Receptor–ligand pair typing and prognostic risk model for papillary thyroid carcinoma based on single-cell sequencing

Abstract

The papillary thyroid carcinoma (PTC) microenvironment consists of various cancer and surrounding cells, and the communication between them is mainly performed through ligand-receptor (LR) interactions. Single-cell RNA sequencing (scRNA-seq) has been performed to investigate the role of intercellular communication networks in tumor progression. In addition, scRNA-seq can accurately identify the characteristics of immune cell subsets, which is of great significance for predicting the efficacy of immunotherapy. In this study, the cell-cell communication network was analyzed through LR pairs, and a new PTC molecular phenotype was developed based on LR pairs. Furthermore, a risk model was established to predict patient response to PD-1 blockade immunotherapy. The scRNA-seq dataset was obtained from GSE184362, and the bulk tumor RNA-seq dataset was obtained from The Cancer Genome Atlas. CellPhoneDB was used for cellular communication analysis. LR pair correlations were calculated and used to identify molecular subtypes, and the least absolute shrinkage and selection operator (Lasso) Cox regression was used to develop a risk model based on LR pairs. The IMvigor210 and GSE78220 cohorts were used as external validations for the LR.score to predict responses to PD-L1 blockade therapy. A total of 149 LR pairs with significant expression and prognostic correlation were included, and three PTC molecular subtypes were obtained from those with significant prognostic differences. Then, five LR pairs were selected to construct the risk scoring model, a reliable and independent prognostic factor in the training set, test set, and whole dataset. Furthermore, two external validation sets confirmed the predictive efficacy of the LR.score for response to PD-1 blockade therapy.