Integrative multi-database analysis of a microbiome-based survival model in gastric cancer
Microbiology Spectrum · 20.11.2026
DOI: 10.1128/spectrum.01533-26 ↗Kurz gefasst
Fachartikel in der Zeitschrift Microbiology Spectrum, online erschienen am 20.11.2026. Verfasst von Xiaoyong Zhou und 5 weiteren Autor:innen. Beteiligte Einrichtung: Department of Gastroenterology, Henan Provincial People's Hospital (People's Hospital of Zhengzhou Univeristy), Zhengzhou, Henan, China. Die Arbeit nennt 32 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
This study evaluated a gastric cancer survival model based on tumor-associated microbial DNA profiles and examined their relationships with bulk host marker-expression scores. Microbiome, host-transcriptome, and clinical data from 118 matched TCMA/TCGA-STAD specimens (109 primary tumors and nine adjacent tissues) were integrated. After random 7:3 split and exclusion of incomplete samples, 68 training cases and 32 validation cases remained. A multi-step screening identified survival-associated microbial features, followed by PCA to construct a microbial risk score (MRS). An integrated model combining MRS, age, and sex was built, with reviewer-requested tumor-only sensitivity analyses for pathologic stage and Helicobacter . Neisseria and Veillonella showed hazard ratio estimates below 1 across multiple transformations, with Actinomyces showing the same direction at a trend-level significance threshold. Within the original age-sex model, MRS was associated with overall survival (HR = 1.73, P = 0.002). The integrated model (C-index 0.707) outperformed the age-sex comparator (0.590) and the microbiological-only model (0.672) in the training set. In the training set, high-risk patients had shorter median survival (457 vs 2,100 days, P = 0.001; 1-/2-/3-year AUCs: 0.683, 0.760, 0.768). In the held-out internal validation subset, the model yielded a C-index of 0.636 ( P = 0.0494). In 109 primary-tumor specimens, genus-level Helicobacter DNA signal correlated with bulk B-cell and mast-cell marker-expression scores after multiple-testing correction (rho = 0.247, q = 0.049; rho = 0.282, q = 0.030); the dendritic-cell association was not retained ( q = 0.271), and Helicobacter did not improve a tumor-only three-genus survival model (likelihood-ratio P = 0.803). The age-sex-microbiome model showed exploratory survival discrimination; a post hoc tumor-only analysis did not demonstrate incremental value of the three genera beyond age, sex, and pathologic stage (likelihood-ratio P = 0.571). The Helicobacter associations with bulk B-cell and mast-cell marker-expression scores require orthogonal validation. IMPORTANCE This study examines microbiome-informed survival stratification in gastric cancer. Using tumor-associated microbial DNA profiles from 118 matched specimens, the researchers built a survival model combining three genus-level features ( Actinomyces , Neisseria , and Veillonella ) with age and sex. In the training set, this model outperformed the age-sex comparator and distinguished high-risk patients (median survival 457 days) from low-risk patients (2,100 days). In the held-out internal validation subset, the model yielded a C-index of 0.636. Additionally, genus-level Helicobacter DNA signal correlated with bulk B-cell and mast-cell marker-expression scores in the tumor-only analysis. These results define an exploratory microbiome-based framework for further prognostic and biological validation in gastric cancer.
Schlagwörter
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Department of Gastroenterology, Henan Provincial People's Hospital (People's Hospital of Zhengzhou Univeristy), Zhengzhou, Henan, China.
- Cancer Center, Henan Provincial People's Hospital (People's Hospital of Zhengzhou University), Zhengzhou, Henan, China.
ORCID-Profile: Xiaoyong Zhou
Zitieren
APA
Zhou, X., Yan, B., Liu, M., Kuang, S., Ding, H., & Luo, X. (2026). Integrative multi-database analysis of a microbiome-based survival model in gastric cancer. Microbiology Spectrum. https://doi.org/10.1128/spectrum.01533-26
Vancouver
Zhou X, Yan B, Liu M, Kuang S, Ding H, Luo X. Integrative multi-database analysis of a microbiome-based survival model in gastric cancer. Microbiology Spectrum. 2026. doi:10.1128/spectrum.01533-26
BibTeX
@article{Zhou2026,
author = {Zhou, Xiaoyong and Yan, Binbin and Liu, Mingyue and Kuang, Shengli and Ding, Hui and Luo, Xiaoying},
title = {Integrative multi-database analysis of a microbiome-based survival model in gastric cancer},
journal = {Microbiology Spectrum},
year = {2026},
doi = {10.1128/spectrum.01533-26}
}
EUVIMED