Single-Cell Map Links Endothelial Cell Genes to Lung Cancer Survival

Researchers have built a prognostic model for lung adenocarcinoma (LUAD) from genes expressed in endothelial cells, the vascular cells that line tumor blood vessels. Using single-cell RNA sequencing data retrieved from the Gene Expression Omnibus database, the team applied hierarchical density-based weighted gene co-expression network analysis (hdWGCNA) and LASSO regression to identify endothelial cell-related differentially expressed genes, then narrowed these to six genes—CX3CL1, MCAM, AKAP12, MGP, CA4, and CAV2—that together form a risk score predicting patient survival.

The study, published open access in Discover Oncology on 28 October 2026 by Chengbo Tang, Pan Deng, Jia Liu, Sixuan Wu, and Feng Guo, rests on the premise that vascular angiogenesis drives tumor growth and metastasis, with endothelial cells regulating intercellular interactions within the tumor microenvironment. Rather than examining bulk tumor tissue, which averages signals across many cell types, the single-cell approach allowed the authors to isolate the molecular signature of the endothelial compartment and ask whether it carries prognostic information.

The model’s performance was supported by several lines of analysis. Risk scores were inversely correlated with patient survival and remained valid across multiple clinical subgroups. A SHapley Additive exPlanations (SHAP) analysis highlighted CAV2 as the gene making the largest contribution to risk prediction. The authors also constructed a nomogram combining the risk score with clinical parameters, which they report showed strong predictive accuracy. Enrichment and immune analyses revealed distinct pathway activations and immune cell distributions between high- and low-risk groups, and high-risk patients exhibited greater tumor mutation burden and dedifferentiation characteristics.

A notable feature of the work is the use of Scissor analysis, a method that links single-cell data to bulk-level clinical outcomes to identify which cell subpopulations are associated with a given phenotype. In this study, Scissor analysis revealed differences in cellular composition and spatial distribution between risk-stratified subgroups. Intercellular communication in the higher-risk subgroup was more complex and active, primarily involving adipose, epithelial, and immune-related cells, whereas overall communication in the lower-risk subgroup was comparatively weaker. This suggests that the tumor microenvironment of high-risk patients is not merely more inflamed but more densely wired, with endothelial cells potentially serving as hubs in these interactions.

The authors also examined drug sensitivity, finding significant differences in predicted therapeutic responses among risk groups, and they validated the expression of the model genes experimentally using quantitative real-time PCR in LUAD cells and tissues. These experiments corroborated the downregulation of the model genes in LUAD samples, consistent with the inverse relationship between risk scores and survival. The team frames the model as a theoretical framework for personalized therapeutic strategies, though as a computational and observational study it identifies associations rather than demonstrating causal mechanisms, and clinical application would require independent external validation. Because the article is being shared early as a peer-reviewed, accepted manuscript carrying a permanent DOI, some wording may still change before the final version of record appears.

Subject of Research: An endothelial cell-related prognostic gene model and tumor microenvironment analysis in lung adenocarcinoma using single-cell RNA sequencing and Scissor analysis.

Article Title: Integrative scRNA-seq and Scissor analysis reveal an endothelial cell-associated prognostic model and tumor microenvironment in lung adenocarcinoma

Article References: Tang, C., Deng, P., Liu, J., Wu, S., & Guo, F. (2026). Integrative scRNA-seq and Scissor analysis reveal an endothelial cell-associated prognostic model and tumor microenvironment in lung adenocarcinoma. Discover Oncology. https://doi.org/10.1007/s12672-026-06160-0

Image Credits: AI Generated

DOI: 10.1007/s12672-026-06160-0

Keywords: lung adenocarcinoma, single-cell RNA sequencing, endothelial cells, prognostic model, tumor microenvironment, LASSO regression, Scissor analysis, CAV2, tumor mutation burden, drug sensitivity, bioinformatics, cancer research