Next-generation sequencing technologies have started a new era of respiratory tract research in recent years. Alterations in the respiratory microbiome between healthy and malignant conditions have been revealed. However, the composition of the microbiome varies among studies, even in similar medical conditions. Also, there is a lack of complete knowledge about lung–gut microbiome interactions in lung cancer patients. The aim of this study was to explore the lung–gut axis in non-small-cell lung cancer (NSCLC) patients and the associations between lung–gut axis microbiota and clinical parameters (CRP, NLR, LPS, CD8, and PD-L1). Lung tissue and fecal samples were used for bacterial 16S rRNA sequencing. The results revealed, for the first time, that the bacterial richness in lung tumor tissue gradually decreased with an increase in the level of PD-L1 expression (p < 0.05). An analysis of β-diversity indicated a significant positive correlation between the genera Romboutsia and Alistipes in both the lung tumor biopsies and stool samples from NSCLC patients (p < 0.05). Survival analysis showed that NSCLC patients with higher bacterial richness in their stool samples had prolonged overall survival (HR: 2.06, 95% CI: 1.025–4.17, p = 0.0426).

Colorectal cancer (CRC) encompasses biologically diverse cellular phenotypes with distinct microRNA (miRNA) expression patterns. Assessing whether cell-model-derived alterations recur in extracellular vesicles (EVs) and human tumors may help prioritize candidates with broader relevance to CRC. We performed small RNA sequencing of non-malignant HCoEpiC cells, differentiated enterocyte-like Caco-2 cells, less differentiated HCT116 cells, and their corresponding EVs. miRNAs showing concordant dysregulation in both CRC cell lines relative to HCoEpiC, with greater absolute log₂ fold changes in HCT116, were prioritized and evaluated in human colorectal tissues and TCGA-COAD/READ. Seven of 29 prioritized cellular miRNAs met the EV selection criteria. Of the 26 candidates measurable in the patient-tissue dataset, eight showed FDR-significant, directionally concordant differences between CRC and healthy colorectal tissues, including five concordantly dysregulated across cells, EVs, and patient tissues. Analysis of 11 matched TCGA tumor–non-tumor pairs reproduced reduced miR-30a-3p, miR-30c-2-3p, and miR-1271-5p and increased miR-140-5p expression. Target-based enrichment and exploratory miRNA-mRNA correlation analyses identified associations with CRC-relevant signaling and MYC-related, cell-cycle, DNA-repair, and metabolic programs. No significant associations with the evaluated clinicopathological features or overall survival were detected. Collectively, these findings identify a reproducible subset of CRC-associated miRNA alterations and highlight reduced miR-30a-3p and miR-30c-2-3p expression as a recurrent family-level pattern warranting mechanistic investigation.