A postoperative recurrence prediction model for intrahepatic cholangiocarcinoma based on multi-omics analysis of adjacent-to-tumor tissues

Background The patients with intrahepatic cholangiocarcinoma (iCCA) are highly susceptible to recurrence after radical resection, while predicting recurrence remains challenging. Adjacent-to-tumor tissues (ATTs), as the main microenvironment for postoperative recurrence, exhibited superior predictive value for recurrence compared with tumor tissues. However, the postoperative recurrence prediction model based on iCCA ATTs characteristics has not been studied. This study aims to construct recurrence prediction model based on iCCA ATTs and discover possible beneficial postoperative treatment options. Methods Consensus clustering was employed to classify the proteome of 116 iCCA ATTs. Multivariate Cox regression was used to construct recurrence prediction model. Tissue microarray containing 88 iCCA ATTs (another independent cohort) and immunohistochemistry were used for validating the expression of target proteins. Results We classified iCCA ATTs into two subtypes (S1 and S2) based on 116 iCCA patients’ proteomic data, where S1 exhibited higher recurrence rates and immune scores than those of S2. We constructed a multivariate Cox regression model based on four molecules (ENO3, HSPA13, POSTN, PTBP3). The expression of POSTN was an independent prognostic factor for recurrence of iCCA. The high-risk group for recurrence exhibited a poorer response to immunotherapy but was more sensitive to certain chemotherapy and targeted therapies. Conclusions We obtained novel molecular subtyping and constructed a postoperative recurrence prediction model based on iCCA ATTs, offering novel perspectives for tumorigenesis and providing some references for postoperative treatment options.