PARP inhibitor sensitivity prediction method based on specific cell cluster and application

Abstract

The invention discloses a PARP inhibitor sensitivity prediction method based on a specific cell cluster and application. The method comprises the following steps: acquiring an ovarian cancer tumor tissue sample of a subject; detecting whether a specific cell cluster exists in the sample or not, and determining the abundance of the specific cell cluster, wherein the specific cell cluster comprises an IFN epithelial cell cluster CN12 taking an IFN tumor cell as a center and / or a TAM enrichment cluster CN29 taking an M2 type tumor related macrophage as a center; according to the detection result, predicting according to at least one of the following rules: if the CN12 abundance before treatment is greater than 0.8%, the CN29 abundance before treatment is greater than 32.7%, the CN12 abundance after treatment is greater than 0.009%, the CN29 abundance after treatment is greater than 2.9%, or the SPP1 cell density before treatment is greater than 1.7%, predicting that the drug resistance is not sensitive or easily generated. The prediction model is established from a cell cluster space tissue level, and the method has the advantages of high prediction accuracy, flexible sample requirements, high operability and the like.

Description

Technical Field

[0001] This application relates to the fields of biomedicine and molecular diagnostics, specifically to a method and application for predicting PARP inhibitor sensitivity based on specific cell clusters; more specifically, this application relates to a method for predicting the sensitivity of ovarian cancer patients to PARP inhibitor treatment by detecting the presence, abundance changes or spatial interactions of specific cell clusters (such as IFN⁺ epithelial cell clusters and tumor-associated macrophage enrichment clusters) in tumor tissue, as well as the use of related kits. Background Technology

[0002] Ovarian cancer is the deadliest gynecological malignancy. Due to its insidious onset and lack of effective early screening methods, most patients are diagnosed at an advanced stage with extensive peritoneal metastasis (Torre et al., 2018). The thoroughness of cytoreductive surgery (R0 resection) is a key factor affecting patient prognosis (Lheureux et al., 2019). However, only about 50% of patients can achieve R0 resection (Vergote et al., 1998, 2010). For patients where R0 resection is difficult to achieve, clinical guidelines recommend neoadjuvant chemotherapy followed by intermittent cytoreductive surgery. Although this strategy improves the R0 resection rate, it has not significantly improved long-term patient outcomes (Fagotti et al., 2016, 2020; Kehoe et al., 2015), which may be related to the maintenance of stem cell line regeneration or acquired resistance induced by neoadjuvant chemotherapy (Eisenhaueret et al., 2008; Liu et al., 2020).

[0003] The advent of poly(ADP-ribose) polymerase (PARP) inhibitors has brought a revolutionary breakthrough in the treatment of ovarian cancer. Their mechanism of action is based on a "synthetic lethality" effect, making them particularly suitable for patients with homologous recombination deficiency (HRD), especially those carrying BRCA1 / 2 gene mutations (Bryant et al., 2005; Farmer et al., 2005). Currently, PARP inhibitors are approved for maintenance therapy in platinum-sensitive recurrent ovarian cancer and as first-line maintenance therapy for newly diagnosed advanced ovarian cancer (González-Martín et al., 2019; Mirza et al., 2016; Moore et al., 2018). However, clinical practice has shown that even in the predominantly HRD-positive population, a significant proportion of patients remain insensitive to PARP inhibitor therapy or develop acquired resistance (Noordermeer & van Attikum, 2019).

[0004] Early studies largely attributed PARP inhibitor resistance to the restoration of DNA damage repair function or enhanced replication fork stability within tumor cells. However, mounting evidence suggests that the tumor microenvironment plays a crucial role in regulating the efficacy of PARP inhibitors. Previous studies have reported a close correlation between PARP inhibitor efficacy and the recruitment of CD8⁺ T cells within tumors (Pantelidou et al., 2019), while the elimination of tumor-associated macrophages (TAMs) or the remodeling of tumor-associated fibroblasts (CAFs) can significantly enhance the antitumor effects of PARP inhibitors (Jin et al., 2023; Mehta et al., 2021). These studies suggest that immune cells and stromal cells in the tumor microenvironment are not passive bystanders, but actively participate in the regulation of drug sensitivity.

[0005] While the aforementioned studies have revealed the role of single cell types (such as CD8⁺ T cells and TAMs) or single molecular pathways in PARP inhibitor resistance, the tumor microenvironment is a complex ecosystem composed of multiple cell types. Within this system, cells form highly ordered "cellular neighborhoods" through physical contact and paracrine signaling. The structure and function of these neighborhoods determine the overall biological behavior of the microenvironment (Schürch et al., 2020; Katzenelenbogen et al., 2020). For example, specific tumor cell subpopulations can spatially approach specific TAM subpopulations, forming pro-cancer or immunosuppressive microniches, thereby driving malignant tumor progression and treatment resistance.

[0006] However, currently used clinical biomarkers for predicting the efficacy of PARP inhibitors (such as BRCA mutation status and HRD score) are all based on the genomic characteristics of tumor cells themselves, failing to consider the complexity and spatial heterogeneity of the tumor microenvironment. More importantly, existing technologies have not yet revealed which specific "cell clusters" and their spatial interactions are directly related to the clinical response to PARP inhibitors, and there is a lack of ways to translate these spatial multi-cell interaction characteristics into actionable predictive methods.

[0007] Therefore, there is an urgent clinical need to develop a new method that can more accurately predict PARP inhibitor sensitivity from the spatial tissue level of the tumor microenvironment, in order to guide patient stratification and achieve individualized treatment. Summary of the Invention

[0008] In view of this, the purpose of this application is to provide a method and application for predicting PARP inhibitor sensitivity based on specific cell clusters, overcoming the shortcomings of existing technologies that rely on single prediction methods and fail to consider the spatial heterogeneity of the tumor microenvironment. This application predicts whether a patient is sensitive to PARP inhibitors by detecting the abundance of specific cell clusters (including the IFN⁺ epithelial cell cluster CN12 and the tumor-associated macrophage enrichment cluster CN29) in tumor tissue samples from ovarian cancer patients and based on abundance thresholds before or after treatment. This application is the first to establish a PARP inhibitor efficacy prediction model at the spatial organizational level of cell clusters, offering advantages such as high predictive accuracy, strong operability, and flexible sample requirements (both pre- and post-treatment samples are acceptable), providing a novel technical means for individualized treatment decisions for ovarian cancer patients.

[0009] To achieve the above objectives, this application provides the following technical solution:

[0010] In a first aspect, this application provides a method for predicting PARP inhibitor sensitivity based on specific cell clusters, comprising the following steps:

[0011] (1) Obtain ovarian cancer tumor tissue samples from the subjects;

[0012] (2) Detect whether a specific cell cluster exists in the tumor tissue sample and determine the abundance of the cell cluster; wherein, the specific cell cluster includes: CN12, an IFN⁺ epithelial cell cluster centered on IFN⁺ tumor cells and surrounded by immunosuppressive stromal cells, and / or CN29, a TAM-enriched cluster centered on M2 tumor-associated macrophages and surrounded by immunosuppressive cells;

[0013] (3) Based on the test results of step (2), predict the subject's sensitivity to PARP inhibitors according to at least one of the following rules:

[0014] (i) If the sample obtained is before treatment and the abundance of CN12 in the sample is higher than the first preset threshold, it is predicted to be insensitive or prone to drug resistance.

[0015] (ii) If the sample obtained is before treatment and the abundance of CN29 in the sample is higher than the second preset threshold, it is predicted to be insensitive or prone to drug resistance.

[0016] (iii) If the sample obtained is a post-treatment sample and the CN12 abundance in the sample is higher than the third preset threshold, it is predicted to be insensitive or prone to drug resistance.

[0017] (iv) If the sample obtained is a post-treatment sample and the abundance of CN29 in the sample is higher than the fourth preset threshold, it is predicted to be insensitive or prone to drug resistance.

[0018] (v) If the sample obtained is a pre-treatment sample and the SPP1⁺ cell density in the sample is higher than the fifth preset threshold, it is predicted to be insensitive or prone to drug resistance.

[0019] In some embodiments, in step (2), CN12 is a cell cluster with a total number of ≥5 cells and containing at least one type of immunosuppressive stromal cell within a radius of 25 μm centered on PanCK⁺IFN-γ⁺ tumor cells.

[0020] In some embodiments, in step (2), CN29 is a cell cluster with a total number of ≥5 cells and containing at least one immunosuppressive cell within a radius of 25 μm, centered on a macrophage with CD68⁺ and CD163⁺ or SPP1⁺.

[0021] In some implementations, the detection in step (2) is performed using spatial single-cell proteomics or multicolor immunofluorescence histochemistry.

[0022] In some implementations, in step (3) rule (i), the first preset threshold is a CN12 abundance of 0.8%.

[0023] In some implementations, in step (3) rule (ii), the second preset threshold is a CN29 abundance of 32.7%.

[0024] In some implementations, in step (3) rule (iii), the third preset threshold is a CN12 abundance of 0.009%.

[0025] In some implementations, in step (3) rule (iv), the fourth preset threshold is a CN29 abundance of 2.9%.

[0026] In some implementations, in step (3) rule (v), the fifth preset threshold is an SPP1⁺ cell density of 1.7%; the SPP1⁺ cell density is the percentage of SPP1⁺ cells to the total number of DAPI⁺ cells.

[0027] In some preferred embodiments, the prediction in step (3) further includes constructing a comprehensive scoring model, which scores and predicts according to the following method:

[0028] (a) Assign values ​​to the following four indicators respectively:

[0029] (i) If the CN12 abundance is greater than 0.8% before treatment, score 1 point; otherwise, score 0 points.

[0030] (ii) If the CN29 abundance is greater than 32.7% before treatment, score 1 point; otherwise, score 0 points.

[0031] (iii) If the CN12 abundance is greater than 0.009% after treatment, 1 point is awarded; otherwise, 0 points are awarded.

[0032] (iv) If the CN29 abundance is greater than 2.9% after treatment, score 1 point; otherwise, score 0 points.

[0033] (v) If the SPP1⁺ cell density is greater than 1.7% before treatment, score 1 point; otherwise, score 0 points.

[0034] (b) Add up the scores of the five indicators in (a) to get the total score. If the total score is ≥3, it is predicted that the subject is insensitive to PARP inhibitors or is prone to developing drug resistance.

[0035] Secondly, this application provides the use of the method described in the first aspect or a kit for detecting CN12 and / or CN29 in the method described in the first aspect in the preparation of products for predicting PARP inhibitor sensitivity.

[0036] Compared with the prior art, this application has at least the following advantages and beneficial effects:

[0037] 1. This application establishes, for the first time, a predictive model for PARP inhibitor efficacy at the spatial organization level of cell clusters, overcoming the limitations of traditional predictions based on a single molecule or cell type. By identifying specific IFN⁺ epithelial cell clusters CN12 and TAM-enriched clusters CN29, and quantifying their abundance changes and spatial distances, this application, for the first time, uses "cell clusters" as a spatial functional unit as a predictive biomarker, significantly improving the accuracy of predicting PARP inhibitor sensitivity.

[0038] 2. This application addresses potential differences in sample types that may arise in clinical practice by establishing prediction rules based on pre-treatment samples (SPP1⁺ macrophage ratio) and paired pre- and post-treatment samples (CN12 and CN29 abundance). Clinicians can choose the appropriate prediction method based on the available sample types, demonstrating high flexibility in clinical application.

[0039] 3. The prediction method in this application has clear and quantifiable interpretation criteria, and its prediction accuracy is high. Validation data is based on the NANT clinical trial cohort, where:

[0040] Based on the prediction rule of CN12 abundance before treatment, with a threshold of 0.8%, the sensitivity for predicting drug resistance is 0.46, the specificity is 1, and the AUC value is 0.69.

[0041] Based on the prediction rule of CN29 abundance before treatment, with 32.7% as the threshold, the sensitivity for predicting drug resistance is 0.77, the specificity is 0.57, and the AUC value is 0.64.

[0042] Based on the prediction rule of CN12 abundance after treatment, with a threshold of 0.009%, the sensitivity for predicting drug resistance is 0.89, the specificity is 1, and the AUC value is 0.94.

[0043] Based on the prediction rule of CN29 abundance after treatment, with a threshold of 2.9%, the sensitivity for predicting drug resistance is 0.53, the specificity is 1, and the AUC value is 0.78.

[0044] Based on the predictive rule of pretreatment SPP1⁺ cell density, the objective response rate of the high-density SPP1⁺ cell group (n=10) was only 30.0%, while the objective response rate of the low-density group (n=21) reached 76.2% (P=0.021), with median progression-free survival of 13.5 months and 28.3 months, respectively (P=0.0006).

[0045] 4. This application also provides a comprehensive scoring model, which further improves the stability and accuracy of prediction. By comprehensively scoring the above five indicators, and using a total score of ≥3 points as the standard for drug resistance prediction, it demonstrates excellent predictive performance.

[0046] In summary, this application provides a novel technical means for individualized treatment decisions using PARP inhibitors for ovarian cancer patients, and has promising clinical translation prospects and application value. Attached Figure Description

[0047] Figure 1 This is a PCF staining diagram. It shows the staining status of each channel in the 39-index PCF.

[0048] Figure 2 This is a graph showing the cell grouping results for PCF cells. In Figure A, 3,796,556 cells from the PCF sample were classified into epithelial cells, fibroblasts, endothelial cells, and CD4+ cells after UMAP dimensionality reduction. + T cells, CD8 + T cells, neutrophils, and other various immune cell subsets are shown, along with the proportion of each cell type. B is a heatmap showing the expression of typical marker genes in different cell classes, indicating that each cell population specifically expresses the corresponding molecular markers.

[0049] Figure 3 This is a diagram showing the subgrouping results of PCF cells.

[0050] Figure 4Figure A shows the results of PCF cell cluster analysis. Figure A illustrates the spatial cell structure pattern of PCF samples, calculated based on spatial windows and divided into cell neighborhoods (CNs) using K-means clustering. Figure B shows that the optimal number of clusters for K-means clustering was determined to be 40 using the elbow rule. Figure C is a heatmap showing the enrichment of different cell neighborhoods in various cell subpopulations, demonstrating the diverse spatial structural features within PCF tissues.

[0051] Figure 5 The diagram shows the cellular composition of the CN12 and CN29 cell clusters. Figure A shows that the CN12 cell neighborhood is mainly composed of Epi_IFNG⁺ epithelial cells, accompanied by a certain proportion of stromal cells and various immune cell subsets. Figure B shows that the CN29 cell neighborhood is dominated by TAM-related subsets, including macrophages with various activated and immune-regulating phenotypes and a small number of T cell components.

[0052] Figure 6 This is a graph showing the relationship between the proportion of major cell types and the efficacy of PARP inhibitors. Figure A shows the pre-treatment efficacy groupings for different RECIST groups (PR vs SD). + Comparison of the proportions of each major cell class among PD patients; Figure B shows the grouping analysis according to pathological reaction.

[0053] Figure 7 This is a graph showing the ratio of cell subgroups and the efficacy of PARP inhibitors. Graph A shows the comparison between different efficacy groups (PR, NR) and before and after treatment. Graph B shows the proportions of key immune cell subsets and cells before and after treatment, as well as between the responding and non-responding groups.

[0054] Figure 8 Figure 1 shows the results of single-cell RNA-seq analysis revealing key cell subpopulations related to the efficacy of PARP inhibitors. Specifically: Figure A: scRNA-seq analysis showed significant differences in the proportion of TAMs in the matrix among different efficacy groups; Figure B: Venn analysis revealed a specific increase in cell subpopulations in the NR group compared to pre-treatment and the response group, and screened out key TAM subtypes with common changes; Figure C: Changes in Log2FC and GSVA scores of differentially expressed TAM subpopulations across different efficacy comparisons, suggesting functional remodeling; Figure E: IFN-responsive TAMs were significantly enriched in the NR group; Figure F: Epi_IFN-responsive cells, CD4⁺ eTregs, and myCAFs showed differential changes before and after treatment and among efficacy groups.

[0055] Figure 9Figure 1 shows the results of the PCF cell cluster and PARP inhibitor efficacy analysis. Figure A shows the spatial cellular structure pattern of the PCF samples, calculated based on spatial windows and divided into cell neighborhoods (CNs) using K-means clustering. Figure B shows the optimal number of clusters for K-means clustering determined using the elbow rule. Figure C is a heatmap showing the enrichment of different cell neighborhoods in various cell subpopulations, illustrating the diverse spatial structural features within PCF tissues.

[0056] Figure 10-13 ROC curve analysis of CN12 abundance before treatment, CN29 abundance before treatment, CN12 abundance after treatment, and CN29 abundance after treatment showed that the optimal cutoff values ​​corresponding to the maximum Youden index were 0.8%, 32.7%, 0.009%, and 2.9%, respectively.

[0057] Figure 14 Figure A shows the association between SPP1⁺ cell density and the efficacy of PARP inhibitors. Figure A: Clinical response rate (complete response [CR] + partial response [PR]) of niraparib monotherapy in the NANT cohort, stratified by low (n=21) and high (n=10) SPP1⁺ cell infiltration density. Two-sided Fisher's exact test was used, P=0.021. Figure B: Kaplan-Meier estimate of progression-free survival (PFS) in the NANT trial, stratified by low (n=28) and high (n=11) SPP1⁺ cell infiltration density (one patient was excluded due to loss to follow-up). Log-rank test was used, P=0.0006.

[0058] Figure 15 The image shows the results of multicolor histochemistry analysis. The abundance differences of TAM, effector Treg, and IFN⁺ tumor cells in sensitive and drug-resistant samples were verified using three 7-color histochemistry protocols (macrophage protocol, T cell protocol, and CAF protocol). Detailed Implementation

[0059] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0060] The materials used in the following embodiments are not limited to those listed below, and other similar materials may be used instead. Unless otherwise specified, the instruments shall be used under conventional conditions or as recommended by the manufacturer. Those skilled in the art should have relevant knowledge of the use of conventional materials and instruments.

[0061] To better understand this teaching and without limiting its scope, all figures and other numerical values ​​used in the specification and claims to express quantities, percentages, or proportions should, in all cases, be understood to be modified by the term "about." Therefore, unless otherwise stated, the numerical parameters set forth in the following specification and appended claims are approximate values ​​that may vary depending on the desired properties sought. At a minimum, each numerical parameter should be interpreted based at least on the reported significant figures and by applying common rounding techniques.

[0062] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the subject matter of this application pertains. Before providing a detailed description of this application, the following terms and definitions are provided to better understand it.

[0063] 1. Cellular Neighborhood (CN): A multicellular structural unit composed of multiple cells of different lineages or functional states within a certain radius centered on a central cell in a tissue space. Cellular clusters reflect the spatial proximity and potential interactions between cells in the tumor microenvironment and are the basic units constituting the functional niche of a tissue.

[0064] 2. IFN⁺ Epithelial Cell Cluster CN12: Refers to a specific cell cluster as defined in this application, centered on tumor cells expressing interferon-related markers. In one specific embodiment, CN12 is defined as: a cell cluster with a total cell count ≥5 within a circular region of 25 μm radius centered on PanCK⁺IFN-γ⁺ tumor cells, and containing at least one type of immunosuppressive stromal cell (including but not limited to tumor-associated macrophages, regulatory T cells, cancer-associated fibroblasts, etc.).

[0065] 3. TAM-enriched cluster CN29: refers to a specific cell cluster as defined in this application, which is centered on a specific subtype of tumor-associated macrophages. In one specific embodiment, CN29 is defined as: a cell cluster with a total number of ≥5 cells in a circular region with a radius of 25μm centered on CD68⁺ and CD163⁺ cells, and the region contains at least one type of immunosuppressive cell (including but not limited to regulatory T cells, cancer-associated fibroblasts, etc.).

[0066] 4. IFN⁺ tumor cells: These refer to epithelial-derived tumor cells exhibiting characteristics of interferon signaling pathway activation. In this application, IFN⁺ tumor cells are defined as PanCK⁺ and IFN-γ⁺ cells, detected by immunohistochemistry or spatial proteomics methods, or as cells positive for other interferon-related markers (such as p-STAT1, IFNGR1, etc.).

[0067] 5. Tumor-Associated Macrophages (TAMs): These are macrophages that infiltrate tumor tissue. In this application, TAMs are identified using macrophage lineage markers, including but not limited to CD68, CD163, CD206, and SPP1. Based on their functional phenotype, TAMs can be further divided into subpopulations such as M2 TAMs (CD68⁺CD163⁺) and SPP1⁺ TAMs (CD68⁺SPP1⁺).

[0068] 6. M2 type tumor-associated macrophages: refers to a subset of tumor-associated macrophages exhibiting M2 polarization, which is typically associated with immunosuppression, tissue repair, and tumor progression. In this application, M2 type TAMs are defined as cells with CD68⁺ and CD163⁺.

[0069] 7. SPP1⁺ Tumor-associated macrophages: refers to a subset of tumor-associated macrophages that highly express secreted phosphoprotein 1 (SPP1, also known as osteopontin). In this application, SPP1⁺ TAM is defined as CD68⁺ cells with SPP1⁺.

[0070] 8. Immunosuppressive stromal cells: These are non-solid cells that have immunosuppressive functions in the tumor microenvironment, including but not limited to tumor-associated macrophages (TAMs), regulatory T cells (Tregs), cancer-associated fibroblasts (CAFs), and myeloid-derived suppressor cells (MDSCs).

[0071] 9. Immunosuppressive cells: In a broad sense, this refers to all immune and non-immune cells with immunosuppressive functions. In this application, it specifically refers to, but is not limited to, regulatory T cells (Tregs), tumor-associated macrophages (TAMs), myeloid-derived suppressor cells (MDSCs), etc.

[0072] 10. Abundance: refers to the proportion of a specific cell cluster or cell type in a sample. For cell clusters, abundance is usually expressed as the percentage of cells in that cluster out of the total number of cells in the sample.

[0073] 11. SPP1⁺ cell density: refers to SPP1 + The proportion of cells to the total number of cells.

[0074] 12. PARP inhibitors: These refer to poly(ADP-ribose) polymerase inhibitors, a class of targeted drugs that inhibit DNA damage repair and induce tumor cell synthesis leading to death. The PARP inhibitors described in this application include, but are not limited to, niraparib, olaparib, and rucaparib.

[0075] Other matters that need to be explained before writing the implementation examples:

[0076] 1. Regarding the detection methods, the detection methods described in this application include, but are not limited to, spatial single-cell proteomics technology (such as the PhenoCycler-Fusion platform) and multicolor immunofluorescence histochemistry (mIHC). Those skilled in the art can select a suitable technology platform to implement this application based on actual conditions and needs, as long as it can achieve spatial localization and biomarker detection at the single-cell level.

[0077] 2. Regarding the adjustability of the threshold

[0078] The specific values ​​given in this application (such as radius 25 μm, cell number ≥ 5, abundance variation threshold, SPP1⁺ cell density ≥ 1.7%) are based on the optimized results of the clinical trial cohort upon which this application is based. In practical applications, these thresholds can be appropriately adjusted according to the specific detection platform, staining conditions, image analysis algorithms, and characteristics of the target population, or the optimal cutoff value can be re-determined through receiver operating characteristic (ROC) curve analysis. All such adjustments fall within the scope of protection of this application.

[0079] 3. Regarding the combined use of forecasting rules

[0080] This application provides multiple prediction rules (based on changes in abundance before and after treatment, based on spatial distance after treatment, based on SPP1⁺ density before treatment, and a comprehensive scoring model). These rules can be used individually or in combination according to clinical needs and sample accessibility. When multiple rules are applied in combination, the accuracy of prediction can be further improved.

[0081] The following are specific examples:

[0082] Example 1: Patient pairing and sample collection

[0083] The samples used in this embodiment came from a multicenter phase II clinical trial, NANT (NCT04507841), which was the first clinical study to use PARP inhibitors as neoadjuvant therapy.

[0084] 1.1 Patient inclusion criteria

[0085] Enrolled patients were women aged 18–75 years with histologically confirmed new-onset high-grade serous or endometrioid ovarian cancer, primary peritoneal cancer, or fallopian tube cancer. Patients were assessed as unresectable at baseline (abdominal CT score ≥3 or laparoscopic Fagotti score ≥8). All patients were positive for homologous recombination deficiency (HRD), defined by a genomic instability score ≥42 or confirmed BRCA1 / 2 mutations.

[0086] 1.2 Treatment Plan

[0087] Enrolled patients received two 28-day cycles of neoadjuvant niraparib monotherapy (200 mg or 300 mg orally once daily) followed by intermittent cytoreductive surgery.

[0088] 1.3 Sample Collection

[0089] During baseline diagnostic laparoscopy, paired tumor specimens were collected from each patient:

[0090] Fresh tumor specimens: immediately after ex vivo, they were placed in tissue preservation solution at 4 ℃. One sample was used for single-cell RNA sequencing, and the other was routinely processed to prepare formalin-fixed paraffin-embedded (FFPE) tissue blocks.

[0091] Post-treatment samples: After completing two cycles of niraparib treatment, tumor resection specimens were obtained during debulking surgery and prepared as FFPE tissue blocks.

[0092] 1.4 Ethical Statement

[0093] All participants signed a written informed consent form before the screening began, and the research protocol has been approved by the Ethics Committee of Tongji Medical College, Huazhong University of Science and Technology (approval number: 2020-S122).

[0094] Example 2: Pathological Remission Assessment

[0095] This embodiment defines the "gold standard" for distinguishing between PARP inhibitor-sensitive and resistant patients. Subsequent embodiments will use this standard to verify the accuracy of the prediction method.

[0096] 2.1 Evaluation Criteria

[0097] This study used a modified histopathological grading system to evaluate post-treatment specimens at the lesion level:

[0098] Pathological remission (sensitive group): defined as complete disappearance of surviving tumors or near-complete regression of tumors (only scattered single tumor cells or nodules with a maximum diameter ≤2 mm remaining).

[0099] Pathological non-remission (drug-resistant group): Specimens that do not meet the above criteria, including specimens that are mainly composed of surviving tumors or show only limited regression.

[0100] 2.2 Grouping Results

[0101] Based on the above criteria, 39 patients with complete matched samples were selected from the NANT clinical trial, among whom:

[0102] PCF: Pre-treatment samples were assessed according to RECIST 1.1: R (PR, 13), NR (SD+PD, 7); Post-treatment samples were assessed according to pathological remission: R (21), NR (18), and so on.

[0103] Sc-RNA: pre(17), R(8), NR(5)

[0104] Example 3: Tissue microarray preparation and PCF detection

[0105] This embodiment describes in detail the specific operation process of space single-cell omics detection.

[0106] 3.1 Tissue microarray fabrication

[0107] The 182 FFPE paraffin blocks collected in Example 1 were prepared into tissue microarrays. One to four representative regions (1.5 mm in diameter) from each paraffin block were arranged in the recipient paraffin block to form 4 μm thick sections.

[0108] 3.2 Main Reagents and Instruments

[0109] The PhenoCycler-Fusion system consists of multiple parts: imaging component: AB000002 PhenoimagerFusion; staining component: INST1000 PhenoCycler Instrument; AB000003 PhenoCycler FusionAccessory; all manufactured by Akoya Biosciences.

[0110] PhenoCycler Sample Kit; the English name of the staining kit is Sample Kit for PhenoCycler-Fusion, the catalog number is 7000017, and the manufacturer is Akoya Biosciences.

[0111] 39 oligonucleotide barcode-labeled antibodies (Akoya Biosciences). The antibody panel includes core markers: PanCK, EpCAM, IFN-γ, CD68, CD163, SPP1, CD3, CD4, CD8, FOXP3, PD-1, α-SMA, FAP, etc.; antibody information is shown in Table 1 below.

[0112] Histochoice clearing agent; Tris-EDTA buffer (pH 9.0, self-prepared); paraformaldehyde.

[0113] Table 1 Antibody Information

[0114]

[0115] 3.3 Staining Procedure

[0116] (1) Dewaxing and hydration: The sections were baked at 55 °C for 1 hour, dewaxed with Histochoice clearing agent, and hydrated with graded ethanol;

[0117] (2) Antigen retrieval: Heat-mediated antigen retrieval was performed in Tris-EDTA buffer (pH 9.0) and cooled to room temperature;

[0118] (3) Blocking: Treat the slides with the pre-blocking solution provided in the PhenoCycler Sample Kit;

[0119] (4) Primary antibody incubation: Mix the 39 oligonucleotide barcode-labeled antibodies at the optimal concentration and incubate them with the slides at 4 °C overnight;

[0120] (5) Fixation: After washing the next day, fixation was performed three times in sequence using paraformaldehyde, methanol, and fixation reagent;

[0121] (6) Imaging: The flow cell was assembled onto the slide, and the working solution for the fluorescent reporter group was prepared. Hybridization, imaging, and dehybridization cycles were automatically performed on the PhenoCycler-Fusion system. Images of DAPI and three markers were acquired in each cycle. Exposure times were: DAPI 1 ms, ATTO550 150 ms, AF647 150 ms, and AF750 150 ms. Full-field QPTIFF images were finally obtained. Figure 1 The diagram shows the PCF staining protocol, illustrating the combination of 39 antibodies and the imaging workflow of the PhenoCycler-Fusion system.

[0122] Example 4: Image Processing and Cell Segmentation

[0123] This embodiment details the processing flow from raw images to single-cell data.

[0124] 4.1 Image Preprocessing

[0125] QPTIFF images were processed using QuPath software (v0.3.2). Defocused areas, tissue folds, debris, and other contaminants were manually removed.

[0126] 4.2 Cell Segmentation

[0127] Cell nucleus segmentation was performed using the StarDist deep learning algorithm (v0.3.2, default parameters). The cytoplasm was defined by extending 5 μm outward through a nuclear mask. The centroid coordinates and average fluorescence intensity of 39 markers for each cell were extracted.

[0128] 4.3 Quality Control

[0129] Low-quality cells should be removed according to the following criteria:

[0130] (1) Cells with a total fluorescence intensity of all markers <10 or >2000 (based on 99.5 percent of the total intensity distribution of all cells).

[0131] (2) Cells with a single marker fluorescence intensity higher than the mean of the marker plus 3 times the standard deviation, and abnormally high signal in three or more markers at the same time; (3) Cells with a DAPI nuclear staining area of ​​<20 pixels² or >500 pixels².

[0132] After quality control, a total of 3,796,556 cells were retained for subsequent analysis.

[0133] Example 5: Annotation of Cell Types and Functional Subpopulations

[0134] This embodiment describes in detail a cell type annotation method based on marker expression patterns.

[0135] 5.1 Annotation of Major Cell Types

[0136] Principal component analysis, UMAP dimensionality reduction, and unsupervised clustering were performed based on 19 cell type markers (CD45, CD20, CD79a, CD38, CD3e, CD4, CD8, CD68, CD14, HLA-DR, CD11c, MPO, CD31, CD34, Podoplanin, Vimentin, αSMA, PanCK, and EpCAM). Major cell types were annotated according to the expression patterns of each cluster marker.

[0137] B cells: CD45⁺CD20⁺; plasma cells: CD79a⁺CD38⁺; T cells: CD45⁺CD3e⁺; tumor-associated macrophages / dendritic cells: CD45⁺CD14⁺CD68⁺; neutrophils: CD45⁺MPO⁺; endothelial cells: CD31⁺; fibroblasts: Vimentin⁺αSMA⁺; epithelial cells: PanCK⁺EpCAM⁺.

[0138] Figure 2 This is a graph showing the results of PCF cell classification by major categories. A shows the 3,796,556 cells in the PCF sample, after dimensionality reduction using UMAP, classified into epithelial cells, fibroblasts, lymphatic endothelial cells, vascular endothelial cells, neutrophils, and various immune cell subpopulations, with the proportion of each cell type displayed. B shows a heatmap of typical marker genes expressed in different cell categories, indicating that each cell population specifically expresses corresponding molecular markers.

[0139] 5.2 Fine-grained annotation of functional subgroups

[0140] Secondary clustering and fine-grained annotation of key functional subgroups:

[0141] IFN⁺ tumor cells: The result of dimensionality reduction clustering classification of PCF data is a subset of epithelial cells;

[0142] SPP1⁺ TAM: Two-color immunohistochemistry, CD68 / SPP1 double staining, double positive indicates SPP1. + Macrophagocytosis;

[0143] M2 type TAM: CD68⁺ and CD163⁺, HLADR-low;

[0144] Effector Tregs: Double positive for FOXP3⁺ and PD-1⁺ in CD3⁺CD4⁺ T cells;

[0145] myCAF: The result of dimensionality reduction clustering classification of PCF data, which is a subpopulation of CAF cells.

[0146] Figure 3 The image shows the subgrouping results of PCF cells, illustrating the results of secondary clustering and fine annotation of key functional subgroups (including IFN⁺ tumor cells, SPP1⁺TAM, M2 type TAM, effector Treg, myCAF, etc.).

[0147] Example 6: Identification and Definition of Cell Clusters

[0148] This embodiment describes in detail a method for identifying specific cell clusters CN12 and CN29 from single-cell spatial data.

[0149] 6.1 Preliminary identification of cell clusters

[0150] A circular window with a radius of 25 μm was defined centered on each cell, and the types of all cells within the window were counted. Windows with less than 5 cells in each window (accounting for 9.96%) were removed. MiniBatchKMeans clustering was performed on the cell type vectors of each window, and the optimal number of clusters was determined to be 40 using the elbow rule, resulting in the identification of 40 cell clusters (CN1-CN40).

[0151] Figure 4 Figure A shows the results of PCF cell cluster analysis. Figure A illustrates the spatial cell structure pattern of PCF samples, calculated based on spatial windows and divided into cell neighborhoods (CNs) using K-means clustering. Figure B shows that the optimal number of clusters for K-means clustering was determined to be 40 using the elbow rule. Figure C is a heatmap showing the enrichment of different cell neighborhoods in various cell subpopulations, demonstrating the diverse spatial structural features within PCF tissues.

[0152] 6.2 Definition and Identification of CN12

[0153] CN12 is defined as an IFN⁺ epithelial cell cluster, and its identification criteria are as follows:

[0154] (1) The central cell is a PanCK⁺IFN-γ⁺ tumor cell (i.e., IFN⁺ tumor cell); (2) The total number of cells in a circular window with a radius of 25 μm centered on the central cell is ≥5; (3) The window contains at least one type of immunosuppressive stromal cell, including but not limited to TAM, Treg, myCAF, PD-L1⁺ DC, exhausted CD8⁺ T cells, etc.

[0155] The identification parameters of a radius of 25 μm and a cell count ≥5 were determined through systematic parameter optimization. Specifically, combinations of different radii (e.g., 20 μm, 25 μm, 30 μm, 35 μm) with different neighbor cell count thresholds (e.g., 2, 3, 4, 5) were tested to evaluate the stability of the identified cell clusters under each parameter combination and their correlation with clinical efficacy. By comparing the clustering effects of different parameter combinations (e.g., homogeneity of cell composition within clusters, discriminability between clusters) and the strength of their association with patient prognosis, a radius of 25 μm and a total cell count ≥5 were ultimately selected as the optimal parameter combination. The central cell and all cells within its window that meet all the above conditions are marked as members of the CN12 cluster.

[0156] like Figure 5 As shown in A, the CN12 cell neighborhood is mainly composed of Epi_IFNG⁺ epithelial cells, accompanied by a certain proportion of stromal cells and various immune cell subsets.

[0157] 6.3 Definition and Identification of CN29

[0158] CN29 is defined as a TAM enrichment cluster, and its identification criteria are as follows:

[0159] (1) The central cell is an M2 type TAM (CD68⁺CD163⁺); (2) The total number of cells in a circular window with a radius of 25 μm centered on the central cell is ≥5; (3) The window contains at least one type of immunosuppressive cell, including but not limited to Treg, myCAF, etc. The central cell and all cells in the window that meet all the above conditions are marked as members of the CN29 cluster.

[0160] like Figure 5 As shown in B, the CN29 cell neighborhood is dominated by TAM-related subsets, including macrophages with various activation and immune-regulating phenotypes and a small number of T cell components.

[0161] Example 7: Association analysis between CN12 and CN29 abundance and therapeutic efficacy

[0162] This embodiment, based on the grouping results of Example 2, analyzes the association between CN12 and CN29 abundance and the efficacy of PARP inhibitors. The results are as follows: Figure 6-9 As shown.

[0163] 7.1 Abundance Calculation Method For each sample, the abundance of CN12 and CN29 is calculated using the following formulas:

[0164] CN12 abundance = (Number of CN12 cell clusters / Total number of epithelial clusters) × 100%;

[0165] CN29 abundance = (Number of CN29 cell clusters / Total number of interstitial clusters) × 100%.

[0166] 7.2 Analysis of Abundance Changes Before and After Treatment

[0167] Figure 6 This is a graph showing the relationship between the proportion of major cell types and the efficacy of PARP inhibitors. Graph A compares the proportions of each cell type before treatment in different RECIST response groups (PR vs SD+PD), as shown in the figure. There was no statistically significant difference in the proportions of each cell type between the different response groups before treatment. Graph B shows the analysis by pathological response group, as shown in the figure. After niraparib treatment, there were significant differences between the response group and the non-response group in the proportions of certain cell subsets, such as fibroblasts, TAMs, and neutrophils.

[0168] Figure 7This is a graph showing the relationship between cell subpopulation proportions and PARP inhibitor efficacy. Figure A compares different efficacy groups (PR, NR) and pre- and post-treatment outcomes, demonstrating differential changes in various cell subpopulations in niraparib treatment-related responses. Figure B shows statistically significant differences in the proportions of key immune cell subpopulations and cells before and after treatment, as well as between the responding and non-responding groups.

[0169] Further analysis of the scRNA-seq data of the samples confirmed the above findings. Figure 8 Figure 1 shows the results of single-cell RNA-seq data analysis. Figure A shows that scRNA-seq analysis revealed significant differences in the proportion of TAMs in the matrix among different efficacy groups; Figure B shows Venn analysis revealing a specific increase in cell subpopulations in the NR group compared to pre-treatment and the response group, and identifying key TAM subtypes with common changes; Figure C shows changes in Log2FC and GSVA scores of differentially expressed TAM subpopulations across different efficacy comparisons, suggesting functional remodeling; Figure E shows significant enrichment of IFN-responsive TAMs in the NR group; Figure F shows differential changes in cell subpopulations such as Epi_IFN-responsive cells, CD4⁺ eTregs, and myCAFs before and after treatment, and between efficacy groups.

[0170] Figure 9 Figure 1 shows the results of the PCF cell cluster and PARP inhibitor efficacy analysis. Figure A shows the spatial cellular structure pattern of the PCF samples, calculated based on spatial windows and divided into cell neighborhoods (CNs) using K-means clustering. Figure B shows the optimal number of clusters for K-means clustering determined using the elbow rule. Figure C is a heatmap showing the enrichment of different cell neighborhoods in various cell subpopulations, illustrating the diverse spatial structural features within PCF tissues.

[0171] Using patient efficacy grouping (sensitive vs. resistant) as the state variable and CN12 and CN29 abundance as the test variables, receiver operating characteristic (ROC) curves were plotted, and the optimal cutoff values ​​were determined as follows:

[0172] Pre-treatment CN12 abundance: ROC curve analysis showed an area under the curve (AUC) of 0.69, with the optimal cutoff value corresponding to the maximum Youden index being 0.8%. Using this threshold to predict drug resistance, the sensitivity was 0.46 and the specificity was 1. Figure 10 As shown.

[0173] Pre-treatment CN29 abundance: ROC curve analysis showed an area under the curve (AUC) of 0.64, with the optimal cutoff value corresponding to the maximum Youden index being 32.7%. Using this threshold to predict drug resistance, the sensitivity was 0.77 and the specificity was 0.57. Figure 11 As shown.

[0174] Post-treatment CN12 abundance: ROC curve analysis showed an area under the curve (AUC) of 0.94, with the optimal cutoff value corresponding to the maximum Youden index being 0.009%. Using this threshold to predict drug resistance, the sensitivity was 0.89 and the specificity was 1. Figure 12 As shown.

[0175] Post-treatment CN29 abundance: ROC curve analysis showed an area under the curve (AUC) of 0.78, with the optimal cutoff value corresponding to the maximum Youden index being 2.9%. Using this threshold to predict drug resistance, the sensitivity was 0.53 and the specificity was 1. Figure 13 As shown.

[0176] Example 8: Correlation Analysis of SPP1⁺ Cell Density and Therapeutic Effect

[0177] This embodiment analyzes the correlation between SPP1⁺ cell density and the efficacy of PARP inhibitors in pre-treatment samples.

[0178] 8.1 SPP1⁺ Cell Density Calculation Method: SPP1⁺ cells (including SPP1⁺ TAM and other SPP1⁺ stromal cells) were identified according to the method in Example 5, and statistical analysis was performed on a single cell density.

[0179] The number of SPP1⁺ cells per mm² and the percentage of SPP1⁺ cells in the total number of cells.

[0180] 8.2 Determination and Grouping of Preset Threshold for SPP1⁺ Cell Density

[0181] Multicolor immunofluorescence histochemistry was used to double-stain tumor tissue sections with CD68 and SPP1. Four representative fields of view were selected, and the number of SPP1⁺ cells and the total number of DAPI⁺ cells in each field of view were counted. The percentage of SPP1⁺ cells to the total number of DAPI⁺ cells was calculated as the SPP1⁺ cell density, and the average value of the four fields of view was taken. If the SPP1⁺ cell density was greater than 1.7%, it was determined to be high density, that is, it met the preset threshold.

[0182] The preset threshold is: multichromatic CD68 / SPP1 double staining; for each slice, four fields of view are taken, and the SPP1 value for each field of view is calculated. + Cell count, DAPI count represents the total number of cells, SPP1 + Cell / DAPI count is SPP1 +Cell density was calculated as the average of four fields of view. To determine the optimal predictive threshold for SPP1⁺ cell density, receiver operating characteristic (ROC) curve analysis was used. Patient efficacy grouping (sensitive vs. resistant) was used as the state variable, and SPP1⁺ cell density as the test variable. ROC curve analysis showed that the optimal cutoff value was 1.7%, corresponding to the maximum Youden's index (sensitivity + specificity - 1). At this threshold, the optimal balance between sensitivity and specificity in predicting resistance was achieved. Based on this, patients were divided into two groups:

[0183] High-density group (n=10): SPP1⁺ cell density >1.7%;

[0184] Low-density group (n=21): SPP1⁺ cell density ≤1.7%; another 8 cases were excluded from the analysis due to sample quality issues.

[0185] 8.3 Correlation Analysis of Therapeutic Effect

[0186] (1) Objective response rate in the two groups: The objective response rate in the high-density group (n=10) was 30.0% (3 / 10), significantly lower than that in the low-density group (n=21) at 76.2% (16 / 21). The difference between the two groups was statistically significant (P=0.021, two-sided Fisher's exact test). Figure 14 As shown in Figure A.

[0187] (2) Progression-free survival: The median progression-free survival in the high-density group (n=11) was 13.5 months (95% CI: 8.2–18.8 months), significantly shorter than that in the low-density group (n=28) at 28.3 months (95% CI: 21.6–35.0 months), with a significant difference between the two groups (P=0.0006, Log-rank test). Figure 14 As shown in B.

[0188] The above results indicate that using 1.7% as the preset threshold for SPP1⁺ cell density can effectively distinguish between PARP inhibitor-sensitive and drug-resistant patients, and this threshold has good predictive efficacy.

[0189] like Figure 8 As shown in C, the changes in Log2FC and GSVA scores of the differential TAM subgroups in different efficacy comparisons suggest functional remodeling, and SPP1⁺ TAM is significantly enriched in the drug-resistant group.

[0190] Example 9: Establishment and Verification of the Comprehensive Scoring Model

[0191] Based on the above findings, this embodiment establishes a comprehensive scoring model and verifies its predictive performance on an independent validation set.

[0192] 9.1 Construction of the Comprehensive Scoring Model

[0193] Assign values ​​to the following five indicators:

[0194] Indicator 1: CN12 abundance greater than 0.8% before treatment, score 1 point; otherwise, score 0 points;

[0195] Indicator 2: CN29 abundance greater than 32.7% before treatment, score 1 point; otherwise, score 0 points;

[0196] Indicator 3: CN12 abundance greater than 0.009% after treatment, score 1 point; otherwise, score 0 points.

[0197] Indicator 4: CN29 abundance greater than 2.9% after treatment, score 1 point; otherwise, score 0 points.

[0198] Indicator 5: SPP1⁺ cell density greater than 1.7% before treatment, score 1 point; otherwise, score 0 points.

[0199] The scores of the five indicators are added together to obtain a total score, ranging from 0 to 5. A total score of ≥3 is used as the criterion for predicting PARP inhibitor resistance.

[0200] 9.2 Independent Validation Set

[0201] An additional 15 patients from the NANT clinical trial with paired samples were included as an independent validation set, among whom:

[0202] Sensitive group: 8 cases (pathological remission);

[0203] Drug-resistant group: 7 cases (pathologically unresolved).

[0204] The prediction was validated using a blind method, meaning that the predictor calculated the comprehensive score and made a prediction without knowing the patient's actual treatment group.

[0205] Example 10: Multichromosome-based verification of changes in key cell subpopulations

[0206] This embodiment uses multicolor immunofluorescence histochemistry to verify the key cell subpopulation changes found in the aforementioned scRNA-seq and PCF analyses at the protein level.

[0207] 10.1 Design of Multicolor Grouping Scheme

[0208] Three 7-color histochemistry schemes were designed for three key cell subpopulations, as shown in Table 2 below.

[0209] Table 2. Component Scheme

[0210]

[0211] 10.2 Staining Procedure (1) Section Preparation: 4 μm thick continuous sections were prepared from the FFPE wax blocks collected in Example 1;

[0212] (2) Dewaxing and hydration: The sections were baked at 60°C for 2 hours, dewaxed with xylene, and hydrated with graded ethanol;

[0213] (3) Antigen retrieval: Autoclaving and thermal retrieval were performed in citrate buffer (pH 6.0);

[0214] (4) Blocking: Block with 10% goat serum at room temperature for 1 hour;

[0215] (5) Primary antibody incubation: Mix the antibody according to the above protocol and incubate overnight at 4°C;

[0216] (6) Secondary antibody incubation: Incubate with secondary antibodies labeled with the corresponding fluorescent dyes (such as AF488, AF555, AF647) at room temperature for 1 hour;

[0217] (7) Nuclear staining and mounting: stain the nuclei with DAPI for 5 minutes, and mount with anti-fluorescence quenching mounting medium;

[0218] (8) Imaging: Images are acquired using a multispectral imaging system, and autofluorescence is removed by spectral splitting.

[0219] 10.3 Image Analysis and Quantitative Analysis

[0220] Cell identification and quantification using image analysis software:

[0221] Identification criteria: Refer to Example 5 (e.g., M2 type TAM is CD68⁺CD163⁺, effect Treg is CD4⁺FOXP3⁺PD1⁺, etc.);

[0222] Statistical methods: Calculate the density (cells / mm²) of each cell subpopulation in the tissue and its proportion of the corresponding cell lineage.

[0223] 10.4 Verification Results

[0224] The multicolor histochemistry results were highly consistent with the scRNA-seq and PCF analysis results:

[0225] 10.4.1 Macrophage Protocol

[0226] In the sensitive group, the density of CD68⁺CD163⁺ M2 type TAM was significantly lower than that in the resistant group (P<0.01).

[0227] In the drug-resistant group, the density of M2 type TAM remained at a high level.

[0228] 10.4.2 T-cell protocol

[0229] In the sensitive group, the density of CD4⁺FOXP3⁺PD1⁺ effector Tregs was significantly lower than that in the resistant group (P<0.01).

[0230] In the drug-resistant group, the density of effector Tregs remained at a high level.

[0231] 10.4.3 CAF Scheme

[0232] In the sensitive group, the density of FAP⁺ACTA2⁺ myCAF was significantly lower than that in the resistant group (P<0.05).

[0233] In the drug-resistant group, myCAF density remained at a high level.

[0234] 10.5 Consistency Conclusions of Multi-omics Data

[0235] Combining the scRNA-seq, PCF, and multichromatographic histochemistry of the previous embodiments with those of this embodiment ( Figure 15 The analysis results of the three independent technology platforms are highly consistent, and together they reveal the pattern shown in Table 3 below.

[0236] Table 3. Consistency conclusions of multi-omics data

[0237]

[0238] The consensus results of the above multi-omics data indicate that the abundance levels of IFN⁺ tumor cells, M2-type / SPP1⁺ TAM, and effector Tregs after PARP inhibitor treatment are reliable indicators for distinguishing between sensitive and resistant patients. These findings provide a solid multidimensional experimental basis for the prediction method based on CN12 and CN29 in this application.

[0239] Example 11 Preparation of the detection kit

[0240] This embodiment describes the preparation of a reagent kit for implementing the application prediction method.

[0241] 11.1 Kit Composition The kit for detecting CN12 and CN29 contains the following components:

[0242] First antibody combination: Antibodies targeting PanCK and IFN-γ, used to recognize IFN⁺ tumor cells;

[0243] The second antibody combination consists of antibodies against CD68, CD163, and SPP1, used to recognize M2 type or SPP1⁺ TAM.

[0244] Auxiliary reagents include blocking solution, washing buffer, antigen retrieval solution, DAPI nuclear staining solution, etc.

[0245] Positive control sections: cell lines or tissue sections known to express the above markers;

[0246] Negative control reagent: isotype control antibody;

[0247] Instruction manual: Describes the testing steps and result interpretation criteria in detail.

[0248] 11.2 Antibody Labeling Methods The antibodies described may be labeled using one of the following methods to adapt to different detection platforms:

[0249] For the PhenoCycler-Fusion platform: oligonucleotide barcode labeling is used, and the staining protocol is as follows: Figure 1 As shown.

[0250] For multicolor immunofluorescence platforms: direct labeling with fluorescent dyes (such as AF488, AF555, AF647, etc.) is used.

[0251] The kit prepared in this embodiment can be used to implement the aforementioned method for predicting PARP inhibitor sensitivity. Specifically, by detecting the presence and abundance of CN12 and / or CN29 in a subject's tumor tissue sample, and according to the aforementioned prediction rules (including CN12 abundance threshold, CN29 abundance threshold, CN12 abundance threshold, CN29 abundance threshold, SPP1⁺ cell density threshold, or a comprehensive scoring model based on pre-treatment samples), the subject's sensitivity to PARP inhibitors is predicted. Therefore, the kit described in this embodiment can be used to prepare products for predicting PARP inhibitor sensitivity, including but not limited to diagnostic kits, detection devices, or analytical systems.

[0252] The present application has been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of the present application. The descriptions of the embodiments above are only for the purpose of helping to understand the present application and its core ideas. It should be noted that those skilled in the art can make several improvements and modifications to the present application without departing from the principles of the present application, and these improvements and modifications also fall within the protection scope of the claims of the present application.

Claims

1. A method for predicting PARP inhibitor sensitivity based on specific cell clusters, characterized in that, Includes the following steps: (1) Obtain ovarian cancer tumor tissue samples from the subjects; (2) Detect whether a specific cell cluster exists in the tumor tissue sample and determine the abundance of the cell cluster; wherein, the specific cell cluster includes: CN12, an IFN⁺ epithelial cell cluster centered on IFN⁺ tumor cells and surrounded by immunosuppressive stromal cells, and / or CN29, a TAM-enriched cluster centered on M2 tumor-associated macrophages and surrounded by immunosuppressive cells; (3) Based on the test results of step (2), predict the subject's sensitivity to PARP inhibitors according to at least one of the following rules: (i) If the sample obtained is before treatment and the abundance of CN12 in the sample is higher than the first preset threshold, it is predicted to be insensitive or prone to drug resistance. (ii) If the sample obtained is before treatment and the abundance of CN29 in the sample is higher than the second preset threshold, it is predicted to be insensitive or prone to drug resistance. (iii) If the sample obtained is a post-treatment sample and the abundance of CN12 in the sample is higher than the third preset threshold, it is predicted to be insensitive or prone to drug resistance. (iv) If the sample obtained is a post-treatment sample and the abundance of CN29 in the sample is higher than the fourth preset threshold, it is predicted to be insensitive or prone to drug resistance. (v) If the sample obtained is before treatment and the density of SPP1⁺ cells in the sample is higher than the fifth preset threshold, it is predicted to be insensitive or prone to drug resistance.

2. The prediction method according to claim 1, characterized in that, In step (2), CN12 is a cell cluster with a total number of ≥5 cells and containing at least one type of immunosuppressive stromal cell within a radius of 25 μm, centered on PanCK⁺IFN-γ⁺ tumor cells.

3. The prediction method according to claim 1, characterized in that, In step (2), CN29 is a cell cluster with a total number of ≥5 cells and containing at least one immunosuppressive cell within a radius of 25 μm, centered on a macrophage with CD68⁺ and CD163⁺ or SPP1⁺.

4. The prediction method according to claim 1, characterized in that, In step (3) rule (i), the first preset threshold is a CN12 abundance of 0.8%.

5. The prediction method according to claim 1, characterized in that, In step (3) rule (ii), the second preset threshold is a CN29 abundance of 32.7%.

6. The prediction method according to claim 1, characterized in that, In step (3) rule (iii), the third preset threshold is a CN12 abundance of 0.009%.

7. The prediction method according to claim 1, characterized in that, In step (3) rule (iv), the fourth preset threshold is a CN29 abundance of 2.9%.

8. The prediction method according to claim 1, characterized in that, In step (3) rule (v), the fifth preset threshold is an SPP1⁺ cell density of 1.7%; the SPP1⁺ cell density is the percentage of SPP1⁺ cells to the total number of DAPI⁺ cells.

9. The prediction method according to claim 1, characterized in that, The prediction in step (3) also includes constructing a comprehensive scoring model, which scores and predicts according to the following method: (a) Assign values ​​to the following four indicators respectively: (i) If the CN12 abundance is greater than 0.8% before treatment, score 1 point; otherwise, score 0 points. (ii) If the CN29 abundance is greater than 32.7% before treatment, score 1 point; otherwise, score 0 points. (iii) If the CN12 abundance is greater than 0.009% after treatment, 1 point is awarded; otherwise, 0 points are awarded. (iv) If the CN29 abundance is greater than 2.9% after treatment, score 1 point; otherwise, score 0 points. (v) If the SPP1⁺ cell density is greater than 1.7% before treatment, score 1 point; otherwise, score 0 points. (b) Add up the scores of the five indicators in (a) to get the total score. If the total score is ≥3, it is predicted that the subject is insensitive to PARP inhibitors or is prone to developing drug resistance.

10. The use of the method of any one of claims 1-9 or the kit for detecting CN12 and / or CN29 in any one of the methods of claims 1-9 in the preparation of a product for predicting PARP inhibitor sensitivity.

Images