Nomogram model for predicting papillary thyroid carcinoma based on clinical-ultrasound characteristics and inflammatory biomarkers: a multicenter study

Objectives: The predictive significance of lymphocyte-related inflammatory biomarkers for papillary thyroid carcinoma (PTC) remains unclear. This study aims to provide a new tool for differentiating PTC from benign thyroid nodule (BTN) by constructing a nomogram model. Methods: Institution A (n=733) was randomly divided into a training cohort (n=513) and an internal validation cohort (n=220) at a 7:3 ratio. Institution B (n=164) served as external validation cohort 1, while Institutions C and D (n=143) were combined as external validation cohort 2. In the training cohort, a nomogram model was constructed by stepwise selection of features through univariate and multivariate logistic regression. The model's discrimination, calibration, and clinical applicability were assessed using the area under the receiver operating characteristic curve (AUC), decision curve analysis (DCA), and clinical impact curve (CIC). Results: The final nomogram integrated the inflammatory biomarker NLR with patient age and Ultrasound(US) features. This model demonstrated excellent predictive performance across the training cohort (AUC 0.841, 95%CI: 0.807-0.872), internal validation cohort (AUC 0.828, 95%CI: 0.772-0.876), external validation cohort 1 (AUC 0.756, 95%CI: 0.683-0.820), and external validation cohort 2 (AUC 0.833, 95%CI: 0.762-0.890). DCA and CIC evaluations further confirmed the model's good calibration and significant net clinical benefit.Additionally,1000 bootstrap resamplings in the entire dataset demonstrated robust diagnostic performance(AUC 0.826, 95%CI: 0.798-0.852).The nomogram maintained robust generalizability and clinical practical value across different centers, populations, and examination equipment. Conclusion: The nomogram model we developed has good diagnostic performance and provides added value for the individualized diagnosis and treatment of PTC.