Development And Validation Of Nomogram And Risk Classification System Predicting Overall Survival In Patients With Stage Ⅳ Siewert Type Ⅱ Adenocarcinoma Of The Esophagogastric Junction

Background:Adenocarcinoma of the esophagogastric junction(AEG)is an adenocarcinoma whose tumor center is between 5cm and 5cm of the esophagogastric junction and involves the esophagogastric junction(EGJ).The incidence of the AEG has been increasing recently.About 63% of AEG patients are diagnosed at an advanced stage with poor prognosis.The 5-year survival rate is between 25% and 47%.AEG is staged using the esophageal cancer TNM staging to guide the clinical treatment and prognosis evaluation.For patients with stage Ⅳ b AEG,the TNM staging system has limited predictive power.Therefore it’s necessary to develop new clinical predictive models,including more independent prognostic factors,to more comprehensively analyze and predict the prognosis of patients with advanced AEG,and to guide the individualized treatment of patients.Methods:From 2010 to 2016,data of ⅣB stage Siewert Ⅱ AEG patients were acquired from SEER database,and were randomly divided into the training cohort and the validation cohort in a 7:3 ratio.Kaplan-Meier analysis was used to analyze the survival of patients with stage Ⅳ b AEG.The prognostic factors for AEG patients were identified by Univariate and multivariate Cox proportional risk models.Based on the results of multivariate Cox analysis,the nomogram was validated to predict the1,2 years’ OS of patients with stage Ⅳ b AEG.Then the C-index was calculated and the ROC curve and calibration curve were drawn to verify the accuracy of the model.Finally,the risk stratification system was established according to the point of the nomogram.Results:1.Clinicopathological characteristics: A total of 1524 patients were included,including 1244 males(81.6%)and 280 females(18.4%),with a median age of 61 years old.The patients with liver,lung,bone and brain metastasis were 1223(80.2%),407(26.7%),303(19.9%)and 75(4.9%),respectively.1139 patients(74.7%)received chemotherapy,462 patients(31.0%)received radiotherapy,and 48 patients(3.1%)received palliative resection.1524 patients were randomly divided into a training cohort(n=1068)and a validation cohort(n=456).2.Survival analysis and prognostic analysis of patients:(1)K-M survival analysis showed that the median survival time of patients with age > 50,50-59,60-69,>69were 12,11,9,and 7 months(P<0.0001);The median survival time of patients with high,medium,low,and undifferentiated tumors were 15,11,9,8 months(P<0.0001);The median survival times for T0,T1-T2,and T3-T4 were 11,11 and 8 months(P<0.0001).The OS in the surgery group(HR:0.577,95%CI 0.429-0.776,P=0.0008)was higher than non-surgery group;The OS in the chemotherapy group(HR 0.435,95%CI 0.361-0.525,P<0.0001)was higher than non-chemotherapy group;Compared with patients without bone,brain and lung metastases,The HR values were 0.719(95%CI 0.603-0.856,P=0.012),0.589(95%CI 0.408-0.849,P=0.002),and 0.837(95%CI 0.716-0.978,P=0.025)respectively.(2)Univariate Cox regression analysis showed that age at diagnosis,T stage,tumor differentiation,primary surgery,chemotherapy,bone,lung,brain metastases and cumulative sites of metastases were related to prognosis(P<0.05).Multivariate Cox regression analysis indicated that age at diagnosis,T stage,degree of tumor differentiation,primary surgery,chemotherapy and brain metastasis were related to prognosis of patients(P<0.05).3.Development and validation of nomogram: Nomogram was constructed based on age at diagnosis,T stage,tumor differentiation,primary surgery,chemotherapy,and brain metastasis.The C-index of the training cohort and validation cohort were 0.691(95%CI: 0.683-0.702)and 0.695(95%CI: 0.681-0.707).The AUC values of training cohort predicted 1 and 2 years of OS were 0.7 and 0.79.The AUC values of validation cohort predicted 1 and 2 years of OS were 0.7 and 0.81.The calibration curve showed a good agreement.4.Development and validation of risk classification system: According to the cut-off value,the risk was stratified into low-risk group 0-90.4,medium-risk group90.5-170.1 and high-risk group 170.2-230.6.In the training cohort,the HR values of the medium and high-risk groups were 2.085(95%CI: 1.777-2.446,P<0.0001),26.98(95% CI 16.13-45.11,P<0.0001)compared with the low-risk group,HR in the high-risk group was 4.535(95%CI,3.064-6.710,P<0.0001)compared with the medium-risk group;In the validation set,the HR values of the medium and high-risk groups were 2.156(95%CI: 1.626-2.858,P<0.0001)and 7.166(95%CI 4.228-12.14,P<0.0001)compared with the low risk group,HR in the high-risk group was 4.439(95% CI 2.850-6.914,P<0.0001)compared with the medium-risk group.Conclusion:1.Receiving primary surgery and chemotherapy are beneficial factors,however,patients with advanced age,poor differentiation,advanced local stage and brain metastasis had a worse prognosis.2.For patients with stage Ⅳ b AEG,the nomogram established in this study can relatively accurately predict 1 and 2 years’ s OS,which was better than the AJCC TNM staging system.3.The risk stratification system constructed in this study can better distinguish high-risk patients and strengthen the clinical management of patients.Patients with a point less than 90.4 had a relatively good long-term prognosis and patients with a point greater than 170.3 have a poor long-term prognosis.