Establishment Of A Diagnostic Model For Lymph Node Metastasis Of Esophageal Cancer Based On ¹⁸F-FDG PET/CT Metabolic Parameter SUVmax

Objective:To explore the relationship between ¹⁸F-FDG PET/CT parameters SUVmax and lymph node metastasis in patients with esophageal cancer,and to establish a lymph node metastasis model of esophageal cancer and evaluate its diagnostic efficacy,so as to provide reference for clinical decision-making.Materials and methods:A retrospective analysis was made on the cases of esophageal cancer treated by surgery in the The First Division of Thoracic Surgery of Fujian Medical University Union Hospital from November 2015 to October 2018 and examined by ¹⁸F-FDG PET/CT before operation.Their demographic characteristics?sex,age,etc.?,smoking history,drinking history,preoperative tumor-related indicators?CEA level,tumor location,tumor length,etc.?,PET/CT metabolic parameters SUVmax and Postoperative pathological data were collected.The patients were divided into positive group and negative group according to whether there were lymph node metastasis or not from the pathological data.Logistic binary regression analysis was used to establish a diagnostic model for lymph node metastasis of esophageal cancer.ROC curve,Delong test and paired chi-square test were used to analyze the difference and consistency between the diagnostic model and PET/CT criteria for lymph node metastasis.Results:According to the inclusion criteria,120 patients were enrolled,60 in the positive group and 60 in the negative group,all of them were squamous cell carcinoma.Univariate analysis showed that there were significant differences between the two groups in drinking history,tumor length and the metabolic parameters SUVmax of PET/CT in primary tumors.We establish a diagnostic model for lymph node metastasis of esophageal cancer,Ln?p/1-p?=-2.304+1.880*SUVmax+0.994*tumor length+0.850*drinking history.P value is the probability of lymph node metastasis in patients with esophageal cancer.The comprehensive test of model coefficients??²=32.1,P<0.001?shows that the regression model has statistical significance.Hosmer and Lemeshow test??²=3.290,P=0.772?,indicating that there is no significant difference between fitting equation and real equation.SUVmax>7.450,tumor length>3.75 cm and drinking history were independent risk factors for lymph node metastasis of esophageal cancer.According to the regression coefficient of the model equation,the score of the model was calculated,and the ROC curve analysis showed that the AUC of the model was 0.782.When the score was 3,the lymph node metastasis rate was more than 50%,and when the score was 4,the lymph node metastasis rate was87.1%.Further analysis showed that the diagnostic model and PET/CT diagnostic criteria had significant difference in AUC?0.782 v.s.0.725,Z=2.402,P=0.016?.The diagnostic efficiency of the diagnostic model was slightly better than that of the PET/CT diagnostic criteria.The sensitivity of the diagnostic model was higher than that of the PET/CT criteria?75.0%v.s.63.3%,?²=2.118,P=0.143?,while the specificity was lower than that of the PET/CT criteria?68.3%v.s.81.7%,?²=2.227,P=0.134?.There was no significant difference between the two groups.Conclusion:This study establishes a diagnostic model of lymph node metastasis of esophageal cancer based on ¹⁸F-FDG PET/CT metabolic parameter SUVmax.The model is more effective than PET/CT in diagnosing lymph node metastasis of esophageal cancer.It has certain predictive and diagnostic ability for lymph node metastasis of esophageal cancer and can provide diagnostic reference for clinic.