Characteristics of gut and lung microbiota in patients with lung masses and their relationship with clinical features

Abstract

Objectives: The exploration of how dysbiosis relates to lung masses is still nascent, with few studies focusing on the microbial characteristics across various sites. Therefore, we categorized the microbiota into feces and bronchoalveolar fluid (BALF) groups to compare microbial characteristics between benign and malignant masses, analyze their clinical correlations, and develop predictive models for lung cancer.

Methods: A total of 238 fecal samples and 34 BALF samples were collected from patients with benign and malignant masses and then analyzed by 16 SrRNA. We explored the distinct composition of the gut and lung microbiota and their associations with clinical features. The diagnostic models were constructed using microbial features identified through two approaches: random forest algorithm with five-fold cross-validation and comparative analysis of significantly differential taxa. The performance evaluation was subsequently conducted using receiver operating characteristic (ROC) analysis.

Results: There was no significant difference in α-and β-diversity between feces and BALF groups. The relative abundance of Lachnospiraceae_NK4A136_group (P = 0.003232) and Erysipelotrichaceae_UCG-003 (P = 0.01316) in feces group and Proteobacteria (P = 0.03654) in BALF group were significantly increased in lung cancer patients. We also found Bacteroides (P = 0.01458) was abundant in NSCLC than those of SCLC in feces group, while the BALF group was dominated by norank_c_Cyanobacteria (P = 0.03384). Smoking history appeared to be related to the distribution of microbiota, with enrichment of Parabacteroides (P = 0.02054) in feces and Prevotella_1 (P = 0.03286) in BALF. Furthermore, the patients with Sellimonas (P = 0.04148) in feces and Alloprevotella (P = 0.04283) in BALF seemed to have better response to chemotherapy combined with immunotherapy. For differentiating benign and malignant masses, the combination of Megasphaera and norank_p__Saccharibacteria in BALF demonstrated superior predictive performance, with an AUC reaching 0.8 (95% CI 0.59-1).

Conclusion: The microbiota composition significantly differed between benign and malignant masses in both fecal and BALF groups, with minimal evidence supporting microbial migration between these two sites. Our findings suggest that BALF microbiota may serve as a more reliable biomarker for lung masses classification, offering valuable insights for early diagnosis and clinical decision-making.

Keywords: Clinical features; Gut microbiota; Lung masses; Lung microbiota.

Fig. 1

The complete methodology

Fig. 2

Differences in α- and β-diversity between the two groups. The α-diversity was assessed using the Ace and Chao1 indices for both feces (A) and bronchoalveolar lavage fluid (BALF) group (B), while the β-diversity was evaluated through Principal Coordinate Analysis (PCoA) and Non-metric Multidimensional Scaling (NMDS) for feces (C) and BALF group (D)

Fig. 3

Distinct composition and microbial characteristics of the two groups at each level. Microbiota composition analysis of feces (A) and bronchoalveolar lavage fluid (BALF) (B) group in patients with benign and malignant masses at genus level. The length of the histogram represents linear discriminant analysis (LDA) scores computed for differentially abundant microbes in feces (C) and BALF (D) of Benign (red) and Malignant (blue) (LDA score > 3). Taxonomic cladogram from linear discriminant analysis effect size (LEfSe) showing differences in feces (E) and BALF group (F). Red dots indicate a significant difference while yellow dots indicate no significant difference. *P < 0.05, **P < 0.01, ***P < 0.001

Fig. 4

Relationship between microbiota and clinical features of lung cancer patients. Analysis of feces (A) and bronchoalveolar lavage fluid (BALF) (B) group differences between small cell lung cancer (SCLC) (red) and non-small cell lung cancer (NSCLC) (blue) at the genus level. Analysis of feces (C) and BALF (D) group differences between smoking (red) and non-smoking (blue) at the genus level. Patients who responded were further analyzed and divided into partial response (PR)_ complete response (CR) (red) and stable disease (SD) (blue) groups, and the difference of feces (E) and BALF (F) group was analyzed at the genus level *P < 0.05, **P < 0.01, ***P < 0.001

Fig. 5

Predictive value of microbiota in lung cancer patients. The random forest model demonstrated robust performance through five-fold cross-validation (A). The ROC curve of the five-fold cross-validation experiment(B). The differential microbiotas of lung cancer patients were selected for prediction model. Prediction results of the feces (C) and bronchoalveolar lavage fluid (BALF) (D) group at the genus level