Analysis Of Risk Factors And Establishment Of Prediction Model For Papillary Thyroid Carcinoma With Central Lymph Node Metastasis

Objective(s): The aim of this study was to investigate the independent risk factors for central lymph node metastasis(CLNM)in patients with papillary thyroid carcinoma(PTC),and to construct a clinical prediction model based on the screened indicators to predict the probability of CLNM in PTC patients and to guide physicians to make individualized treatment decisions for PTC patients with negative clinical lymph node metastasis.Methods: The clinical data of 317 patients with PTC who visited our hospital from January 2016 to September 2022 were retrospectively analyzed.They were divided into CLNM-positive and CLNM-negative groups according to the postoperative pathological Central District lymph node findings.Clinical data such as general clinicopathological characteristics,laboratory test indexes and ultrasound indexes of PTC patients were used as predictors for univariate analysis to screen out statistically significant influencing factors(p<0.05),and then multi-factor logistic regression analysis was performed to further screen out independent risk factors for CLNM in PTC patients.The model that was constructed based on the screened independent predictors was presented in the form of nomogram.Finally,ROC curves,calibration curves,and decision analysis curves were applied to evaluate the model in three dimensions of discrimination,calibration,and clinical usefulness.Results: Among 317 patients with PTC,155 were positive for CLNM and 162 were negative for CLNM,and the incidence of CLNM was 48.9%.In the univariate regression analysis,age(p=0.007),gender(p=0.003),bilaterality(p=0.038),tumor size(p<0.001),and tumor location(p=0.012)were significantly different(p<0.05)and included in the multifactorial analysis.The multifactorial analysis showed that age,gender,tumor size,and tumor location were independent risk factors for CLNM.A prediction model based on the above four indicators was established and presented as nomogram.The model showed good predictive ability with an AUC of 0.733.Calibration curves indicated that the predicted and actual probabilities of CLNM were in a good agreement,and the mean absolute error between actual and predicted risk was 0.023 after 1000 replicate sampling.Decision analysis curves indicated that the model could provide a good net clinical benefit for PTC patients.Conclusion(s):1.In this study,age,gender,tumor size,and tumor location were derived as independent risk factors for CLNM in PTC patients by univariate and multifactorial.2.The CLNM risk prediction model constructed based on the above indicators has good discrimination,calibration and net clinical benefit,which can individually assess the probability of CLNM in PTC patients before surgery and help guide clinical decision making.