Breast cancer immunotherapy prediction method, system and equipment

Abstract

The invention discloses a breast cancer immunotherapy prediction method, system and equipment. The invention provides a construction method of a breast cancer immunotherapy curative effect prediction model, a corresponding prediction model, a system, equipment and a computer readable storage medium, and the method and the model can realize accurate prediction of the immunotherapy curative effect of a breast cancer patient, and can provide valuable guidance for a treatment strategy.

Description

Technical Field

[0001] The present invention belongs to the field of intelligent medical treatment, and specifically relates to a method, system and device for predicting breast cancer immunotherapy. Background Art

[0002] Breast cancer is a global disease, with the highest incidence rate among female malignant tumors worldwide. Data from the Global Cancer Survey show that new cases of breast cancer account for 24.2% of all new malignant tumors worldwide, and the mortality rate of breast cancer accounts for 15% of the mortality rate of malignant tumors, and shows a trend of continuous growth.

[0003] With the continuous development of immunology and biology, immunotherapy has long become another important treatment for breast cancer after traditional chemotherapy, radiotherapy and surgery. In the tumor microenvironment, during the activation of T lymphocytes stimulated by antigens, the presence of multiple immune checkpoints leads to the inability of T cells to be effectively activated. Immune checkpoints play a key role in the occurrence and development of tumors. The identification of immune checkpoint inhibition pathways has led to significant improvements in immunotherapy. At present, most tumor treatments are mainly targeted at related immune checkpoint molecules such as lymphocyte activation gene-3 (LAG-3), cytotoxic T lymphocyte-associated antigen-4 (CTLA-4) and programmed death receptor-1 (PD-1 / PD-L1). Although immune checkpoint inhibitors (ICI) (e.g., anti-PD-1 / PD-L1 antibodies) have achieved remarkable results in the treatment of breast cancer, they are limited to a small number of tumor patients, and a considerable proportion of tumor patients do not respond to anti-PD1 / PD-L1 antibodies. Therefore, it is very important to predict and screen patients who can benefit from immunotherapy. Summary of the invention

[0004] To make up for the deficiencies of the prior art, the present invention provides a method, system and device for predicting breast cancer immunotherapy.

[0005] To achieve the above purpose, the present invention adopts the following technical solution.

[0006] The first aspect of the present invention provides a method for constructing a breast cancer immunotherapy prediction model, the construction method comprising: Obtain patient samples; According to the patient efficacy (6-month PFS), the patients were divided into R (responders) and NR (non-responders) groups; Cluster analysis was performed on the main immune cell subsets to obtain differential immune cell subsets; Constructing a predictive model for breast cancer immunotherapy based on differential immune cell subsets.

[0007] Furthermore, the sample is selected from a peripheral blood sample.

[0008] Furthermore, the major immune cell subsets include T cells and myeloid cells.

[0009] Furthermore, T cells were clustered according to the expression of CD3.

[0010] Furthermore, myeloid cells were clustered according to the expression of CD11b.

[0011] Furthermore, the R group includes patients with PFS ≥ 6 months, and the NR group includes patients with PFS < 6 months.

[0012] Furthermore, the breast cancer is advanced HER2-negative breast cancer treated with immunotherapy combined with chemotherapy.

[0013] Furthermore, the immunotherapy is PD-1 inhibitor treatment.

[0014] A second aspect of the present invention provides a method for predicting breast cancer immunotherapy, the method comprising: Obtaining samples to be tested; Detecting a baseline level of target immune cells in a sample to be tested, wherein the target immune cells include CD39+Tregs and / or CCR2+moDCs; Classification predictions are made based on the baseline level / risk score of target immune cells to obtain prediction results for breast cancer immunotherapy.

[0015] Furthermore, the prediction method based on the baseline level of target immune cells includes: based on the baseline level of target immune cells, obtaining a prediction result of breast cancer immunotherapy; if the baseline level of CD39+Tregs is increased and / or the baseline level of CCR2+moDCs is decreased, obtaining a prediction result of poor treatment with breast cancer immunotherapy; if the baseline level of CD39+Tregs is decreased and / or the baseline level of CCR2+moDCs is increased, obtaining a prediction result of good treatment with breast cancer immunotherapy.

[0016] Furthermore, the target immune cells also include CD161+CD8+Tem.

[0017] Furthermore, the calculation formula of the risk score prediction model is h(t|X)=h0(t)exp(0.3376*[CD39+Treg]-0.0303*[CCR2+moDC]).

[0018] Furthermore, the prediction method based on the risk score of target immune cells includes: obtaining a prediction result of breast cancer immunotherapy based on the risk score of target immune cells; if the risk score is higher than 0.84, obtaining a prediction result of poor treatment with breast cancer immunotherapy; if the risk score is lower than 0.84, obtaining a prediction result of good treatment with breast cancer immunotherapy.

[0019] Furthermore, the breast cancer is advanced HER2-negative breast cancer treated with immunotherapy combined with chemotherapy.

[0020] Furthermore, the immunotherapy is PD-1 inhibitor treatment.

[0021] The third aspect of the present invention provides a breast cancer immunotherapy prediction model, which is constructed by the breast cancer immunotherapy prediction model construction method described in the first aspect of the present invention or the breast cancer immunotherapy prediction method described in the second aspect of the present invention.

[0022] Furthermore, the calculation formula of the prediction model is h(t|X)=h0(t)exp(0.3376*[CD39+Treg]-0.0303*[CCR2+moDC]).

[0023] Furthermore, the breast cancer is advanced HER2-negative breast cancer treated with immunotherapy combined with chemotherapy.

[0024] Furthermore, the immunotherapy is PD-1 inhibitor treatment.

[0025] A fourth aspect of the present invention provides a breast cancer immunotherapy prediction system, the system comprising: Acquisition unit: used to obtain the baseline level of immune cells in the sample to be tested; Processing unit: used for inputting the baseline level of immune cells of the sample to be tested into the prediction model constructed by the method for constructing a breast cancer immunotherapy prediction model according to the first aspect of the present invention to obtain a breast cancer immunotherapy prediction result.

[0026] Furthermore, the breast cancer is advanced HER2-negative breast cancer treated with immunotherapy combined with chemotherapy.

[0027] Furthermore, the immunotherapy is PD-1 inhibitor treatment.

[0028] A fifth aspect of the present invention provides a breast cancer immunotherapy prediction system, the system comprising: Acquisition unit: used to obtain the baseline level of CD39+Tregs and / or CCR2+moDCs in the sample to be tested; Processing unit: Based on the baseline level / risk score of target immune cells, a prediction result of breast cancer immunotherapy is obtained; if the baseline level of CD39+Tregs is increased and / or the baseline level of CCR2+moDCs is decreased, a prediction result of poor treatment with breast cancer immunotherapy is obtained; if the baseline level of CD39+Tregs is decreased and / or the baseline level of CCR2+moDCs is increased, a prediction result of good treatment with breast cancer immunotherapy is obtained.

[0029] Furthermore, the target immune cells also include CD161+CD8+Tem.

[0030] Furthermore, the calculation formula of the risk score prediction model is h(t|X)=h0(t)exp(0.3376*[CD39+Treg]-0.0303*[CCR2+moDC]).

[0031] Furthermore, the prediction method based on the risk score of target immune cells includes: obtaining a prediction result of breast cancer immunotherapy based on the risk score of target immune cells; if the risk score is higher than 0.84, obtaining a prediction result of poor treatment with breast cancer immunotherapy; if the risk score is lower than 0.84, obtaining a prediction result of good treatment with breast cancer immunotherapy.

[0032] Furthermore, the breast cancer is advanced HER2-negative breast cancer treated with immunotherapy combined with chemotherapy.

[0033] Furthermore, the immunotherapy is PD-1 inhibitor treatment.

[0034] A sixth aspect of the present invention provides a breast cancer immunotherapy prediction device, the device comprising: Memory: used to store program instructions; Processor: used to call program instructions. When the program instructions are executed, the steps of the method for constructing a breast cancer immunotherapy prediction model described in the first aspect of the present invention or the method for predicting breast cancer immunotherapy described in the second aspect of the present invention are executed.

[0035] The seventh aspect of the present invention provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the method for constructing a breast cancer immunotherapy prediction model described in the first aspect of the present invention or the steps of the breast cancer immunotherapy prediction method described in the second aspect of the present invention.

[0036] The eighth aspect of the present invention provides the use of a reagent for detecting immune cells in the preparation of a product for predicting immunotherapy for breast cancer, wherein the immune cells include CD39+Tregs and / or CCR2+moDCs.

[0037] Furthermore, the immune cells also include CD161+CD8+Tem.

[0038] Furthermore, the breast cancer is advanced HER2-negative breast cancer treated with immunotherapy combined with chemotherapy.

[0039] Furthermore, the immunotherapy is PD-1 inhibitor treatment.

[0040] The ninth aspect of the present invention provides the use of immune cells in constructing a predictive model for predicting breast cancer immunotherapy, wherein the immune cells include CD39+Tregs and / or CCR2+moDCs.

[0041] Furthermore, the immune cells also include CD161+CD8+Tem.

[0042] Furthermore, the breast cancer is advanced HER2-negative breast cancer treated with immunotherapy combined with chemotherapy.

[0043] Furthermore, the immunotherapy is PD-1 inhibitor treatment.

[0044] The tenth aspect of the present invention provides an application of the breast cancer immunotherapy prediction model described in the third aspect of the present invention in constructing a system / device / computer-readable storage medium for predicting breast cancer immunotherapy.

[0045] Furthermore, the breast cancer is advanced HER2-negative breast cancer treated with immunotherapy combined with chemotherapy.

[0046] Furthermore, the immunotherapy is PD-1 inhibitor treatment.

[0047] Advantages and beneficial effects of the present invention: The present application provides a method for constructing a breast cancer immunotherapy efficacy prediction model and a corresponding prediction model, system, device and computer-readable storage medium. The method and model can achieve accurate prediction of the efficacy of immunotherapy for breast cancer patients and provide valuable guidance for treatment strategies. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 It is a flow chart of the method for constructing a predictive model for breast cancer immunotherapy; Figure 2 It is a flow chart of the prediction method of breast cancer immunotherapy; Figure 3 This is a schematic diagram of the prediction system for breast cancer immunotherapy; Figure 4 It is a schematic diagram of a predictive device for breast cancer immunotherapy; Figure 5It is an immune cell atlas of patients with advanced breast cancer, where 5A is a schematic diagram of the study, 5B is a Kaplan-Meier survival analysis of the overall survival of the responder (R) and non-responder (NR) groups, 5C is a distribution diagram of the main immune cells between the two groups, 5D is an expression diagram of lineage markers used to identify the main types of immune cells, and 5E-5F are bar and box plots of the differences in the proportions of the main immune cell subsets between the two groups; Figure 6 The mass cytometry analysis revealed the immune cell profile of patients with advanced breast cancer, where 6A is a schematic diagram of the gate of mass spectrometry data, 6B is a Kaplan-Meier curve of the PFS comparison between the R group and the NR group, and 6C is a tSNE map of the distribution of major immune cells; Figure 7 Figure 7 is an analysis of the proportion of various cell subsets and marker expression in the T cell compartment, among which, 7A is a heat map of the standardized expression of 42 pre-designed markers of 28 subsets, 7B is a tSNE map of the distribution of major T cell subsets, 7C is a ratio map of major T cells in different groups of patients, 7D is a difference map of marker expression on responder (R) and non-responder (NR) T cells, 7E-7F are the difference maps of the proportion of T cells in T-C07 (CD39+Treg) and T-C15 (CD161+CD8+Tem) between the two groups, 7G-7H are comparison maps of the expression of CD39 (upper) on T-C07 and CD161 (lower) on T-C15 in the two groups, 7I-7J are the ratio maps of T-C08 (PD-1+CD4+Tex) and PD-1 expression on T-C08 in the two groups; Figure 8 It is an in-depth analysis of the T cell compartment, among which 8A is a tSNE diagram of the expression of the main markers used to distinguish the major categories on T cells, 8B is the difference between T-C07 (CD39+Treg) and T-C15 (CD161+CD8+Tem) in different treatment groups, 8C is the difference between the 28 cell subsets in the T cell compartment between the two groups, and 8D is the difference between the expression of various proteins on T cells between the two groups; Fig. 9 It is the correlation and longitudinal analysis diagram of specific immunosuppressive T cell subsets, among which 9A is the correlation diagram between two specific T cells and other T cell subsets, 9B-9C are the correlation diagrams between different T cell clusters in responders (R) and non-responders (NR), and 9D is the statistical analysis diagram of the longitudinal analysis of CD39 and PD-1 expression before and after treatment and the proportion of T-C07 and T-C08; Fig.10It is a dynamic analysis diagram of the T cell compartment, among which 10A is a correlation diagram between different T cell clusters in all patients, 10B is a statistical analysis diagram of the longitudinal analysis of PD-1 expression and T-C08 proportion at three treatment time points, and 10C is a statistical analysis diagram of the longitudinal analysis of PD-1 expression and T-C08 proportion at three treatment time points; Fig.11 It is an analysis diagram of the proportion of various cell subsets and marker expression in the myeloid cell compartment, among which 11A-11B are tSNE diagrams and heat maps of the distribution and marker expression of 25 myeloid cell clusters, 11C is a bar graph of the difference in the proportions of major myeloid cell subsets between responders (R) and non-responders (NR), 11D-11E are the proportions of M-C12 (CCR2+moDC) in myeloid cell subsets, 11F-11G are correlation diagrams between different myeloid cell clusters in the two groups of patients, and 11H-11I are correlation diagrams between different T cell clusters and myeloid cell clusters in the two groups of patients; Fig.12 It is an in-depth analysis of the myeloid cell compartment, among which 12A is a diagram showing the difference in the expression of various proteins on myeloid cells between the two groups, 12B is a diagram showing the difference in the expression of PD-L1 on myeloid cells between the two groups, 12C is a diagram showing the difference in M-C12 (CCR2+moDC) between the two groups in different treatment groups, 12D is a diagram showing the correlation between different myeloid cell clusters in all patients, 12E is a diagram showing the significant correlation between M-C12 and other myeloid subsets, 12F is a diagram showing the statistical analysis of the proportions of several important myeloid cell subsets before and after treatment, and 12G is a diagram showing the difference in 25 cell subsets between the two groups in the myeloid cell compartment; Fig.13 is a graph of the prognostic value of the breast cancer immunotherapy prediction model, where 13A is a receiver operating characteristic (ROC) curve graph of the model for classification of 6-month, 12-month, and 18-month progression-free survival, and 13B is a Kaplan-Meier survival analysis graph of progression-free survival by risk score group; Fig.14 14A is the model validation graph, among which, 14B is the receiver operating characteristic (ROC) curve graph of the model for classifying the 6-month, 12-month, and 18-month progression-free survival in the three treatment groups, 14B is the tSNE graph of the main immune cell categories in the external validation set, 14C is the tSNE graph of the main T cell subsets in the validation set, 14D is the gene expression graph of the main T cell subsets in the validation set, 14E is the tSNE graph of the main myeloid cell subsets in the validation set, 14F is the gene expression graph of the main myeloid cell subsets in the validation set, and 14G is the ROC curve graph of the model's predictive efficacy of immunotherapy in the external validation set. DETAILED DESCRIPTION

[0049] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0050] In some of the processes described in the specification and claims of the present invention and the above-mentioned figures, multiple operations that appear in a specific order are included, but it should be clearly understood that these operations may not be executed in the order in which they appear in this article or executed in parallel. The serial numbers of the operations, such as 101, 102, etc., are only used to distinguish between different operations, and the serial numbers themselves do not represent any execution order. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions of "first", "second", etc. in this article are used to distinguish different messages, devices, modules, etc., do not represent the order of precedence, and do not limit the "first" and "second" to be different types.

[0051] Figure 1 : is a flow chart of a method for constructing a breast cancer immunotherapy prediction model provided by an embodiment of the present invention. Specifically, the method comprises the following steps: 101: Obtaining Patient Samples In some embodiments, a sample may refer to a biological sample, typically a clinical sample, and includes, for example, blood and other body fluids, including but not limited to peripheral blood, serum, plasma, urine, and saliva; and solid tissue samples, such as biopsy specimens, especially those containing cancer cells. Samples also include samples that have been manipulated or processed in any appropriate manner after purchase, including but not limited to centrifugation, filtration, precipitation, dialysis, chromatography, treatment with reagents, washing, or enrichment of a certain component of the sample, such as a cell population.

[0052] In a specific embodiment, the sample is selected from a peripheral blood sample. Specifically, this study (NCT04389073) included peripheral blood samples from 57 patients with advanced HER2-negative breast cancer who received PD-1 inhibitors combined with chemotherapy.

[0053] 102: According to PFS, patients were divided into R group and NR group In a specific embodiment, patients are divided into two groups according to 6-month PFS (progression-free survival): responders (R) are defined as patients with PFS ≥ 6 months, and non-responders (NR) are patients with PFS < 6 months.

[0054] 103: Cluster analysis of the main immune cell subsets to obtain differential immune cell subsets This application designed a comprehensive 42-marker mass cytometry panel to analyze the peripheral blood of 57 patients with advanced HER2-negative breast cancer who received anti-PD-1 combination therapy. Patients were divided into responders and non-responders based on 6-month progression-free survival (PFS), and then phenotypical and functional comparisons were performed to identify candidate predictive biomarkers. Longitudinal analysis of paired samples further revealed the dynamic changes of these specific subpopulations, and the specific experimental results are as follows: Identification of immune cells associated with immunotherapy efficacy in patients with HER2-negative breast cancer: A total of 57 patients with advanced HER2-negative breast cancer who were treated with PD-1 inhibitors combined with chemotherapy participated in this study (NCT04389073). The patients were divided into two groups based on their 6-month PFS: responders (R) with PFS ≥ 6 months and non-responders (NR) with PFS < 6 months. A comprehensive analysis of immune cell subsets between the two groups was then performed ( Figure 5 A). The baseline clinicopathological characteristics of the patients were first evaluated, showing no significant correlation between the remission group and these characteristics. The PFS and overall survival of patients in the R group were significantly better ( Figure 5 B; Figure 6 B).

[0055] PBMCs were collected from patients and analyzed by CyTOF to detect various immune cell subsets. After quality filtering and marker identification, multiple immune cell subsets were identified, including T cells, B cells, natural killer (NK) cells, granulocytes, dendritic cells (DCs), and monocytes ( Figure 5 C, 5D; Figure 6 C). Responders had a higher proportion of T cells in peripheral blood (p = 0.082; adjusted p = 0.257) and fewer monocytes (p = 0.086; adjusted p = 0.257), but the differences were not statistically significant ( Figure 5 E, 5F).

[0056] To elucidate specific immune cell subsets that can predict treatment effects, two major immune cell subsets in peripheral blood: T cells and myeloid cells were manually gated, followed by clustering and dimensionality reduction for further in-depth analysis.

[0057] Activated Tregs are a robust predictor of poor clinical outcomes of immunotherapy: Given the key role of T cells in immunotherapy, this application first manually gated the T cell compartment based on the expression of CD3. Using unsupervised clustering and manual annotation, 28 T cell subsets were identified, including CD4+T cells and CD8+T cells at different differentiation stages, such as naive (Tn), central memory (Tcm), effector memory (Tem), effector (Teff), exhausted (Tex), and regulatory T (Treg) cells, as well as double-positive T (DPT), double-negative T (DNT), and gdT cells ( Figure 7 A, 7B; Figure 8 A). Overall, responders tended to have more CD4+Tcm cells and fewer Tregs and CD8+Tem cells ( Figure 7 C). In terms of the expression of all T cell markers, the levels of CD39 and HLA-DR were significantly increased in non-responders, while the expression levels of PD-1 and cytotoxic T lymphocyte antigen-4 (CTLA-4) were not significantly different ( Figure 7 D).

[0058] Further analysis of specific cell subsets showed that non-responders had higher baseline levels of T-C07 (CD39+Treg) (p<0.001; adjusted p=0.031) and T-C15 (CD161+CD8+Tem) (p=0.006; adjusted p=0.097) compared to responders. Figure 7 E, 7F). These two subgroups showed similar trends in all three treatment cohorts ( Figure 8 B). In addition, these patients also showed increased expression of CD39 on CD39+ Treg cells and CD161 on CD161+CD8+ TEM cells ( Figure 7 G, 7H).

[0059] In particular, this application investigated the relationship between PD-1 expression and treatment efficacy. A subpopulation of PD-1+CD4+T cells (T-C08) was identified, called CD4+Tex. The baseline frequency of this population tended to be higher in responders, but it was not statistically significant (p=0.377; corrected p=1) ( Figure 7 I). In addition, the intensity of PD-1 expression on these cells was comparable between the two groups ( Figure 7 J). Other T cell subsets and surface marker expressions were not significantly correlated with efficacy ( Figure 8 C, 8D).

[0060] The data in this application demonstrate a potential correlation between elevated levels of CD39+ Tregs in the T cell compartment and worse outcomes in patients receiving anti-PD-1 therapy, suggesting promising potential for predicting treatment efficacy.

[0061] Immunosuppressive T cells cannot be blocked by PD-1 inhibitors, leading to primary resistance: Correlation analysis between different T cell subsets was performed to gain a deeper understanding of the potential immunosuppressive mechanisms of these cells ( Fig.10 A). Studies have shown that the frequency of T-C07 (CD39+Treg) is significantly positively correlated with T-C15 (CD161+CD8+Tem) and negatively correlated with C13 (initial CD8+T cells) ( Fig. 9 A; Fig.10A). Notably, these negative correlations were mainly observed in the NR group ( Fig. 9 In addition, a significant negative correlation was detected between T-C15 (CD161+CD8+Tem) and T-C23 (CD8+Teff) in the NR group, but not in the R group ( Fig. 9 B, 9C). These findings confirm that activated immunosuppressive T cell subsets may affect naive CD8+ T cells and CD8+ Teff in systemic immunity.

[0062] Interestingly, this application found that in the NR group, T-C07 (CD39+Tregs) and T-C08 (PD-1+CD4+T cells) were significantly positively correlated, and T-C15 and T-C08 also showed a positive correlation trend ( Fig. 9 C). These findings suggest that these cells may work synergistically with PD-1+ T cells to exert immunosuppressive effects.

[0063] In order to study the causes of immunotherapy resistance, this application analyzed the dynamic changes of these immune cell subsets in peripheral blood (baseline, after 2 cycles of treatment). At the same time, blood samples were collected from 9 patients at the progression stage to further explore their potential role in acquired resistance. It was found that the overall expression of PD-1 on T cells and the frequency of PD-1+CD4+Tex (T-C08) were significantly reduced after treatment and at progression compared with baseline, indicating the effect of PD-1 inhibitors ( Fig. 9 D; Fig.10 B). However, after 2 cycles of treatment, the expression of CD39 and the frequency of T-C07 (CD39+Tregs) increased significantly ( Fig. 9 D; Fig.10 C). There was no significant change in CD161 expression and T-C15 after treatment ( Fig.10 C). These results suggest that these patients may not benefit from blockade of the PD-1 / PD-L1 axis alone and may require additional targeted therapies, such as antagonists of CD39 and CD161, to achieve better outcomes.

[0064] CCR2+ monocyte-derived dendritic cells (moDCs) in the myeloid cell compartment are associated with immunotherapy response: Next, this application focused on the ability of the myeloid cell compartment to predict the efficacy of patients receiving immunotherapy for advanced breast cancer. By clustering cells gated by CD11b, 25 subsets were identified, including DCs, macrophages, monocytes, neutrophils, and eosinophils ( Fig.11 A, 11B). Overall, responders tended to have more DCs and monocytic myeloid-derived suppressor cells (M-MDSCs) and fewer classical monocytes ( Fig.11C). Expression analysis in bone marrow cells did not reveal any statistically significant molecules predictive of PFS ( Fig.12 A). PD-L1 expression tended to be higher in nonresponders, although this difference was not statistically significant (p=0.135) ( Fig.12 B).

[0065] Among these subpopulations, the baseline level of M-C12 (activated CCR2+moDC) was higher in the R group than in the NR group (p = 0.001; corrected p = 0.037) ( Fig.11 D, 11E). This trend was observed in all three treatment cohorts ( Fig.12 C). Correlation analysis found that this subpopulation was significantly positively correlated with M-C08 (pDC) and significantly negatively correlated with monocytes ( Fig.12 D, 12E). Separate correlation analysis showed that in the R group, M-C12 (CCR2+moDC) and M-C11 (CCR6+CXCR3+cDC) were significantly positively correlated ( Fig.11 F), M-C12 increased significantly after 2 cycles of treatment and at progression compared with baseline ( Fig.12 F). This finding confirms that the collaboration between two different types of DCs in peripheral blood may be an important factor in the effectiveness of immunotherapy. In contrast, in the NR group, M-C12 was significantly positively correlated with M-C20 (M-MDSCs), a group of immunosuppressive cells whose abundance decreased significantly after treatment and at progression compared with baseline ( Fig.11 G, Fig.12 F). These results suggest that in the NR group, the M-C12 cell population may be suppressed by MDSCs, leading to a lack of efficacy. Other bone marrow cell subsets were not significantly correlated with patient efficacy ( Fig.12 G).

[0066] Since bone marrow cells mainly affect anti-tumor immunity by directly or indirectly affecting T cells, this application further explored the interaction between bone marrow cells and T cell subsets in peripheral blood. It was found that in the R group, M-C12 (CCR2+moDC) was significantly negatively correlated with T-C21 (CD8+Tem), T-C22 (CD8+Tem), T-C23 (CD8+Teff), and T-C24 (CD8+Teff) ( Fig.11 H). These four subpopulations were characterized by CD38 and CXCR3 negativity, indicating that the cell populations in a non-chemotactic, non-activated state tended to terminally differentiate. In contrast, in the NR group, the M-C12 subpopulation was significantly negatively correlated with T-C01 (naive CD4+ T cells) ( Fig.11 I), which may be detrimental to the activation of T cell immunity.

[0067] 104: Constructing a predictive model for breast cancer immunotherapy based on differential immune cell subpopulations Combined parameters in peripheral blood can effectively predict PFS after immunotherapy: This application further studied the predictive potential of these efficacy-related immune cell subsets for patients' PFS. By integrating the two immune cell subsets (CD39+Tregs and CCR2+moDCs) identified in the previous analysis into a prediction model called "Breast Cancer Immunotherapy Prediction Score", it was found that the comprehensive risk score can also effectively predict PFS, with AUC values ​​of 0.809, 0.787, and 0.726 for 6-month, 12-month, and 18-month PFS, respectively ( Fig.13 A), whose predictive efficacy was independent of the treatment group. Using a risk score cutoff of 0.84, this application describes the PFS curves for the low-risk and high-risk groups, clearly distinguishing the two groups in terms of PFS (log rank, p < 0.001; Fig.13 B).

[0068] Model formula: h(t|X)=h0(t)exp(0.3376*[T-C07]-0.0303*[M-C12]), that is, h(t|X)=h0(t)exp(0.3376*[CD39+Treg]-0.0303*[CCR2+moDC]).

[0069] Among them, h(t, X) is the risk function of the observed subject's progression-free survival at time t; X=([T-C07], [M-C12]) are two independent variables related to progression-free survival, which are the proportion of T-C07 cell subsets in the T cell compartment and the proportion of M-C12 in the myeloid cell compartment; h0(t) is the risk function at time t when [T-C07]=[M-C12]=0, which is called the basic risk function; (0.3376, -0.0303) is the regression coefficient of the Cox model.

[0070] The predictive performance of the model was robustly validated in three treatment cohorts: the cisplatin cohort (AUC values ​​at 6, 12, and 18 months were 0.717, 0.833, and 0.688, respectively), the bevacizumab cohort (AUC values ​​at 6, 12, and 18 months were 0.750, 0.808, and 0.808, respectively), and the VEX cohort (AUC values ​​at 6, 12, and 18 months were 0.883, 0.720, and 0.619, respectively) ( Fig.14 A).

[0071] To further evaluate the robustness of the model, the single-cell dataset GSE189125 was selected for validation, which includes 14 melanoma patients treated with immune checkpoint inhibitors (ICIs). First, the main clusters of T cells and myeloid cells were identified ( Fig.14B). Subclustering and analysis of expression signatures revealed that clusters 15 and 23 in T cells corresponded to T-C07 in cyTOF analysis ( Fig.14 C, 14D), clusters 5 and 15 in myeloid cells correspond to M-C12 in cyTOF analysis ( Fig.14 E, S5F). Using the proportions of these subpopulations and their respective model coefficients as input, we predicted the probability of ICI response (5 non-responders and 9 responders) and obtained an AUC value of 0.64 for the single-cell external trial cohort ( Fig.14 G). These results confirm that the combined risk score is a robust predictor of prognosis and treatment response to immunotherapy.

[0072] In summary, this application demonstrates that immune monitoring of CD39+Tregs and moDCs in peripheral blood can predict the response to anti-PD-1 therapy in patients with advanced HER2-negative breast cancer. A model was developed based on the ratio of these two cell subsets, which can serve as a biomarker to predict the treatment outcome of these patients. The findings of this application are clinically significant because they support the identification of peripheral immune predictive biomarkers for anti-PD-1 therapy, thereby guiding treatment selection in clinical practice and providing a basis for new treatment strategies for patients with different systemic immune status.

[0073] The specific experimental methods are as follows: Study Design and Sample Collection Blood samples were collected from patients in a prospective, phase II, randomized adaptive clinical trial (NCT04389073). Eligible patients had histologically or cytologically confirmed HER2-negative metastatic breast cancer and had received no more than 1 prior standard chemotherapy. Patients were randomly assigned to different treatment cohorts by Bayesian adaptive randomization to receive different combinations of chemotherapy and PD-1 inhibitors. Patients included in this study were from 3 immunotherapy cohorts with peripheral blood samples collected at baseline: (1) antiangiogenic bevacizumab, vinorelbine (NVB), and toripalimab (n = 15); (2) conventional cisplatin, NVB, and toripalimab (n = 17); and (3) cyclophosphamide, capecitabine, NVB metronomic therapy, and toripalimab (VEX cohort; n = 25). This study was approved by the Ethics Committee of the Cancer Hospital, Chinese Academy of Medical Sciences / National Cancer Center and complied with all relevant ethical regulations. All patients provided written informed consent before the study.

[0074] Peripheral blood (4 mL / person) was collected from each subject using EDTA-coated vacuum blood collection tubes at three time points: before treatment, after two treatment cycles, and at disease progression.

[0075] Preparation of cell suspension Blood samples were immediately sent to the laboratory for processing, and peripheral PBMCs were separated by Ficoll density gradient centrifugation. The cell pellet was resuspended in 5 mL of pre-cooled PBS and centrifuged at 400 × g for 5 min at 4°C. The supernatant was discarded, the cells were counted, and the single-cell suspension was frozen in liquid nitrogen cell freezing solution containing DMSO for large-scale cell staining.

[0076] CyTOF antibody staining and data acquisition Purified antibodies were obtained from BioLegend, eBioscience, BioXcell, R&D Systems, and BD Biosciences. Antibodies were labeled with the indicated metal tags using the MaxPAR Antibody Labeling Kit (Fluidigm). Bound antibodies were titrated to optimal concentrations before use.

[0077] Thaw the single cell suspension and perform cell counts. Samples for subsequent analysis met the following quality criteria: cell counts not less than 3×10^6 and viability greater than 85%. Wash the cells with 1×PBS and then stain with 100 μL 250 nM cisplatin (Fluidigm) on ice for 5 minutes to exclude dead cells, and then incubate with Fc receptor blocking solution and surface antibody mixture on ice for 30 minutes. Wash the cells twice with FACS buffer (1×PBS + 0.5% BSA) and fix overnight in 200 μL insertion solution (Maxpar Fix and Perm Buffer containing 250 nM 191 / 193Ir, Fluidigm). After fixation, wash the cells once with FACS buffer, then wash the cells with perm buffer (eBioscience) and stain with intracellular antibody mixture on ice for 30 minutes. The cells were washed and suspended in deionized water, 20% EQ beads (Fluidigm) were added, and data were acquired on a mass cytometer (Helios, Fluidigm).

[0078] Clustering and Dimensionality Reduction Data from each sample were de-coded based on the raw data using a double filtering protocol with unique mass tag barcodes. FCS files generated from different batches were normalized using a bead-based normalization method. Manual gating was performed using FlowJo software to exclude debris, dead cells, and doublets, leaving live single cells ( Figure 6A). Manual gating was further used to delineate the cell subsets of interest, including immune cells, T cells, and myeloid cells, and then the X-shift clustering algorithm was used to classify cells into different phenotypes based on marker expression levels. Dimensionality reduction was performed using t-SNE to visualize the high-dimensional data in two dimensions, showing the distribution of each cluster and marker expression, as well as the differences between groups or sample types. The cell type of each cluster was annotated on the heat treatment based on its marker expression pattern.

[0079] CyTOF statistical analysis The differences in cell cluster frequency and protein expression intensity at baseline between the R and NR groups were compared using independent t-tests or Wilcoxon tests (depending on the normality of the data distribution by the Shapiro-Wilk test). Depending on the normality of the data, paired t-tests or paired Wilcoxon tests were used to explore the dynamic changes before and after immunotherapy. The P values ​​for multiple comparisons between different clusters were adjusted by the Benjamini-Hochberg (BH) procedure to control the false discovery rate. For patients with paired samples at the time of disease progression, Friedman tests were performed, followed by post hoc pairwise comparisons using the Dunn-Bonferroni method to explore the dynamic changes of cell subpopulations and molecules of interest at three time points with different biological characteristics. In addition, the Spearman test was used to measure the correlation between the abundance of each subpopulation in the paired data.

[0080] Construction and validation of prognostic models The significantly different markers predicting treatment efficacy were included in the multivariate Cox regression model to test their prognostic value for progression-free survival (PFS). The predictive performance of the model was evaluated by the area under the receiver operating characteristic curve (AUC). The optimal critical value was then determined by maximizing statistical significance through the log-rank test, and the patients were divided into high-risk and low-risk groups for Kaplan-Meier survival analysis.

[0081] Based on the coefficients derived from the model, it was tested using a completely independent single-cell dataset (GSE189125). Unsupervised clustering and dimensionality reduction were performed according to the standard workflow in Seurat. After two rounds of clustering, the gene expression corresponding to the cyTOF markers within each cluster was detected, and the AUC of the response was used as the evaluation criterion to determine the cell subpopulations of interest for model validation.

[0082] Statistical analysis Quantitative variables are presented as medians, and categorical variables are presented as proportions. Comparison of baseline clinicopathological parameters between the two patient groups was performed using the Fisher exact probability method. Survival outcomes were estimated using the Kaplan-Meier method, and differences between groups were analyzed using the log-rank test. All statistical analyses and visualizations were performed using R (version 4.4.0). A two-sided p value < 0.05 was considered statistically significant.

[0083] Figure 2 The present invention provides a flow chart of a method for predicting breast cancer immunotherapy. Specifically, the method comprises the following steps: 201: Get the sample to be tested 202: Detecting the baseline level of target immune cells in the sample to be tested, wherein the target immune cells include CD39+Tregs and / or CCR2+moDCs 203: Classification prediction based on the baseline level / risk score of target immune cells to obtain the prediction results of breast cancer immunotherapy.

[0084] Based on the baseline level of target immune cells, a prediction result of breast cancer immunotherapy is obtained; if the baseline level of CD39+Tregs increases and / or the baseline level of CCR2+moDCs decreases, a prediction result of poor treatment with breast cancer immunotherapy is obtained; if the baseline level of CD39+Tregs decreases and / or the baseline level of CCR2+moDCs increases, a prediction result of good treatment with breast cancer immunotherapy is obtained.

[0085] Alternatively, based on the risk score of the target immune cells, a prediction result of breast cancer immunotherapy is obtained; if the risk score is higher than 0.84, a prediction result of poor treatment with breast cancer immunotherapy is obtained; if the risk score is lower than 0.84, a prediction result of good treatment with breast cancer immunotherapy is obtained.

[0086] Figure 3 Schematic diagram of a breast cancer immunotherapy prediction system provided by the present invention, the system comprising: Acquisition unit 301: used to obtain the baseline level of immune cells in the sample to be tested Processing unit 302: inputs the baseline level of immune cells of the sample to be tested into the prediction model constructed by the above-mentioned method for constructing a breast cancer immunotherapy prediction model to obtain a breast cancer immunotherapy prediction result Or, the acquisition unit 301 is used to obtain the baseline level of CD39+Tregs and / or CCR2+moDCs in the sample to be tested Processing unit 302: Based on the baseline level of target immune cells, a prediction result of breast cancer immunotherapy is obtained; if the baseline level of CD39+Tregs increases and / or the baseline level of CCR2+moDCs decreases, a prediction result of poor treatment with breast cancer immunotherapy is obtained; if the baseline level of CD39+Tregs decreases and / or the baseline level of CCR2+moDCs increases, a prediction result of good treatment with breast cancer immunotherapy is obtained.

[0087] Alternatively, based on the risk score of the target immune cells, a prediction result of breast cancer immunotherapy is obtained; if the risk score is higher than 0.84, a prediction result of poor treatment with breast cancer immunotherapy is obtained; if the risk score is lower than 0.84, a prediction result of good treatment with breast cancer immunotherapy is obtained.

[0088] Figure 4 Schematic diagram of a breast cancer immunotherapy prediction device provided by an embodiment of the present invention, the device comprising: Memory: used to store program instructions Processor: used to call program instructions, and when the program instructions are executed, the method for constructing the breast cancer immunotherapy prediction model or the steps of the breast cancer immunotherapy prediction method are executed.

[0089] An embodiment of the present invention further provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the method for constructing the above-mentioned breast cancer immunotherapy prediction model or the steps of the above-mentioned breast cancer immunotherapy prediction method.

[0090] The validation results of this validation example show that assigning intrinsic weights to indications can moderately improve the performance of the method relative to the default setting.

[0091] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0092] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0093] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0094] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0095] A person skilled in the art may understand that all or part of the steps in the various methods of the above embodiments may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, and the storage medium may include: a read-only memory (ROM), a random access memory (RAM), a disk or an optical disk, etc.

[0096] A person skilled in the art can understand that all or part of the steps in the above-mentioned embodiment method can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable storage medium. The above-mentioned storage medium can be a read-only memory, a disk or an optical disk, etc.

[0097] The present invention also provides the use of a reagent for detecting immune cells in the preparation of a product for predicting breast cancer immunotherapy, wherein the immune cells include CD39+Tregs and / or CCR2+moDCs.

[0098] The target immune cells also include CD161+CD8+Tem.

[0099] In some embodiments, the reagents include reagents for detecting markers of CD39+Tregs, CCR2+moDCs, and CD161+CD8+Tem.

[0100] The above is a detailed introduction to a computer device provided by the present invention. For a person skilled in the art, according to the concept of the embodiments of the present invention, there may be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as limiting the present invention.

Claims

1. A method for constructing a breast cancer immunotherapy prediction model, characterized in that: The construction method comprises: Obtain patient samples; According to PFS, patients were divided into R and NR groups; Cluster analysis was performed on the main immune cell subsets to obtain differential immune cell subsets; Construct a predictive model for breast cancer immunotherapy based on differential immune cell subsets; Preferably, the sample is selected from a peripheral blood sample; Preferably, the major immune cell subsets include T cells and myeloid cells; Preferably, clustering analysis is performed on T cells based on CD3 expression; Preferably, myeloid cells are clustered according to the expression of CD11b; Preferably, the R group is patients with PFS ≥ 6 months, and the NR group is patients with PFS < 6 months; Preferably, the breast cancer is advanced HER2-negative breast cancer treated with immunotherapy combined with chemotherapy; Preferably, the immunotherapy is PD-1 inhibitor treatment.

2. A method for predicting breast cancer immunotherapy, characterized in that: The method comprises: Obtaining samples to be tested; Detecting a baseline level of target immune cells in a sample to be tested, wherein the target immune cells include CD39+Tregs and / or CCR2+moDCs; Classification prediction based on the baseline level / risk score of target immune cells to obtain the prediction results of breast cancer immunotherapy; Preferably, the prediction method based on the baseline level of target immune cells comprises: based on the baseline level of target immune cells, obtaining a prediction result of breast cancer immunotherapy; if the baseline level of CD39+Tregs increases and / or the baseline level of CCR2+moDCs decreases, obtaining a prediction result of poor treatment with breast cancer immunotherapy; if the baseline level of CD39+Tregs decreases and / or the baseline level of CCR2+moDCs increases, obtaining a prediction result of good treatment with breast cancer immunotherapy; Preferably, the target immune cells also include CD161+CD8+Tem; Preferably, the calculation formula of the risk score prediction model is h(t|X)=h0(t)exp(0.3376*[CD39+Treg]-0.0303*[CCR2+moDC]); Preferably, the prediction method based on the risk score of the target immune cells comprises: obtaining a prediction result of breast cancer immunotherapy based on the risk score of the target immune cells; if the risk score is higher than 0.84, obtaining a prediction result of poor treatment with breast cancer immunotherapy; if the risk score is lower than 0.84, obtaining a prediction result of good treatment with breast cancer immunotherapy; Preferably, the breast cancer is advanced HER2-negative breast cancer treated with immunotherapy combined with chemotherapy; Preferably, the immunotherapy is PD-1 inhibitor treatment.

3. A breast cancer immunotherapy prediction model, characterized in that: The prediction model is constructed by the method for constructing a breast cancer immunotherapy prediction model according to claim 1 or the method for predicting breast cancer immunotherapy according to claim 2; Preferably, the calculation formula of the prediction model is h(t|X)=h0(t)exp(0.3376*[CD39+Treg]-0.0303*[CCR2+moDC]); Preferably, the breast cancer is advanced HER2-negative breast cancer treated with immunotherapy combined with chemotherapy; Preferably, the immunotherapy is PD-1 inhibitor treatment.

4. A breast cancer immunotherapy prediction system, characterized in that: The system comprises: Acquisition unit: used to obtain the baseline level of immune cells in the sample to be tested; A processing unit: used for inputting the baseline level of immune cells of the sample to be tested into the prediction model constructed by the method for constructing a breast cancer immunotherapy prediction model according to claim 1 to obtain a breast cancer immunotherapy prediction result; Preferably, the breast cancer is advanced HER2-negative breast cancer treated with immunotherapy combined with chemotherapy; Preferably, the immunotherapy is PD-1 inhibitor treatment.

5. A breast cancer immunotherapy prediction system, characterized in that: The system comprises: Acquisition unit: used to obtain the baseline level of CD39+Tregs and / or CCR2+moDCs in the sample to be tested; Processing unit: based on the baseline level / risk score of the target immune cells, a prediction result of breast cancer immunotherapy is obtained; if the baseline level of CD39+Tregs increases and / or the baseline level of CCR2+moDCs decreases, a prediction result of poor treatment with breast cancer immunotherapy is obtained; if the baseline level of CD39+Tregs decreases and / or the baseline level of CCR2+moDCs increases, a prediction result of good treatment with breast cancer immunotherapy is obtained; Preferably, the target immune cells also include CD161+CD8+Tem; Preferably, the calculation formula of the risk score prediction model is h(t|X)=h0(t)exp(0.3376*[CD39+Treg]-0.0303*[CCR2+moDC]); Preferably, the prediction method based on the risk score of the target immune cells comprises: obtaining a prediction result of breast cancer immunotherapy based on the risk score of the target immune cells; if the risk score is higher than 0.84, obtaining a prediction result of poor treatment with breast cancer immunotherapy; if the risk score is lower than 0.84, obtaining a prediction result of good treatment with breast cancer immunotherapy; Preferably, the breast cancer is advanced HER2-negative breast cancer treated with immunotherapy combined with chemotherapy; Preferably, the immunotherapy is PD-1 inhibitor treatment.

6. A breast cancer immunotherapy prediction device, characterized in that: The device comprises: Memory: used to store program instructions; Processor: used to call program instructions, and when the program instructions are executed, the steps of the method for constructing a breast cancer immunotherapy prediction model according to claim 1 or the method for predicting breast cancer immunotherapy according to claim 2 are executed.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method for constructing a breast cancer immunotherapy prediction model according to claim 1 or the method for predicting breast cancer immunotherapy according to claim 2 are implemented.

8. Use of a reagent for detecting immune cells in the preparation of a product for predicting breast cancer immunotherapy, wherein the immune cells include CD39+Tregs and / or CCR2+moDCs; Preferably, the immune cells further include CD161+CD8+Tem; Preferably, the breast cancer is advanced HER2-negative breast cancer treated with immunotherapy combined with chemotherapy; Preferably, the immunotherapy is PD-1 inhibitor treatment.

9. Application of immune cells in constructing a predictive model for predicting breast cancer immunotherapy, wherein the immune cells include CD39+Tregs and / or CCR2+moDCs; Preferably, the immune cells further include CD161+CD8+Tem; Preferably, the breast cancer is advanced HER2-negative breast cancer treated with immunotherapy combined with chemotherapy; Preferably, the immunotherapy is PD-1 inhibitor treatment.

10. Use of the breast cancer immunotherapy prediction model of claim 3 in constructing a system / device / computer-readable storage medium for predicting breast cancer immunotherapy; Preferably, the breast cancer is advanced HER2-negative breast cancer treated with immunotherapy combined with chemotherapy; Preferably, the immunotherapy is PD-1 inhibitor treatment.

Images