| Objective:To establish more accurate-effective survival model to predict prognosis in M0 stage gallbladder cancer patients after surgical resection.Methods:A retrospective analysis was conducted in patients with M0 stage gallbladder cancer who underwent surgical resection and registered in the Surveillance,Epidemiology,and End Results(SEER)database from 2010 to 2014.Univariate survival analysis and multivariate analysis was performed for the related independent factors that might affect the gallbladder cancer specifc survival(GCSS).Loading the above independent factors into the R 3.4.1 software,a prognostic gallbladder cancer specifc survival(GCSS)model was formulated by using the nomogram tool of survival package to predict the probablity for GCSS.According to the GCSS-model,getting the scores of each factors to establish the gallbladder cancer specifc survival(GCSS)scale.When beginning to predict the survival of the gallbladder carcinoma patients,we should first get the total score by the GCSS-scale and then make the total score corresponding to the position of the ruler below by the GCSS-model,to get the 1 or 3 years cancer specifc survival(CSS)respectively.Bootstrap method was used to calculate the c-index of GCSS-model and 7th AJCC gallbladder carcinoma TNM staging respectively.And comparing the c-index of the two terms,the higher the c-index,the more accuracy of the model.The calibration curves were drawn to test the GCSS-model's prediction and actual observation of the 1-year CSS rate or 3-year CSS rate,and the red line is closer to the diagonal blue dotted line,the more consistent the results are.Results: A total of 1422 M0 stage gallbladder cancer patients were identified.The overall median survival time was 13 months(1-59 months),the median survival time was 18.17 months,1 or 3 years of CSS were 82.3% or 56.7% respectively.Based on the univariate and multivariate analysis,the independent prognostic factors of CSS include age,gender,lymph node dissection,postoperative adjuvant chemotherapy,tumor size,histological grading,p T staging and p N staging.Based on the independent prognostic factors,the GCSS-model and GCSS-scale for non-metastatic gallbladder cancer were established.The GCSS-scale showed that < 60 years old scored 2.20,> 60 years old scored 0,female scored 1.15,male scored 0,lymph node dissection scored 4.13 points,none lymph node dissection scored 0 points,postoperative chemotherapy scored 1.86,none chemotherapy scored 0,the tumor size less than 3cm scored 1.90,the tumor size >3 cm scored 0,and histological grade(I scored 3.72,II scored 1.78,III/ IV scored 0),p T stage(T1a stage scored 10,T1 b stage scored 8.33,T2 stage scored 6.10 T3 scored 2.55,T4 scored 0 points),p N stage(N0 scored 3.55,N1 scored 0.25,N2 scored 0).It is known from the model that the higher the total score,the higher 1 or 3 years CSS respectively.The discrimination test suggested that the GCSS-model's C-index was 0.775,which was higher than the seventh edition of AJCC staging(C-index = 0.649).The calibration test showed that the red line was near the diagonal blue dotted line in 1 or 3 years CSS calibration curve,indicated that the correlation between prediction and actual CSS is good.Conclusions:This study retrospectively analyzed the SEER database and established the GCSS-model based on 8 independent factors,which affecting postoperative survival of patients with M0 stage gallbladder cancer.Compared with the 7th edition AJCC stage of gallbladder carcinoma,this model had refined the contribution of the contribution of p T staging and p N staging and other related factors and proved to be more accurate and reliable prediction.More importantly,the model is useful for clinicians to screen poor prognosis patients for closer follow-up or adjuvant treatment. |
PDF Full Text Request