An Online Model for Central Lymph Node Metastases in Papillary Thyroid Carcinoma With BRAF V600E Mutation

Abstract

Background: To construct a predictive model to direct the dissection of the central lymph nodes in papillary thyroid cancer (PTC) with BRAF V600E mutation by identifying the risk variables for central lymph node metastases (CLNM). Methods: Data from 466 PTC patients with BRAF V600E mutations underwent thyroid surgery was collected and analyzed retrospectively. For these patients, we conducted univariate and multivariate logistic regression analysis to find risk variables for CLNM. To construct a nomogram, the independent predictors were chosen. The calibration, discrimination, and clinical utility of the predictive model were assessed by training and validation data. Results: CLNM was present in 323/466 PTC patients with BRAF V600E mutations. By using univariate and multivariate logistic regression, we discovered that gender, age, tumor size, multifocality, and pathological subtype were all independent predictors of CLNM in PTC patients with BRAF V600E mutations. A predictive nomogram was created by combining these variables. In both training and validation groups, the nomogram demonstrated great calibration capacities. The training and validation groups’ areas under the curve (AUC) were 0.772 (specificity 0.694, sensitivity 0.728, 95% CI: 0.7195-0.8247) and 0.731 (specificity 0.778, sensitivity 0.653, 95% CI: 0.6386-0.8232) respectively. According to the nomogram’s decision curve analysis (DCA), the nomogram might be beneficial. As well, an online dynamic calculator was developed to make the application of this nomogram easier in the clinic. Conclusion: An online nomogram model based on the 5 predictors included gender, age, pathological subtype, multifocality, and tumor size was confirmed to predict CLNM and guide the central lymph nodes dissection in PTC patients with BRAF V600E mutations.