| Objective:Lateral lymph node metastasis(LLNM)affects the prognosis and survival of patients with thyroid carcinoma.To explore the influencing factors of lateral lymph node metastasis of thyroid papillary carcinoma and provide reference for surgical diagnosis and treatment.Methods:Patients who underwent surgical treatment in the Second Affiliated Hospital of Dalian Medical University from October 2019 to December 2020,and who were pathologically confirmed as papillary thyroid carcinoma after surgery,with complete clinical data were selected.Randomly divided into training group and validation group according to the ratio of 1:1.The training group is used to build the model,and the validation group is used to validate the model.The relationship between lateral lymph node metastasis and 11 clinical characteristics including age,gender,Hashimoto’s disease,maximum diameter of tumor,tumor location,calcification,proximity to the dorsal side of thyroid,central lymph node metastasis(CLNM),multifocality,capsular invasion,vascular invasion were retrospectively analyzed.SPSS software was used for single factor and multiple factor logistic regression analysis to screen out relevant risk factors,and R software was used to draw a visual nomogram.The discrimination,calibration and clinical practicability of the nomogram were evaluated by receiver operating characteristic curve(ROC),calibration plot and decision curve analysis(DCA).Results:After screening by inclusion and exclusion criteria,a total of 370 cases were collected and randomly divided into training group and validation group,with 185 cases each.By comparison between groups,we found that there was no statistical difference in clinical and pathological indicators between the training group and the validation group.This shows that the data of training group and validation group are comparable.Univariate logistic regression analysis,with P < 0.2 as the standard for preliminary screening,showed that six factors,including the maximum diameter of the tumor,central lymph node metastasis,multifocality,capsule invasion,proximity to the dorsal side,and the location of the tumor,were different between the group with cervical lymph node metastasis and the group without cervical lymph node metastasis.With P< 0.05 as the criterion,multiple logistic regression analysis showed that maximum tumor diameter(OR=1.157,95%CI: 1.078-1.242,P<0.001),CLNM(OR=10.291,95%CI: 2.584-40.992,P=0.001),multifocality(OR=4.351,95%CI: 1.247-15.182,P=0.021),proximity to the dorsal side(OR=4.498,95%CI: 1.455 ～ 13.906,P=0.009),the location of the tumor(the upper part was the reference,the middle OR=0.127,95%CI: 0.033 ～ 0.489,P=0.003;The lower OR=0.056,95%CI: 0.011-0.298,P=0.001)were independent risk factors for lymph node metastasis in the lateral cervical region of thyroid cancer.The nomogram model was constructed based on these 5 factors.In the ROC curve,the AUC(Area Under Curve)of the model is 0.902,and the AUC of the validation group is 0.869.The AUC values of LLNM were 0.76,0.648,0.616,0.697,0.629,respectively,as predicted by risk factors such as maximum tumor diameter,central lymph node metastasis,multifocal,proximity to the dorsal thyroid and tumor location.In the calibration curve,the C-index of the model was 0.902,the corrected C-index was 0.875,the C-index of the verification group was0.869(95%CI: 0.785 ～ 0.841),and the corrected C-index was 0.842.The bias-corrected curve is in good coincidence with the actual curve.In the DCA curve,the threshold probability ranges from 2% to 83%.The threshold probability of the verification group ranges from 5% to 96%.The model has good differentiation,calibration and clinical practicability.Conclusions:Based on five factors including maximum tumor diameter,central lymph node metastasis,multifocal,proximity to the dorsal thyroid,and tumor location,the graph can intuitively and accurately predict the risk of cervical lymph node metastasis in patients with papillary thyroid carcinoma. |
PDF Full Text Request