| ObjectiveWith the advancement of imaging technology and the popularization of health examinations,the detection rate of early or mid-term lung cancer is getting higher and higher.The preoperative staging of non-small cell lung cancer(NSCLC),especially the mediastinal lymph node staging,is the key to determining whether or not the operation can be performed and the prognosis of the patient.To investigate the independent risk factors of mediastinal lymph node metastasis in early or mid-term NSCLC patients with diameter ?5 cm and to construct a mathematical prediction model for the diagnosis of mediastinal lymph node metastasis,providing data support for preoperative prediction of mediastinal lymph node metastasis in patients with clinically advanced NSCLC patients with diameter<5 cm.In order to improve the accuracy of preoperative mediastinal lymph node staging,assist clinicians in making rational diagnosis and treatment decisions.MethodsWe retrospectively analyzed 608 NSCLC patients meeting the inclusion criteria admitted to the Department of Thoracic Surgery,Jiangsu Provincial People's Hospital between January 2012 and August 2017.The enrolled cases were randomly divided into two groups at a ratio of 3:1.The former is a modeling group(group A)for establishing a model;the latter is a verification group(group B)for external validation of the model and comparison with other models.Summarize the relevant clinical history data,imaging data and pathological conditions of 456 patients in group A.Univariate and multivariate analysis were used to select independent risk factors associated with mediastinal lymph node metastasis in early or mid-term NSCLC patients with a diameter of ? 5 cm and to establish a binary Logistic regression mathematical model.The H-L test was used to evaluate the fit of the model.The area under the receiver operating characteristic(ROC)curve(AUC)was used to assess the accuracy of the prediction model and determine the optimal cutoff value.The data of 152 patients in group B were used to determine the consistency of the actual metastasis of the mediastinal lymph nodes by chi-square test,and the consistency rate was evaluated by Kappa value.Substituting Group B data into each model,comparing the predictive ability of each mathematical prediction model,and judging its clinical application value.ResultsMultivariate analysis showed that five factors,age,tumor diameter,tumor location,carcinoma pathology,and pleural pull sign,were independent risk factors for the diagnosis of mediastinal lymph node metastasis in patients with early or mid-term NSCLC with a diameter of? 5cm.Through the correlation analysis of group A data,the mathematical predictive model established is the possibility of lymph node metastasis P=ex/(1+ex),where x=-2.831 +(0.825 tumor diameter)+(1.53 tumor site)+(0.779 pleural traction sign)+(1.883 pathological type)-(0.06 age),where e is the natural logarithm.The H-L test showed that there was no significant difference between the predicted and observed values.The area under the ROC curve was 0.763(95%CI:0.697-0.829).The results of external validation show that this model predicts the consistency of mediastinal lymph node metastasis in patients with early or mid-term NSCLC with a diameter of ? 5 cm,with a sensitivity of 68.75%and a specificity of 92.65%.Compared with VA model and Fudan model,this research model has a wider scope of application and higher accuracy.Conclusion1.The age,tumor size,tumor location,carcinoma pathology and pleural traction sign were the independent risk factors of mediastinal lymph node metastasis in early or mid-term NSCLC with diameter ? 5 cm.2.The mathematical prediction model of mediastinal lymph node metastasis in early or mid-term NSCLC with diameter ? 5 cm was P=ex/(1+ex),x =-2.831 +(0.825 tumor diameter)+(1.53 tumor site)+(0.779 pleural traction sign)+(1.883 pathological type)-(0.06 age),where e is the natural logarithm.3.The mathematical model established in this study has high sensitivity and specificity for the diagnosis of mediastinal lymph node metastasis of NSCLC ? 5 cm in diameter,and its prediction ability and accuracy are higher than other similar models.4.Combined with our prediction model,it is recommended to perform a further examination of patients with high risk of mediastinal lymph node metastasis to clarify the nature of lymph nodes. |
PDF Full Text Request