Machine learning algorithm integrates bulk and single-cell transcriptome sequencing to reveal immune-related personalized therapy prediction features for pancreatic cancer.

Abstract

Pancreatic cancer (PC) is a digestive malignancy with worse overall survival. Tumor immune environment (TIME) alters the progression and proliferation of various solid tumors. Hence, we aimed to detect the TIME-related classifier to facilitate the personalized treatment of PC. Based on the 1612 immune-related genes (IRGs), we classified patients into Immune_rich and Immune_desert subgroups via consensus clustering. Patients in distinct subtypes exhibited a difference in sensitivity to immune checkpoint blockers (ICB). Next, the immune-related signature (IRS) model was established based on 8 IRGs (SYT12, TNNT1, TRIM46, SMPD3, ANLN, AFF3, CXCL9 and RP1L1) and validated its predictive efficiency in multiple cohorts. RT-qPCR experiments demonstrated the differential expression of 8 IRGs between tumor and normal cell lines. Patients who gained lower IRS score tended to be more sensitive to chemotherapy and immunotherapy, and obtained better overall survival compared to those with higher IRS scores. Moreover, scRNA-seq analysis revealed that fibroblast and ductal cells might affect malignant tumor cells via MIF-(CD74+CD44) and SPP1-CD44 axis. Eventually, we identified eight therapeutic targets and one agent for IRS high patients. Our study screened out the specific regulation pattern of TIME in PC, and shed light on the precise treatment of PC.