Markers for diagnosis and treatment of gastric cancer

Abstract

The invention discloses a marker for diagnosing and treating gastric cancer. The invention provides a novel non-invasive method which is used for early detection of peritoneal metastasis of gastric cancer patients. An RSA model combines 4-mRNA groups (BUB1, SPC25, CT83 and MMP3) and clinical characteristics, in addition, the invention also provides application of 4-mRNA in treatment of gastric cancer, provides a new direction for improving risk stratification, treatment plan and monitoring of gastric cancer patients and effective treatment of gastric cancer, and has a wide application prospect.

Description

Technical Field

[0001] The present invention belongs to the field of biomedicine, and in particular relates to markers for diagnosing and treating gastric cancer. Background Art

[0002] Gastric cancer remains one of the leading causes of cancer-related deaths worldwide and is particularly common in East Asian countries. Despite advances in diagnostic and therapeutic strategies, the prognosis of patients with locally advanced gastric cancer (LAGC) remains poor due to the high incidence of metastasis and recurrence. Among them, peritoneal metastasis is particularly occult, often presenting as occult in the early stages and undetectable with conventional imaging techniques. Early and accurate detection of occult peritoneal metastasis is crucial for optimizing treatment plans and improving the overall survival of patients with LAGC. Summary of the invention

[0003] To make up for the deficiencies of the prior art, the present invention provides markers for diagnosing and treating gastric cancer.

[0004] To achieve the above object, the present invention adopts the following technical solution:

[0005] The first aspect of the present invention provides the use of a reagent for detecting the expression level of a biomarker in a sample in the preparation of a product for diagnosing gastric cancer / diagnosing gastric cancer metastasis / predicting the prognosis of gastric cancer, wherein the biomarker includes any one or more of BUB1, SPC25, CT83 and MMP3.

[0006] Furthermore, the gastric cancer metastasis is gastric cancer peritoneal metastasis.

[0007] Furthermore, the prediction of gastric cancer prognosis includes predicting gastric cancer survival, predicting gastric cancer treatment efficacy, predicting gastric cancer recurrence, and predicting gastric cancer cachexia.

[0008] Furthermore, the survival period includes overall survival rate (OS) and disease-free survival rate (DFS).

[0009] Furthermore, the gastric cancer treatment efficacy also includes the success rate of conversion therapy.

[0010] Furthermore, the gastric cancer recurrence is peritoneal recurrence of gastric cancer.

[0011] Furthermore, the reagents include oligonucleotide probes that specifically recognize the biomarker gene, primers that specifically amplify the biomarker gene, or binding agents that specifically bind to proteins encoded by the biomarker gene.

[0012] Furthermore, the primer sequences for specifically amplifying BUB1 are shown in SEQ ID NO: 1-2.

[0013] Furthermore, the primer sequences for specifically amplifying SPC25 are shown in SEQ ID NO: 3-4.

[0014] Furthermore, the primer sequences for specifically amplifying CT83 are shown in SEQ ID NO: 5-6.

[0015] Furthermore, the primer sequences for specifically amplifying MMP3 are shown in SEQ ID NOs: 7-8.

[0016] Furthermore, the sample includes tissue and blood.

[0017] Furthermore, the blood includes peripheral blood and serum.

[0018] Furthermore, the gastric cancer includes locally advanced gastric cancer (LAGC).

[0019] The second aspect of the present invention provides a product for diagnosing gastric cancer / diagnosing gastric cancer metastasis / predicting the prognosis of gastric cancer, the product comprising a reagent for detecting the expression level of a biomarker in a sample, the biomarker comprising any one or more of BUB1, SPC25, CT83 and MMP3.

[0020] Furthermore, the reagent also includes a detectable substance.

[0021] Furthermore, the product includes a test kit and a test paper.

[0022] The third aspect of the present invention provides the use of biomarkers in constructing a model for diagnosing gastric cancer metastasis, wherein the biomarkers include any one or more of BUB1, SPC25, CT83 and MMP3.

[0023] Furthermore, the indicators of the model also include T stage, N stage, and pathological type.

[0024] Furthermore, the gastric cancer metastasis is gastric cancer peritoneal metastasis.

[0025] A fourth aspect of the present invention provides a model for diagnosing gastric cancer metastasis, wherein the indicators of the model include the levels of biomarkers, and the biomarkers include any one or more of BUB1, SPC25, CT83 and MMP3.

[0026] Furthermore, the indicators of the model also include T stage, N stage, and pathological type.

[0027] Furthermore, the model is a nomogram model.

[0028] Furthermore, the calculation formula of the model includes:

[0029] The probability of gastric cancer metastasis = [(4.855×4-mRNA)+(4.139×T stage)+(4.140×N stage)+(3.147×pathological type)+(-9.961)], or

[0030] The probability of gastric cancer metastasis = [(6.558×4-mRNA)+(8.467×T stage)+(8.054×N stage)+(8.164×pathological type)-14.876].

[0031] Furthermore, the gastric cancer metastasis is gastric cancer peritoneal metastasis.

[0032] The fifth aspect of the present invention provides the use of a biomarker inhibitor in the preparation of a pharmaceutical composition for treating gastric cancer, wherein the biomarker comprises any one or more of BUB1, SPC25, CT83 and MMP3.

[0033] Furthermore, the inhibitors include nucleic acid inhibitors and protein inhibitors.

[0034] Furthermore, the nucleic acid inhibitors include siRNA, shRNA, and ribozymes.

[0035] Furthermore, the nucleic acid inhibitor is selected from siRNA.

[0036] Furthermore, the siRNA sequence of BUB1 is shown in SEQ ID NO:9 or SEQ ID NO:10.

[0037] Furthermore, the siRNA sequence of SPC25 is shown in SEQ ID NO:11 or SEQ ID NO:12.

[0038] Furthermore, the siRNA sequence of CT83 is shown in SEQ ID NO:13 or SEQ ID NO:14.

[0039] Furthermore, the siRNA sequence of MMP3 is shown in SEQ ID NO:15 or SEQ ID NO:16.

[0040] Furthermore, the treatment of gastric cancer includes inhibiting gastric cancer metastasis.

[0041] Furthermore, the inhibition of gastric cancer metastasis includes inhibiting gastric cancer peritoneal metastasis.

[0042] A sixth aspect of the present invention provides a pharmaceutical composition comprising an inhibitor of a biomarker, wherein the biomarker comprises any one or more of BUB1, SPC25, CT83 and MMP3.

[0043] Furthermore, the pharmaceutical composition also includes a pharmaceutically acceptable carrier.

[0044] The seventh aspect of the present invention provides the use of biomarkers as targets in screening candidate drugs for treating gastric cancer, wherein the biomarkers include any one or more of BUB1, SPC25, CT83 and MMP3.

[0045] Furthermore, the method for screening candidate drugs for treating gastric cancer includes: testing the effect of the test substance on the expression level of biomarkers in the sample, wherein, after using the test substance, the biomarker level decreases, indicating that the test substance is a candidate drug for treating gastric cancer.

[0046] The eighth aspect of the present invention provides the use of biomarker inhibitors in the preparation of products for regulating gene / protein levels, wherein the biomarkers include any one or more of BUB1, SPC25, CT83 and MMP3, and the genes / proteins include any one or more of MMP9, N-cadherin, vimentin, E-cadherin.

[0047] Furthermore, the inhibitors include nucleic acid inhibitors and protein inhibitors.

[0048] Furthermore, the nucleic acid inhibitors include siRNA, shRNA, and ribozymes.

[0049] Furthermore, the nucleic acid inhibitor is selected from siRNA.

[0050] A ninth aspect of the present invention provides any of the following methods:

[0051] (1) A method for screening a candidate drug for treating gastric cancer, the method comprising: testing the effect of a test substance on the expression level of a biomarker in a sample, wherein after using the test substance, a decrease in the level of the biomarker indicates that the test substance is a candidate drug for treating gastric cancer, and the biomarker comprises any one or more of BUB1, SPC25, CT83 and MMP3;

[0052] (2) A method for inhibiting gastric cancer cell metastasis in vitro, the method comprising administering an inhibitor of a biomarker, wherein the biomarker comprises any one or more of BUB1, SPC25, CT83 and MMP3;

[0053] (3) A method for regulating gene / protein expression in vitro, the method comprising administering an inhibitor of a biomarker, wherein the biomarker comprises any one or more of BUB1, SPC25, CT83 and MMP3, and the gene / protein comprises any one or more of MMP9, N-cadherin, vimentin, E-cadherin.

[0054] Advantages and beneficial effects of the present invention:

[0055] This application proposes a novel non-invasive method for early detection of peritoneal metastasis in patients with gastric cancer. The RSA model combines the 4-mRNA group (BUB1, SPC25, CT83 and MMP3) and clinical characteristics. In addition, it also provides the application of 4-mRNA in the treatment of gastric cancer, which provides a new direction for improving the risk stratification, treatment planning and monitoring of gastric cancer patients, as well as the effective treatment of gastric cancer, and has broad application prospects. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 This is a flow chart of the study design to discover and validate a 4-mRNA combination to predict occult peritoneal metastasis in LAGC patients;

[0057] Figure 2 Figure 2 is the identification and preliminary validation of biomarkers for LAGC peritoneal metastasis using public databases and transcriptome data, where 2A is a Venn diagram of differentially expressed genes from TCGA, GSE62254, and paired mRNA sequencing data, 2B is a volcano diagram of differentially expressed genes from GSE62254 data, 2C is a volcano diagram of differentially expressed genes from TCGA data, 2D is a volcano diagram of differentially expressed genes from paired mRNA sequencing data, 2E is an expression diagram of 4-mRNA in 42 matched cancer tissues of patients with and without peritoneal metastasis, 2F is an expression diagram of 4-mRNA in peripheral blood of 33 pairs of matched patients, 2G is a protein immunoblot diagram of cancer tissues of patients with and without peritoneal metastasis, and 2H is Figure 2 is the immunohistochemical analysis of 4-mRNA in matched cancer tissue samples, 2I is the quantitative immunohistochemical results of 4-mRNA, 2J-2M are the correlation diagrams of five-year DFS with high and low mRNA expressions, 2N is the correlation diagram between 4-mRNA expression levels and 41 clinicopathological characteristics, 2O is the overall survival analysis diagram of 4-mRNA, 2P is the protein-protein interaction (PPI) network diagram of 4-mRNA, 2Q-T is the results of Western blot quantitative analysis of 4-mRNA in primary cancer tissues of patients with and without peritoneal metastasis, 2U-X are the overall survival curves of 4-mRNA genes in formalin-fixed specimens;

[0058] Figure 3Figure 3 is a model of 4-mRNA in fresh frozen tissue samples for training and validation, and detection of peritoneal metastasis in LAGC patients. 3A is a multivariate logistic regression analysis of factors affecting peritoneal metastasis in LAGC patients, 3B is a nomogram of 4-mRNA features and clinical features for predicting peritoneal metastasis in LAGC patients, 3C is a ROC curve of various predictive variables in the training set, 3D is a ROC curve of various predictive variables in the validation set, 3E is a confusion matrix of various prediction models in the training and validation sets, 3F is a calibration curve of the RSA model in the training set, 3G is a calibration curve of the RSA model in the validation set, and 3H is the performance indicators of different prediction models in the training set. Radar chart, 3I is a radar chart of performance indicators of different prediction models in the validation set, 3J is a clinical double-layer concentric circle chart of various prediction models in the training and validation sets, 3K is a log-rank test survival curve chart of patients in the training set, 3L is a log-rank test survival curve chart of patients in the validation set, 3M is a comparison chart of the AUC of different prediction models in the stratified analysis based on the expression of molecular markers HER2 and PDL1, 3N is a comparison chart of the specificity of different prediction models in the stratified analysis based on TNM stage and the expression of molecular markers HER2 and PDL1, 3O is a comparison chart of the sensitivity of different prediction models in the stratified analysis based on TNM stage and the expression of molecular markers HER2 and PDL1;

[0059] Figure 4 Figure 4 is a graph of conversion therapy results and cachexia predicted by the 4-mRNA model and fresh frozen tissue specimens in patients with peritoneal metastasis of gastric cancer. Among them, 4A is a graph of conversion therapy regimen, 4B is a waterfall graph of CT efficacy evaluation after conversion therapy, 4C is a graph of cachexia grading defined by the consensus of the Asian Cachexia Working Group, 4D is a graph comparing the pathological regression grade of primary lesions after conversion therapy in high-risk and low-risk patients according to the RSA model, 4E is a graph of CT evaluation of the changes in primary lesions before and after conversion therapy in the high-risk and low-risk groups in the RSA model, 4F is a graph of changes in peritoneal metastases evaluated by laparoscopic exploration before and after conversion therapy in the high-risk and low-risk groups in the RSA model, and 4G 4H is a comparison of clinical efficacy and cachexia incidence between different models, 4I is a comparison of the efficacy of CT evaluation between different models, 4J is a comparison of the incidence of 43 cachexias between different models, 4K is a comparison of the pathological regression grading of different models, 4L is a ROC curve comparison of the effectiveness of different models in predicting CT evaluation results, 4M is a ROC curve evaluation of the accuracy of different models in predicting the success rate of conversion therapy, 4N is a ROC curve for predicting the occurrence of cachexia by different models, and 4O is a three-year OS curve for patients receiving conversion therapy;

[0060] Figure 5Figure 5 is a graph of the 4-mRNA-based model used to validate fresh frozen specimens to detect peritoneal free cancer cells in patients with LAGC. 5A is a representative graph of free cancer cells in peritoneal lavage fluid, 5B is a ROC curve graph of various predictive variables, 5C is a calibration curve graph of the RSA model, 5D is a confusion matrix graph of various predictive models, 5E is a radar graph of performance indicators of different predictive models, 5F is a double-layer concentric circle graph of the clinical advantages of various predictive models, 5G is a log-rank test survival curve graph, and 5H is a prospective clinical study (ChiCTR1800014817 ) is a waterfall diagram of the treatment effect of patients divided into high-risk group and low-risk group according to the RSA model, 5I is a comparison diagram of the incidence of cachexia in the high-risk group and low-risk group in the RSA model in a prospective clinical study (ChiCTR1800014817), 5J is a waterfall diagram of the treatment effect of patients in the high-risk and low-risk groups in the RSA model before and after conversion treatment in a prospective clinical study 44 (NCT03718624), 5K is a diagram analyzing the incidence of cachexia in high-risk and low-risk patients in the RSA model in a prospective clinical study (NCT03718624);

[0061] Figure 6 Figure 6 is a graph showing the non-invasive prediction and detection of peritoneal metastasis in LAGC patients using a model based on peripheral blood 4-mRNA. Figure 6A is a graph showing the changes in the A260 / 280 ratio of peripheral blood samples at different time points. Figure 6B is a nomogram showing the combination of 4-mRNA profiles and clinical features to predict peritoneal metastasis in LAGC patients. Figure 6C is a graph showing the ROC curves of various prediction models in the training set. Figure 6D is a graph showing the ROC curves of various prediction models in the validation set. Figure 6E is a graph showing the confusion matrix of different prediction models in the training and validation sets. Figure 6F is a graph showing the calibration curve of the RSA model in the training set. , 6G is a calibration curve diagram of the RSA model in the validation set, 6H is a radar diagram of the performance indicators of different prediction models in the training set, 6I is a radar diagram of the performance indicators of different prediction models in the validation set, 6J is a double-layer concentric circle diagram of the clinical benefits of various prediction models in the training and validation sets, 6K is a log-rank test survival curve diagram of patients in the training set, 6L is a log-rank test survival curve diagram of patients in the validation set, 6N is a comparison diagram of CT evaluation effects between different models, 6O is a comparative analysis diagram of pathological regression grading between models, and 6P is a comparison diagram of cachexia incidence between different prediction models;

[0062] Figure 77A is a ROC curve diagram of different prediction models in the validation set of GC-CY1 patients, 7B is a calibration curve diagram of the RSA model used for GC-CY1 validation, 7C is a radar diagram comparing the performance indicators of various prediction models for GC-CY1 patients, 7D is a log-rank test survival curve diagram for GC-CY1 patients, 7E is a double-layer concentric circle diagram of the clinical benefit of clinical characteristics and 4-mRNA data in predicting GC-CY1, 7F is a confusion matrix diagram of the combination of clinical characteristics and 4-mRNA data in predicting GC-CY1, 7G is a prediction diagram of peritoneal recurrence after radical surgery for LAGC patients by the RSA model, 7H is a ROC curve diagram of different models for predicting peritoneal metastasis in patients with negative tumor markers, 7I is a ROC curve diagram of different models for predicting peritoneal recurrence in LAGC patients after radical surgery, and 7J is a comparison of different prediction models in 7K is a radar chart of the performance indicators in patients with negative tumor markers, 7L is a double-layer concentric circle chart of the clinical benefit of the RSA model in predicting postoperative peritoneal recurrence in 46 patients with negative tumor markers, 7M is a confusion matrix chart of the RSA model in predicting postoperative peritoneal recurrence in patients with negative tumor markers, 7N is a comparison chart of the consistency between the RSA model and laparoscopic exploration in predicting peritoneal metastasis in a prospective observational study (NCT06478394), 7O is a ROC curve chart of various prediction models for predicting peritoneal metastasis in a prospective observational study (NCT06478394), 7P is a graph of the overall survival of patients with peritoneal recurrence after radical resection of LAGC in the validation group, 7Q is a graph of the overall survival of patients with negative tumor markers in the validation group, 7R is a calibration curve chart for predicting peritoneal recurrence after radical resection of LAGC in the validation group, 7S is a calibration curve chart for predicting the prognosis of patients with negative tumor markers in the validation group;

[0063] Figure 8Figure 8 is the effect of 4-mRNA on GC cell metastasis in vivo and in vitro, where 8A-D are the migration ability of HGC-27 cells after knocking down BUB1, SPC25, CT83 and MMP3, 8E-H are the invasion and metastasis potential of HGC-27 cells after knocking down BUB1, SPC25, CT83 and MMP3, 8I is a representative image of peritoneal metastatic tumors after intraperitoneal injection of gene-specific knockdown HGC-27 cells in mice, and 8J is a representative image of peritoneal metastatic tumors after intraperitoneal injection of BUB1, SPC25, CT83 and MMP3 knockdown HGC-27 cells in mice. 8K-L are the immunohistochemistry (IHC) images of peritoneal metastatic tumors in mice injected with HGC-27 cells with BUB1 and SPC25 knockdown (left), and the IHC staining quantification images of MMP9, N-cadherin, E-cadherin and vimentin in each group (right), 8M-8P are the immunohistochemistry (IHC) images of peritoneal metastatic tumors in mice injected with HGC-27 cells with CT83 (8M) and MMP3 (8O) knockdown, and the IHC staining quantification images of MMP9, N-cadherin, E-cadherin and vimentin in each group (8N, 8P). DETAILED DESCRIPTION

[0064] The definitions of some terms used in this specification are provided below. Unless otherwise defined, all technical and scientific terms used herein generally have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0065] The present invention provides the use of a reagent for detecting the expression level of a biomarker in a sample in the preparation of a product for diagnosing gastric cancer / diagnosing gastric cancer metastasis / predicting the prognosis of gastric cancer, wherein the biomarker comprises any one or more of BUB1, SPC25, CT83 and MMP3.

[0066] In some embodiments, BUB1, SPC25, CT83 or MMP3 includes wild type, mutant or fragment thereof. The term encompasses full-length, unprocessed BUB1, SPC25, CT83 or MMP3, as well as any form of BUB1, SPC25, CT83 or MMP3 derived from processing in cells. The term encompasses naturally occurring variants (e.g., splice variants or allelic variants) of BUB1, SPC25, CT83 or MMP3. The term encompasses, for example, BUB1, SPC25, CT83 or MMP3 of human and any other vertebrate origin, including mammals, such as BUB1, SPC25, CT83 or MMP3 of primates and rodents (e.g., mice and rats), BUB1 gene ID: 699, SPC25 gene ID: 57405, CT83 gene ID: 203413, MMP3 gene ID: 4314.

[0067] The reagent also includes a detectable substance.

[0068] In some embodiments, a detectable substance refers to any substance that can be detected by fluorescent, spectroscopic, photochemical, biochemical, immunological, electrical, optical or chemical means. Preferably, such substances can be suitable for immunological detection (e.g., enzyme-linked immunosorbent assay, radioimmunoassay, fluorescent immunoassay, chemiluminescent immunoassay, etc.). Such substances are well known in the art and include, but are not limited to, enzymes (e.g., horseradish peroxidase, alkaline phosphatase, β-galactosidase, urease, glucose oxidase, etc.), radionuclides (e.g., 3 H. 125I , 35 S. 14 C or 32 P), fluorescent dyes (e.g., fluorescein isothiocyanate (FITC), fluorescein, tetramethylrhodamine isothiocyanate (TRITC), phycoerythrin (PE), Texas Red, rhodamine, quantum dots or cyanine dye derivatives (e.g., Cy7, Alexa 750)), acridinium ester compounds, magnetic beads, calorimetric labels such as colloidal gold or colored glass or plastic (e.g., polystyrene, polypropylene, latex, etc.) beads, and biotin for binding to avidin (e.g., streptavidin) modified with the above labels.

[0069] The present invention provides a model for diagnosing gastric cancer metastasis, wherein the indicators of the model include the levels of biomarkers, and the biomarkers include any one or more of BUB1, SPC25, CT83 and MMP3.

[0070] The indicators of the model also include T stage, N stage, and pathological type.

[0071] The model is a nomogram model.

[0072] In some embodiments, the nomogram model, also known as the Nomogram model, is a clinical prediction model that integrates multiple prediction indicators and then uses scaled line segments to draw them on the same plane at a certain ratio, so as to express the relationship between the various variables in the prediction model and evaluate the disease risk.

[0073] In some embodiments, a logistic regression formula is used to construct a model, the sample is a fresh frozen sample, and the calculation formula of the model is: probability of gastric cancer metastasis = [(4.855×4-mRNA)+(4.139×T stage)+(4.140×N stage)+(3.147×pathological type)+(-9.961)], and the model is visualized in the form of a nomogram. The nomogram model includes 7 straight lines, which are arranged from top to bottom and are parallel to each other. Each straight line represents a ruler with scales.

[0074] The first line is a score scale (Points), with a value range of 0-100.

[0075] The second line is the pathology scale (Pathology). When the pathology type is high / medium (High / Median), the score is 0, and when the pathology type is low / none (Low / None), the score is 74.

[0076] The third line is the T stage scale (T.stage). When the T stage is T2 / T3, the score is 0, and when the T stage is T4, the score is 92.

[0077] The fourth line is the N stage scale (N.stage). When the N stage is N0, the score is 0, and when the N stage is N+, the score is 92.

[0078] The fifth row is the 4-mRNA scale (X4.panel). When the 4-mRNA expression level is low (Low), the score is 0, and when the 4-mRNA expression level is high (High), the score is 100.

[0079] The sixth line is the total score scale (Total Points), with a value range of 0-400.

[0080] The seventh line is the risk of metastasis scale (Risk of metastasis), with a value range of 0-0.95.

[0081] In some embodiments, a logistic regression formula is used to construct a model, the sample is a peripheral blood sample, and the calculation formula of the model is: probability of gastric cancer metastasis = [(6.558×4-mRNA)+(8.467×T stage)+(8.054×N stage)+(8.164×pathological type)-14.876]. The model is visualized in the form of a nomogram. The nomogram model includes 7 straight lines, which are arranged from top to bottom and are parallel to each other. Each straight line represents a ruler with scales.

[0082] The first line is a score scale (Points), with a value range of 0-100.

[0083] The second line is the pathology scale (Pathology). When the pathology type is high / medium (High / Median), the score is 0, and when the pathology type is low / none (Low / None), the score is 97.

[0084] The third line is the T stage scale (T.stage). When the T stage is T2 / T3, the score is 0, and when the T stage is T4, the score is 100.

[0085] The fourth line is the N stage scale (N.stage). When the N stage is N0, the score is 0, and when the N stage is N+, the score is 98.

[0086] The fifth row is the 4-mRNA scale (X4.panel). When the 4-mRNA expression level is low (Low), the score is 0, and when the 4-mRNA expression level is high (High), the score is 90.

[0087] The sixth line is the total score scale (Total Points), with a value range of 0-450.

[0088] The seventh line is the risk of metastasis scale (Risk of metastasis), with a value range of 0.1-0.45.

[0089] The present invention provides a pharmaceutical composition, which comprises an inhibitor of a biomarker, wherein the biomarker comprises any one or more of BUB1, SPC25, CT83 and MMP3.

[0090] The pharmaceutical composition also includes other drugs.

[0091] In some embodiments, other drugs include other drugs for treating gastric cancer, including but not limited to chemotherapy drugs and targeted drugs. Chemotherapy drugs include oxaliplatin and fluorouracil preparations, such as 5-fluorouracil, or capecitabine, and S-1. In addition, targeted therapy drugs for treating gastric cancer include apatinib.

[0092] The pharmaceutical composition also includes a pharmaceutically acceptable carrier.

[0093] In some embodiments, a pharmaceutically acceptable carrier refers to a non-toxic material that does not interact with the action of the active ingredient of the pharmaceutical composition. The pharmaceutically acceptable carrier refers to a natural or synthetic, organic or inorganic component that is used in conjunction with the active ingredient to facilitate application. In some embodiments, a pharmaceutically acceptable carrier includes one or more compatible solid or liquid fillers, diluents or encapsulating materials, and the carrier is suitable for administration to a patient. The components of the pharmaceutical composition of the present application generally do not interact significantly to affect the desired drug efficacy.

[0094] The invention will be further described below in conjunction with specific examples. It should be understood that the specific embodiments described herein are presented by way of example and are not intended to limit the invention. The main features of the invention may be used in a variety of embodiments without departing from the scope of the invention.

[0095] Example

[0096] 1. Materials and Methods

[0097] Screening for mRNA biomarkers

[0098] Figure 1 The biomarker discovery and validation process for peritoneal metastasis of gastric cancer performed in this study is outlined. First, several datasets were analyzed for biomarker identification, including the GSE62254 cohort from the Gene Expression Omnibus (GEO) database, which consisted of 300 gastric cancer patients (121 with peritoneal metastasis and 179 without peritoneal metastasis). In addition, data from the Cancer Genome Atlas (TCGA) cohort, which included 412 gastric cancer tissues and 36 normal tissues, were utilized. In addition, transcriptome sequencing data of 17 LAGC patients (5 with peritoneal metastasis and 12 with primary lesions but without peritoneal metastasis) from the Fourth Hospital of Hebei Medical University (FHHMU) were collected.

[0099] To further validate the clinical applicability of these markers as predictors of peritoneal metastasis in gastric cancer, various specimens (fresh-frozen gastroscopic biopsies and peripheral blood samples) from multiple large clinical centers in China were utilized. Forty-two pairs of fresh-frozen cancer tissue specimens and 33 peripheral blood samples were collected from patients with and without peritoneal metastasis at the Fourth Hospital of Hebei Medical University (FHHMU) using a 1:1 propensity score matching method. The expression of selected mRNAs was evaluated using qRT-PCR. In addition, 36 pairs of formalin-fixed specimens, also matched by 1:1 propensity score, were analyzed to further elucidate the expression of 4-mRNA by immunohistochemistry. These samples were collected at different time periods at FHHMU.

[0100] Validation of mRNA biomarkers

[0101] First, a cohort of 313 gastric cancer patients from three treatment centers (FHHMU, Shijiazhuang People's Hospital (SJZPH), and Hengshui People's Hospital (HSCPH)) was used as a training set for predicting peritoneal metastasis. The external validation cohort included 131 patients from three other centers: Baoding Central Hospital (BDCH), Nanjing University Jinling Hospital (NJJLH), and Wuhan University People's Hospital (WHPH). All patients underwent laparoscopic exploration and peritoneal metastasis was confirmed by biopsy. In addition, data from 43 gastric cancer patients who received conversion therapy from the above six centers, including hyperthermic intraperitoneal chemotherapy (HIPEC) combined with neoadjuvant intraperitoneal and systemic (NIPS) paclitaxel, were collected. This was done to evaluate whether the prediction model can identify patients who may benefit from this conversion therapy.

[0102] To evaluate the ability of the nomogram to predict patients with gastric cancer positive for peritoneal exfoliated cells (GC-CY1), data from 159 patients from two centers, FHHMU and SJZPH, were analyzed. All patients underwent laparoscopic exploration and peritoneal exfoliated cytology, and GC-CY1 cases were confirmed by the presence of free cancer cells in the peritoneal cavity.

[0103] To transition the mRNA panel from tissue specimens to non-invasive liquid biopsies, serum samples from LAGC patients from six centers were retrospectively analyzed. The training cohort included 215 LAGC patients from three treatment centers (FHHMU, SJZPH, and HSCPH), while the validation cohort included serum samples from 127 patients from the remaining three centers (BDCH, NJJLH, WHPH). In addition, data from 41 patients treated with HIPEC combined with NIPS paclitaxel at six centers were collected to evaluate the predictive accuracy of treatment efficacy.

[0104] A prospective clinical observational study (NCT06478394) is also registered to evaluate the effectiveness of the mRNA panel in predicting peritoneal metastasis. This study will begin recruiting patients from January to June 2024. Peripheral blood samples will be collected from all participants for qRT-PCR analysis before laparoscopic exploration. The possibility of peritoneal metastasis will be predicted using a calibrated histogram, which will then be compared with the results of laparoscopic exploration.

[0105] All patients were monitored for recurrence or progression according to the gastric cancer treatment guidelines, using laboratory tests, endoscopy, and abdominal and pelvic CT. Tissue specimens from all patient groups were frozen in liquid nitrogen and stored at −80°C. Surgical specimens were processed according to the guidelines of the Chinese Society of Clinical Oncology, and tumor and lymph node staging was performed according to the 8th edition of the AJCC. All procedures adhered to the Declaration of Helsinki, and written informed consent was obtained from all participants and approved by the institutional review boards of the participating institutions.

[0106] Evaluation of treatment response

[0107] In this study, all patients with newly diagnosed gastric cancer underwent abdominal CT scans to assess tumor size. Except for those with pyloric obstruction, patients were instructed to drink 800-1000 ml of water 30 minutes before the scan. Scopolamine butylbromide was injected intramuscularly to reduce gastrointestinal motility, fully expand the gastric cavity, and prevent gastric wall thickening, thereby improving lesion localization and observation capabilities. This method can accurately identify target lesions on CT scans, so that tumor diameter and thickness can be accurately measured to evaluate changes after neoadjuvant therapy. For each patient, the axial image showing the largest lesion was selected to measure the longest tumor diameter, and perigastric lymph node lesions with a short diameter >1.5 cm were recorded. All measurements were performed by senior radiologists. The sum of the diameters of all target lesions at baseline (the longest diameter of non-lymph node lesions and the shortest axis of lymph node lesions) was used as a reference for subsequent evaluations, and the final measurement was the average of three measurements, which was used to evaluate the treatment effect before and after conversion therapy.

[0108] The tumor efficacy evaluation criteria were the "Effectiveness Evaluation Criteria in Solid Tumors" (RECIST) version 1.1, which divided the efficacy into complete remission (CR), partial remission (PR), stable disease (SD), and progressive disease (PD). CR was defined as the disappearance of all target lesions; PR was defined as a decrease of at least 30% in the sum of target lesion diameters compared with baseline; PD was defined as an increase of at least 20% in the sum of target lesion diameters compared with baseline; SD was a change that did not meet the PR or PD criteria. The objective response rate (ORR) is the proportion of patients achieving CR and PR, and the disease control rate (DCR) includes patients achieving CR, PR, or SD.

[0109] Pathological sections of patients who underwent radical surgical resection were reviewed by two pathologists, and tumor regression was graded using the tumor regression grade (TRG) scale according to the AJCC / CAP guidelines. TRG 0 was defined as the absence of residual tumor cells in multiple consecutive sections under the microscope; TRG 1 was defined as a few discrete clusters of tumor cells under the plasma membrane; TRG 2 was defined as the presence of fibrosis and residual tumor cell debris within the lesion; and TRG 3 was defined as the presence of stable nonfibrotic cell populations within the lesion.

[0110] Diagnostic criteria for tumor cachexia

[0111] The diagnostic criteria for tumor cachexia are as follows: (1) unintentional weight loss of more than 5% within 6 months; (2) European and American body mass index (BMI) < 20 kg / m 2 , Chinese body mass index <18.5kg / m 2 , and any weight loss >2% within 6 months; (3) Limb skeletal muscle index indicates sarcopenia (male <7.26kg / m 2 , female <5.45kg / m 2) and weight loss >2% (according to the criteria of the European Association for the Study of Palliative Care); (4) decreased food intake and / or systemic inflammation.

[0112] RNA extraction and gene expression analysis

[0113] Total RNA was extracted from fresh frozen surgical tissue using TRIzol reagent (Invitrogen, Frederick, MA, USA) according to the manufacturer's instructions. For serum samples, total RNA was isolated using the PAXgene Blood RNA kit (Qiagen, Hilden, Germany). The extracted RNA was then reverse transcribed into cDNA using the GoScript Reverse Transcription System kit (Promega) according to the manufacturer's protocol. Quantitative reverse transcription PCR (qRT-PCR) was then performed. 2^ -ΔΔCT Methods The relative expression levels of target genes were calculated and normalized with GAPDH as an internal reference, where ΔCT represents the difference between the CT values ​​of the target gene and GAPDH.

[0114] Protein-protein interaction (PPI) network analysis

[0115] The human PPI network was constructed using the STRING database (https: / / string-db.org). Gene set and pathway analysis was performed using Enrichr (https: / / maayanlab.cloud / Enrichr / ) on the four-gene candidate list.

[0116] Wound healing assay

[0117] The HGC-27 cell monolayer transfected with siRNA was scratched with a 10 μL pipette tip, and the scratch healing migration distance was measured under a microscope 24-48 hours later. The image taken at 0 hour was used as the control for normalization and calculation of the relative migration rate.

[0118] Transwell migration and invasion assays

[0119] Transwell inserts were prepared according to the manufacturer's instructions, either uncoated (for migration assays) or Matrigel-coated (for invasion assays). Transfected HGC-27 cells (4 × 10 4 ) were seeded into the upper chamber and incubated for 24-48 hours. Cells that had migrated or invaded to the lower surface were counted in five randomly selected microscopic fields.

[0120] Western Blotting

[0121] The transfected HGC-27 cells were lysed with RIPA buffer containing 1% PMSF, and the total protein was extracted. The denatured proteins were separated by 10% SDS-PAGE and transferred to PVDF membrane (Seven, Beijing). The membrane was detected with BUB1, SPC25, CT83, MMP3, and GAPDH antibodies, and the immunoreactive bands were detected with ECL Plus reagent (Solarbio).

[0122] Immunohistochemistry (IHC)

[0123] Tissue sections were dewaxed at 60°C for 2 h and treated with xylene. Antigen retrieval was performed using EDTA, and endogenous peroxidase activity was quenched with 3% hydrogen peroxide. Sections were then blocked with 5% bovine serum albumin (BSA) and incubated with primary antibodies overnight at 4°C. The next day, secondary antibodies were applied, and antigen detection was performed using diaminobenzidine (DAB), followed by nuclear counterstaining with hematoxylin. The percentage of positive cells was calculated using ImageJ software. The scoring criteria were based on the percentage of positive cells: 1 for ≤25%, 2 for 26%-50%, 3 for 51%-75%, and 4 for >75%. The staining intensity was also scored: 1 for no staining, 2 for light brown, 3 for medium brown, and 4 for dark brown. The final immunohistochemistry (IHC) score was obtained by multiplying the percentage of coverage and the staining intensity score.

[0124] Animal experiments

[0125] The in vivo experiments were approved by the Animal Laboratory Animal Care Committee (IACUC) of the Fourth Affiliated Hospital of Hebei Medical University (IACUC approval number: 20240255), and BALB / c mice were obtained from Beijing Spaifu Biotechnology Co., Ltd. To evaluate the effect of gene knockdown on peritoneal metastasis, HGC-27 cells transfected with lentivirus encoding siRNA targeting specific genes were intraperitoneally injected into mice. When there was a significant difference between the experimental group and the control group, the mice were killed and the results were statistically analyzed. Each mouse received only one treatment.

[0126] Statistical analysis

[0127] Statistical analyses were performed using IBM SPSS (v23), R (v3.6.3), and GraphPad Prism (v8.0). Significant clinicopathological variables and mRNA classifiers were identified by univariate and multivariate logistic regression analysis. Variables that showed significance in univariate analysis were included in multivariate regression. The Wilcoxon rank sum test with Bonferroni correction was used to examine differential gene expression between peritoneal and nonperitoneal metastasis groups in the discovery phase. In the clinical validation phase, a gene-based risk score was developed using logistic regression with backward elimination. Model performance was evaluated using receiver operating characteristic (ROC) curves and area under the curve (AUC) values, calculated by the pROC package in R. ROC curve comparisons were performed using the DeLong test. The sensitivity, specificity, positive and negative predictive values, precision, and accuracy of the 4-mRNA biomarker panel in all cohorts were determined using the reportROC package, and the results were presented in the form of a confusion matrix. The optimal ROC curve cutoff was determined using the Youden index. Individuals were divided into high-risk or low-risk groups using the Youden index and the median risk score for the prediction of peritoneal metastasis. Disease-free survival (DFS) was analyzed using the Kaplan-Meier method and was defined as the time from radical surgery to disease recurrence or death due to disease progression. Patients were reviewed at the 5-year milestone if they were alive and without recurrence, whereas patients who were lost to follow-up without evidence of recurrence within 5 years were assessed at the last visit. Statistical significance was set at P < 0.05.

[0128] 2. Experimental results

[0129] Identification of LAGC peritoneal metastasis-related mRNA

[0130] After analysis, four differentially expressed genes were identified: BUB1, SPC25, CT83, and MMP3 ( Figure 2 A). These genes were upregulated in cancer tissues with peritoneal metastasis compared with cancer tissues without peritoneal metastasis ( Figure 2 BD).

[0131] To this end, a pilot cohort was established at FHHMU to further validate the expression of 4-mRNA (BUB1, SPC25, CT83, and MMP3) in tissue and peripheral blood samples. To minimize the influence of clinical variables, a 1:1 propensity score matching method was used to select 42 pairs of primary cancer tissue samples with and without peritoneal metastasis. qRT-PCR results confirmed that the four candidate mRNAs were highly expressed in cancer tissues with peritoneal metastasis ( Figure 2 E), and their expression levels were significantly correlated with the clinical pathological characteristics of patients ( Figure 2 N). Similar results were observed in 33 matched peripheral blood samples ( Figure 2 F). In addition, Western blot analysis showed that these four genes were significantly upregulated in cancer tissues with peritoneal metastasis ( Figure 2 G, Figure 2 QT).

[0132] In addition, IHC staining of 36 pairs of patients (1:1 propensity score matching) was analyzed, and the results showed that the mRNA expression level in cancer tissues with peritoneal metastasis was significantly higher than that in cancer tissues without peritoneal metastasis (P<0.05) ( Figure 2 Follow-up data further showed that patients with high IHC expression had worse 5-year overall survival (OS) and disease-free survival (DFS) compared with patients with low IHC expression ( Figure 2 JM, Figure 2 UX). Kaplan-Meier survival analysis (https: / / kmplot.com / analysis / ) confirmed these results ( Figure 2 O). In addition, the protein-protein interaction network was constructed using the STRING database (https: / / www.string-db.org / ) and visualized using Cytoscape 3.9.1, which elucidated the potential roles of these mRNAs in gastric cancer ( Figure 2 P).

[0133] 4-mRNA predicts peritoneal metastasis in fresh-frozen samples of LAGC patients

[0134] First, qRT-PCR was used to quantitatively evaluate the expression of 4-mRNA (BUB1, SPC25, CT83, and MMP3) in fresh-frozen samples of 313 patients from three centers, FHHMU, SJZPH, and HSCPH, which constituted the training set (the 4-mRNA primer sequences are shown in Table 1). Multivariate logistic regression analysis showed that these four mRNAs were independent risk factors for peritoneal metastasis in patients with LAGC (all P<0.05). The effects of other clinical characteristics and 4-mRNA groups on the risk of peritoneal metastasis in patients with LAGC were further analyzed. The results showed that the 4-mRNA combination (OR=4.855, 95%CI: 2.508-9.396, P<0.001), T stage (OR=4.139, 95%CI: 1.389-12.329, P=0.011), N stage (OR=4.140, 95%CI:

[0135] 1.193-14.365, P = 0.025) and pathological type (OR = 3.147, 95% CI: 1.021-9.696, P = 0.046) were independent risk factors for peritoneal metastasis of LAGC ( Figure 3A). The probability of peritoneal metastasis was calculated using the logistic regression formula: [(4.855×4-mRNA)+(4.139×T stage)+(4.140×N stage)+(3.147×pathological type)+(-9.961)], with a threshold of 0.1373112. The model was visualized in the form of a nomogram, providing an intuitive method for predicting the risk of peritoneal metastasis in LAGC patients ( Figure 3 B).

[0136] Table 1 4-mRNA primer sequences

[0137] Name Sequence BUB1_F TGTCCTTCAATACATACAGTGGGT(SEQ ID NO:1) BUB1_R GGAACTCTCTGGGTTCAGCC(SEQ ID NO:2) SPC25_F AGCGAATGCAGAGAGGTTGA(SEQ ID NO:3) SPC25_R CCTCAAGATGAGGGGCACTA(SEQ ID NO:4) CT83_F ACTCCTAGCGAGCAGCATTC(SEQ ID NO:5) CT83_R CCCGAGAGAGGTCGTAGACT(SEQ ID NO:6) MMP3_F TGAGGACACCAGCATGAACC(SEQ ID NO:7) MMP3_R ACTTCGGGATCCAGGAAAG(SEQ ID NO:8)

[0138] Combining the 4-mRNA panel with clinical variables, a prediction model, called the risk stratification assessment (RSA) model, was developed, which showed strong predictive ability for peritoneal metastasis with an AUC of 0.836 (95% CI: 0.789-0.884, P < 0.001; Figure 3 C, 3E, 3H). DeLong test showed that the AUC of the RSA model was significantly higher than that of the clinical model in the training set (0.836 vs. 0.711; P = 0.001). In addition, the calibration curve of the model further emphasized its predictive accuracy ( Figure 3 F). According to the threshold determined by the Youden index, the patients were divided into low-risk and high-risk groups. The 3-year OS of high-risk patients was significantly worse than that of low-risk patients (38.6% vs. 57.2%, P < 0.001; Figure 3 K).

[0139] These findings were validated in a separate validation set of 131 LAGC patients from three additional centers, using the same statistical parameters as the training set. The RSA model based on 4-mRNA and clinical features had better predictive performance for peritoneal metastasis than either 4-mRNA or clinical features alone (RSA model: AUC = 0.882, 95% CI: 0.817-0.948, P < 0.001; 4-mRNA: AUC = 0.811, 95% CI: 0.714-0.907, P < 0.001; clinical features: AUC = 0.788, 95% CI: 0.702-0.873, P < 0.001), consistent with the training set results ( Figure 3 D, 3E, Figure 3 I).

[0140] Calibration curve analysis further confirmed the predictive accuracy of the model in the validation set. It is worth noting that the RSA model effectively distinguished the 3-year OS between the high-risk group and the low-risk group (41.4% vs. 60.0%, P < 0.001), supporting the results of the training set ( Figure 3L). In addition, stratified analysis based on HER2 and PDL1 expression levels of all patients in the training and validation sets showed that the RSA model was consistently superior to models based on clinical features or the 4-mRNA panel alone in predicting peritoneal metastasis ( Figure 3 MO).

[0141] In order to more clearly illustrate the clinical significance of the RSA model in predicting peritoneal metastasis, the results are presented in a double-layer concentric circle graph. In the training set, the graph shows that 50.8% of LAGC patients had a high risk of peritoneal metastasis, while 49.2% of patients had a low risk of peritoneal metastasis. After laparoscopic exploration, 15.3% (48 of 313) of high-risk patients and 3.5% (11 of 313) of low-risk patients were found to have peritoneal metastasis. It is worth noting that using the RSA model, only 2.2% (7 of 313) of low-risk patients were found to have peritoneal metastasis, while the incidence of high-risk patients was 16.6% (52 of 313) ( Figure 3 J), more low-risk patients were identified than the clinical feature model (55.3% vs. 49.2%). Consistent results were observed in the external validation set ( Figure 3 J below), further confirming the effectiveness of the RSA model in predicting peritoneal metastasis in patients with LAGC and its clinical relevance.

[0142] 4-mRNA predicts conversion therapy efficacy and cachexia in patients with gastric cancer peritoneal metastasis

[0143] This study systematically analyzed the treatment strategies of all patients with peritoneal metastasis in the training set and validation set. Specifically, 43 patients with peritoneal metastasis of gastric cancer who were treated with HIPEC combined with neoadjuvant peritoneal infusion and systemic chemotherapy (NIPS) were studied. The detailed treatment process is as follows Figure 4 As shown in A. Among these 43 patients, 16 (37.2%) were classified as low risk by the RSA model, and 27 (62.8%) were classified as high risk. The waterfall chart showed that the objective response rate (ORR) in the high-risk group was 3.7% (1 / 27), and the disease control rate (DCR) was 48.1% (13 / 27). In contrast, the ORR in the low-risk group was 62.5% (10 / 16), and the DCR was 87.5% (14 / 16), and the difference was statistically significant (ORR: P<0.001, DCR: P=0.006) ( Figure 4 B, 4G, 4I). It is worth noting that the success rate of conversion therapy in the low-risk group was higher than that in the high-risk group ( Figure 4 H), the proportion of patients reaching TRG 1 was higher in the low-risk group ( Figure 4 K). Figure 4DF summarized the comparison of pathological regression grade, changes in local lesions on abdominal CT, and changes in peritoneal metastasis in the two risk groups before and after conversion therapy based on the RSA model. ROC curve analysis further confirmed that the RSA model was effective in predicting the short-term efficacy of conversion therapy (AUC = 0.866, Figure 4 L) and the success rate of conversion therapy (AUC = 0.770, Figure 4 M) was superior to the clinical features and 4-mRNA groups.

[0144] Several studies have shown that patients with advanced gastric cancer and distant metastasis are prone to cachexia during treatment, which has a significant impact on clinical efficacy. Therefore, it is crucial to identify patients with peritoneal metastasis of gastric cancer who are at risk of cachexia during conversion therapy. To this end, the existing cachexia classification consensus guidelines were used to measure the skeletal muscle of 43 patients with peritoneal metastasis of gastric cancer who underwent conversion therapy ( Figure 4 C). The cachexia rate of high-risk patients identified by the RSA model was 63.0% (17 / 27), while the cachexia rate of low-risk patients was 31.3% (5 / 16), and the difference was statistically significant (P = 0.044) ( Figure 4 GH, 4J). Further ROC curve analysis supported these findings, indicating that the RSA model performed better than clinical features (AUC = 0.680) and the 4-mRNA panel (AUC = 0.714) in predicting cachexia (AUC = 0.837) ( Figure 4 N). In addition, the 3-year OS of high-risk patients in the RSA model was significantly worse than that of low-risk patients (14.8% vs. 37.5%, P = 0.021; Figure 4 O).

[0145] 4-mRNA predicts peritoneal free cancer cells in fresh frozen specimens of LAGC patients

[0146] Gastric cancer with positive abdominal lavage fluid cytology (GC-CY1) is a unique type of peritoneal metastasis, currently classified as stage IV in the UICC / AJCC TNM staging system, 8th edition, and the Japanese gastric cancer classification, 15th edition. GC-CY1 is characterized by the absence of visible peritoneal metastasis in the abdominal cavity, while cancer cells are detected in the peritoneal lavage fluid ( Figure 5 A). It was hypothesized that the RSA model could also predict GC-CY1 cases. To test this, 159 LAGC patients who underwent laparoscopic exploration and abdominal exfoliative cytology were recruited. The performance of the RSA model in predicting GC-CY1 patients (AUC = 0.888, 95% CI: 0.834-0.941) was significantly better than clinical features (AUC = 0.761, 95% CI: 0.684-0.838) and 4-mRNA (AUC = 0.830, 95% CI: 0.751-0.909) ( Figure 5 B).

[0147] The confusion matrix and radar plots further showed that the RSA model had superior discrimination ability, sensitivity, and specificity in predicting GC-CY1 compared with clinical features and the 4-mRNA panel ( Figure 5 DE). Calibration curve analysis also verified the prediction accuracy of the model in the validation set ( Figure 5 C). The two-layer concentric circle plot showed that the RSA model identified more patients in the low-risk group (61.6% vs. 53.3%) and a higher proportion of GC-CY1 patients in the high-risk group (18.9% vs. 17.4%) compared with the clinical model ( Figure 5 F). Consistent with the previous cohort follow-up results, the patients were stratified into high-risk group and low-risk group according to the nomogram. The results showed that the 3-year OS of the high-risk group was significantly lower than that of the low-risk group (41.0% vs. 62.2%, P = 0.0019) ( Figure 5 G).

[0148] To verify that the RSA model can also predict the efficacy of conversion therapy in GC-CY1 patients, two prospective neoadjuvant treatment cohorts (NCT03718624 (n=36) and ChiCTR1800014817 (n=38)) conducted by FHHMU were included. In both cohorts, the results showed that patients classified as high risk by the RSA model had significantly worse prognosis in terms of ORR, DCR, and conversion therapy success rate compared with low-risk patients (ChiCTR1800014817: ORR 23.5% vs. 76.2%, P = 0.004; DCR 76.5% vs. 100.0%, P = 0.032; conversion success rate 11.6% vs. 66.7%, P = 0.001; NCT03718624: ORR 57.1% vs. 95.5%, P = 0.008; DCR 85.7% vs. 100.0%, P = 0.144; conversion success rate 50.0% vs. 95.5%, P = 0.003) ( Figure 5 H, 5J). In addition, this difference still existed in the later follow-up, and the incidence of cachexia in high-risk patients in both groups was higher than that in low-risk patients (ChiCTR1800014817: 70.6% vs. 33.3%, P = 0.022, Figure 5 I; NCT03718624: 78.6% vs. 27.3%, P = 0.003, Figure 5 K).

[0149] 4-mRNA predicts peritoneal metastasis in peripheral blood samples of LAGC patients

[0150] First, 42 LAGC peripheral blood samples were randomly selected for quality control analysis to confirm that the A260 / 280 and A260 / 230 ratios were within the normal range at different time points and suitable for qRT-PCR analysis ( Figure 6 A). Using a training set of 215 samples from FHHMU, SJZPH, and HSCPH, multivariate logistic regression analysis based on 4-mRNA and clinical characteristics was applied to construct a nomogram for predicting peritoneal metastasis in LAGC patients ( Figure 6 B). The logistic regression coefficient was used to calculate the risk probability of peritoneal metastasis: [(6.558×4-mRNA group)+(8.467×T stage)+(8.054×N stage)+(8.164×pathological type)-14.876], with a threshold of 0.1838748.

[0151] The ROC curve analysis of the training set showed that the AUC for predicting peritoneal metastasis by clinical features was 0.746 (95% CI: 0.677-0.816), and the AUC for predicting peritoneal metastasis by the 4-mRNA group was 0.778 (95% CI: 0.694-0.863), both lower than the AUC of the RSA model of 0.857 (95% CI: 0.803-0.911) ( Figure 6 C). Confusion matrix and radar plot further showed that the RSA model was superior to the clinical and 4-mRNA models in terms of sensitivity, specificity, and accuracy in predicting peritoneal metastasis ( Figure 6 E, 6H). Calibration curve analysis confirmed the good fit of the RSA model ( Figure 6 F). Double-layer concentric circle analysis showed that the RSA model had greater clinical benefit than the single model. The detection rate of no peritoneal metastasis in the high-risk group decreased from 36.8% to 21.9%, and the detection rate of no peritoneal metastasis in the low-risk group increased from 45.1% to 60.0% ( Figure 6 J). When divided into low-risk group and high-risk group according to Youden index, the 3-year OS of patients in the high-risk group was significantly lower than that in the low-risk group (29.6% vs. 68.7%, P<0.001) ( Figure 6 K).

[0152] These findings were further validated in an external validation set from three additional diagnostic and treatment centers. ROC curve analysis showed that the RSA model had the highest AUC among the three models (AUC = 0.883, 95% CI: 0.824-0.941, P < 0.001) ( Figure 6 D). Radar plot and confusion matrix analysis confirmed that the RSA model was superior to the 4-mRNA and clinical feature models in terms of sensitivity, specificity, and accuracy in predicting peritoneal metastasis ( Figure 6 E, 6I). The calibration curve analysis also demonstrated the excellent predictive performance of the RSA model ( Figure 6 G). These results suggest that 4-mRNA, originally developed for tissue analysis, can be effectively applied to liquid biopsy testing of peripheral blood samples. The RSA model combines serum 4-mRNA data with clinical features to effectively predict peritoneal metastasis in patients with LAGC. The double-layer concentric circle diagram further confirms the potential of the RSA model in enhancing clinical decision-making and optimizing patient management ( Figure 6 J below), in addition, follow-up data showed that the 3-year OS rate in the high-risk group was significantly lower than that in the low-risk group (34.0% vs. 62.2%, P = 0.0028) ( Figure 6 L).

[0153] To further evaluate the potential of the RSA model as a tool to predict the efficacy of conversion therapy in patients with peritoneal metastasis from gastric cancer, 41 patients who received HIPEC combined with NIPS were analyzed. Figure 6 M. Notably, models based on clinical features or 4-mRNA alone were unable to effectively identify patients who might benefit from conversion therapy. In contrast, the RSA model successfully distinguished high-risk from low-risk groups and identified the risk of cachexia ( Figure 6 NP).

[0154] 4-mRNA predicts GC-CY1 and peritoneal recurrence in peripheral blood samples of LAGC patients

[0155] 76 patients from FHHMU were analyzed. Of these, 15 patients (19.7%) were diagnosed with GC-CY1 after laparoscopic exploration and peritoneal lavage fluid examination. Using the nomogram developed from the training set, the RSA model (AUC = 0.817, 95% CI: 0.714-0.921) was validated to be superior to the clinical feature model (AUC = 0.686, 95% CI: 0.548-0.825) and 4-mRNA (AUC = 0.789, 95% CI: 0.667-0.911) in predicting GC-CY1. Figure 7 A). Calibration curve ( Figure 7 B) and the confusion matrix ( Figure 7 C, 7F) further confirmed that the RSA model had higher sensitivity, specificity, and accuracy compared with the clinical feature model and the 4-mRNA group. In addition, the RSA model was more effective than other models in accurately identifying high-risk patients ( Figure 7 E). In addition, the Log-rank test showed that the 3-year OS of the high-risk group was significantly lower than that of the low-risk group (36.4% vs. 55.8%, P = 0.023) ( Figure 7 D).

[0156] To verify that the RSA model can also predict postoperative peritoneal recurrence in patients with LAGC, data from 148 patients from six diagnostic and treatment centers were collected, and all patients underwent abdominal CT or PET-CT examinations after surgery. Figure 7 As shown in G, the longitudinal dynamic peripheral blood test of a patient showed that the RSA model could indicate a high risk of peritoneal recurrence 16 months in advance. ROC curve analysis showed that the AUC value of the RSA model for predicting peritoneal recurrence (AUC = 0.827, 95% CI: 0.757-0.897) was superior to the clinical feature model (AUC = 0.758, 95% CI: 0.653-0.862) and the 4-mRNA combination (AUC = 0.721, 95% CI: 0.633-0.810) ( Figure 7 I). The calibration curve also showed good model fit ( Figure 7 R).

[0157] Radar plots and confusion matrices confirmed that the RSA model was significantly superior to the clinical feature model and 4-mRNA ( Figure 7 K, 7M). The double-layer concentric circle diagram further emphasizes the potential of the RSA model to enhance clinical decision-making and optimize patient management ( Figure 7 L). In addition, follow-up data showed that the 3-year OS rate of the high-risk group identified by the RSA model was significantly lower than that of the low-risk group (34.6% vs. 54.2%, P = 0.0029) ( Figure 7 P).

[0158] The performance of the RSA model in predicting peritoneal metastasis in patients with negative peripheral blood tumor markers (CA19-9, CA72-4, CEA) was also evaluated. Data from all tumor marker-negative patients from six diagnostic and treatment centers were pooled, with a total of 92 LAGC patients. Compared with a single model, the RSA model showed the highest AUC value in predicting peritoneal metastasis (AUC = 0.782, 95% CI: 0.683-0.881) ( Figure 7 H). Further analysis showed that the RSA model was superior to the clinical features and 4-mRNA model in terms of sensitivity, specificity, and accuracy ( Figure 7 J, 7M). The calibration curves also confirmed the excellent predictive performance of the RSA model ( Figure 7 S). Clinical benefit analysis showed that compared with the clinical characteristics model, the RSA model increased the detection rate of peritoneal metastasis in the high-risk group from 15.2% to 18.4%, and reduced the detection rate of peritoneal metastasis in the low-risk group from 3.3% to 0.0% ( Figure 7 In addition, long-term follow-up showed that the 3-year OS of high-risk patients identified by the RSA model was significantly lower than that of low-risk patients (29.6% vs. 56.2%, P = 0.0022) ( Figure 7 Q).

[0159] In addition, an observational prospective clinical study (NCT06478394) was launched to further validate the performance of the 4-mRNA panel in predicting peritoneal metastasis. This study included 120 LAGC patients enrolled from January to June 2024. Peripheral blood samples were collected from all patients for 4-mRNA testing before laparoscopic exploration. Among them, 20 patients (16.7%) were classified as high risk by the RSA model, and 100 patients (83.3%) were classified as low risk. Laparoscopic exploration later confirmed that 28 patients (23.3%) had peritoneal metastasis. In the high-risk group, only 3 patients (15.0%) had no peritoneal metastasis, while 11 patients (11.0%) in the low-risk group were found to have peritoneal metastasis ( Figure 7 N). Further ROC curve analysis showed that in this prospective study, the AUC of the RSA model for predicting peritoneal metastasis was higher than that of a single model ( Figure 7 O).

[0160] Inhibition of 4-mRNA inhibits gastric cancer metastasis

[0161] Based on these findings, the clinical potential of the 4-mRNA panel for predicting peritoneal metastasis in LAGC patients was explored, and the biological roles of the 4-mRNAs (BUB1, SPC25, CT83, and MMP3) that comprise this panel were analyzed in depth. Specific siRNA sequences targeting these genes were utilized (Table 2). Wound healing and transwell assays showed that silencing BUB1, SPC25, CT83, and MMP3 significantly reduced the migration and invasion abilities of HGC-27 cells ( Figure 8 AH).

[0162] Table 2 siRNA sequences of 4-mRNA

[0163]

[0164] In addition, knockdown experiments targeting BUB1, SPC25, CT83, and MMP3 were performed in BALB / c nude mice to examine the effects of these genes on the peritoneal metastasis of HGC-27 cells ( Figure 8 I). Compared with the control group, the peritoneal metastasis in the siRNA treatment group was significantly reduced ( Figure 8 J). In addition, IHC staining showed that silencing these genes resulted in decreased expression of metastatic (MMP9) and mesenchymal markers (N-cadherin, vimentin) and increased expression of epithelial markers (E-cadherin) ( Figure 8 KP).

[0165] Overall, these in vitro and in vivo findings suggest that four genes in the 4-mRNA group promote gastric cancer cell migration, invasion, and intraperitoneal metastasis.

[0166] The description of the above embodiments is only used to understand the method and core idea of ​​the present invention. It should be pointed out that, for those skilled in the art, several improvements and modifications can be made to the present invention without departing from the principle of the present invention, and these improvements and modifications will also fall within the scope of protection of the claims of the present invention.

Claims

1. Use of a reagent for detecting the expression level of a biomarker in a sample in the preparation of a product for diagnosing gastric cancer / diagnosing gastric cancer metastasis / predicting the prognosis of gastric cancer, wherein the biomarker comprises any one or more of BUB1, SPC25, CT83 and MMP3; Preferably, the gastric cancer metastasis is gastric cancer peritoneal metastasis; Preferably, the prediction of gastric cancer prognosis includes prediction of gastric cancer survival, prediction of gastric cancer treatment efficacy, prediction of gastric cancer recurrence, and prediction of gastric cancer cachexia; Preferably, the survival period includes overall survival rate and disease-free survival rate; Preferably, the gastric cancer treatment efficacy also includes the success rate of conversion therapy; Preferably, the gastric cancer recurrence is peritoneal recurrence of gastric cancer.

2. The use according to claim 1, characterized in that: The reagents include oligonucleotide probes that specifically recognize the biomarker gene, primers that specifically amplify the biomarker gene, or binding agents that specifically bind to proteins encoded by the biomarker gene; Preferably, the primer sequence for specifically amplifying BUB1 is as shown in SEQ ID NO: 1-2; Preferably, the primer sequence for specifically amplifying SPC25 is as shown in SEQ ID NO:3-4; Preferably, the primer sequences for specific amplification of CT83 are shown in SEQ ID NO: 5-6; Preferably, the primer sequences for specific amplification of MMP3 are shown in SEQ ID NO:7-8; Preferably, the sample includes tissue or blood; Preferably, the blood includes peripheral blood and serum.

3. A product for diagnosing gastric cancer / diagnosing gastric cancer metastasis / predicting gastric cancer prognosis, characterized in that: The product includes reagents for detecting the expression level of biomarkers in a sample, wherein the biomarkers include any one or more of BUB1, SPC25, CT83, and MMP3; Preferably, the reagent further comprises a detectable substance; Preferably, the product includes a test kit and a test paper.

4. Application of biomarkers in constructing a model for diagnosing gastric cancer metastasis, wherein the biomarkers include any one or more of BUB1, SPC25, CT83 and MMP3; Preferably, the indicators of the model also include T stage, N stage, and pathological type; Preferably, the gastric cancer metastasis is gastric cancer peritoneal metastasis.

5. A model for diagnosing gastric cancer metastasis, characterized in that: The indicators of the model include the levels of biomarkers, and the biomarkers include any one or more of BUB1, SPC25, CT83 and MMP3; Preferably, the indicators of the model also include T stage, N stage, and pathological type; Preferably, the model is a nomogram model; Preferably, the calculation formula of the model includes: The probability of gastric cancer metastasis = [(4.855×4-mRNA)+(4.139×T stage)+(4.140×N stage)+(3.147×pathological type)+(-9.961)], or The probability of gastric cancer metastasis = [(6.558 × 4-mRNA) + (8.467 × T stage) + (8.054 × N stage) + (8.164 × pathological type) - 14.876]; Preferably, the gastric cancer metastasis is gastric cancer peritoneal metastasis.

6. Use of a biomarker inhibitor in the preparation of a pharmaceutical composition for treating gastric cancer, wherein the biomarker comprises any one or more of BUB1, SPC25, CT83 and MMP3; Preferably, the inhibitor includes a nucleic acid inhibitor and a protein inhibitor; Preferably, the nucleic acid inhibitor includes siRNA, shRNA, ribozyme; Preferably, the nucleic acid inhibitor is selected from siRNA; Preferably, the siRNA sequence of BUB1 is shown in SEQ ID NO:9 or SEQ ID NO:10; Preferably, the siRNA sequence of SPC25 is shown in SEQ ID NO: 11 or SEQ ID NO: 12; Preferably, the siRNA sequence of CT83 is shown in SEQ ID NO: 13 or SEQ ID NO: 14; Preferably, the siRNA sequence of MMP3 is shown in SEQ ID NO: 15 or SEQ ID NO: 16; Preferably, the treatment of gastric cancer includes inhibiting gastric cancer metastasis; Preferably, the inhibiting gastric cancer metastasis includes inhibiting gastric cancer peritoneal metastasis.

7. A pharmaceutical composition, characterized in that The pharmaceutical composition includes an inhibitor of a biomarker, wherein the biomarker includes any one or more of BUB1, SPC25, CT83, and MMP3; Preferably, the pharmaceutical composition further comprises a pharmaceutically acceptable carrier.

8. Use of biomarkers as targets in screening candidate drugs for treating gastric cancer, wherein the biomarkers include any one or more of BUB1, SPC25, CT83 and MMP3; Preferably, the method for screening candidate drugs for treating gastric cancer comprises: The effect of the test substance on the expression level of the biomarker in the sample is tested, wherein after using the test substance, the level of the biomarker decreases, indicating that the test substance is a candidate drug for treating gastric cancer.

9. Use of a biomarker inhibitor in the preparation of a product for regulating gene / protein levels, wherein the biomarker comprises any one or more of BUB1, SPC25, CT83 and MMP3, and the gene / protein comprises any one or more of MMP9, N-cadherin, vimentin and E-cadherin; Preferably, the inhibitor includes a nucleic acid inhibitor and a protein inhibitor; Preferably, the nucleic acid inhibitor includes siRNA, shRNA, ribozyme; Preferably, the nucleic acid inhibitor is selected from siRNA.

10. Any of the following methods: (1) A method for screening candidate drugs for treating gastric cancer, characterized in that: The method comprises: testing the effect of a test substance on the expression level of a biomarker in a sample, wherein after using the test substance, the level of the biomarker decreases, indicating that the test substance is a candidate drug for treating gastric cancer, and the biomarker comprises any one or more of BUB1, SPC25, CT83 and MMP3; (2) A method for inhibiting gastric cancer cell metastasis in vitro, characterized in that the method comprises administering an inhibitor of a biomarker, wherein the biomarker comprises any one or more of BUB1, SPC25, CT83 and MMP3; (3) A method for regulating gene / protein expression in vitro, characterized in that the method comprises administering an inhibitor of a biomarker, wherein the biomarker comprises any one or more of BUB1, SPC25, CT83 and MMP3, and the gene / protein comprises any one or more of MMP9, N-cadherin, vimentin, and E-cadherin.

Images