Diagnostic model for predicting gastric cancer CAR-T cellular immunotherapy prognosis and application thereof

Abstract

The invention relates to the field of medical detection, and particularly provides a diagnostic model for predicting prognosis of gastric cancer CAR-T cellular immunotherapy and application of the diagnostic model. According to the present invention, the content and the spatial distribution map of the CLDN18.2 positive gastric cancer tissue sample microenvironment component are systematically analyzed by integrating the multiple fluorescence immunohistochemical detection and the full-slice digital imaging quantitative analysis method for the first time; through multi-dimensional evaluation of a tumor microenvironment under the treatment background of a targeted CLDN18.2 chimeric antigen receptor T cell immunotherapy CT041, a diagnosis model for predicting the prognosis of the gastric cancer CAR-T cell immunotherapy is established for the first time, accurate screening of benefited people and risk stratification before treatment are realized, treatment selection is guided, and the treatment effect is improved.

Description

Technical Field

[0001] The present invention relates to the field of medical detection. Specifically, the present invention relates to a diagnostic model for predicting the prognosis of gastric cancer CAR-T cell immunotherapy and its application. Background Art

[0002] Chimeric antigen receptor T cell (CAR-T) therapy is a revolutionary individualized treatment innovation technology that genetically engineers patients' autologous T cells to specifically recognize and target tumor cells. With its remarkable efficacy and durable anti-tumor ability (especially in hematological malignancies), CAR-T therapy has become a research hotspot in recent years. However, the application of CAR-T therapy in solid tumors is still in its infancy, and the treatment effect is extremely limited, showing a significant gap compared with hematological tumors.

[0003] Claudin18.2 (CLDN18.2), a member of the tight junction protein family, is restricted in expression and only localized to gastric mucosal tissue in a healthy state. However, the disruption of the tight junction structure between cells during tumorigenesis leads to the abnormal exposure of CLDN18.2 epitopes on the cell surface (especially in digestive tract tumors). This specific expression pattern makes it a potential therapeutic target, which has rapidly promoted the research and development of therapies such as antibody-drug conjugates, monoclonal / bispecific antibodies, etc. However, the limited clinical benefits of existing CLDN18.2-targeted therapies, the strict inclusion criteria (such as the high expression threshold of CLDN18.2), and the poor tolerance caused by treatment-related adverse events (such as vomiting) all restrict clinical applications, and innovative strategies are urgently needed to optimize the efficacy.

[0004] Our team previously reported the positive efficacy and controllable safety of CLDN18.2-specific CAR-T therapy (hereinafter referred to as CT041) in the treatment of digestive tract tumors in a phase I clinical trial (NCT03874897). For refractory / recurrent advanced gastric cancer, the objective response rate and disease control rate of CT041 reached 54.9% and 96.1% respectively, significantly superior to existing therapies, and the median progression-free survival and overall survival reached 5.8 months and 9.0 months respectively. The success of CT041 achieved a breakthrough in the treatment of solid tumors with CAR-T therapy and established its potential as a new strategy for the treatment of solid tumors.

[0005] Although the efficacy of CT041 is clear, some patients still have primary drug resistance, and about 50% of patients experience progression within 6 months after obtaining initial disease control. The highly heterogeneous and complex tumor microenvironment (TME) of CLDN18.2-positive gastric cancer may be a key factor affecting treatment response and prognosis. However, there has been no study on the multi-dimensional analysis of the cellular components and spatial relationships of the TME in patients receiving CT041 treatment. Traditional immunobiological analysis still remains at the stage of analyzing the content of single / few markers. More importantly, there is currently a lack of methods or combined indicators for predicting the efficacy and prognosis of patients receiving CT041 treatment, which hinders the improvement of treatment efficacy and patient prognosis. Summary of the Invention

[0006] To fill the gaps in the prior art, the present invention provides the following technical solutions: In the first aspect of the present invention, a method for constructing a diagnostic model for predicting the prognosis of CAR-T cell immunotherapy for gastric cancer is provided. The method includes the following steps: 1) Collect sufficient gastric cancer tissue specimen samples from patients with gastric cancer treated with CAR-T cells as the development cohort group; 2) Follow up the patients in the development cohort group for more than 166 days, and evaluate whether the patients are treatment responders or non-responders through imaging; 3) Detect the differences in the tumor microenvironment of the patients in the development cohort group by mIHC, use univariate Cox proportional hazards regression analysis to evaluate the relationship between cellular components and progression-free survival, and determine the percentage of TGF-β1+IQGAP3+ cells, the percentage of CD8+TCF1+GZMK+ cells, and the effective score between TGF-β1+IQGAP3+ cells and FAP+ cells as the calculation variables for the CAR-T treatment efficacy and prognosis model after screening; 4) Calculate the weights for each index through the univariate Cox proportional hazards regression algorithm, and construct a prediction model for CAR-T treatment efficacy and prognosis by means of weighted sum: CTRS = [Ln(3.42) (Normalized percentage of TGF-β1+IQGAP3+ cells)+ Ln(0.21) (Normalized percentage of CD8+TCF1+GZMK+ cells)+ Ln(1.46) (Normalized effective score between TGF-β1+IQGAP3+ cells and FAP+ cells)] (-1).

[0007] In one embodiment, patients with a progression-free survival of ≥166 days were defined as treatment responders, and the rest were treatment non-responders.

[0008] In one embodiment, the effective score was defined as the ratio of the number of pairs of central cells within a radius of 20 micrometers to peripheral cells to the total number of central cells.

[0009] In one embodiment, the CTRS threshold / cut-off value was -0.68. When CTRS was greater than the threshold, it was determined to be responsive to CAR-T treatment; conversely, when CTRS was less than the threshold, it was determined to be non-responsive to CAR-T treatment.

[0010] In one embodiment, CTRS was divided into high and low score groups according to the cut-off value. The progression-free survival and overall survival of the high CTRS score group were both higher than those of the non-response group.

[0011] The second aspect of the present invention provides a model for predicting the prognosis of CAR-T cell immunotherapy for gastric cancer, and the model is: CTRS = [Ln(3.42) (Normalized percentage of TGF-β1+IQGAP3+ cells)+ Ln(0.21) (Normalized percentage of CD8+TCF1+GZMK+ cells)+ Ln(1.46) (Normalized effective score between TGF-β1+IQGAP3+ cells and FAP+ cells)] (-1); wherein, CAR-T is the treatment response score (CAR-T Response Score, CTRS); Ln(X) is the logarithm of X with the mathematical constant e as the base.

[0012] In one embodiment, the CTRS threshold / cut-off value was -0.68. When CTRS was greater than the threshold, it was determined to be responsive to CAR-T treatment; conversely, when CTRS was less than the threshold, it was determined to be non-responsive to CAR-T treatment.

[0013] The third aspect of the present invention provides the use of the prognostic model constructed according to the foregoing construction method for evaluating the prognosis of CAR-T cell immunotherapy for gastric cancer.

[0014] The fourth aspect of the present invention provides the use of the foregoing model for evaluating the prognosis of CAR-T cell immunotherapy for gastric cancer.

[0015] Compared with the prior art, the present invention has the following remarkable progress: 1. For the first time, the present invention systematically analyzed the content and spatial distribution map of the microenvironment components in CLDN18.2-positive gastric cancer tissue samples by integrating mIHC and whole-slide digital imaging quantitative analysis methods.

[0016] 2. For the first time, the present invention constructed a model based on the spatial omics characteristics of CLDN18.2-positive gastric cancer TME to predict the prognosis of patients receiving CAR-T therapy.

[0017] 3. Through the multidimensional evaluation of the TME under the background of the targeted CLDN18.2 chimeric antigen receptor T cell immunotherapy CT041, the present invention for the first time established a set of ideas and methods for screening gastric cancer patients responsive to CT041, realizing the precise screening of the benefited population before treatment, guiding treatment selection, and improving the treatment efficacy. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The drawings are used to provide a further understanding of the present invention, and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention, but do not constitute a limitation to the present invention. In the drawings: Figure 1 . Schematic diagram for calculating the CTRS score and multi-channel and single-channel staining results of multiplex fluorescence immunohistochemical staining; Figure 2 . Cell counting results of multiplex fluorescence immunohistochemical staining; Figure 3 . Image analysis process for multiplex fluorescence immunohistochemical detection; Figure 4 . Spatial analysis principle (A) and example for calculating the effective score (B); Figure 5 . Difference in CTRS scores between the response group and the non-response group; Figure 6 . Waterfall plot visualizing the CTRS score of each patient and its association with the CAR-T treatment response; Figure 7 . Evaluating the predictive efficacy of the CTRS score for the CAR-T treatment effect based on the receiver operating characteristic curve and the area under the curve; Figure 8 . Predictive efficacy of the CTRS score for the progression-free survival of gastric cancer patients receiving CAR-T therapy; Figure 9 . Predictive efficacy of the CTRS score for the overall survival of gastric cancer patients receiving CAR-T therapy. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0019] The preferred embodiments of the present invention will be described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only for the purpose of illustrating and explaining the present invention, and are not intended to limit the present invention.

[0020] Example 1 Content and spatial distribution of microenvironment components in CLDN18.2-positive gastric cancer tissue samples were analyzed by the system (discovery set) 1. Specimen collection

[0021] The inclusion criteria were as follows: (1) histologically confirmed advanced gastric cancer; (2) positive expression of CLDN18.2 (immunohistochemical staining intensity ≥2+ and positive tumor cell ratio ≥40%); (3) patients who had failed at least first-line standard treatment.

[0022] The exclusion criteria included: (1) target lesions ≥4 cm in size at the time of single lymphocyte collection; (2) active infection; (3) extensive metastatic lesions in the lungs, liver, and brain.

[0023] A total of 48 formalin-fixed paraffin-embedded (FFPE) gastric cancer tissue specimens from 48 histologically confirmed gastric cancer patients collected from March 2019 to April 2021 were finally included. Among the 48 included patients, 42 (87.5%) had received at least second-line treatment or at least triple treatment with fluoropyrimidine, oxaliplatin, and paclitaxel, and 20 (41.7%) had previously received anti-PD-1 / PD-L1 antibody treatment. Among them, 24 tissue specimens were randomly selected as discovery set samples, and 24 tissue specimens were used as validation set samples. This study was approved by the Ethics Committee of Peking University Cancer Hospital, and all participants or their legal guardians signed informed consent forms.

[0024] Primary tumor samples were obtained by surgical resection or endoscopic biopsy before CAR-T treatment. The efficacy was evaluated according to the RECIST 1.1 criteria. Progression-free survival (PFS) was defined as the time from the first infusion to disease progression for any reason, and overall survival (OS) was the time from the first infusion to death. Patients with PFS ≥166 days (the median PFS of the cases included in this study) were defined as treatment responders, and the rest were treatment non-responders.

[0025] 2. Multiplex fluorescence immunohistochemistry (mIHC): A. Tumor immune panel setting:

[0026] a. Structural composition of tumor tissues and cells: Used for tumor and stroma segmentation, including cytokeratin for cell membrane labeling and 4',6-diamidino-2-phenylindole for nuclear labeling; b. Markers for CD8+ T cell subsets: Cell identity marker CD8, immune checkpoint markers LAG-3 and PD-1, T cell killing marker GZMK, T cell activity marker CXCL13, T cell exhaustion markers TCF1 and CD39.

[0027] B. Technical process: To study the infiltration characteristics of each component in the TME, we quantitatively analyzed the content and spatial distribution relationship of cell components in 24 discovery set FFPE specimens by mIHC staining. The specific steps are as follows: a. Baking the slides: Place the FFPE samples in an oven at 65 °C for 20 minutes; b. Deparaffinization: Use xylene for deparaffinization, 5 minutes each time, repeat 3 times; c. Hydration: Infiltrate the slides with gradient ethanol, 5 minutes in 100% ethanol, 5 minutes in 95% ethanol, and 2 minutes in 70% ethanol respectively; d. Antigen retrieval: Immerse the slides in an alkaline antigen retrieval solution (PANOVUE, Cat.0019020500), heat to boiling with high power in a microwave and then switch to low power to maintain for 15 minutes, and cool naturally to room temperature; e. Blocking non-specific binding sites: Dropwise add a primary antibody blocking solution (PANOVUE; Cat.0018001120), incubate with shaking at room temperature for 10 minutes; f. Incubation with primary antibody: Taking GZMK as an example, use a GZMK monoclonal antibody (ab282703, diluted 1:200), incubate at room temperature for 30 minutes; g. Incubation with secondary antibody: Dropwise add a horseradish peroxidase (HRP)-conjugated secondary antibody working solution (PANOVUE; Cat.0013001010), incubate with shaking at room temperature for 10 minutes; h. Tyramide signal amplification (TSA): Add PPD520 fluorescently labeled tyramide (diluted 1:100), react with shaking at 25 °C for 10 minutes. Wash with TBST buffer (pH 7.6) by shaking, 5 minutes each time, repeat 3 times; i. Cyclic antigen retrieval and antibody labeling: Through microwave-mediated cyclic antigen epitope retrieval technology (pH 9.0, 95 °C for 10 minutes), antigen activity regeneration between multiple rounds of antibody labeling is achieved. After each round of retrieval, it is rinsed 3 times by shaking with Tris-EDTA buffer (pH 8.0) to eliminate cross-reactions; j. Nuclear counterstaining: Add DAPI staining solution (0.1 μg / mL, containing anti-fluorescence quencher) and incubate at room temperature for 5 minutes in a dark and humid box. Subsequently, wash it 3 times by shaking with TBST buffer (containing 0.1% Tween-20, pH 7.4), 5 minutes each time; k. Mounting: Wash the slides with sterile water for 2 minutes. Wait for the slides to be slightly dry, then use a pipette to add a super anti-quenching mounting agent on the slides to immerse the sample area. Cover with a cover slip and seal it peripherally with neutral resin; l. Slide reading: Use a laser confocal microscope to collect and interpret multi-channel fluorescence images; m. Antibody Panel optimization and verification: Based on the antibody combination optimization system established in the preliminary experiment (Table 1), use orthogonal experimental design to screen the best labeling order and fluorescence spectrum matching scheme.

[0028] The steps of the preliminary experiment are as follows: (1) Optimize and clarify the dilution ratios of the primary antibody and secondary antibody through conventional immunohistochemistry; (2) Set up experimental controls to confirm the antibody titer and the reliability of the detection system. Include positive controls, using homologous and heterologous tissue sections that have been confirmed to contain the target antigen; negative controls, using tissues that have been confirmed to not contain the target antigen; and negative reagent controls, i.e., blank controls; (3) Determine the matching principle between the TSA fluorescent dye and the antigen. Antigens with high abundance are matched with TSA dyes with weak fluorescence intensity, and antigens with low abundance are matched with TSA dyes with high fluorescence intensity; (4) Determine the multi-label staining order, and set the primary antibody staining order as cell membrane - cytoplasm - nucleus.

[0029] Table 1. Antibody information for detection markers

[0030] 3. Multispectral microscopy imaging and quantitative analysis: A. Use a Vectra 3.0 multispectral imaging system (PerkinElmer) to complete panoramic scanning of multi-round stained samples. The spectral channel configuration is as follows: B. CY5 channel: Covers signal acquisition of DAPI, PPD650, and PPD690; C. FITC channel: Specifically captures PPD520 fluorescence signal (excitation 488 nm / emission 525 nm); D. Texas Red channel: Analyze the emission spectrum of PPD620 (excitation 587 nm / emission 615 nm); Results: The results of multi-channel and single-channel staining are as Figure 1 shown. It can be seen that CLDN18.2 and IQGAP3 are mainly expressed in the gastric epithelial region, COL3A1 and FAP are mainly expressed in the stromal region, and there is spatial co-localization of the remaining markers, which is in line with the experimental expectations. Moreover, the percentages of COL3A1+ cells and CD39+ cells among the detected cell types are significantly higher ( Figure 2 ).

[0031] 4. Multispectral imaging: A. Use the Mantra quantitative pathology imaging system of PerkinElmer to capture images; B. Two experienced pathologists are responsible for selecting representative fields of view and obtaining multiple fields of view under a 20x objective lens for subsequent in-depth analysis; C. Selection of the tumor center: Based on the previously obtained whole-slide scan images, use a fixed-size marking tool in the phenotype map of PerkinElmer to accurately select the tumor center. Select as many feasible regions as possible in each sample while avoiding overlap between regions as much as possible. All processed data have undergone strict quality control by pathologists; D. Analyze the multispectral images using inForm image analysis software 2.4 (the analysis process is as Figure 3 shown). First, construct a spectral library based on the single-channel staining images of each fluorescent dye, and use unstained sections to extract the autofluorescence spectrum of the tissue. Subsequently, identify each DAPI-stained cell based on the fluorophore characteristics and cell morphological features related to the DAPI signal. Use inForm software to perform tissue component segmentation on the obtained images to distinguish tumor cells and stromal regions. As Figure 3 shown, the results of tissue and cell segmentation of multispectral images by inForm software have a high consistency with H&E (hematoxylin and eosin staining) pathological sections, indicating the accuracy of the method.

[0032] 5. Digital analysis of the tissue microenvironment Use inForm software (Perkin Elmer) to perform basic splitting of multispectral images. Perform tissue segmentation of tumors and stroma based on cytokeratin expression, and perform cell segmentation based on DAPI staining. Perform further data analysis of multi-marker co-expression according to the staining results as follows: (1) Cell ratio = number of phenotypically positive cells counted / total number of cells counted (2) Cell density = number of phenotypically positive cells counted / area of the effective detection region, with the unit of cells / mm² (3) Distance: Cell coordinate data: The single-cell spatial coordinate matrix is derived from the cell_seg_data file exported by the inform software. Among them, the two columns of Cell X Position and Cell Y Position are the abscissa and ordinate of the cell in the field-of-view image respectively; Intercellular distance data: The two-dimensional Euclidean space coordinate system is used to analyze the intercellular distance data. The coordinate reference is based on the rectangular image acquisition area to generate the original pixel-level spatial positioning data; the spatial scale calibration is based on the magnification of the optical objective lens and applies the pixel-micron conversion coefficient. The formula is physical length = number of pixels / system calibration parameter, with the unit of pixels / μm. Finally, the following Euclidean distance calculation formula is used:

[0033] (4) Effective score: It is defined as the ratio of the number of pairs between the central cell and the surrounding cells within a range of 20 microns of the central cell radius to the total number of cells of the central cell. Examples of spatial analysis and calculation of the effective score are shown as Figure 4 shown.

[0034] Example 2 Construction of the CAR-T treatment efficacy and prognosis model Univariate Cox proportional hazards regression analysis was used to evaluate the relationship between cell components and progression-free survival (PFS) (Table 2). It was found that the percentage of TGF-β1+IQGAP3+ cells among the detected cells in panel1 and the effective score between TGF-β1+IQGAP3+ cells and FAP+ cells were associated with shorter PFS after CAR-T treatment. While the percentage of CD8+TCF1+GZMK+ cells was associated with longer PFS after CAR-T treatment. And after multivariate adjustment, the original direction and effect intensity were maintained.

[0035] Table 2. Univariate Cox proportional hazards regression analysis of the association between cell components and progression-free survival in patients with CLDN18.2-positive gastric cancer treated with CAR-T

[0036] a: The multivariate Cox proportional hazards regression model included three variables: peritoneal metastasis status (present vs. absent), Lauren classification (diffuse vs. mixed vs. intestinal type), and the positive rate of Claudin18.2 protein detected by immunohistochemistry for joint analysis; b-d: The cut-off values for high and low are as follows: 5.63%, 0.01%, 0.03%; Abbreviations: HR, hazard ratio. CI, confidence interval.

[0037] Based on this, the percentages of TGF-β1+IQGAP3+ cells, CD8+TCF1+GZMK+ cells, and the effective score between TGF-β1+IQGAP3+ cells and FAP+ cells were selected as the calculation variables for the CAR-T treatment efficacy and prognosis model. The weight was calculated for each index by the univariate Cox proportional hazards regression algorithm, and the prediction model for CAR-T treatment efficacy and prognosis was constructed by the weighted sum method: CTRS = [Ln(3.42) (Normalized percentage of TGF-β1+IQGAP3+ cells)+ Ln(0.21) (Normalized percentage of CD8+TCF1+GZMK+ cells)+ Ln(1.46) (Normalized effective score between TGF-β1+IQGAP3+ cells and FAP+ cells)] (-1).

[0038] Among them, CAR-T is the treatment response score (CAR-T Response Score, CTRS); Ln(X) is the natural logarithm of X with the mathematical constant e as the base.

[0039] The CTRS threshold / cutoff value is -0.68. When the CTRS is greater than the threshold, it is determined that the patient responds to CAR-T treatment; conversely, when the CTRS is less than the threshold, it is determined that the patient does not respond to CAR-T treatment.

[0040] Example 3 Validation of the Model's Effectiveness To further verify the effectiveness of the model, the efficacy of 24 patients in the validation set after CAR-T treatment was evaluated, and the performance of the CTRS score was detected: The results showed that in the CAR-T treatment response group, the CTRS score was significantly higher than that in the non-response group ( Figure 5 ); by showing the CTRS score of each patient separately, it can be found through the waterfall plot ( Figure 6 ) that the treatment responders are concentrated in the group with higher CTRS scores on the left, further proving that the CTRS score is related to the efficacy of CAR-T treatment in gastric cancer patients; the receiver operating characteristic curve (ROC) and the area under the curve (AUC) were used to evaluate the predictive performance of the CTRS score for CAR-T treatment response ( Figure 7). It was found that the maximum AUC value was 0.889, the 95% confidence interval was 0.752 - 1.000, the specificity was 0.833, the sensitivity was 1.000, the accuracy was 0.875, and the Youden index was 0.833.

[0041] The CTRS was divided into high - and low - score groups according to the cut - off value. It was found that the progression - free survival (PFS) (HR 0.39, 95%CI 0.17 - 0.92, P = 0.032) and overall survival (OS) (HR 0.36, 95%CI 0.13 - 0.96, P = 0.041) in the high - CTRS score group were both higher than those in the non - response group, and the CTRS could effectively predict the prognosis ( Figure 8-9 ).

[0042] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalent technologies, the present invention also intends to include these changes and modifications.

Claims

1. A method for constructing a diagnostic model for predicting the prognosis of gastric cancer CAR-T cell immunotherapy, characterized in that: The construction method includes the following steps: 1) Enroll sufficient gastric cancer tissue specimens from patients undergoing gastric cancer CAR-T cell immunotherapy as a development cohort group; 2) Follow up patients in the development cohort for more than 166 days and evaluate whether they are responders or non-responders through imaging; 3) The differences in tumor microenvironment of patients in the development cohort group were detected by mIHC, and the relationship between cell components and progression-free survival was evaluated using univariate Cox proportional hazard regression analysis. After screening, the percentage of TGF-β1+IQGAP3+ cells, the percentage of CD8+TCF1+GZMK+ cells, and the effective score between TGF-β1+IQGAP3+ cells and FAP+ cells were determined as calculation variables for CAR-T treatment efficacy and prognostic model; 4) The weights for each indicator were calculated using the univariate Cox proportional hazard regression algorithm, and a model for predicting the efficacy and prognosis of CAR-T therapy was constructed by weighted sum: CTRS = [Ln(3.42) (Normalized percentage of TGF-β1+IQGAP3+ cells) + Ln(0.21) (Normalized percentage of CD8+TCF1+GZMK+ cells) + Ln(1.46) (Normalized validity score between TGF-β1+IQGAP3+ cells and FAP+ cells)] (-1).

2. The construction method according to claim 1, characterized in that Patients with progression-free survival ≥166 days were defined as treatment responders, and the rest were defined as treatment non-responders.

3. The construction method according to claim 1, characterized in that: The effective score was defined as the ratio of the number of pairings between the central cell and surrounding cells within a 20-μm radius to the total number of central cells.

4. The construction method according to claim 1, characterized in that: The CTRS threshold / cutoff value is -0.

68. When CTRS is greater than the threshold, it is judged as a response to CAR-T treatment; conversely, when CTRS is less than the threshold, it is judged as no response to CAR-T treatment.

5. The construction method according to claim 4, characterized in that: CTRS was divided into high and low score groups according to the cutoff value. The progression-free survival and overall survival of the high CTRS score group were higher than those of the non-response group.

6. A model for predicting the prognosis of gastric cancer CAR-T cell immunotherapy, characterized in that: The model is: CTRS = [Ln(3.42) (Normalized percentage of TGF-β1+IQGAP3+ cells) + Ln(0.21) (Normalized percentage of CD8+TCF1+GZMK+ cells) + Ln(1.46) (Normalized validity score between TGF-β1+IQGAP3+ cells and FAP+ cells)] (-1); Wherein, CAR-T is the treatment response score (CAR-T Response Score, CTRS); Ln(X) is the logarithm of X with the mathematical constant e as the base.

7. The model according to claim 6, characterized in that The CTRS threshold / cutoff value is -0.

68. When CTRS is greater than the threshold, it is judged as a response to CAR-T treatment; conversely, when CTRS is less than the threshold, it is judged as no response to CAR-T treatment.

8. Use of the prognostic model constructed according to any one of the construction methods described in claims 1-5 for evaluating the prognosis of gastric cancer CAR-T cell immunotherapy.

9. Use of the model according to claim 6 or 7 for evaluating the prognosis of gastric cancer CAR-T cell immunotherapy.

Images