Abstract
Background: Tumor stem cells (TSCs) have been widely reported to play a critical role in tumor progression and metastasis. We explored the role of tumor stemness in intrahepatic cholangiocarcinoma (iCCA) and established a prognostic risk model related to tumor stemness for prognosis prediction and clinical treatment guidance in iCCA patients. Materials and Methods: The expression profiles of iCCA samples (E-MTAB-6389 and GSE107943 cohorts) were used in the study. One-class logistic regression algorithm calculated the mRNA stemness index (mRNAsi). The mRNAsi-related genes were used as a basis for the identification of mRNAsi-related molecular subtypes through consensus clustering. The immune characteristics and biological pathways of different subtypes were assessed. The mRNAsi-related risk model was constructed with differentially expressed genes (DEGs) between subtypes. Results: The patients with high mRNAsi had longer overall survival than that with low mRNAsi. Two subtypes were identified with that C2 had higher mRNAsi and better prognosis than C1. Tumor-related pathways such as TGF-β and epithelial-mesenchymal transition (EMT) were activated in C1. C1 had higher enrichment of cancer-associated fibroblasts and tumor-associated macrophages, as well as higher immune response and angiogenesis score than C2. We screened a total 98 prognostic DEGs between C1 and C2. Based on the prognostic DEGs, we constructed a risk model containing three genes (ANO1, CD109, and CTNND2) that could divide iCCA samples into high- and low-risk groups. The two groups had distinct prognosis and immune characteristics. Notably, the risk score was negatively associated with mRNAsi (R = -0.53). High-risk group had higher enrichment score of T cell inflamed GEP, INF-γ, and cytolytic activity, and lower score of estimated IC50 of 5-fluorouracil and cisplatin than low-risk group. Conclusions: This study clarified the important role of tumor stemness in iCCA and developed an mRNAsi-related risk model for predicting the prognosis and supporting the clinical treatment in iCCA patients. The three genes (ANO1, CD109, and CTNND2) may serve as potential targets for iCCA treatment.