A combination of plasma circulating microbial biomarkers, reagents, kits, risk prediction models, and their applications for esophageal adenocarcinoma.

Abstract

This invention discloses a combination of circulating microbial biomarkers in plasma for esophageal adenocarcinoma, reagents, kits, a risk prediction model, and their applications, belonging to the interdisciplinary field of liquid biopsy and microbiome. The circulating microbial biomarker combination for esophageal adenocarcinoma described in this invention includes Enterobacteriaceae, Trichophyton, Prevotella, Akkermansia, Lactobacillus fermentum, and Streptococcus. The circulating microbial biomarker combination for esophageal adenocarcinoma was validated in esophageal adenocarcinoma patients and healthy controls. The AUC was 0.952, sensitivity was 100%, specificity was 80.00%, accuracy was 91.43%, positive predictive value was 86.96%, and negative predictive value was 100%, indicating that the circulating microbial biomarker combination for esophageal adenocarcinoma can effectively distinguish between esophageal adenocarcinoma patients and healthy individuals, is suitable for early screening of esophageal adenocarcinoma, and has significant clinical screening application value.

Description

Technical Field

[0001] This invention belongs to the interdisciplinary field of liquid biopsy and microbiome, and particularly relates to a combination of circulating microbial biomarkers in plasma for esophageal adenocarcinoma, reagents, kits, risk prediction models and their applications. Background Technology

[0002] Esophageal adenocarcinoma (EAC) has seen a significant increase in incidence worldwide in recent years. As a highly malignant gastrointestinal tumor, EAC typically develops gradually from Barrett's esophagus (BE) caused by chronic gastroesophageal reflux disease (GERD), a process that often takes many years or even decades to evolve. Due to the lack of obvious early symptoms and effective screening methods, more than 90% of EAC patients are diagnosed at an intermediate or advanced stage, resulting in a five-year survival rate of less than 20% and extremely poor clinical prognosis.

[0003] Currently, screening for esophageal adenocarcinoma (EAC) primarily relies on endoscopy and biopsy. While this method is considered the "gold standard" for diagnosis, it has many limitations. First, endoscopy is invasive, requiring specialized equipment and anesthesia, and may lead to complications such as esophageal perforation and bleeding. Second, this method is costly and requires substantial medical resources, making it difficult to widely implement in the general population. More importantly, the sensitivity of endoscopy decreases significantly in early-stage lesions; its detection rate for early Barrett's esophagus (BE) low-grade dysplasia, a precancerous lesion of esophageal adenocarcinoma, is less than 60%, resulting in a large number of high-risk patients being missed.

[0004] Alternative screening technologies that have emerged in recent years, such as balloon cytology (e.g., EsoGuard) and methylation-based liquid biopsies, while reducing invasiveness to some extent, still have significant limitations. Balloon cytology still requires instrumental intervention, resulting in a poor patient experience; and existing circulating cell-free DNA (cfDNA) methylation tests have a sensitivity of only 50%-70% in early-stage acute exacerbations (EACs), with high testing costs. These limitations severely restrict the early detection and intervention of EACs. Therefore, exploring more stable and specific biomarkers has become one of the current research priorities.

[0005] In recent years, liquid biopsy technology has become a research hotspot in the field of early cancer screening due to its advantages such as being non-invasive and reproducible. While next-generation sequencing (NGS)-based circulating tumor DNA (ctDNA) detection methods (such as targeted panel sequencing and WGBS) can provide more comprehensive genomic information, they also face many challenges in clinical application. First, ctDNA is present in extremely low concentrations in blood, accounting for only 0.01%-1% of cfDNA. Therefore, NGS testing typically requires ultra-high depth sequencing (usually >10000X), resulting in high costs and hindering widespread adoption. Second, the ctDNA signal released by early-stage tumors is weak, limiting detection sensitivity and easily leading to false negatives. Furthermore, non-tumor factors such as clonal hematopoiesis (CHIP) may introduce gene mutation noise, leading to false positives and further affecting the accuracy of the test.

[0006] Therefore, there is an urgent need to find a highly sensitive, highly specific, highly accurate, low-cost, and non-invasive technical solution that can be used for early screening of esophageal adenocarcinoma to make up for the shortcomings of existing technologies. Summary of the Invention

[0007] In view of this, the purpose of this invention is to provide a combination of plasma circulating microbial markers for esophageal adenocarcinoma, which has high sensitivity, specificity, and accuracy, and has the advantages of being non-invasive and low in detection cost, making it suitable for large-scale early screening of esophageal adenocarcinoma.

[0008] Another object of the present invention is to provide a reagent.

[0009] Another object of the present invention is to provide a reagent kit.

[0010] Another object of the present invention is to provide the application of the combination of plasma circulating microbial markers for esophageal adenocarcinoma, or the reagent or the kit, in constructing a risk prediction model for esophageal adenocarcinoma.

[0011] Another objective of this invention is to provide a risk prediction model for esophageal adenocarcinoma.

[0012] Another object of the present invention is to provide an application of the combination of plasma circulating microbial markers for esophageal adenocarcinoma, or the reagent, or the kit, or the esophageal adenocarcinoma risk prediction model in the preparation of esophageal adenocarcinoma screening and diagnostic products.

[0013] To achieve the above-mentioned objectives, the present invention provides the following technical solution:

[0014] This invention provides a combination of circulating microbial biomarkers in plasma for esophageal adenocarcinoma, comprising Enterobacteriaceae, Leptotrichia spp., Prevotella spp., Akkermansia muciniphila, Lactobacillus fermentum, and Streptococci spp.

[0015] The present invention also provides a reagent comprising a reagent for detecting the content or abundance of the combination of circulating microbial markers in plasma of the esophageal adenocarcinoma.

[0016] The present invention also provides a kit comprising the reagents.

[0017] The present invention also provides the application of the combination of plasma circulating microbial markers for esophageal adenocarcinoma, or the reagent or kit, in constructing a risk prediction model for esophageal adenocarcinoma.

[0018] The present invention also provides a risk prediction model for esophageal adenocarcinoma, the risk prediction model comprising:

[0019] Risk score = 1 / (1+e^-(13.698+0.138×Enterobacteriaceae+0.113×Leptotrichia spp.+0.246×Prevotella spp.-0.070×Akkermansia muciniphila+0.352×Lactobacillus fermentum-3.343×Streptococci spp.));

[0020] Among them, Enterobacteriaceae, Leptotrichia spp., Prevotella spp., Akkermansiamuciniphila, Lactobacillus fermentum, and Streptococci spp. represent the actual abundance of each microorganism in the sample to be tested.

[0021] Preferably, a positive result for esophageal adenocarcinoma is output when the risk score is ≥0.59.

[0022] Preferably, the sample to be tested includes peripheral blood of the subject.

[0023] The present invention also provides the application of the combination of plasma circulating microbial markers for esophageal adenocarcinoma, or the reagent, or the kit, or the esophageal adenocarcinoma risk prediction model in the preparation of esophageal adenocarcinoma screening and diagnostic products.

[0024] Preferably, the method for screening and diagnosing esophageal adenocarcinoma includes: obtaining peripheral blood cfDNA samples from the subject and sequencing them; determining the abundance of each microorganism in the plasma circulating microbial marker combination for esophageal adenocarcinoma; and determining the risk of esophageal adenocarcinoma based on the esophageal adenocarcinoma risk prediction model.

[0025] Preferably, the product is used for early screening and diagnosis of esophageal adenocarcinoma.

[0026] The beneficial effects of this invention are:

[0027] This invention, through large-scale clinical sample analysis, identified six microorganisms as a combination of circulating microbial biomarkers for esophageal adenocarcinoma (ESAC), which can be used for early screening of ESAC with high sensitivity, specificity, and accuracy. A high-precision logistic regression risk prediction model was constructed using this combination of ESAC plasma circulating microbial biomarkers, achieving an AUC of 0.952 on the independent validation set. Using this risk prediction model for screening and diagnosing ESAC can improve the detection rate of early ESAC. The use of low-pass whole-genome sequencing (Low Pass WGS) allows for analysis at extremely low sequencing depths, significantly reducing costs and making early detection of ESAC based on plasma cmDNA clinically feasible. This has the potential to fundamentally change the current predicament of ESAC screening and achieve true early diagnosis and treatment. Attached Figure Description

[0028] Figure 1 The flowchart of bioinformatics analysis and decontamination process in Example 1 are shown below;

[0029] Figure 2 The combination and original weights of plasma circulating microbial markers for esophageal adenocarcinoma in Example 1;

[0030] Figure 3 The ROC curves of the weighted optimized logistic regression model and LDA model in Example 1 are shown.

[0031] Figure 4 Microbial weights (coefficient β) optimized for the logistic regression model in Example 1 i );

[0032] Figure 5 The cutoff value for the logistic regression model in Example 1 is determined;

[0033] Figure 6 This is the ROC curve of the validation set in Example 1. Detailed Implementation

[0034] This invention provides a combination of circulating microbial biomarkers in plasma for esophageal adenocarcinoma, comprising Enterobacteriaceae, Leptotrichia spp., Prevotella spp., Akkermansia muciniphila, Lactobacillus fermentum, and Streptococci spp.

[0035] This invention involved nucleic acid extraction, library construction, and sequencing of peripheral venous blood samples from 90 pathologically confirmed esophageal adenocarcinoma patients and 60 healthy individuals. Bioinformatics analysis identified six microorganisms significantly associated with esophageal adenocarcinoma. The esophageal adenocarcinoma plasma circulating microbial biomarker combination of this invention was validated in 40 pathologically confirmed esophageal adenocarcinoma patients and 30 healthy individuals. The AUC was 0.952, sensitivity was 100%, specificity was 80.00%, and accuracy was 91.43%, indicating that the six microorganisms screened by this invention can effectively distinguish esophageal adenocarcinoma patients from healthy individuals, and that the esophageal adenocarcinoma plasma circulating microbial biomarker combination has significant clinical screening application value.

[0036] The present invention also provides a reagent comprising a reagent for detecting the content or abundance of the combination of circulating microbial markers in plasma of the esophageal adenocarcinoma.

[0037] In this invention, the reagent preferably includes primers, probes, aptamers, or antibodies that are specific to the six microorganisms in the plasma circulating microbial marker combination for esophageal adenocarcinoma.

[0038] In this invention, the detection sample of the reagent preferably includes peripheral blood of the subject, and more preferably includes plasma DNA in the peripheral blood of the subject.

[0039] The present invention also provides a kit comprising the reagents.

[0040] The present invention also provides the application of the combination of plasma circulating microbial markers for esophageal adenocarcinoma, or the reagent or kit, in constructing a risk prediction model for esophageal adenocarcinoma.

[0041] The present invention also provides a risk prediction model for esophageal adenocarcinoma, the risk prediction model comprising:

[0042] Risk score = 1 / (1+e^-(13.698+0.138×Enterobacteriaceae+0.113×Leptotrichia spp.+0.246×Prevotella spp.-0.070×Akkermansia muciniphila+0.352×Lactobacillus fermentum-3.343×Streptococci spp.));

[0043] Wherein, Enterobacteriaceae, Leptotrichia spp., Prevotella spp., Akkermansiamuciniphila, Lactobacillus fermentum and Streptococci spp. represent the actual abundance of each microorganism in the sample to be tested, and the actual abundance is preferably the proportion of the number of reads obtained from sequencing each microorganism to the total number of reads obtained from sequencing all microorganisms.

[0044] In this invention, a positive result for esophageal adenocarcinoma is output when the risk score is preferably ≥0.59, and a negative result for esophageal adenocarcinoma is output otherwise.

[0045] In this invention, the sample to be tested preferably includes peripheral blood of the subject.

[0046] This invention constructs a logistic regression model through weighted optimization and retrains the model parameters to optimize the microbial weights using L1 regularized weighted optimization (10-fold cross-validation). The resulting esophageal adenocarcinoma risk prediction model exhibits high sensitivity and specificity. Validation was performed on 40 pathologically diagnosed esophageal adenocarcinoma patients and 30 healthy individuals, showing an AUC of 0.952, sensitivity of 100%, specificity of 80.00%, accuracy of 91.43%, positive predictive value (PPV) of 86.96%, and negative predictive value (NPV) of 100%. This demonstrates that the esophageal adenocarcinoma risk prediction model of this invention can significantly distinguish between esophageal adenocarcinoma patients and healthy individuals, highlighting its potential as a non-invasive screening tool for esophageal adenocarcinoma.

[0047] The present invention also provides the application of the combination of plasma circulating microbial markers for esophageal adenocarcinoma, or the reagent, or the kit, or the esophageal adenocarcinoma risk prediction model in the preparation of esophageal adenocarcinoma screening and diagnostic products.

[0048] In this invention, the preferred method for screening and diagnosing esophageal adenocarcinoma includes: obtaining peripheral blood cfDNA samples from the subject and sequencing them; determining the abundance of each microorganism in the plasma circulating microbial marker combination for esophageal adenocarcinoma; and determining the risk of esophageal adenocarcinoma based on the esophageal adenocarcinoma risk prediction model.

[0049] In this invention, the product is preferably used for early screening and diagnosis of esophageal adenocarcinoma, wherein early stage preferably includes stage I or stage II esophageal adenocarcinoma.

[0050] The verification experiments of this invention show that, among all included esophageal adenocarcinoma samples, the proportion of early-stage (stage I / II) esophageal adenocarcinoma samples reached 66.92%, indicating that the combination of plasma circulating microbial markers for esophageal adenocarcinoma of this invention can serve as a reliable tumor marker and is suitable for non-invasive early screening of esophageal adenocarcinoma.

[0051] The technical solutions provided by the present invention will be described in detail below with reference to the embodiments, but they should not be construed as limiting the scope of protection of the present invention.

[0052] Unless otherwise specified, the following embodiments are all conventional methods.

[0053] Unless otherwise specified, all materials and reagents used in the following examples are commercially available.

[0054] Example 1

[0055] 1. Sample screening and grouping

[0056] Peripheral venous blood samples were collected from 130 patients with pathologically confirmed esophageal adenocarcinoma (who had not received any anti-tumor treatment before surgery) and 90 healthy controls (preserved using Streck cfDNA blood collection tubes). All cancer cases were confirmed by imaging, laboratory tests, and pathology. Healthy controls were from a normal population undergoing routine physical examinations. The samples were randomly divided into training and validation sets, with no statistically significant differences in age or sex between the two groups.

[0057] Training set: 90 cancer patients (25 stage I, 36 stage II, and 29 stage III-IV) and 60 healthy controls;

[0058] Validation set: 40 cancer patients (11 in stage I, 15 in stage II, and 14 in stages III-IV) and 30 healthy controls.

[0059] 2. Low-pass whole-genome sequencing (WGS) experimental workflow

[0060] (1) Low Pass WGS Wet Test Procedure

[0061] a. Sample Collection: Collect 10 mL of peripheral blood using a cell-free DNA blood collection tube (Streck tube). Transport the sample at 6℃~26℃ and complete plasma separation within 72 hours. Severely hemolyzed samples should not be used for the experiment.

[0062] b. Plasma separation: Plasma samples were obtained using a two-step separation method. The first step was low-speed centrifugation at 4℃ and 1600×g for 15 min to remove whole blood cells and collect the supernatant plasma. The second step was high-speed centrifugation at 4℃ and 16000×g for 15 min to remove platelets, cell debris, and apoptotic bodies.

[0063] c. Cell-free nucleic acid extraction: The QIAamp Circulating Nucleic Acid Kit (Kaijie, 55114) was used, with the addition of vector RNA to improve the recovery rate of small nucleic acid fragments and enhance the enrichment of microbial nucleic acids. The concentration of cfDNA was determined using Qubit. If the total extracted amount >300 ng, fragment quality control using Agilent 2100 was required. If the extract contained only genomic DNA and lacked small cfDNA fragments, further experimentation was not recommended.

[0064] d. Library Construction: The xGen Prism DNA Library Prep Kit, KAPA HyperPure magnetic beads, HiFi Hotstart ReadyMix, and PCR Index Primer were used for library construction. The amount of cfDNA used was 50-200 ng. The experimental steps included end repair, magnetic bead purification, ligation reaction 1 (20℃ 15 min → 65℃ 15 min → 4℃ maintenance), ligation reaction 2 (65℃ 30 min → 4℃ maintenance), magnetic bead purification, library amplification (10-12 cycles), and obtaining the library sample through magnetic bead purification. An 8 bp paired-end UMI adapter was added during library construction.

[0065] e. Sequencing: The sequencing equipment platform was the MGISEQ-2000RS (MGI Tech) platform, and the sequencing strategy was PE150 (150bp paired ends), operated according to the MGI Tech manual. Sterile water was added as a control during each sequencing run to monitor reagent / environmental microbial contamination.

[0066] (2) Low Pass WGS Bioinformatics Analysis Workflow

[0067] Raw sequencing data, after being filtered by FASTP quality control, underwent further filtering using Bowtie2 to remove human sequences (GRCh38 reference genome) and BBmap to remove repetitive sequences and mitochondrial contamination. Microorganisms were classified and their abundance corrected using Kraken2 and Bracken, and contamination was removed using Decontam and negative controls (filtering low-abundance, low-frequency species, and exogenous interference). The final microbial characteristic data were used for machine learning modeling to construct a disease diagnostic prediction model. This workflow, through multiple quality control and decontamination steps, significantly improves the specificity and reliability of the detection. See [link to bioinformatics analysis and decontamination workflow] for details. Figure 1.

[0068] 3. Screening of plasma circulating microbial biomarkers for esophageal adenocarcinoma and construction of a risk prediction model

[0069] Based on joint analysis of the training set WGS data and public databases including The Cancer Genome Atlas (TCGA), Gene Expression Omnibus (GEO), and International Cancer Genome Consortium (ICGC), six microorganisms significantly associated with esophageal adenocarcinoma were identified, along with their original weights. These included: Enterobacteriaceae (original weight 0.23), Leptotrichia spp. (original weight 0.18), Prevotella spp. (original weight 0.21), Akkermansia muciniphila (original weight 0.12), and Lactobacillus fermentum (original weight 0.16), all enriched in esophageal adenocarcinoma; and Streptococci spp. (original weight -0.20), significantly reduced in esophageal adenocarcinoma, as shown in Table 1. Figure 2 As shown.

[0070] Table 1. Original weights of 16 microorganisms and optimized weights from the two models.

[0071] Microbial name Original weights Logistic Regression Weights LDA weights Enterobacteriaceae 0.23 0.138 6.521 Leptotrichiaspp. 0.18 0.113 9.550 Prevotellaspp. 0.21 0.246 -1.807 Akkermansiamuciniphila 0.12 -0.070 2.051 Lactobacillusfermentum 0.16 0.352 19.651 Streptococcispp. -0.20 -3.343 -68.680

[0072] Based on the aforementioned six microbial biomarkers, logistic regression and linear discriminant analysis (LDA) models were constructed using the self-developed training set WGS data (90 cancer patients / 60 healthy individuals) through weighted optimization. However, the original weights performed poorly in both models (AUC < 0.85). Therefore, the model parameters were retrained using L1 regularized weighted optimization (10-fold cross-validation) to optimize the microbial weights, obtaining the logistic regression weights and LDA weights (see Table 1). After weighted optimization, the ROC curves of both models showed good performance (see Table 1). Figure 3 ):

[0073] The LDA model had an AUC of 0.966, a sensitivity of 100%, a specificity of 78.33%, an accuracy of 91.33%, a positive predictive value (PPV) of 87.38%, and a negative predictive value (NPV) of 100%.

[0074] The logistic regression model had an AUC of 0.970, a sensitivity of 100%, a specificity of 80.00%, an accuracy of 92%, a positive predictive value (PPV) of 88.24%, and a negative predictive value (NPV) of 100%. The logistic regression model improved specificity while maintaining high sensitivity. Therefore, the logistic regression model was chosen for screening esophageal adenocarcinoma. The constructed logistic regression model is as follows:

[0075] Logistic regression model outputs risk score (Risk) LR Risk LR It is a probability value between 0 and 1, representing the predicted probability of "belonging to cancer". Its model formula is:

[0076]

[0077] Where, β0=13.698, β i Logistic regression weights (see Table 1 and ...) Figure 4 ), A i The actual abundance of each microorganism (i.e., the number of reads of each microorganism obtained from sequencing divided by the total number of reads of microorganisms obtained from sequencing), where i is 1-6, representing 6 microorganisms.

[0078] By maximizing the Youden exponent, the optimal cutoff value was determined to be 0.59. (See...) Figure 5 .

[0079] 4. Validation of plasma circulating microbial biomarker combination and risk assessment model for esophageal adenocarcinoma

[0080] The logistic regression model was validated using an independent validation set (40 patients with esophageal adenocarcinoma and 30 healthy controls). The results showed that the logistic regression model maintained stable performance: AUC 0.952, sensitivity 100%, specificity 80.00%, accuracy 91.43%, positive predictive value (PPV) 86.96%, and negative predictive value (NPV) 100%. Figure 6 ).

[0081] The above performance demonstrates that the combination of plasma circulating microbial biomarkers and risk prediction model for esophageal adenocarcinoma of this invention can significantly distinguish between esophageal adenocarcinoma patients and healthy individuals, highlighting its potential as a non-invasive screening tool for esophageal adenocarcinoma.

[0082] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for constructing a risk prediction model for esophageal adenocarcinoma, characterized in that, Based on a combination of plasma circulating microbial biomarkers for esophageal adenocarcinoma, a risk prediction model for esophageal adenocarcinoma was constructed. The plasma circulating microbial markers for esophageal adenocarcinoma consist of Enterobacteriaceae, Leptotrichia spp., Prevotella spp., Akkermansia muciniphila, Lactobacillus fermentum, and Streptococcis pp.; The risk prediction model includes: Risk score = 1 / (1+e^-(13.698+0.138×Enterobacteriaceae+0.113×Leptotrichiaspp.+0.246×Prevotella spp.-0.070×Akkermansia muciniphila+0.352×Lactobacillus fermentum-3.343×Streptococci spp.)); Among them, Enterobacteriaceae, Leptotrichia spp., Prevotella spp., Akkermansiamuciniphila, Lactobacillus fermentum, and Streptococci spp. represent the actual abundance of each microorganism in the sample to be tested.

2. The construction method according to claim 1, characterized in that, When the risk score is ≥0.59, a positive result for esophageal adenocarcinoma is output.

3. The construction method according to claim 1, characterized in that, The test sample includes the subject's peripheral blood.

4. The application of a reagent or kit in the preparation of esophageal adenocarcinoma screening and diagnostic products, characterized in that, The reagents include those for detecting the content or abundance of a combination of circulating microbial markers in plasma from esophageal adenocarcinoma; the kit includes the reagents. The plasma circulating microbial markers for esophageal adenocarcinoma consist of Enterobacteriaceae, Leptotrichia spp., Prevotella spp., Akkermansia muciniphila, Lactobacillus fermentum, and Streptococcis pp.; The method for screening and diagnosing esophageal adenocarcinoma includes: obtaining peripheral blood cfDNA samples from the subject and sequencing them; determining the abundance of each microorganism in the plasma circulating microbial marker combination for esophageal adenocarcinoma; and determining the risk of esophageal adenocarcinoma based on the esophageal adenocarcinoma risk prediction model constructed according to any one of claims 1 to 3.

5. The application according to claim 4, characterized in that, The product is used for early screening and diagnosis of esophageal adenocarcinoma.

Images