Method for early diagnosis and risk prediction of lung cancer using oral microbial information, or composition therefor

Abstract

The present invention relates to a composition for diagnosing or predicting the risk of developing lung cancer, comprising a preparation capable of detecting the oral microbiome; a kit comprising the same; and a method for providing information for diagnosing, predicting the risk of developing, or predicting the prognosis of lung cancer. By analyzing the oral microbiome derived from saliva, it is possible to rapidly and accurately diagnose early lung cancer, and predict the risk of developing, or prognosis, and the composition can be usefully used for the prevention or treatment of lung cancer.

Description

[Technical field]

[0001] The present invention relates to providing information for early diagnosis and prediction of risk or prognosis of lung cancer using oral microbial information, and a composition therefor. [Background technology]

[0002] Lung cancer is one of the most common cancers, with approximately 22.07 million patients worldwide in 2020, and is the most common cause of cancer death, with 17.96 million deaths from lung cancer worldwide in 2020 (Non-Patent Document 1).

[0003] Lung cancer is classified into small cell lung cancer (SCLC) and non-small cell lung cancer (NSCLC) based on histological type, accounting for 10-15% and 80-85% of all lung cancer cases, respectively (Non-Patent Document 2). NSCLC is subdivided into three subtypes: adenocarcinoma, squamous cell carcinoma, and large cell carcinoma. Of these, lung adenocarcinoma is the most commonly observed subtype of lung cancer, accounting for approximately 40% of all lung cancer cases (Non-Patent Document 2; and Non-Patent Document 3).

[0004] Smoking is the main risk factor for lung cancer (Non-Patent Document 4), but 15-20% of male lung cancer patients and more than 50% of female lung cancer patients are non-smokers, and it can be assumed that risk factors other than smoking also play an important role in the development of lung cancer (Non-Patent Document 5). In particular, in Asia, the proportion of non-smokers among female lung cancer patients has been reported to be as high as 60-80% (Non-Patent Document 6). The most common subtype of lung cancer in non-smokers is lung adenocarcinoma (Non-Patent Document 2). Risk factors that increase the development of lung cancer in non-smokers include exposure to radon, air pollution, occupational hazards, secondhand smoke, and infection or inflammation (Non-Patent Document 7). However, such risk factors cannot fully explain the development of lung cancer in non-smokers.

[0005] Recently, changes in the microbiome have been recognized as one of the risk factors for lung cancer (Non-Patent Document 8). Previous studies have shown that changes in the oral microbiome are associated with lung cancer. Yan et al. reported that the abundance of Neisseria decreased and the abundance of Capnocytophaga, Selenomonas, and Veillonella increased in the saliva of lung cancer patients (Non-Patent Document 9). Another study reported that the abundance of Sphingomonas and Blastomonas microorganisms was significantly increased, and the abundance of Acinetobacter and Streptococcus microorganisms was reduced in the saliva of female non-smoking lung cancer patients (Non-Patent Document 10). It has also been reported that a decrease in oral microbiome diversity and an increase in Lactobacillales abundance in the oral microbiome are associated with an increased risk of lung cancer (Non-Patent Document 11).

[0006] Previous studies have shown that the types of lung cancer-associated microbes are not consistent across all studies. Such discrepancies are likely due to limited sample numbers and differences in patient characteristics. Despite these limitations, the study showed that the microbiomes of lung cancer patients are significantly different from those of healthy controls.

[0007] Recently, the possibility has been presented that microbes themselves or substances capable of regulating microbes can be used as new anticancer therapeutic agents (Non-Patent Document 12). In addition, side effects of anticancer chemotherapy such as mucositis, diarrhea, and constipation are thought to be associated with changes in the human microbiome, and methods for alleviating the side effects of anticancer therapy through microbiome regulation are also being actively explored. In particular, it has been reported that changes in the oral microbiome are associated with chemotherapy-induced oral mucositis in patients with malignant tumors (Non-Patent Document 13), and that the effects of anticancer chemotherapy on lung cancer are associated with an increase or decrease in specific intestinal microbiome species (Non-Patent Document 14).

[0008] This study indicates that the human microbiome has various applications, such as monitoring the progression of cancer, developing anti-cancer therapeutics, and minimizing the side effects of anti-cancer therapy. Therefore, the inventors compared the oral microbiomes of non-smokers with those of healthy individuals through massive parallel sequencing to investigate the microorganisms associated with lung adenocarcinoma and to confirm whether lung cancer patients can be distinguished from healthy individuals based on microbiome information. In addition, the differences in the oral microbiome between treated and untreated patients were investigated to confirm whether the survival rate of lung cancer patients can be predicted using microbiome data. [Prior art documents] [Non-patent literature]

[0009] [Non-Patent Document 1] International Agency for Research on Cancer.Global Cancer Observatory:Cancer Today.Fact sheets on cancer from IARC

Non-licensed Document 2

Non-licensed Document 4

Non-licensed Document 5

Non-licensed Document 6

Non-licensed Document 7

Non-licensed literature 9

Non-licensed literature 10

Non-licensed Document 11

Non-licensed Document 12

Non-licensed Document 13

Non-licensed Document 14

[0010] The object of the present invention is to provide a method for providing information for diagnosing, predicting the risk of onset, or predicting the prognosis of lung cancer, comprising: (a) sequencing the 16S rRNA gene of DNA extracted from saliva; and (b) comparing the abundance of the oral microbiome through the sequencing with a reference value.

[0011] Another object of the present invention is to provide a composition for diagnosing or predicting the risk of developing lung cancer, comprising a preparation capable of detecting the oral microbiome.

[0012] It is still another object of the present invention to provide a kit for diagnosing or predicting the risk of developing lung cancer, comprising the composition. [Means for solving the problem]

[0013] In order to achieve the above object, the present invention provides a method for detecting 16S DNA extracted from saliva. The present invention provides a method for providing information for diagnosing, predicting the risk of developing, or predicting the prognosis of lung cancer, comprising: (a) performing sequencing of rRNA genes; and (b) comparing the abundance of the oral microbiome with a reference value through the sequencing.

[0014] In one embodiment of the present invention, the saliva may be collected from a non-smoker.

[0015] In one embodiment of the present invention, the sequence analysis may be performed by next generation sequencing (NGS), but is not limited thereto.

[0016] In one embodiment of the present invention, when the abundance of one or more oral microbiome species selected from the group consisting of Mogibacterium, Variovorax, Ralstonia, and Bulleidia is increased in the oral microbiome, it can be determined that the patient is not suffering from lung cancer or that the risk of developing lung cancer is low.

[0017] In one embodiment of the present invention, an increase in the abundance of one or more oral microbiome species selected from the group consisting of Veillonella, Actinomyces, Slackia, and Schwartzia in the oral microbiome can be determined to be indicative of lung cancer or a high risk of developing lung cancer.

[0018] In one embodiment of the present invention, an increase in the abundance of one or more oral microbiome species selected from the group consisting of Streptococcus and Megasphaera in the oral microbiome can be determined to be associated with a poor prognosis for lung cancer.

[0019] In one embodiment of the present invention, if the oral microbiome abundance is compared with a reference value and has a high similarity, the prognosis for lung cancer can be determined to be good.

[0020] In one embodiment of the present invention, the lung cancer is non-small cell lung cancer, small cell lung cancer, The cancer may be selected from the group consisting of, but is not limited to, adenocarcinoma, squamous cell carcinoma, and large cell carcinoma.

[0021] The present invention also provides a composition for diagnosing or predicting the risk of developing lung cancer, comprising a preparation capable of detecting the oral microbiome.

[0022] In one embodiment of the invention, the formulation may be selected from the group consisting of, but is not limited to, oral microbiome specific primers, probes, antisense oligonucleotides, aptamers and antibodies.

[0023] In one embodiment of the invention, the oral microbiome may be selected from the group consisting of Mogibacterium, Variovorax, Ralstonia, Bulleidia, Veillonella, Actinomyces, Slackia, and Schwartzia.

[0024] In one embodiment of the present invention, the lung cancer is non-small cell lung cancer, small cell lung cancer, The cancer may be selected from the group consisting of, but is not limited to, adenocarcinoma, squamous cell carcinoma, and large cell carcinoma.

[0025] The present invention also provides a kit for diagnosing or predicting the risk of developing lung cancer, comprising the composition.

[0026] In one embodiment of the present invention, the kit may be selected from the group consisting of, but is not limited to, an RT-PCR kit, a DNA chip kit, an ELISA kit and a rapid kit. Effect of the Invention

[0027] The composition or method according to the present invention enables rapid and accurate early diagnosis of lung cancer and prediction of the risk of onset or prognosis by confirming the abundance of the oral microbiome from a sample isolated from saliva, and can be usefully used for the prevention or treatment of lung cancer. [Brief description of the drawings]

[0028] [Figure 1A]Figure 1A shows the results showing the oral microbiome diversity between healthy individuals (control group) and lung cancer patients. The left panel shows the Shannon diversity results, and the right panel shows the results of PCoA (principal coordinate analysis) plot, comparing the control group and lung cancer patients (*p<0.05, **p<0.01). [Figure 1B] Figure 1B shows the results showing the oral microbiome diversity between healthy individuals (control group) and lung cancer patients. The left panel shows the Shannon diversity results, and the right panel shows the results of PCoA (principal coordinate analysis) plot, comparing the control group, non-treatment group (Treatment-), and treatment group (Treatment+) (*p<0.05, **p<0.01). [Figure 2A] Figure 2A shows the results of identifying differentially abundant genera in the oral microbiome between healthy individuals (control group) and lung cancer patients. The left panel shows the log10 transformed relative abundance results, and the right panel shows the generalized fold change results of differentially abundant genera between all lung cancer patients and controls (q<0.1). [Figure 2B] Figure 2B shows the results of identifying differentially abundant genera in the oral microbiome between healthy individuals (control group) and lung cancer patients. The left panel shows the Log10 transformed relative abundance results, and the right panel shows the generalized fold change results of differentially abundant genera between the non-treated and control groups (q<0.1). [Figure 2C] Figure 2C shows the results of identifying differentially abundant genera in the oral microbiome between healthy individuals (control group) and lung cancer patients. The left panel shows the Log10 transformed relative abundance results, and the right panel shows the generalized fold change results of differentially abundant genera between the treatment and control groups (q<0.1). [Figure 3A]Figure 3A shows the results of lung cancer prediction using the oral microbiome, and shows the receiver operating characteristic (ROC) curve of the predictive model distinguishing between lung cancer patients, treatment groups, non-treatment groups, and control groups. [Figure 3B] Figure 3B shows the results of lung cancer prediction using the oral microbiome, showing the normalized coefficients of the prediction model (Bayesian Ridge) (genera with coefficients in the top and bottom 5%). [Figure 4A] Figure 4A shows the results of confirming the prognostic impact of oral microbiome profiles in lung cancer patients, showing a PCoA plot of the oral microbiome (lung cancer patients were classified into two groups based on Bray-Curtis distances). [Figure 4B] Figure 4B shows the results of confirming the prognostic impact of oral microbiome profiles in lung cancer patients, showing a density plot of oral microbiome communities based on distance to the most similar control group. [Figure 4C] Figure 4C shows the results of confirming the prognostic impact of oral microbiome profiles in lung cancer patients, showing survival curves for two populations of oral microbiome profiles. [Figure 4D] Figure 4D shows the results of confirming the prognostic impact of oral microbiome profiles in lung cancer patients, showing the receiver operating characteristic (ROC) curve of the predictive model distinguishing between Cluster 1 and Cluster 2. [Figure 4E] Figure 4E shows the results of confirming the prognostic impact of oral microbiome profiles in lung cancer patients, showing the normalized coefficients of the predictive model (Bayesian Ridge). DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0029] The present invention will be described in detail below.

[0030] The present invention provides a method for providing information for diagnosing, predicting the risk of onset, or predicting the prognosis of lung cancer, comprising: (a) sequencing the 16S rRNA gene of DNA extracted from saliva; and (b) comparing the abundance of the oral microbiome through the sequencing with a reference value.

[0031] The term "oral microbiome" refers to the microbial community including all of the microorganisms and their genetic information that inhabit or coexist in the oral cavity.

[0032] The term "diagnosis" refers to determining the presence or characteristics of a pathological condition. For purposes of the present invention, it refers to determining the presence or absence of lung cancer, and also includes determining the progression and progression of lung cancer.

[0033] The term "prognosis" means a forecast of future symptoms or progress determined by diagnosing a disease. In cancer patients, prognosis usually means the presence or absence of metastasis or survival time within a certain period after the onset of cancer or surgery. Prediction of prognosis is a very important clinical issue since it provides clues regarding the future direction of cancer treatment, including the presence or absence of chemotherapy.

[0034] In the present invention, the saliva may be collected from a non-smoker.

[0035] In the present invention, the sequence analysis may be performed by next generation sequencing (NGS), but is not limited thereto.

[0036] In the present invention, the oral microbiome may be one or more species selected from the group consisting of Mogibacterium, Variovorax, Ralstonia, Bulleidia, Veillonella, Actinomyces, Slackia, and Schwartzia.

[0037] In addition, when the abundance of one or more oral microbiome species selected from the group consisting of Mogibacterium, Variovorax, Ralstonia, and Bulleidia increases in the oral microbiome, it can be determined that the patient does not have lung cancer or has a low risk of developing lung cancer. Preferably, when the abundance of two or more oral microbiome species selected from the group consisting of Mogibacterium, Variovorax, Ralstonia, and Bulleidia increases in the oral microbiome, it can be determined that the patient does not have lung cancer or has a low risk of developing lung cancer.

[0038] In addition, when the abundance of one or more oral microbiome species selected from the group consisting of Veillonella, Actinomyces, Slackia, and Schwartzia increases in the oral microbiome, it can be determined that the patient has lung cancer or has a high risk of developing lung cancer. Preferably, when the abundance of two or more oral microbiome species selected from the group consisting of Veillonella, Actinomyces, Slackia, and Schwartzia increases in the oral microbiome, it can be determined that the patient has lung cancer or has a high risk of developing lung cancer.

[0039] In the present invention, the oral microbiome may be selected from the group consisting of Streptococcus and Megasphaera, and an increase in the abundance of the oral microbiome selected from the group consisting of Streptococcus and Megasphaera is judged to be a poor prognosis for lung cancer. An increase in the abundance of the genus Haemophilus can be judged as a good prognosis for lung cancer.

[0040] In the present invention, when the oral microbiome abundance is highly similar to a reference value, the prognosis for lung cancer can be judged to be good, and conversely, when the oral microbiome abundance is not highly similar to a reference value, the prognosis for lung cancer can be judged to be poor.

[0041] In the present invention, the lung cancer may be selected from the group consisting of non-small cell lung cancer, small cell lung cancer, adenocarcinoma, squamous cell carcinoma, and large cell carcinoma.

[0042] The present invention also provides a composition for diagnosing or predicting the risk of developing lung cancer, comprising a preparation capable of detecting the oral microbiome.

[0043] In the present invention, the detectable preparation may be selected from the group consisting of primers, probes, antisense oligonucleotides, aptamers and antibodies that specifically bind to the oral microbiome, and preferably the 16S of the oral microbiome. It may be a primer, a probe or an antisense oligonucleotide specific to rRNA.

[0044] The term "primer" refers to a short base sequence having a short free 3' hydroxyl group, which can form a base pair with a complementary template and act as a starting point for template strand copying. A primer can initiate DNA synthesis in the presence of a polymerization reaction reagent (i.e., DNA polymerase or reverse transcriptase) and four different nucleoside triphosphates at an appropriate buffer solution and temperature. PCR conditions and the lengths of sense and antisense primers can be appropriately selected by techniques generally known in the field to which the present invention belongs.

[0045] The term "probe" refers to a nucleic acid fragment such as RNA or DNA that can specifically bind to a gene, and is labeled to confirm the presence or absence and expression level of a specific gene. The probe can be prepared in the form of an oligonucleotide probe, a single strand DNA probe, a double strand DNA probe, an RNA probe, etc. Selection of an appropriate probe and hybridization conditions can be appropriately selected according to techniques generally known in the field to which the present invention belongs.

[0046] The term "antisense oligonucleotide" refers to DNA or RNA or derivatives thereof that contain a nucleic acid sequence complementary to the sequence of a specific mRNA, and acts to bind to a complementary sequence in the mRNA and inhibit the translation of the mRNA into a protein. The antisense oligonucleotide sequence refers to a DNA or RNA sequence that is complementary to the mRNA of the gene and can bind to the mRNA.

[0047] The term "aptamer" refers to a type of polynucleotide consisting of a special type of single-stranded nucleic acid (DNA, RNA or modified nucleic acid) that has a stable tertiary structure by itself and has the characteristic of being able to bind to a target molecule with high affinity and specificity. Aptamers are similar to antibodies. Aptamers can be used instead of antibodies because they can specifically bind to antigenic substances in the same way, are more stable than proteins, have a simple structure, and are composed of polynucleotides that are easy to synthesize. Aptamers can be easily produced using techniques that are widely known in the field to which the present invention belongs, and commercially produced antibodies can be used.

[0048] The term "antibody" refers to a specific immunoglobulin directed against an antigenic site and includes polyclonal antibodies, monoclonal antibodies, recombinant antibodies, and any combination thereof. The term also includes polyclonal antibodies, monoclonal antibodies, recombinant antibodies, and complete forms having two full-length light chains and two full-length heavy chains, as well as functional fragments of antibody molecules, such as Fab, F(ab'), F(ab')2, and Fv. Antibodies can be easily produced using techniques widely known in the field to which the present invention belongs, and commercially produced antibodies can be used.

[0049] In the present invention, the lung cancer may be selected from the group consisting of non-small cell lung cancer, small cell lung cancer, adenocarcinoma, squamous cell carcinoma, and large cell carcinoma, but is not limited thereto.

[0050] In the present invention, the oral microbiome may be selected from the group consisting of Mogibacterium, Variovorax, Ralstonia, Bulleidia, Veillonella, Actinomyces, Slackia, and Schwartzia. Preferably, the oral microbiome may be two or more species selected from the group consisting of Mogibacterium, Variovorax, Ralstonia, Bulleidia, Veillonella, Actinomyces, Slackia, and Schwartzia, and more preferably, the oral microbiome may be three or more species selected from the group consisting of Mogibacterium, Variovorax, Ralstonia, Bulleidia, Veillonella, Actinomyces, Slackia, and Schwartzia.

[0051] The present invention also provides a kit for diagnosing or predicting the risk of developing lung cancer, comprising the composition.

[0052] The kit may include detection agents such as primers, probes, antisense oligonucleotides, aptamers, or antibodies for detecting the oral microbiome, as well as one or more other component compositions, solutions, or devices suitable for an analytical method.

[0053] In the present invention, the kit may be, but is not limited to, an RT-PCR kit, a DNA chip kit, an ELISA kit or a rapid kit.

[0054] The RT-PCR kit includes a test tube or other suitable container, primers, reaction buffer, The reaction solution may contain, for example, deoxynucleotides (dNTPs), enzymes such as Taq-polymerase and reverse transcriptase, DNase, RNAse inhibitors, DEPC-water, or sterile water.

[0055] The DNA chip kit may include a substrate to which cDNA or oligonucleotides corresponding to genes or fragments thereof are attached, and reagents, preparations, enzymes, etc. for producing fluorescently labeled probes. The substrate may also include cDNA or oligonucleotides corresponding to control genes or fragments thereof.

[0056] The ELISA kit includes an antibody specific to the protein. The antibody has high specificity and affinity to each marker protein and has little cross-reactivity to other proteins, and may be a monoclonal antibody, a polyclonal antibody, or a recombinant antibody. The ELISA kit may also include an antibody specific to a control protein. In addition, the ELISA kit may include a reagent capable of detecting the bound antibody, such as a labeled secondary antibody, a chromophores, an enzyme (e.g., conjugated with the antibody) and its substrate or other substance capable of binding to the antibody.

[0057] The rapid kit is a kit containing essential elements necessary for performing a rapid assay that can provide analysis results within 5 minutes, and includes an antibody specific to a protein. The antibody has high specificity and affinity for each marker protein and has almost no cross-reactivity with other proteins, and is a monoclonal antibody, a polyclonal antibody, or a recombinant antibody. The rapid kit may also include an antibody specific to a control protein. The rapid kit may also include other substances such as a reagent capable of detecting the bound antibody, for example, a nitrocellulose membrane on which a specific antibody and a secondary antibody are immobilized, a membrane bound to beads with an antibody, an absorbent pad, and a sample pad. EXAMPLES

[0058] The present invention will be described in more detail below with reference to examples. These examples are intended to more specifically explain the present invention, and the scope of the present invention is not limited to these examples.

[0059] Example 1. Experimental Method 1.1. Experimental design and sample collection In the present study, a total of 100 patients with no smoking history who were diagnosed with lung adenocarcinoma were recruited at the National Cancer Center (Goyang, Korea) between April 2018 and August 2019. The histological type of cancer was determined by the 2004 World Health Organization (WHO) classification for lung tumors. The TNM (tumor-node-metastasis) stage was based on the 8th edition of the lung cancer staging system (Non-Patent Document 15). The present study was conducted under the approval of the Institutional Audit Committee of the National Cancer Center of Korea (approval number: NCC2016-0208). All participants provided written informed consent.

[0060] Saliva samples were collected after patients were asked to refrain from eating, drinking, and brushing their teeth for 1 hour prior to sample collection. Patients' saliva was collected by rinsing their mouths with water and then expectorating 2–5 ml of saliva into a 50 ml tube. Saliva samples were stored at -80°C until use.

[0061] The control group was selected from existing Korean microbiome studies, with never-smoking subjects, age- and sex-matched with the patients; subjects with a history of various chronic diseases, clinically significant diseases, major gastrointestinal diseases, or cancer; and subjects with a history of 18-month follow-up period before sample collection. Subjects who had received antibiotics within the past 3 months were not included in the Korea Microbiome Study. The entire process of saliva sampling for 16S rRNA gene sequencing in the control group was the same as that used for patient samples.

[0062] To validate the prediction model, 41 Korean patients with non-small cell lung cancer (NSCLC) were included. An independent dataset was constructed using oral microbiome data from patients with pulmonary lung cancer and 612 healthy Korean subjects, excluding the control group used for training.

[0063] 1.2. DNA extraction, 16S rRNA gene sequencing and sequencing data analysis Total DNA from saliva samples was extracted using the QIAamp DNA Microbiome Kit (Qiagen) according to the manufacturer's instructions. Total DNA extracted from saliva samples was subjected to 16S rRNA V3-V4 library preparation according to the 16S Metagenomic Sequencing Library Preparation Illumina protocol (Part # 15044223 Rev.B, Illumina, San Diego, CA, USA). Sequencing was performed using an Illumina MiSeq instrument (Illumina). Initial data of 16S rRNA gene sequencing data were processed using QIIME2 (2019.10 version) (Non-Patent Document 16). Sequence quality control and table construction were performed with the q2-dada2 plugin (Non-Patent Document 17). Alpha and beta diversity matrices were calculated using the q2-diversity plugin with a uniform sampling depth of 10,000 sequences per sample. Taxonomy was assigned to amplicon sequence variants (ASVs) using a pre-trained naive Bayes classifier on Greengenes 13_8 99% OTUs with the q2-feature-classifier plugin (18; and 19).

[0064] 1.3.Statistical analysis Significant differences between lung adenocarcinoma patients and healthy controls were assessed using Student's t-test or Fisher's exact test. Shannon diversity of the microbiome between groups was compared using the Wilcoxon rank sum test or Kruskal-Wallis test. Associations between beta diversity and lung cancer were assessed using the Bray-Curtis The differences were assessed by permutational multivariate analysis of variance (PERMANOVA) using the Adonis function in the vegan R package for distance. Differential genus abundance between subject groups was confirmed using the Wilcoxon rank sum test, and experiments were performed on genera with a maximum abundance higher than 0.001. P values ​​were adjusted using the Benjamini-Hochberg FDR method, and microbial features with q < 0.1 were reported.

[0065] 1.4. Prediction model for cancer status Cancer status was recorded as 1 if the individual belonged to the non-disease population and 0 if the individual belonged to the cancer patient population. The relative abundance of microorganisms was transformed using a logarithmic function with a downscale of 10 after adding 0.00005. The Bayesian Ridge method in the Scikit-learn package (Version 0.22.1) was used to predict cancer status using the transformed microbial abundance (Non-Patent Document 20).

[0066] A five-fold cross-validation method was used to evaluate the predictive power. The cross-validation method was repeated five times to reduce the deviation exhibited by splitting the dataset into training and test sets. The false positive rate and positive rate of each test set were used to measure the receiver operating characteristic (ROC) curve and average area under curve (AUC). To evaluate the contribution of each microorganism to each predictive model, the coefficients of the Bayesian Ridge model were extracted and normalized by dividing by the maximum absolute coefficient of the model.

[0067] For the measurement of predictive power in the independent dataset, the Bayesian Ridge model was saved to a file. The models were used to predict cancer status in the independent dataset, and the ROC AUC was measured for each predictive model to measure the performance of the model. The relative abundance data of the independent dataset was used as input for the models after matching the microorganisms to those used to train the models.

[0068] 1.5. Prognostic Analysis Lung cancer patients were divided into two groups based on Bray-Curtis distances using the partition around medoids (PAM) algorithm in the PyClustering software package (v.0.10.1) (21). Kaplan-Meier curves and log-rank tests were used to analyze survival rates using the survminer R packages. Prediction models were performed on the dataset for the two lung cancer patient groups.

[0069] Example 2. Characteristics of the subjects A total of 100 people were enrolled in the present invention, including a total of 91 patients (1 discontinued; 8 failed quality control). Most patients were female (n=84, 92.3%) and had stage 4 lung cancer (n=81, 89.0%). At the time of microbiome sampling, 39 patients had not received treatment for lung cancer, and 52 patients had received chemotherapy or targeted drug therapy for lung cancer at least once after being first diagnosed with lung cancer. The 91 control group consisted of healthy non-smoking Koreans who were age- and sex-matched with the patients. The characteristics of the subjects are shown in Table 1 below.

[0070] [Table 1]

[0071] Example 3. Decreased oral microbiome diversity in lung cancer patients The inventors investigated the oral microbiome diversity in lung cancer patients and healthy individuals (control group). As a result, it was confirmed that the Shannon diversity of the patient group was significantly lower than that of the control group (p = 0.015) (Figure 1A). In addition, the structure of the whole oral microbiome of the cancer patients was statistically significantly different from that of the control group (Adonis, R 2 = 0.014, P < 0.001) (Figure 1A).

[0072] We also examined whether cancer treatment was associated with changes in the oral microbiome. Patients were divided into a chemotherapy or targeted drug treatment group (Treatment+, n=52) and a non-treatment group (Treatment-, n=39). As a result, we confirmed that the Shannon diversity of the treatment group (Treatment+) was significantly lower than that of the control group (Figure 1B). In addition, the microbiome of the non-treatment group (Treatment-) was less diverse than the control group, but this was not statistically significant. The medoids of the treatment group (Treatment+) and non-treatment group (Treatment-) were located close to each other in a PCoA (principal coordinate analysis) plot (Figure 1B).

[0073] The results confirmed that oral microbiome diversity was reduced in lung cancer patients compared to healthy individuals. In addition, lung cancer patients who received lung cancer treatment such as chemotherapy or targeted drug therapy (Treatment+) showed a tendency to have reduced diversity compared to healthy individuals and lung cancer patients who did not receive treatment (Treatment-).

[0074] Example 4. Relationship between lung cancer and oral microorganisms The present inventors identified oral microorganisms whose abundance was altered in lung cancer patients. The relative abundance of oral microorganisms was compared at the genus level between all lung cancer patients and the control group. The results showed that the relative abundance of Veillonella was higher in lung cancer patients than in the control group, while the relative abundance of Mogibacterium, Butyrivibrio, Variovorax, Ralstonia, Catonella, and Bacillus subtilis was higher in lung cancer patients than in the control group. The relative abundance of 15 genera, including Bulleidia and Oribacterium, was significantly lower (Wilcoxon rank sum test, FDR < 0.1) (Fig. 2A).

[0075] In addition, when the treatment group (Treatment+) and the non-treatment group (Treatment-) were compared with the control group, both groups showed significantly lower abundances of the seven genera mentioned above compared to the control group (Figure 2B and Figure 2C). Some taxa were further specifically associated with the treatment condition. Olsenella was more abundant in the treatment group (Treatment+), whereas Veillonella was more abundant in the non-treatment group (Treatment-) (Figure 2B and Figure 2C). Campylobacter was less abundant in the treatment group (Treatment+), whereas Paludibacter, Treponema, Tannerella, and SHD-231 were less abundant in the non-treatment group (Treatment-) (Figure 2B and Figure 2C). All of the genera of microorganisms that were differentially abundant in the comparison between the untreated group and the control group were included in the genera of microorganisms that were differentially abundant in the comparison between all cancer patients and the control group (Figures 2A and 2B).

[0076] Example 5. Prediction of lung cancer using oral microbiome profiles We then investigated whether the composition of the oral microbiome could predict cancer status. We used a machine learning approach to predict cancer status using the composition of the oral microbiome in controls and all lung cancer patients. As a result, we confirmed that the ROC AUC was 0.95, and that the prediction model was able to distinguish lung cancer patients from controls (Figure 3A).

[0077] The present inventors also evaluated the performance of the prediction model using an independent dataset that included 41 advanced NSCLC patients and 612 healthy Koreans. As a result, the predictive power in the independent dataset was confirmed to be very high, with a ROC AUC of 0.88. This confirmed that the prediction model can distinguish lung cancer patients from healthy individuals.

[0078] The prediction model showed that the abundance of genera Mogibacterium, Variovorax, Ralstonia, and Bulleidia was important for prediction, and samples with high abundance of these genera were predicted to be healthy individuals (Figure 3B). Meanwhile, high abundance of genera Veillonella, Actinomyces, Slackia, and Schwartzia confirmed that they were highly likely to be predicted as lung cancer patients (Figure 3B).

[0079] The prediction model trained on the oral microbiome profiles of the control and treatment groups (Treatment+) had a ROC AUC of 0.96, which was better able to distinguish between cancer status than the prediction model trained on the oral microbiome profiles of the control and non-treatment groups (Treatment-) (ROC AUC = 0.88) (Figure 3A). Even though there were differences in predictive power between the two prediction models (Control vs Treatment-; and Control vs Treatment+), a similar set of microorganisms was used in both models, indicating that the coefficients in the models were similar to each other (Figure 3B).

[0080] Example 6. Prediction of lung cancer prognosis by microbiome profile The present inventors analyzed the survival outcomes of lung cancer patients to evaluate the prognostic impact of microbiome profiles. Patients were divided into two groups according to the similarity of their oral microbiomes (Figure 4A). The first group (hereinafter referred to as "Cluster 1") had a microbiome profile that was more dissimilar to the control group than the second group (hereinafter referred to as "Cluster 2") (Figure 4B). In addition, in the non-treatment group (Treatment-), Cluster 2 patients showed a tendency to survive longer than Cluster 1 (p=0.14) (Figure 4C). However, in the treatment group (Treatment+), there was no significant difference in survival rate between Cluster 1 and Cluster 2 (Figure 4C).

[0081] Furthermore, we identified an oral microbial taxonomy that could separate Cluster 1 and Cluster 2. The results showed that Cluster 1, the group with worse survival outcomes, had more abundant Streptococcus and Megasphaera microorganisms, whereas Haemophilus microorganisms were less abundant, compared to Cluster 2 (Figures 4D and 4E).

[0082] From the above results, it was confirmed that among patients not undergoing treatment for lung cancer, if the oral microbiome profile is highly similar to that of a control group (healthy individuals), the prognosis for lung cancer can be predicted to be good, and if the similarity is low, the prognosis for lung cancer can be predicted to be poor.

[0083] They also found that lung cancer patients who had a higher abundance of Streptococcus and Megasphaera microorganisms and a lower abundance of Haemophilus microorganisms could be predicted to have a poor prognosis for lung cancer.

Claims

1. (a) sequencing the 16S rRNA gene on DNA extracted from the saliva; (b) comparing the abundance of the oral microbiome to a reference value through said sequence analysis; A method for providing information for diagnosing lung cancer, predicting a risk of developing lung cancer, or predicting a prognosis thereof, comprising: When the abundance of two or more oral microbiome species selected from the group consisting of Mogibacterium, Variovorax, Ralstonia, and Bulleidia in the oral microbiome is increased compared to the reference value, it is determined that the subject does not have lung cancer or that the risk of developing lung cancer is low; The reference value is the richness of the oral microbiome in a healthy control group. Information provision method.

2. The information providing method according to claim 1 , wherein the saliva is collected from a non-smoker.

3. The information providing method according to claim 1 , wherein the sequence analysis is performed by next generation sequencing (NGS).

4. 2. The method of claim 1, wherein the lung cancer is selected from the group consisting of non-small cell lung cancer, small cell lung cancer, adenocarcinoma, squamous cell carcinoma, and large cell carcinoma.

5. A composition for diagnosing or predicting a risk of developing lung cancer, comprising a preparation capable of detecting an oral microbiome, In the oral microbiome, Mogibacterium, Variovorax, Ralstonia When the abundance of two or more oral microbiome species selected from the group consisting of spp. (a) and Bulleidia increases compared to the reference value, it is determined that the patient does not have lung cancer or has a low risk of developing lung cancer; The reference value is the richness of the oral microbiome in a healthy control group. composition.

6. The composition of claim 5 , wherein the formulation is selected from the group consisting of primers, probes, aptamers and antibodies specific to the oral microbiome.

7. The composition of claim 5, wherein the lung cancer is selected from the group consisting of non-small cell lung cancer, small cell lung cancer, adenocarcinoma, squamous cell carcinoma and large cell carcinoma.

8. A kit for diagnosing or predicting the risk of developing lung cancer, comprising the composition according to any one of claims 5 to 7.

9. The kit according to claim 8, wherein the kit is selected from the group consisting of an RT-PCR kit, a DNA chip kit, and an ELISA kit.

Images