Microbial Signatures in Head and Neck versus Gastrointestinal Tumors: Identification and Prognostic Modeling.

This study identified key intra-tumor microbial signatures distinguishing head and neck cancers from gastrointestinal cancers and explored their diagnostic and prognostic potential. Intra-tumor microbial data of five cancer types were obtained from the Cancer Microbiome Atlas, and corresponding clinical data were retrieved from the Cancer Genome Atlas. The Wilcoxon test was used to analyze differences in microbial populations. Univariate logistic regression, least absolute shrinkage and selection operator, and recursive feature elimination were sequentially applied to screen optimal microbial markers, and a support vector machine classification model was constructed. A nomogram model and Kaplan-Meier curves were used to validate the predictive and prognostic value of the optimal microbes, respectively. Overall, 463 tumor samples and 47 controls were included. Twenty-three microbes showed significant differences in distribution between head and neck and gastrointestinal tumors; among these, eight overlapping microbes were selected as optimal markers. The SVM model based on these eight microbes achieved AUCs of 0.937 and 0.856 in the training and validation datasets, respectively. The nomogram model constructed with these markers showed high predictive accuracy (C-index = 0.8944 in training, 0.8023 in validation). Kaplan-Meier analysis revealed that high abundance of