Imune Cell Infiltration Hierarchical Clustering Based Molecular Subtyping for Predicting Prognosis and Immunotherapy Efficacy in Esophageal Squamous Cell Carcinoma

LIU Min; CHENG Haodong; CHEN Yuanchao; CUI Yuelong; GAO Shaowei; SUN Liling; WANG Dong; NIU Hongxing; QI Yijun

doi:10.15926/j.cnki.issn2096-7381.2026.01.003

Abstract：Objective To construct an immune molecule-based prognostic model for predicting esophageal squamous cell carcinoma(ESCC)prognosis and immunotherapy efficacy based on estimates of the types and characteristics of immune infiltrating cells within the tumor microenvironment(TME)in ESCC.construct an immune molecule-based prognostic model for predicting ESCC prognosis and immunotherapy efficacy.Methods The mRNA transcriptomic data of 179 ESCC samples(GSE53625)were retrieved from the Gene Expression Omnibus(GEO)databases.The proportions of immune cell infiltration in 179 ESCC tissue were calculated using the CIBERSORT algorithm;unsupervised hierarchical clustering was performed based on 22 immune infiltrating cell levels to identify key immune molecules associated with prognosis in different IIC groups.Based on the differential expression between IIC groups,a molecular prognostic TPM model was constructed,and the prognostic risk values(RS)of TPMpatients in each group were calculated.The predictive effect of TPMmodel on immune therapy response was validated further through multiple datasets(GSE 135222,GSE91061,IMvigor210).Kaplan-Meier curves were used to plot survival curves,and the Log-rank test was employed to assess differences in survival times between groups,with a test level of α=0.05.Results Among 179 ESCC samples,101 samples showed significantly infiltrated immune cells(P＜0.05).Through unsupervised hierarchical clustering analysis,101 patients were divided into IIC-1,IIC-2,and IIC-3 groups.Among them,the overall survival time of patients in the IIC-3 group was significantly shorter than that of patients in the IIC-2 and IIC-1 groups at 5 years.Further analysis identified differentially expressed immune related molecules in the IIC 3 group,including CD8A,CD79A,and CXCL9,to construct a TPM model.Among the 101 patients,the high-risk RS group(32 cases)had the shortest overall survival time;the 26 independently validated ESCC samples showed a significant negative correlation between RS and patient prognosis(P=0.002).Conclusion The immune infiltrating cells within the ESCC tumor microenvironment exhibit heterogeneity,represented by CD8A,CD79A,and CXCL9,dictate anti-tumor immunity of ESCC patients,and are correlated with long-term outcome of ESCC patients,which could provide guidance for individualized immunotherapy of ESCC patients.