Foundation And Verification Of A Dynamic Prognostic Prediction Model For Gallbladder Cancer Based On B7-H3 And TIL Immune Stratification Scoring System

Objective:Gallbladder cancer is a highly fatal malignancy of the biliary tract.Immunotherapy has emerged as a promising strategy in the treatment of gallbladder cancer.Tumor infiltrating lymphocytes(TIL)play an important role in the tumor and are associated with the efficacy of immunotherapy.CD8~+T cells are a key factor in the immunotherapeutic response and their high density is usually associated with a good prognosis of the tumor.Foxp3~+T cells suppress the anti-tumor immune response and their high density is usually associated with a poor prognosis.B7-H3(CD276)is involved in tumor immune escape and its high expression is associated with a poor prognosis.In this research,we explore the correlation between the number of CD8~+and Foxp3~+TIL and the expression of B7-H3 with clinical parameters and prognosis of gallbladder cancer.To explore the relationship between B7-H3 expression and the number of CD8~+,Foxp3~+TIL and its correlation with prognosis in gallbladder cancer.More importantly,for the establishment of an immune stratification scoring system based on the amount of B7-H3expression,CD8~+and Foxp3~+TIL,the construction of a dynamic gallbladder cancer prognostic prediction model and its validation to facilitate clinical decision making and improve the efficacy of individualized treatment.Methods:A total of 198 patients who attended Shengjing Hospital of China Medical University from January 2010 to October 2020 and met the inclusion criteria were retrospectively analyzed and randomly divided 1:1 into a model and validation group.Immunohistochemically staining was performed on case specimens to assess the number of CD8~+,Foxp3~+TIL and B7-H3 expression.Data analysis was performed using the statistical software SPSS 26.0 and R-3.6.3.Cox proportional risk regression analysis was used for univariate and multivariate analysis and a p-value of<0.05 was considered statistically significant.Kaplan-Meier curves were used to describe OS at different biomarker expression levels.Column line plots(Nomogram)were made using R-3.6.3software to demonstrate the clinical prediction model and to validate its accuracy.Results:High expression of B7-H3 was associated with low density of CD8~+TIL(P<0.05).In contrast,no significant difference was found between expression of B7-H3 and density of Foxp3~+TIL(P>0.05).Univariate analysis by Cox regression showed that B7-H3,CD8and Foxp3 were all associated with overall survival(P<0.05).In multivariate analysis,density of CD8~+TIL,density of Foxp3~+TIL,distant metastasis,tumor size and surgical approach were all independent risk factors for overall survival(P<0.05),while expression of B7-H3 was not independently associated with overall survival(P>0.05).Immune subgroups and immune scores were shown to be effective in predicting the prognosis of gallbladder cancer.Prediction models(6 in total)were developed based on immune factors,immune subgroups,immune scores,different pathological staging systems and other clinically relevant risk factors,all of which had good predictive power and dynamic nomogram plots.Conclusion:1.High expression of B7-H3 and high density of Foxp3~+TIL suggest a poor prognosis,whereas high density of CD8~+TIL suggest a good prognosis.2.There is a correlation between expression of B7-H3 and density of CD8~+T cell.Although no correlation was found between expression of B7-H3 and density of Foxp3~+TIL,patients with high expression of B7-H3,high density of Foxp3~+TIL and low density of CD8~+TIL had the poorest prognosis.These results suggest that simultaneous detection of expression of B7-H3 and the amount of CD8~+and Foxp3~+TIL may be a more accurate predictor of prognosis than either molecule alone.3.A dynamic predictive model of the immune scoring system based on the expression of B7-H3,density of CD8~+TIL and Foxp3~+TIL was developed and confirmed to have better predictive ability in predicting prognosis of gallbladder cancer ability,and to build dynamic nomogram plots.