A prediction model for evaluating the benefit of immune checkpoint therapy for ovarian cancer patients and application thereof

Abstract

The present application relates to the field of medical detection, and specifically, the present application provides a diagnostic model for evaluating the benefit of immune checkpoint therapy for ovarian cancer patients and application thereof. The present application first discovers that the expression amount of PD-L1 in the ascites exosome of the ovarian cancer patient is correlated with the benefit of the immune checkpoint therapy of the patient, and the ratio of the expression levels of PD-L1 in the tumor tissue sample and the ascites sample of the patient before receiving the immune checkpoint therapy can effectively predict the benefit effect of the patient after receiving the immune checkpoint therapy. The present application realizes the precise screening of the benefit population and risk stratification before treatment, guides the treatment selection, and improves the treatment efficacy.

Description

Technical Field

[0001] This invention relates to the field of medical testing, specifically to a predictive model for assessing the benefit of immune checkpoint therapy in ovarian cancer patients and its application. Background Technology

[0002] Immune checkpoint blockade (ICB) with PD-1 and PD-L1 is a new generation of immunotherapy strategies following surgery, radiotherapy, and chemotherapy, bringing hope to patients with recurrent, metastatic, and refractory cancers. Among these, PD-L1 levels, tumor mutational burden (TMB), and microsatellite instability (MSI) in cancer patients have significant guiding value in predicting immune checkpoint therapy. Currently approved PD-L1 diagnostic kits use immunohistochemistry to detect tissue-level PD-L1 (tumor cell membrane PD-L1, tumor-PD-L1).

[0003] However, according to a study published in the *New England Journal of Medicine*, only 8% of ovarian cancer patients responded to PD-1 inhibitors alone. This suggests that existing tumor-PD-L1 companion diagnostics alone are insufficient for screening patients suitable for ICB therapy, necessitating the development of more new companion diagnostic markers for precise patient screening and improved ICB treatment benefit rates. Recent reports in *Nature*, *Cell*, and other journals indicate an inverse relationship between exosome PD-L1 (exo-PD-L1) levels in cancer patients and the benefit of immune checkpoint inhibitor therapy; lower exo-PD-L1 levels are more suitable for immune checkpoint inhibitor therapy. Therefore, assessing a patient's suitability for immune checkpoint antibody therapy should involve testing both tumor-PD-L1 and exo-PD-L1 levels. Based on this, this project aims to utilize a method developed in our laboratory for direct detection of exo-PD-L1 in plasma and ascites fluid without relying on exosome extraction to investigate the relationship between exo-PD-L1 levels in ascites fluid of ovarian cancer patients and the benefit of immunotherapy. Simultaneously, a ratio assignment model for tumor-PD-L1 / exo-PD-L1 was proposed and established to help screen whether patients are suitable for immune checkpoint inhibitor therapy. The higher the ratio, the more favorable it is; patients with a low ratio are deemed unsuitable for immune checkpoint blockade therapy. Furthermore, the prognosis is determined based on the change in exo-PD-L1 levels before and after immune checkpoint therapy.

[0004] In summary, the novel companion diagnostic method for immunotherapy established based on the tumor-PD-L1 / ascites exo-PD-L1 ratio model is more conducive to the accurate screening of patients suitable for immune checkpoint therapy in ovarian cancer. This model has the potential to enable the detection of exo-PD-L1 in LDT laboratory tests, assisting in the selection of medication regimens for postoperative ovarian cancer patients, and has good clinical translational value and prospects.

[0005] Although there are PD-L1 detection kits on the market, this detection method cannot distinguish the ratio of PD-L1 expressed in exosomes to soluble PD-L1; it can only detect the total expression of both forms of PD-L1. To detect the expression level of PD-L1 in exosomes alone, an exosome separation process is required (10000g ultracentrifugation or polymer precipitation kit method), which is time-consuming, requires a large number of samples, and is complex and costly. Summary of the Invention

[0006] To overcome the shortcomings of existing technologies, this invention establishes a tumor-PD-L1 / exo-PD-L1 = H-score / PHR-Score prediction model by screening biomarkers. The tumor-PD-L1 / exo-PD-L1 value is used to comprehensively evaluate the immune checkpoint therapy benefit in ovarian cancer patients, improving the accuracy of using immunohistochemical detection of tumor-PD-L1 values ​​alone in patients suitable for companion diagnosis of ovarian cancer immune checkpoint therapy. The technical solution provided by this invention does not require exosome extraction; it can rapidly quantify exo-PD-L1 using only 20 μL of sample.

[0007] To achieve the above-mentioned technical effects, the present invention provides the following technical solution:

[0008] Firstly, in a first aspect, the present invention provides a method for constructing a predictive model for evaluating the benefit of immune checkpoint therapy in ovarian cancer patients, wherein the method comprises the following steps:

[0009] 1) Collect sufficient tumor tissue and ascites samples from ovarian cancer patients to serve as a development cohort group;

[0010] 3) Follow up with patients in the development cohort for more than 180 days and evaluate whether patients are treatment responders or non-responders using imaging.

[0011] 3) The calculation of Tumor PD-L1 score (H-score) and Exo-PD-L1 entity expression score (exo-PD-L1-Score) was performed by detecting tumor tissue sections by mIHC and exosomes in ascites samples by ELISA.

[0012] 4) Predictive models for the benefit of immune checkpoint therapy in ovarian cancer patients:

[0013] tumor-PD-L1 / exo-PD-L1=tumor PD-L1 H-score / exo-PD-L1-Score.

[0014] In one embodiment, the immune checkpoint therapy refers to administration of a monoclonal antibody targeting the programmed death (PD-1) receptor.

[0015] In one implementation, a treatment response is defined as no disease progression or tumor shrinkage at 6 months, based on the imaging results of a radiological examination.

[0016] In one implementation, the Tumor PD-L1 score (H-score) is calculated as follows: H-score = ∑Pi(i+1), where Pi is the percentage of PD-L1 positive cells under different staining intensities (i); the Exo-PD-L1 entity expression score (exo-PD-L1-Score) is calculated as follows: the expression level of PD-L1 in exosomes in ascites is detected by ELISA, and the OD value is measured by a microplate reader. This OD value is counted as the exo-PD-L1 entity expression score.

[0017] In one implementation, the model threshold / cutoff value is 1.514. When the tumor-PD-L1 / exo-PD-L1 ratio is greater than the threshold, it is determined that there is a response to immune checkpoint therapy; conversely, when the tumor-PD-L1 / exo-PD-L1 ratio is less than the threshold, it is determined that there is no response to immune checkpoint therapy.

[0018] A second aspect of the present invention provides a predictive model for evaluating the benefit of immune checkpoint therapy in patients with ovarian cancer, characterized in that the model is:

[0019] tumor-PD-L1 / exo-PD-L1=tumor PD-L1 H-score / exo-PD-L1-Score;

[0020] The Tumor PD-L1 score (H-score) is calculated as follows: H-score = ∑Pi(i+1), where Pi is the percentage of PD-L1 positive cells under different staining intensities (i). The Exo-PD-L1 entity expression score (exo-PD-L1-Score) is calculated as follows: the expression level of PD-L1 in exosomes in ascites is detected by ELISA, and the OD value is measured by a microplate reader. This OD value is counted as the exo-PD-L1 entity expression score.

[0021] In one implementation, the optimal cut-off value for the model is 1.514; when the tumor-PD-L1 / exo-PD-L1 ratio is greater than the threshold, it is determined that there is a response to immune checkpoint therapy; conversely, when the tumor-PD-L1 / exo-PD-L1 ratio is less than the threshold, it is determined that there is no response to immune checkpoint therapy.

[0022] A third aspect of the invention provides the use of a predictive model constructed according to the above-described construction method for evaluating whether immune checkpoint therapy is beneficial to ovarian cancer patients.

[0023] In a fourth aspect, the present invention provides the use of the above-described model for evaluating whether immune checkpoint therapy is beneficial in patients with ovarian cancer.

[0024] Compared with the prior art, the present invention has the following significant advantages:

[0025] 1) Improve the specificity of exo-PD-L1 detection: The specially designed plate with specific grooves and folds is conducive to the enrichment of exosomes and will not lose exosomes even after washing.

[0026] 2) Reduced detection time and cost: This method directly enriches exosomes on ELISA plates without the need for ultracentrifugation, density gradient centrifugation or nanoparticle enrichment. Compared with existing methods (>24 hours), the detection time is reduced by more than 50%, meeting the needs of rapid clinical testing.

[0027] 3) Small sample requirement and high throughput: This test only requires 20 microliters of sample to meet the testing needs; the 96-well plate can meet the testing needs of 5-10 parallel plates, and can cover rapid screening of 500-1000 samples in a single test.

[0028] 4) Rapid diagnosis and quantification of exosomal PD-L1 through body fluid samples can quickly and accurately predict the level of exosomal PD-L1 in the body fluids of cancer patients that inhibits T-cell immunity, providing a new and reliable strategy for predicting the benefits of immunotherapy for patients. Attached Figure Description

[0029] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0030] Figure 1 A schematic diagram of the structure of a customized ELISA plate;

[0031] Figure 2 ROC curves predicting immune checkpoint therapy response rates for (A) tumor-PD-L1 alone and (B) exo-PD-L1 alone.

[0032] Figure 3 Training set ROC curves for predicting immune checkpoint therapy response using the Tumor-PD-L1 / exo-PD-L1 ratio;

[0033] Figure 4ROC curves for the validation set of Tumor-PD-L1 / exo-PD-L1 ratio to predict immune checkpoint therapy response. Detailed Implementation

[0034] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0035] The reagents used in the examples are as follows:

[0036]

[0037] Example 1: Prospective Study and Establishment of Prognostic Prediction Model

[0038] In a prospective study, ascites from high-grade serous ovarian cancer patients (pre-treatment ascites collected from ovarian cancer patients in the Department of Obstetrics and Gynecology, Peking University Third Hospital; post-ICB treatment ascites collected from the Department of Oncology Chemotherapy and Radiation Oncology; inclusion criteria: 1) age: ≥25 years and ≤65 years; 2) voluntary participation in this trial and signing of informed consent; 3) high-grade serous ovarian cancer; exclusion criteria: 1) patients diagnosed with other cancers; 2) pregnant or lactating patients) was analyzed. The results showed that before ICB (immune checkpoint blockade) treatment, the exo-PD-L1 level in the tumor ascites group was 2–3 times higher than that in the non-tumor group (confirmed non-ovarian cancer patients), and the difference was significant. However, the total PD-L1 protein level did not differ significantly, suggesting that exosomal PD-L1 may play a more prominent role in tumor immunity than soluble PD-L1.

[0039] Plasma exosome analysis in two patients who received ICB treatment revealed elevated exo-PD-L1 levels compared to pre- and post-treatment levels. This was accompanied by blockade of immune checkpoint brake molecules (PD-1 / PD-L1) and T-cell activation, leading to increased IFN-γ secretion, which may induce tumor cells to produce more exo-PD-L1. This elevation may be one indicator of ICB immunotherapy response. However, the current results only reflect changes in plasma levels in two patients, and more cases are needed to further determine the companion diagnostic value of exo-PD-L1 in immunotherapy response and benefit.

[0040] Furthermore, we enrolled more than 60 ovarian cancer patients who received immune checkpoint therapy (received from the Department of Oncology Chemotherapy and Radiology, Peking University Third Hospital, all of whom had signed informed consent forms) to verify the above results.

[0041] Inclusion criteria: Ovarian cancer patients ≤65 years of age; no other tumors or complications; high-grade serous ovarian cancer; patients who have previously undergone tumor debulking surgery or have tissue paraffin blocks or electronic immunohistochemical staining results.

[0042] All patients received ICB immune checkpoint therapy (nivolumab monotherapy; 240 mg / 2 weeks infusion, treatment cycle 2-3 months). Prior to immune checkpoint blockade therapy, tumor tissue sections and ascites samples were collected from each enrolled patient. Immunohistochemical staining (IHC) was performed on the tumor tissue sections, and the staining results were analyzed and scored. The Tumor PD-L1 score (H-score) and Exo-PD-L1 solid expression score (exo-PD-L1-Score) were calculated for each enrolled patient.

[0043] Patients were followed up with tumor imaging results at 6 and 12 months after treatment. At 6 months, the radiological imaging results were used by 2-3 experts to determine whether the tumor signal was present or not. No progression or tumor shrinkage at 6 months was considered a treatment response. Based on clinical outcomes, 20 patients who responded to treatment and 20 patients who did not respond to treatment were randomly selected from the enrolled patients as the training set, and another 15 patients who responded to treatment and 15 patients who did not respond to treatment were selected as the validation set.

[0044] Criteria for determining treatment response: Tumor tissue sections and ascites samples were collected from patients at month 6 after immune checkpoint blockade therapy. Immunohistochemical staining (IHC) was performed on the tumor tissue sections of each enrolled patient. The staining results were used for panoramic scanning and scoring to semi-quantitatively assess molecular expression levels.

[0045] The Tumor PD-L1 score (H-score) is calculated by considering both staining intensity and the proportion of positive cells (tumor cells expressing PD-L1). Staining intensity is categorized as 0 (no staining), 1 (weak staining), 2 (moderate staining), and 3 (strong staining), with the proportion of positive cells expressed as a percentage. The formula is H-score = ∑Pi(i+1), where Pi is the percentage of positive cells at different staining intensities (i). For example, if weakly stained cells account for 30%, moderately stained cells account for 20%, and strongly stained cells account for 10%, then H-score = 10% × 1 + 20% × 2 + 10% × 3 = 0.6.

[0046] The Exo-PD-L1 entity expression score (exo-PD-L1-Score) is calculated as follows: Exosomes are collected using a specially customized grooved ELISA plate. The customized ELISA plate has 200nm grooves within the capture channels on its upper surface. Figure 1As shown, the capture well 6 has a diameter of approximately 0.69 cm and a height of 1.4 cm. The spacing of the grooves 7 within the well is approximately 200 nm. Through the design of the nanoscale grooves 7, exosomes can more easily enter and stably bind to these grooves. Soluble proteins, due to their small size, are less likely to specifically bind to the grooves, thus reducing the occupation and consumption of the plate-site capture antibody by soluble proteins. This allows the characteristic marker antibody for capturing exosomes to be used more extensively to bind to exosomes, significantly improving the enrichment and capture efficiency of exosomes. This structure effectively avoids the occupation and consumption of the plate-site capture antibody by soluble proteins, while reducing the loss of exosomes caused by the washing process, providing more suitable binding sites for exosomes, thereby greatly improving the exosome capture efficiency, reducing detection errors caused by low capture efficiency, and meeting the requirements of high-sensitivity detection.

[0047] Anti-CD9 antibodies were coated onto a custom-designed groove ELISA plate to specifically capture CD9-expressing exosomes. The specific steps are as follows:

[0048] (1) Customized ELISA plate coating: Dilute mouse anti-human CD9 antibody 1:200 with R&D coating buffer, add 100 μL to each well of an ELISA 96-well plate, and incubate overnight at 4°C.

[0049] (2) ELISA plate blocking: Remove the coating antibody, add 300 μL of washing buffer per well, wash 3 times, 1 minute each time; then add 300 μL of washing buffer / well, incubate at 37°C for one hour;

[0050] (3) Washing: Discard the blocking solution, use 300 μL of washing solution per well, wash for 1 minute each time, and wash 3 times;

[0051] (4) Add the ascites or plasma sample to be tested: 50 μL / well, incubate at 37°C for 2 hours, and remove cells and cell debris from the sample beforehand by centrifugation;

[0052] (5) Washing: Discard the test solution, add 300 μL of washing solution to each well, wash for 1 minute each time, and wash 3 times;

[0053] (6) Add antibody: Dilute rabbit anti-human PD-L1 antibody 1:1000 with 1X blocking buffer, 100 μL / well, and incubate at 37°C for 2 hours;

[0054] (7) Washing: Discard the test solution, add 300 μL of washing solution to each well, wash for 1 minute each time, and wash 3 times;

[0055] (8) Biotin-labeled IgG: Biotin-labeled goat anti-rabbit IgG antibody was diluted 1:1000 with 1X blocking buffer, 100 μL / well, and incubated at 37°C for 1 hour.

[0056] (9) Washing: Discard the test solution, add 300 μL of washing solution to each well, wash for 1 minute each time, and wash 3 times;

[0057] (10) Horseradish peroxidase coupled with streptavidin (Streptavidin-HRP): Dilute Streptavidin-HRP 1:1000 with 1X blocking buffer, 100 μL / well, wash; incubate at 37°C for 30 minutes.

[0058] (11) Substrate addition: Mix substrate solution A and substrate solution B in a 1:1 ratio, 100 μL / well, HRP catalytic reaction;

[0059] (12) Stop color development; 50 μL / well stop solution;

[0060] (13) ELISA reader measurement; OD450 was measured using a Deakin ELISA reader.

[0061] (14) Standard curve plotting: A standard curve is established based on the OD values ​​of serially diluted body fluid samples;

[0062] (15) Quantify the sample based on the OD value of the sample to be tested.

[0063] This OD value is the entity representation score of exo-PD-L1.

[0064] The detection results of the training set are shown in Table 1-2:

[0065] Table 1. PD-L1 scores of tumor tissues in 40 patients (training set)

[0066]

[0067]

[0068] Table 2. Exo-PD-L1 scores of ascites in 40 patients (training set)

[0069]

[0070] As shown in Tables 1-2, patients responding to immune checkpoint therapy had significantly lower exo-PD-L1 levels compared to non-responders. The AUCs for predicting immune checkpoint therapy response with tumor-PD-L1 alone and exo-PD-L1 alone were 0.8346 and 0.8450, respectively. Figure 2 ).

[0071] During the study, we unexpectedly discovered that the tumor-PD-L1 / exo-PD-L1 ratio was more effective than the original tumor-PD-L1 score or exo-PD-L1 score in predicting the prognosis of immune checkpoint therapy.

[0072] Based on the above findings, we further calculated the tumor-PD-L1 / exo-PD-L1 ratio in 40 patients in the training set, and the prognostic prediction model was calculated using the following formula:

[0073] tumor-PD-L1 / exo-PD-L1=tumor PD-L1 H-score / exo-PD-L1-Score;

[0074] The results are shown in Table 3:

[0075] Table 3. Ratio of tumor-PD-L1 / exo-PD-L1 in 40 cancer patients (training set)

[0076]

[0077] The model's AUC for predicting immune checkpoint therapy response was 0.9230, with an optimal cutoff value of 1.514. At the optimal cutoff value, the model's sensitivity and specificity for predicting immune checkpoint therapy response in ovarian cancer patients were 90.0% and 96.5%, respectively.

[0078] Performance verification of the model in Example 3 (verification set)

[0079] Following the same method described in Example 2, the predictive model for the immune checkpoint therapy response in ovarian cancer patients was validated using a validation set of 30 patients. The results are shown in Tables 4-6.

[0080] Table 4. PD-L1 scores of tumor tissues in 30 patients (validation set)

[0081]

[0082]

[0083] Table 5. Exo-PD-L1 scores of ascites in 30 patients (validation set)

[0084]

[0085] Table 6. Ratio of tumor-PD-L1 / exo-PD-L1 in 30 cancer patients (validation set data)

[0086]

[0087]

[0088] The results showed that the AUCs for predicting immune checkpoint therapy response were 0.8311 and 0.8434 for tumor-PD-L1 alone and exo-PD-L1 alone, respectively. Figure 2 ).

[0089] The AUC of the Tumor-PD-L1 / exo-PD-L1 ratio in predicting immune checkpoint therapy response was 0.9143 in the validation set, with a sensitivity of 88.8% and a specificity of 94.5%, indicating that the Tumor-PD-L1 / exo-PD-L1 ratio is a more accurate prognostic marker for predicting the benefit of immune checkpoint therapy in cancer patients.

[0090] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A method for constructing a predictive model to assess the benefit of immune checkpoint therapy in ovarian cancer patients, characterized in that, The construction method includes the following steps: 1) Collect sufficient tumor tissue and ascites samples from ovarian cancer patients as a development cohort before conducting immune checkpoint therapy; 2) Follow up with patients in the development cohort for more than 180 days and evaluate whether patients are treatment responders or non-responders using imaging. 3) Tumor PD-L1 score (Tumor PD-L1 H-score) and Exo-PD-L1 entity expression score (exo-PD-L1-Score) were calculated by detecting tumor tissue sections by mIHC and exosomes in ascites samples by ELISA. 4) Predictive models for the benefit of immune checkpoint therapy in ovarian cancer patients: tumor-PD-L1 / exo-PD-L1=tumor PD-L1 H-score / exo-PD-L1-Score, The Tumor PD-L1 score is calculated as follows: H-score = ∑Pi(i+1), where Pi is the percentage of PD-L1 positive cells at different staining intensities i. The exo-PD-L1 entity expression score is calculated as follows: the expression level of PD-L1 in exosomes in ascites is detected by ELISA, and the OD value is measured by a microplate reader. This OD value is used as the exo-PD-L1 entity expression score.

2. The construction method as described in claim 1, characterized in that, The immune checkpoint therapy refers to treatment with monoclonal antibodies targeting the programmed death receptor.

3. The construction method as described in claim 1, characterized in that, At 6 months, based on the imaging results of radiological examinations, medical experts determine that if the disease has not progressed or the tumor has shrunk after 6 months, it is considered a treatment response.

4. The construction method as described in claim 1, characterized in that, The model threshold / cutoff value is 1.

514. When the tumor-PD-L1 / exo-PD-L1 ratio is greater than the threshold, it is considered that there is a response to immune checkpoint therapy; conversely, when the tumor-PD-L1 / exo-PD-L1 ratio is less than the threshold, it is considered that there is no response to immune checkpoint therapy.

5. A predictive model for evaluating the benefit of immune checkpoint therapy in ovarian cancer patients, constructed using the method described in any one of claims 1 to 4, characterized in that... The model is as follows: tumor-PD-L1 / exo-PD-L1=tumor PD-L1 H-score / exo-PD-L1-Score; The Tumor PD-L1 score (H-score) is calculated as follows: H-score = ∑Pi(i+1), where Pi is the percentage of PD-L1 positive cells at different staining intensities i. The Exo-PD-L1 entity expression score (exo-PD-L1-Score) is calculated as follows: the expression level of PD-L1 in exosomes in ascites is detected by ELISA, and the OD value is measured by a microplate reader. This OD value is counted as the exo-PD-L1 entity expression score. The optimal cut-off value of the model is 1.

514. When the tumor-PD-L1 / exo-PD-L1 ratio is greater than the threshold, it is determined that there is a response to immune checkpoint therapy; conversely, when the tumor-PD-L1 / exo-PD-L1 ratio is less than the threshold, it is determined that there is no response to immune checkpoint therapy.

6. The use of a predictive model constructed according to any one of claims 1 to 4 for evaluating whether immune checkpoint therapy is beneficial for ovarian cancer patients.

Images