Construction of a Prognostic Model for Hepatocellular Carcinoma Based on Cell Death-related Genes and Characterization of Immune Microenvironment

Wei  Wu; Yingji  Wang; Xiaocheng  Zhao; Ke  Dong; Maode  Li; Xiang  An; Yingyan  Xu; Shuai  Wang; Dexin  Li

doi:10.18502/ijaai.v25i3.21259

Wei Wu Department of Anesthesia and Surgery Center, Sichuan Provincial People’s Hospital, University of Electronic Science and Technology of China, Chengdu, Chin
Yingji Wang Department of Geriatric Endocrinology, Sichuan Provincial People's Hospital, University of Electronic Science and Technology of China, Chengdu, China
Xiaocheng Zhao Department of Hepatobiliary Surgery, Sichuan Provincial People's Hospital, University of Electronic Science and Technology of China, Chengdu, China
Ke Dong Department of Hepatobiliary Surgery, Sichuan Provincial People's Hospital, University of Electronic Science and Technology of China, Chengdu, China
Maode Li Department of Hepatobiliary Surgery, Sichuan Provincial People's Hospital, University of Electronic Science and Technology of China, Chengdu, China
Xiang An Department of Hepatobiliary Surgery, Sichuan Provincial People's Hospital, University of Electronic Science and Technology of China, Chengdu, China
Yingyan Xu Department of Hepatobiliary Surgery, Sichuan Provincial People's Hospital, University of Electronic Science and Technology of China, Chengdu, China
Shuai Wang Department of Anesthesia and Surgery Center, Sichuan Provincial People’s Hospital, University of Electronic Science and Technology of China, Chengdu, China
Dexin Li Department of Hepatobiliary Surgery, Sichuan Provincial People's Hospital, University of Electronic Science and Technology of China, Chengdu, China

DOI: https://doi.org/10.18502/ijaai.v25i3.21259

Keywords: Hepatocellular carcinoma; Immune infiltration; Prognostic genes; Single-cell analysis; 25 types of cell death

Abstract

Hepatocellular carcinoma (HCC) is the predominant type of primary liver cancer. This study aimed to elucidate the involvement of genes associated with 25 cell-death modalities in HCC development and progression. HCC transcriptomic datasets were integrated with curated cell death-related genes. Candidate genes were screened by differential expression analysis and protein– protein interaction network construction. Prognostic genes were identified using univariate Cox regression, proportional hazards assumption testing, and stepwise multivariate Cox regression. A risk score model and a nomogram were established, followed by risk stratification and analyses of immune infiltration, immune checkpoints, somatic mutations, and in silico drug sensitivity. Single-cell RNA sequencing was used to identify key cell types, infer temporal dynamics, and characterize intercellular communication, and findings were validated by quantitative real-time PCR (qRT-PCR). MAPT, CDKN2A, NQO1, CHGA, SERPINE1, and RET were identified as prognostic genes, and the risk model and nomogram showed good prognostic performance. Immune profiling revealed significant differences in multiple immune cell subsets between risk groups, including activated CD4+ T cells. Notably, CDKN2A correlated with activated CD4+ T cells, NQO1 with natural killer cells, RET with CD4+ central memory cells, and SERPINE1 with activated dendritic cells; RET also showed the strongest positive correlation with HAVCR2. Mutation spectra differed across risk groups, and ten drugs displayed significant predicted IC50 differences; all six genes were negatively correlated with KIN001.135. Single-cell analyses highlighted hepatocytes as a key cell type with strong hepatocyte–epithelial communication. qRT-PCR confirmed higher MAPT, CDKN2A, NQO1, and SERPINE1 expression in HCC tissues than in normal tissues.