thyroid papillary carcinoma (PTC), establish a nomograph prediction model to predict the probability of LLNM before operation, reduce the risk of recurrence, avoid secondary surgery, and achieve the goal of accurate treatment. Methods: In this retrospective study, 242 patients were enrolled and randomly divided into two groups according to the ratio of 7:3. The preoperative clinical data of patients (including ultrasonic characteristics and selected clinical characteristics) were collected, and least absolute shrinkage and selection operator (Lasso) in the training cohort was used to screen the influencing factors. Based on the results of multivariable logistic regression, the nomograph model is established, and its performance is evaluated by its identification, calibration and clinical usefulness in training and validation cohorts. Results: This study showed that age>45 years (OR 0.340; 95%CI 0.129-0.856), abnormal serum TG (OR 3.626; 95%CI 1.397-10.187), thyroid tumer location1 (OR 0.250; 95%CI 0.083-0.702), US-reported positive central cervical lymph node (CLN) (OR 3.116; 95%CI 1.152-8.601), and US-reported positive LLN (OR 8.426; 95%CI 3.192-23.779) were independent risk factors for LLNM of PTC. The nomogram constructed based on the above factors has good resolution and calibration in predicting the LLNM of PTC. The Concordance index (C-index) of the experimental group and the validation group are 0.880 and 0.937, respectively. Conclusion: The nomogram based on ultrasound and clinical features constructed in this study has good performance in non-invasive predicting LLNM of PTC and is expected to promote clinical decision-making.