A sixteen-genus lung bacterial profile is associated with early-stage non-small cell lung cancer progression

Lung carcinomas evolve through complex interactions with the tumor microenvironment (TME), where intratumoral bacteria are being increasingly recognized as biologically relevant components. We investigated the microbiota signatures of normal and malignant tissues from patients with early-stage non-small cell lung cancer (NSCLC) and explored their associations with the clinical outcomes. We analyzed paired tumor and normal tissues from patients with early-stage NSCLC with favorable or unfavorable outcomes who were matched by stage, nodal status, and histology. The microbiota was profiled by 16​S rRNA sequencing and analyzed using DADA2/QIIME 2, centered log-ratio (CLR)-based diversity tests, MiRKAT, and PICRUSt2. Immune cell infiltration and tumor mutations were assessed by performing IHC and targeted NGS. Among a total of 155 paired tumor and normal tissues, the differential abundance analysis detected 10 genera in tumor tissues, 8 in normal tissues, and 3 in paired tumor–normal tissues, for a total of 16 genera associated with the outcome in at least one tissue type. The performance of the composite U16 score in discriminating deceased patients from surviving patients had an AUC of 0.819 (95% CI: 0.753–0.885), which was confirmed by a pooled out-of-fold AUC of 0.692 (95% CI: 0.608–0.775). The U16 score was significantly associated with the number of FOXP3⁺/CD4⁺ TILs and specific tumor mutations and remained independently associated with outcomes in multivariate models including age. These findings support a distinct tumor-associated bacterial signature present in early-stage NSCLC that stratifies patient outcomes independently of molecular features and is linked to an immunosuppressive TME.