Development of a MVI associated HCC prognostic model through single cell transcriptomic analysis and 101 machine learning algorithms.
Journal:
Scientific reports
PMID:
40055377
Abstract
Hepatocellular carcinoma (HCC) is an exceedingly aggressive form of cancer that often carries a poor prognosis, especially when it is complicated by the presence of microvascular invasion (MVI). Identifying patients at high risk of MVI is crucial for personalized treatment strategies. Utilizing the single-cell RNA-sequencing dataset (GSE242889) of HCC, we identified malignant cell subtypes associated with microvascular invasion (MVI), in conjunction with the TCGA dataset, selected a set of MVI-related genes (MRGs). We developed an optimal prognostic model comprising 11 genes (NOP16, YIPF1, HMMR, NDC80, DYNLL1, CDC34, NLN, KHDRBS3, MED8, SLC35G2, RAB3B) based on MVI-related signature genes by integrating single-cell transcriptomic analysis with 101 machine learning algorithms. This model is meticulously crafted to forecast the prognosis of individuals afflicted with hepatocellular carcinoma (HCC). Additionally, we affirmed the predictive precision and superiority of our model through a meta-analysis against existing HCC models. Furthermore, we explored the differences between high- and low-risk groups through mutation and immune infiltration analyses. Lastly, we investigated immunotherapy responses and drug sensitivities between risk groups, providing novel therapeutic insights for liver cancer.
Authors
Keywords
Biomarkers, Tumor
Carcinoma, Hepatocellular
Clinical Decision-Making
Datasets as Topic
Gene Expression Regulation, Neoplastic
Humans
Immune Checkpoint Proteins
Immunotherapy
Kaplan-Meier Estimate
Liver
Liver Neoplasms
Machine Learning
Meta-Analysis as Topic
Microvessels
Mutation
Neoplasm Invasiveness
Precision Medicine
Prognosis
Risk Assessment
Single-Cell Gene Expression Analysis
Treatment Outcome