Exploring the role of the oral microbiome in saliva, sputum, bronchoalveolar fluid, and lung cancer tumor tissue: A systematic review.

Introduction
Cancer forms in lung tissues, typically in the cells that line the air passages. This neoplasia has two main types: small-cell lung cancer and non-small-cell lung cancer (NSCLC) [1]. Globally, lung cancer (LC) was the leading cause of cancer-related deaths in 2018, with an incidence of over 2 million cases, and was the cause of 1.76 million deaths worldwide [2]. This is the most prevalent type of cancer in men and occurs more frequently in Asia and Europe [2]. In the United States, 235,760 new cases were recorded in 2021, reflecting a mortality rate of 36.7 per 100,000 inhabitants [3].
Among the risk factors are activities of daily living, such as smoking tobacco (including cigarettes and pipes), which is the leading risk factor. However, LC also affects non-smokers due to exposure to secondhand smoke, occupational hazards, and environmental factors (e.g., asbestos, radon, and certain chemicals) [3]. The risk of developing this type of cancer and its progression has also been associated with lung infections primarily caused by bacterial components or other microbes. Initially believed to be sterile, sequencing studies have shown that healthy lungs host a microbiome closely resembling the microbiome of saliva, possibly due to lifelong micro-aspirations [4].
The oral microbiome is the collective genome of the ecological community of symbiotic, commensal, and pathogenic microbial residents in the oral cavity [5]. It forms an ecosystem that supports health under symbiotic conditions [6]. Conversely, oral dysbiosis is characterized by an imbalance in the relative abundance (quantitative and qualitative) of microbial species inhabiting the oral habitat. This imbalance can increase proteolytic, anaerobic, acidophilic, acidogenic, and saccharolytic species and decrease aerobic species, leading to several pathologies [7].
Several researchers have studied the association between the oral microbiome and LC and observed that a higher risk of LC is associated with lower alpha diversity of the oral microbiome. An altered risk was related to the abundance of specific taxa: reduced LC risk with Bacteroidetes and Spirochaetes and increased risk with Firmicutes. These findings suggest that decreased alpha diversity and specific taxa abundance influence LC pathogenesis [6,8].
The oral microbiome has received considerable attention as an LC diagnosis and progression biomarker. Current evidence consists of narrative reviews describing the relationship between the oral microbiome and LC [9]. However, few systematic reviews evaluate this relationship using diverse sources, including the oral microbiome [10,11]. This study aims to review and summarize primary research on the association of the oral microbiome, including saliva, sputum, bronchoalveolar lavage fluid (BALF), and tumor tissue, with LC diagnoses and progression, including immune checkpoint inhibitors (ICIs) therapy response. It also examines the biological mechanisms linking the oral microbiome to LC.

Methods
This systematic review followed the PRISMA protocol [12]: PROSPERO Register CRD 42,024,603,034

Define the review questions

•Are healthy versus lung cancer patients' microbial components and alpha diversity differences in saliva, sputum, BALF, and tumor samples?

•Are oral microorganisms associated with LC, LC recurrence, or progression, and ICIs therapy response?

•What are the biological mechanisms or pathways of the oral microbiome in lung cancer development?

Selection of keywords and search strategy (see supplementary material)

Determine the inclusion and exclusion criteria
Inclusion criteriaStudies exploring metagenomic sequencing techniques that evaluated the presence of oral bacteria in LC.1)Observational studies: cross-sectional, case and control, and cohort studies.Clinical studies that evaluated the presence and relationship of the oral microbiome with LC and its mechanisms.

2)Studies that compare the oral microbiome of patients with LC with a control group.

3)Studies that compare the oral microbiome in one or more cohorts of LC evaluating the tumor progression.

4)Studies that compare the oral microbiome in responders or non-responders to ICIs therapy.

5)Studies consider collecting and comparing saliva, sputum, bronchoalveolar lavage, brushing by bronchoscopy, and tumoral samples.

Exclusion criteria
Studies that have not established bacterial genera only report the phylum.
Incomplete data.

Quality analysis (Mod NOS scale)
The Newcastle-Ottawa Scale (NOS) was adapted to evaluate the quality of the studies [13]. The scoring considered: 1) selection, 2) comparability, 3) results, and 4) statistics domains. The individual items were evaluated by 1 or 0, respectively, to indicate whether each item met the quality, which yielded an ideal score of 14 points; the presence of the domains accounted for the risk of bias (Fig. S1).

Data extraction and analysis
Two independent reviewers (JAG; AW) reviewed the titles, abstracts, and full texts using Rayyan; any disagreement was resolved through discussion or third-party arbitration (GIL). The following data were extracted per article: citation, author, publication status, year of publication, country, type of study, type of cancer, sequencing technique, unit of analysis, alpha diversity, beta diversity, location of the sample, the microorganisms found, results, and conclusions. Alpha diversity was evaluated using Shannon, Simpson, Chao 1, and ACE indices through a meta-analysis of non-parametric data, employing the Quantile Estimation (QE) method in the meta-median package in R. The meta-analyses were conducted using Shannon's index by a random-effects model with the REML estimator for the between-study variance (τ²). The 95% confidence intervals (95% CIs) were adjusted using the Hartung–Knapp method, and for each study, CIs were calculated using the standard Wald approach (ES ± 1.96 SE). Heterogeneity was assessed by quantifying I², complemented by Cochran's Q test for overall heterogeneity. In subgroup analyses, the Q_between statistic was applied to test for differences between sampling technique categories. The certainty of the evidence was obtained using the GRADE system (Table S2). However, after reviewing the available databases, it was not possible to perform a meta-analysis of the microbiome sequencing results due to the heterogeneity of the analyses, selective reporting (only LC data are reported, and control data are absent), and lack of sample identification.

Results
1916 articles were identified from January 2015 to January 2025, with 36 duplicates removed, leaving 1880 articles. After screening the titles and abstracts, 1858 articles were excluded. Subsequently, 54 full-text articles were assessed for eligibility, of which four were excluded, resulting in 50 articles being selected for inclusion (Fig. 1) (Table 1).

Overview of studies
Among the selected studies, 29 were cross-sectional designs [8,14,16,19,[22], [23], [24],28,29,31,34,37,38,40,[45], [46], [47],49,51,52,55,[57], [58], [59], [60], [61]], seven case-control studies [15,17,18,21,25,26,32,41], and fourteen co 201hort studies [20,27,30,35,36,39,[42], [43], [44],48,50,53,54,56]. Six studies evaluated the ICIs therapy [20,27,35,36,43,62]. Of the 50 studies evaluated, twenty-three originated from China [8,[14], [15], [16],22,23,25,27,29,31,32,[35], [36], [37], [38], [39], [40],48,51,53,57,[60], [61], [62]], nine from the United States [17,18,21,41,42,45,48,49,56], three from Russia [26,28,50], two each from Portugal [32,34], India [19,55] and Korea [44,46,54], and one each from France [18], Japan [20], Switzerland [43], Taiwan [52], United Kingdom [24] and Spain [59].

Type of samples and localization
Saliva samples were the most frequently studied [[14], [15], [16], [17], [18], [19], [20], [21], [22], [23],62]. Nine studies also evaluated sputum samples [8,[24], [25], [26], [27], [28], [29], [30]]. Other studies evaluated BALF [[31], [32], [33], [34], [35], [36], [37], [38], [39]], brushing samples obtained by bronchoscopy [[40], [41], [42], [43], [44], [45], [46]], or lung tissue biopsies [[47], [48], [49], [50], [51], [52], [53], [54], [55], [56], [57]]. Some studies compared saliva, BALF, and tumoral samples [58,59] or compared the saliva samples with precancerous or tumoral tissue inside the studio [60,61].
The cross-sectional and case-control studies in saliva, sputum, BALF, and brushing by bronchoscopy compared the microbiome between LC and healthy controls [[14], [15], [16], [17], [18], [19],[22], [23], [24], [25], [26],28,31,32,40,41] or with benign lesions [34,38,39,46]. However, other studies compared the microbiome between different histological types in large LC [33] or contralateral or adjacent samples [37]. In cohort studies, respondents were compared with no respondents who considered recurrence or progression [20,27,30,35,36] or by different prognostic factors [42,45], or compared with an untreated cohort in the ICIs therapy [43]. The samples of tissue were compared with adjacent healthy tissue [47,49,50,54], contralateral tissue [51,59], prognostic factors [48,54,56], or responders and non-responders to ICIs therapy [20,27,35,36,43,62], and also with other samples within the study [58,60,61] (Table 1).

Outcomes
The most relevant results included the proportion of microorganisms identified in each sample by 16S rRNA gene sequencing, and evaluating alpha diversity [[15], [16], [17], [18],[20], [21], [22],[24], [25], [26], [27],29,[31], [32], [33], [34], [35], [36], [37], [38], [39], [40],42,45,46,[48], [49], [50], [51],54,[57], [58], [59], [60], [61]] and beta diversity compared to samples of LC and controls in all studies. Progression and survival were assessed in eleven cohort studies [30,35,36,39,40,43,44,48,53,54,56], and six focused primarily on ICIs therapy [20,27,30,35,36,43,62]. The biological mechanisms were evaluated concomitantly in nine studies [16,19,22,27,38,41,42,47,51].

Quality of evidence, limitations, and possible biases
Zhang et al. (2021) [27] and Cheng et al. (2024) [39] scored 12 points (High). Nine of the articles scored 11 points [20,23,28,31,38,40,54,55,59], while 13 articles scored 10 points (Moderate) [17,18,26,30,32,34,35,41,44,45,49,50,61]. However, the majority had low scores (<10 points) (Low). The lack of uniformity among studies and insufficient data contributed significantly to the risk of bias. The studies' representativeness was also limited due to their small sample sizes (Fig. S1).

Oral microbiome alpha diversity analysis in lung cancer diagnosis
Ten articles [17,20,24,26,[32], [33], [34],49,50,59] nullified changes in alpha diversity in the oral microbiome in LC compared with controls or between histological LC [45] or in responders or non-responders to ICIs therapy [27,35], PD-L1 levels [28,57], or according to the progression of LC based on different diversity indexes [42]. However, changes in alpha diversity were decreased in patients with lung cancer in saliva [15,16,18,21,22], sputum [25], BALF [31,38,39], brushing by bronchoscopy [40], and tumoral tissue [54]. The alpha diversity was higher in LC in a few studies, including brushing by bronchoscopy and LC tissue [46,49,51,58] (Table 1). Fig. 2 displays the forest plots for alpha diversity of the different samples for diagnostic markers, as well as the response to ICIs therapy in BALF samples. The studies with BALF yielded the most information. The online PlotDigitizer extracted data from the box plot graph. None of the samples showed a significant difference in alpha diversity between cancer cases and controls, as measured by the alpha diversity index, evaluated by the Shannon index (Fig. 2A). The certainty of the evidence by GRADE was very low (Table S2).

Comparison of the distribution of the phylum of microorganisms of the oral microbiome and lung cancer

Firmicutes
Overgrowth of the Firmicutes phylum in saliva in LC compared with control individuals or lung tissues was detected in 9/14 (64.3%) articles [14,16,18,21,[58], [59], [60], [61]]. The genera Streptococcus [14,16,21,[58], [59], [60], [61]] and Veillonella [14,16,19,[58], [59], [60], [61]] were the most abundant. Likewise, Streptococcus was the most common Firmicutes detected in sputum in 4/8 samples (50%). Interestingly, Granulicatella adiacens [8,24,29], and Gemella [28,31,59] were identified in three studies (37.5%). In BALF, Veillonella [31,33,35,38] was the most abundant in 4/10 (40%), followed by Streptococcus [35,39]. Likewise, Streptococcus [[40], [41], [42],[44], [45], [46]] was identified in samples from brushing by bronchoscopy except one [46], followed by Veillonella [41,42,45]. However, the Firmicutes genus in tumoral tissue, such as Streptococcus [53,[58], [59], [60], [61]] and Veillonella [[58], [59], [60], [61]], was less frequent than the other samples evaluated (Fig. 3).

Bacteroidetes
A few oral microorganisms of this phylum showed a positive association with LC in saliva samples and BALF. Capnocytophaga was associated with LC with 84.6% sensitivity and 86.7% specificity in distinguishing patients with SCC [14]. Prevotella was also associated with LC [[15], [16], [17], [18], [19]], and Treponema and Filifactor were identified as potential biomarkers in BALF samples in LC [31]. However, a higher abundance of Spirochaetes and Bacteroidetes in saliva, sputum, and BALF were associated with a reduced risk of lung cancer [18,26,46]. An overgrowth of Bacteroidetes, such as Prevotella [61], Prevotella oralis [61], and Porphyromonas [59], in LC tissues, was associated with a poor prognosis (Table 1).
Fusobacterium:
F. nucleatum has been strongly linked to carcinogenesis. In LC, overgrowth of Fusobacterium was associated with LC in several studies [22,23,37,55] and was related to the poor response to the ICIs therapy [62].

Oral microbiome in ICI therapy
A poor response to ICI therapy resulted in a reduction in alpha diversity as measured by the Shannon index in BALF samples (Fig. 2B). However, the certainty was very low (Table S2). For microbiome sequencing, it was not possible to perform a meta-analysis using data from different studies, and data were only available from one published study [30]. Oral and non-oral microorganisms associated with patients responding to and non-responding to ICIs therapy, as reported in different studies, are summarized in Table 2. In responders, Veillonella dispar and Actinomyces species were the oral microorganisms associated with a good response to ICIs therapy [35]; other microorganisms of respiratory or oral origin present in saliva, such as streptococci, Neisseria, and Lactobacillus fermentum, are also associated with responders [62]. In non-responders, Fusobacterium nucleatum [36] and Streptococcus oralis were the oral microorganisms associated with a poor response to ICIs therapy [62] (Table S2).

Mechanisms found in the literature on the relationship between the oral microbiome and LC
Several mechanisms using cell culture, animal experiments in vivo, and in different samples from saliva, sputum, BALF, and tumoral tissue using transcriptomic, immunoassays, and metabolomic technology (Table 3).

Immune modulation
Many commensals and pathogens of the oral microbiome modulate immune responses within the tumor microenvironment, promoting inflammation or immune escape, which can either promote or inhibit tumorigenesis. Veillonella activates inflammasome pathways, including NLRP1, IL-1β, IL-18, and CASP1, inducing a pro-inflammatory environment and tumor progression [41]. It also promotes the expression of PD-L1 on T cells, which predicts responsiveness to anti-PD-1/PD-L1 immunotherapy [22]. In turn, Streptococcus mitis stimulates Th17 cells and increases PD-L1 expression on CD8+ T lymphocytes [27]. While this promotion of inflammation suggests enhanced immunotherapy potential, Streptococcus has also been associated with poor responses to immune checkpoint inhibitors (ICIs) [62]. Granulicatella adiacens causes pro-inflammatory TME and resistance to ICIs, resulting in shorter progression-free survival, while Neisseria and Actinomyces are enriched in immunotherapy responders and are associated with lipid metabolites that correlate with PD-L1 expression and better treatment outcomes [62]. F. nucleatum modulates TME by suppressing immune response, primarily through T-cell repression, cytokine dysregulation (IL-6, IL-8), immune evasion mechanisms, and cell redox homeostasis in LC, correlating with resistance to PD-L1-targeted therapies [63]. P. gingivalis promotes immunosuppressive shifts, including increased Treg cells, imbalance in Th17 responses, and polarization of macrophages toward the M2 phenotype, which collectively enhances tumor growth and inhibit anti-tumor immunity [65] (Table 3, Fig. 4).

Epithelial signaling
Oral microorganisms have been associated with the dysregulation of key epithelial cell signaling pathways in lung cancer. Veillonella initiates signaling cascades, including those involving IL-17, PI3K, ERK, p53, EGFR, and VEGF, via NF-κB activation and the NOD2/CCN4 axis, contributing to epithelial cell proliferation and resistance to apoptosis [15,42]. Oral Streptococcus and Prevotella increase ERK and PI3K signaling, which boosts tumor growth and cell survival in LC [62,63]. P. gingivalis modulates the p53 pathway and apoptotic responses in epithelial cells, similar to Streptococcus intermedius [63,64]. Likewise, F. nucleatum invades epithelial cells and activates oncogenic signaling pathways. It interacts with E-cadherin and activates β-catenin signaling, both of which are involved in the process of oncogenesis [66] (Table 3, Fig. 4).

Metabolic reprogramming
Oral microorganisms in saliva may influence metabolic reprogramming by tumor cells. They have been associated with enhanced xenobiotic biodegradation, altered amino acid utilization [16], and changes in sugar and iron metabolism. These microbial-driven metabolic alterations may support tumor progression, immune evasion, and therapeutic resistance in lung cancer. Bacterial genes from oral commensals were linked to glycosyltransferase activity, peptidases, amino sugar and nucleotide sugar metabolism, as well as starch and sucrose metabolism in LAC [19]. Additionally, in sputum samples, shifts in the microbiome have been associated with increased expression of genes involved in polyamine metabolism and iron siderophore receptor pathways, both of which are known to support tumor growth and survival [24].

Discussion

Summary of the main results
Oral microorganisms such as Streptococcus and Veillonella exhibit consistent overgrowth in the saliva of patients with LC, regardless of histological type, similar to what is observed in BALF. Sputum shows a more specific microbiota with microorganisms such as Granuicatella adiacens and Gemella. In tumoral tissue, the presence of oral microorganisms is lower than that observed in bronchoscope brushing samples. Oral microorganisms found in peritumoral tissues may not frequently infiltrate tumor tissues. The presence of species from the Bacteroidetes phylum (excluding Prevotella), Spirochetes, and Synergistetes in saliva was associated with a reduced risk of LC. However, in cancer tissue, these microorganisms have been linked to a poor prognosis. In ICI therapy, a reduction in alpha diversity was observed, and Veillonella dispar was associated with responding; Veillonella elevates PD-1 expression in T cells and enhances the immunotherapy response. Oral microbial communities in saliva were related to the biodegradation of xenobiotics, as well as the metabolism of amino acids, sugars, starch, sucrose, and lipids. These communities also inhibited immune responses and disrupted cellular redox homeostasis. Additionally, they regulated pathways such as IL-17, PI3K, MAPK, and ERK in the airway transcriptome. Some microorganisms induced significantly higher infiltration of CD8 T cells and CD4 T cells, potentially enhancing the efficacy of ICIs in LC.

Agreements and disagreements with previous studies
The overall results of this systematic review suggest that microorganisms are commonly associated with lung cancer, highlighting the relevance of the oral microbiome. Most studies agree on the overgrowth of the Firmicutes phylum, especially Streptococcus and Veillonella [8,14,16,18,21,24,29,31,[33], [34], [35],38,[40], [41], [42],[44], [45], [46],[58], [59], [60], [61]] and diminished association with the Bacteroidetes phylum [16,18,21,26,46,[59], [60], [61]]. Skallevold et al. (2021) published a narrative review providing an overview of salivary biomarkers for the diagnosis and prognosis of lung cancer. Veillonella, Streptococcus, and Capnocytophaga demonstrated sensitivity ranges from 78% to 88% [15,40]. However, Capnocytophaga was not frequently observed in other studies that can validate its diagnostic value.
The genus Streptococcus is the most abundant in the respiratory tract and oral cavity and plays a crucial role in shaping the oral microbiome through its presence in saliva. It is possible that these microorganisms can quickly enter the lungs by microaspiration [6]. However, these microorganisms can overregulate ERK and PI3K signaling pathways in airway epithelial cells, similar to Prevotella and Veillonella [41]. Likewise, Streptococcus mitis can induce inflammation, activate Th17 cells, and promote PD-L1 expression, contributing to tumorigenesis [27]. The findings of this systematic review support that oral microbiome dysbiosis by the overgrowth of these commensals could be a biomarker in lung cancer.
Veillonella is relevant in the oral microbiome, especially for its ability to reduce nitrates to nitrites, directly affecting nitric oxide production and favoring cardiovascular health. In the review by Huang and Huang [11], Veillonella emerged as a primary biomarker in lung cancer patients, consistent with the findings of this study. Transcriptomic analyses of the lower respiratory tract in lung cancer patients reveal a significant association between Veillonella enrichment and oncogenic pathways [14,16,19,31,33,35,38]. Recently, a study has explored the role of Veillonella parvula in promoting lung adenocarcinoma proliferation via the nucleotide oligomerization domain 2 (Nod2)/cellular communication network factor 4 (CCN4)/nuclear factor kappa B (NF-κB) signaling pathway [38]. Using 16S rRNA sequencing, V. parvula was significantly enriched in lung adenocarcinoma samples. In murine models, V. parvula suppressed the infiltration of tumor-associated and peripheral T cells, thereby facilitating tumor growth [38]. Additionally, V. parvula activated the NF-κB pathway through the Nod2/CCN4 signaling axis [38]. The overgrowth of Veillonella has been linked to increased recruitment of Th17 cells and neutrophils, enhanced IL-17 production, and elevated PD-1 expression in T cells [42]. These immune responses may explain the positive effects of Veillonella on ICIs therapy [35]. To date, carcinogenic mechanisms have been described in the literature, including suppressing the P53 gene, which directly impacts lung tissues [15].
P. gingivalis has been linked to progressive lesions in LC when colonizing lung cancer cells through micro-aspirations or the bloodstream and promoting tumor proliferation [11]. It enhances epithelial cell proliferation and tumor growth by dysregulating PI3K/Akt signaling and suppressing p53 function [67]. Additionally, lipopolysaccharide (LPS) from P. gingivalis is recognized by TLR4, activating the MyD88 pathway, which triggers NF-κB and induces proinflammatory cytokine production, facilitating cancer progression [69]. Furthermore, P. gingivalis induces epithelial-mesenchymal transition (EMT) in normal epithelial cells via Snail and Slug transcription factors, activating β-catenin–LEF-1 complexes, which promote EMT, epithelial proliferation, and migration [68] (Fig. 3).
Periodontopathic microorganisms are infrequently identified in saliva in LC, possibly because they thrive in anaerobic environments like the subgingival microbiome, which is less frequently studied in LC. P. gingivalis is frequently observed in bacteremia patients with periodontitis from the subgingival microbiome inside periodontal pockets, reaching different tissues at a distance [69]. This could be another way to access lung tumor tissue. In individuals with LC, the smoking habit is associated with a history of periodontitis and increasing tooth loss. Smoking is a known predictor of microbiome imbalances in clinical medicine [70]. It is possible that at the time of LC diagnosis, patients may have already experienced significant tooth loss due to smoking, resulting in a reduced microbial load in saliva. We hypothesize that salivary biofilm dysbiosis in LC is caused by an overgrowth of commensal microbiota, which displaces pathogenic bacteria associated with periodontitis in the saliva. However, future studies should better evaluate these mechanisms.
Oral microorganism communities in the saliva have also been linked to the biodegradation of xenobiotics [16]. Certain bacteria in the phylum Bacteroidetes produce proteases that can degrade xenobiotic compounds, such as polycyclic aromatic hydrocarbons (PAHs) found in environmental pollutants, or be reduced by these microorganisms. While direct evidence is limited, some studies support the hypothesis that these bacteria (mainly from the class Bacteroidia and order Bacteroidales) may contribute to a decreased risk of LC due to their xenobiotic degradation capacity [16].
Geographic and demographic characteristics of the studies showed similarities in salivary results between samples from China [15,16] and the USA [17,18,21], but differed from other Chinese populations [22,23], which was similar to the Indian population [19], where a significant number of LC-associated taxa were present in saliva. Various factors, including geographic, periodontal status, and tobacco use, influence the composition of the oral microbiome in saliva. In India [19], patients with poor oral conditions were excluded, and this overgrowth of microorganisms might not be due to social status. However, smoking can also alter the microbiome composition in saliva and the gut microbiome [71]. In India, the prevalence of tobacco use is still high and is among the five countries with the highest tobacco production [72]. Otherwise, in China, several factors may be influencing these differences; higher education in the population is associated with less severe lung cancer than in more socioeconomically deprived areas [73]. In China, changes in the gut microbiome have been observed between rural and urban populations, as well as during the transition from a traditional rural lifestyle to a Westernized one, influencing the oral and gut microbiome composition and associated with an increased risk of dietary cancer [74].
Although saliva studies were the most homogeneous in design, studies using sputum, BALF, bronchoscopy, and tumor tissue showed many differences. Since these were more invasive techniques, establishing the control group was challenging, which made their comparison difficult. Only 15 studies reported their data in public repositories, and some of these had restricted access. Nine databases could be reviewed; however, it was not possible to unify them for integrative analysis. Poor sample identification and difficulty in identifying control groups were factors that contributed to the challenges in performing the meta-analysis.
We obtained data from a significant number of studies to perform meta-analyses for the different diversity indices. The Shannon index was the most frequently evaluated in studies comparing cancer samples and controls. Although there was a trend toward reduced alpha diversity in individuals with cancer, none of the meta-analyses were able to demonstrate this. However, a reduction in alpha diversity was found in patients who did not respond to ICI therapy. These findings are relevant because they illustrate an overgrowth of some bacteria in non-responding patients. Immune checkpoint inhibitors (ICIs) in LC work by blocking the action of proteins produced by T cells to promote tumor cell destruction, such as programmed cell death protein 1 (PD-1) and programmed cell death protein ligand 1 (PD-L1). However, not all patients respond adequately to this therapy, which requires tumors with high levels of PD-L1 for a monotherapy indication [75]. Veillonella and its PD-L1 activation support this microorganism as a possible biomarker of response to ICIs therapy in LC [22]. Other microorganisms, such as F. nucleatum, are related to poor immunotherapy responses, although the underlying mechanisms remain unclear [63]. A recent study found that Neisseria and Actinomyces species have a positive relationship with lipids associated with PD-L1 expression in responders to ICI therapy; however, others, such as G. adiacens and S. oralis, were associated with low PD-L1 expression and abnormal fat metabolism. This suggests that the specific oral microbiota can affect the normal regulation of lipid metabolism signals and reduce the efficacy of lung cancer immunotherapy [62]. This mechanism should be further studied in future research on the oral microbiome and its impact on the response to immunotherapy. Recently, extensive research supported several promising therapies for modulating oral biofilm that could be used in immunotherapy to reduce or stimulate specific phylotypes using antimicrobial peptides (AMPs) with targets in different cellular components and functions of bacteria, including Veillonella, Porphyromonas, and other anaerobic microorganisms [76].
The oral-gut-lung axis represents a communication network connecting oral, gastric, intestinal, and pulmonary microbiota. The oral microbiome is translocated to the intestine through the systemic circulation, altering intestinal homeostasis and promoting an immune response [77]. Oral dysbiosis triggers an inflammatory stage in the intestinal compartment, and oral microorganisms, cytokines, and other metabolites are disseminated through the gut-lung axis, thereby exacerbating the LC [78]. Likewise, oral commensals such as S. sanguinis and S. vestibularis have been reported in the gut microbiome in fecal samples associated with shorter PFS in patients with LC [79]. Hang et al. [80] evaluated several databases and identified Enterococcus, Lactobacillus, Escherichia, and Streptococcus as the main markers of LC obtained in samples from BALF and gut microbiome. However, they observed a relevant bacterial interaction network associated with tumor lesions, which included microorganisms from the oral-gut-lung axis, such as Actinomycetes and Veillonella, that contribute to the pathogenesis and progression of LC. Due to the importance of this association, the oral-gut-lung axis should be further investigated to establish the interaction between the oral-gut-lung axis in LC in the future.

Conclusion
The oral-lung axis represents a pathway of oncogenic activity in the lung. Commensal microorganisms such as Streptococcus and Veillonella are consistently implicated, showing overgrowth in saliva, BALF, and lung tissue. The mechanisms by which oral microorganisms contribute to lung cancer development primarily involve activating inflammation and cell proliferation. Despite these insights, a deeper understanding of how the oral microbiome influences the response to ICI therapy is needed. Monitoring changes in microbial communities before and after therapies (e.g., immunotherapy) is also recommended, as limited evidence currently addresses these interactions and relationships.

Fundings
This research received funding from the internal call for research at El Bosque University. PCI-2022–11,049.

Available data
All additional results and analyses are available in the supplementary material. You can obtain other required information from the corresponding author.

Abbreviatures
LC=Lung cancer; LAC=Lung adenocarcinoma; SCLC =Small Cell Lung; NSCLC=not small cells lung carcinoma; LSCC= Lung squamous cell carcinoma; CLC: Central lung cancer; SSN=Semi-solid nodules; SN= Solid nodules; 16 s rRNA=16S ribosomal RNA; ROC= Receiver operating characteristic; qPCR =A real-time polymerase chain reaction; ICs= Immune Checkpoint Inhibitor Therapy; COPD=chronic obstructive pulmonary disease; NA= No available; OUT=Operational Taxonomic Unit; ASV = Amplicon sequence variant; PCoA= Principal Coordinates Analysis; BALF=Bronchoalveolar lavage fluid.

CRediT authorship contribution statement
Adrianna Michalina Kwiatkowska: Writing – review & editing, Writing – original draft, Methodology, Investigation, Conceptualization. Jaime Andrés Guzmán: Writing – review & editing, Writing – original draft, Investigation, Formal analysis, Data curation, Conceptualization. Gloria Inés Lafaurie: Writing – review & editing, Writing – original draft, Validation, Supervision, Resources, Project administration, Methodology, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization. Diana Marcela Castillo: Writing – review & editing, Writing – original draft, Project administration, Funding acquisition, Formal analysis. Andrés F. Cardona: Writing – review & editing, Writing – original draft, Validation, Supervision, Resources, Project administration, Investigation, Funding acquisition, Formal analysis, Conceptualization.

Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.