A kit for early risk assessment of placental implantation disorder

Abstract

This invention discloses a kit for early risk assessment of placenta accreta. It includes biomarkers for early assessment of the risk and / or severity of placenta accreta, wherein the biomarkers are selected from one or more of PARVB, JADE3, and G-3-P. The application of reagents for detecting the biomarkers in the preparation of a kit or system for early assessment of the risk and / or severity of placenta accreta is also described. A kit for early assessment of the risk and / or severity of placenta accreta includes reagents for detecting the biomarkers in blood. This invention provides a non-invasive detection method based on blood samples, enabling risk assessment of placenta accreta in early or mid-pregnancy; the combined application of specific protein and metabolite biomarkers improves the accuracy of early identification of placenta accreta.

Description

Technical Field

[0001] This invention belongs to the field of biological testing and relates to a kit for early risk assessment of placental implantation disease. Background Technology

[0002] Placenta accreta spectrum (PAS) is a group of diseases characterized by abnormal placental adhesion or invasion of the myometrium. Based on the depth of placental villi invading the myometrium, it is classified into three categories: placental adhesion, placenta accreta, and penetrating placenta accreta. After delivery, the placenta cannot detach normally, causing massive bleeding at the placental separation site. It is a leading cause of obstetric emergencies and even perinatal death, including emergency hysterectomy, multiple organ failure, disseminated intravascular coagulation (DIC), shock, etc., with its incidence increasing three to five times in the past thirty years. Statistically, it occurs in approximately 1 in 300 to 400 pregnancies. Perinatal outcomes of PAS are closely related to its classification; the incidence of serious complications (including massive hemorrhage, hysterectomy, shock, and intensive care unit admission) is significantly higher in cases of placenta accreta or penetrating placenta accreta than in cases of PA.

[0003] Prenatal screening for prenatal gestational hypertension (PAS) relies primarily on ultrasound. While widely used, ultrasound is highly dependent on the operator and prone to inter-observer variability. Magnetic resonance imaging (MRI) can be used as an adjunct diagnostic tool but is not recommended for routine screening. Consequently, over half of PAS cases are not diagnosed until delivery, and the misdiagnosis rate for PAS grading is as high as 36%, exposing patients to unexpected bleeding and severe maternal mobidity (SMM). Furthermore, typical ultrasound signs are not apparent in the early to mid-pregnancy stages, highlighting the urgent need for objective, repeatable, and highly sensitive early risk screening methods.

[0004] Currently, there are no sensitive and specific early biomarkers available for clinical application. Early and mid-pregnancy fetal chromosome screening is based on placental blood molecules, including: ① early and mid-pregnancy serological indicators for Down syndrome screening—pregnancy-associated plasma protein-A, free β-human chorionic gonadotropin, and alpha-fetoprotein; ② plasma cell-free DNA, cell-free fetal DNA concentration, and fetal fragments from non-invasive prenatal DNA testing. However, both prospective and retrospective cohort studies have shown that the above indicators cannot accurately predict PAS, are difficult to apply clinically, and cannot determine whether they are independent of the presence of high-risk factors (J Obstet Gynaecol Can, 2024.46(11):p.102663; Geburtshilfe Neonatol, 2022.226(2):p.92-97; Prenat Diagn, 2025.45(9):p.1160-1166). In other words, current technologies have not yet fully identified plasma biomarkers for early prediction of PAS occurrence, and there is a lack of technical methods to objectively differentiate PAS grades and risk stratifications. Therefore, if we can screen and validate characteristic PAS proteins and metabolites with predictive value in high-risk populations, elucidate their sources and associated mechanisms, and develop new early screening and diagnostic strategies, it is hoped that this will promote the implementation of personalized pregnancy management, surgical timing planning, and multidisciplinary collaborative diagnosis and treatment, thereby improving pregnancy outcomes for PAS patients. Summary of the Invention

[0005] The purpose of this invention is to address the aforementioned shortcomings of the prior art by providing biomarkers for early assessment of the risk and / or severity of placental implantation disease.

[0006] Another object of the present invention is to provide a reagent application for detecting this biomarker.

[0007] Another object of the present invention is to provide a kit for early assessment of the risk and / or severity of placental implantation disease.

[0008] The fourth objective of this invention is to provide an early assessment system for the risk and / or severity of placental implantation disease.

[0009] The objective of this invention can be achieved through the following technical solutions:

[0010] Biomarkers used for early assessment of the risk and / or severity of placenta accreta, wherein the biomarkers are selected from any of the following:

[0011] (1) β-Parvenin (PARVB);

[0012] (2) Jade family PHD refers to protein 3 (JADE3);

[0013] (3) β-Parvenin (PARVB) and Jade family PHD finger protein 3 (JADE3);

[0014] (4) One or two of β-Parvenin (PARVB) and Jade family PHD finger protein 3 (JADE3) combined with Sn-glycerol-3-phosphate G-3-P;

[0015] PARVB and JADE3 are selected from their natural forms, fragments with the same or equivalent detection properties, variants, or functional equivalents.

[0016] Among them, the protein biomarker β-parvin (PARVB) is a protein encoded by the PARVB gene, and its standard number in the UniProt database is Q9ES46.

[0017] The protein biomarker is Jade family PHD finger 3 (JADE3), which is a protein encoded by the JADE3 gene. Its standard number in the UniProt database is Q92613.

[0018] The metabolite biomarker Sn-glycerol 3-phosphate (G-3-P) is an endogenous metabolite that can be identified by CAS registry number 17989-41-2.

[0019] The application of reagents for detecting the biomarkers in the preparation of kits or systems for early assessment of the risk and / or severity of placental implantation disease.

[0020] The risk assessment includes, but is not limited to, assessment of the risk of placental implantation, assessment of the probability of occurrence, and assessment of the severity or grading of the disease.

[0021] Preferably, the detection method for the biomarker is selected from immunological detection methods, enzyme detection methods, mass spectrometry detection methods, or equivalent techniques.

[0022] Preferably, the PARVB or JADE3 is detected by enzyme-linked immunosorbent assay (ELISA), and the G-3-P is quantitatively detected by enzymatic or mass spectrometry methods.

[0023] A kit for early assessment of the risk and / or severity of placental implantation disease, the kit comprising reagents for detecting the biomarkers described therein in the blood.

[0024] Preferably, the reagents for detecting the biomarkers in blood include antibodies, enzymes, substrates, buffer solutions, markers, or their functional equivalents for detecting the biomarkers.

[0025] Preferably, the kit further includes calibrators, controls, and / or instructions for use, which are used to guide the assessment of the risk of placental implantation and / or the severity of the disease based on the test results.

[0026] A system for early assessment of the risk and / or severity of placenta accreta, characterized in that it comprises:

[0027] (1) Sample collection unit: containing reagents or devices for collecting blood samples;

[0028] (2) Detection unit: containing reagents and apparatus for detecting the biomarkers described in the blood;

[0029] (3) Data processing unit: software and equipment for processing the raw data obtained by the detection unit to obtain the results;

[0030] (4) Result output unit: Outputs visual results.

[0031] Preferably, the reagents for detecting the biomarkers in blood include antibodies, enzymes, substrates, buffer solutions, markers, or their functional equivalents for detecting the biomarkers.

[0032] Preferably, the data processing unit performs a logarithmic transformation (log10) on the concentrations of PARVB, JADE3, and G-3-P in the plasma of each pregnant woman, denoted as X1, X2, and X3. Using the occurrence and severity of placenta accreta as dependent variables, a logistic regression equation of the following form is established:

[0033] logit(P) = β0 + β1·X1 + β2·X2 + β3·X3

[0034] Where P represents the predicted probability, and β0–β3 are the model parameters obtained by fitting known samples.

[0035] Beneficial effects:

[0036] Compared with the prior art, the present invention has at least the following beneficial effects:

[0037] This invention provides a non-invasive detection method based on blood samples, which can be used to assess the risk of placenta accreta during early or mid-pregnancy.

[0038] The combined application of specific protein and metabolite biomarkers has improved the accuracy of early identification of placental implantation disease;

[0039] This invention can not only predict the risk of placenta accreta, but also assess the severity of the disease, which is beneficial for the stratified management of high-risk pregnant women.

[0040] The present invention further provides a reagent kit that is compatible with the method, which has good feasibility and industrial transformation potential. Attached Figure Description

[0041] Figure 1 Plasma proteomics features associated with PAS. (A) PCA plot of protein expression in the implantation / penetration group, adhesion group, and control group. (B) Volcano plot showing differentially expressed proteins between the implantation / penetration group and control group (left) and between the adhesion group and control group (right). (C) Hierarchical clustering dendrogram based on protein module associations, with module colors shown below. (D) Heatmap of correlations between protein modules and clinical and demographic variables. Color scale indicates correlation coefficient. Significance thresholds defined by p-values ​​are: *, <0.05; **, <0.01; ***, <0.001. (E) Overlap between the implantation / penetration group and control group, the adhesion group and control group, and the WGCNA protein module. (F) Heatmap of differentially expressed proteins. (GH) GSEA analysis of the implantation / penetration group and the adhesion group.

[0042] Figure 2 Plasma metabolomics features related to PAS. (A) PLS-DA plot showing the distinction between the implantation / penetration group, adhesion group, and control group. (B) Permutation test of PLS-DA to assess the robustness of the distinction between groups. (C) Volcano plot showing differentially expressed metabolites between the implantation / penetration group and control group (left), and between the adhesion group and control group (right). (D) Venn diagram showing the overlap of metabolites between the implantation / penetration group and control group, and between the adhesion group and control group based on differential expression. (E) Heatmap of differentially expressed metabolites. (F) Pathway enrichment analysis of differentially expressed metabolites.

[0043] Figure 3To identify potential proteins and metabolites in the early stratification of PAS using machine learning. (A) Regularized LASSO coefficient plot showing the optimal log(λ) value selected through cross-validation. (B) Feature contribution and AUC performance plot highlighting the top four features (red) selected using RFE-SVM. (C) Box plot ranking of PAS predicted feature importance based on Boruta feature selection. (D) Overlap between protein features selected by the three machine learning methods. (E) Regularized LASSO coefficient plot showing the optimal log(λ) value selected through cross-validation. (F) Box plot ranking of PAS predicted feature importance based on Boruta feature selection. (G) Feature contribution and AUC performance plot highlighting the top 14 features (red) selected using RFE-SVM. (H) Overlap between metabolite features selected by the three machine learning methods.

[0044] Figure 4 Performance of variable combinations in early PAS diagnostic models. The performance of all 2 to 5 variable combinations involving two proteins and eight metabolites is shown, with combinations demonstrating an accuracy greater than 0.75.

[0045] Figure 5 Predictive performance of three biomarker combinations. (A) Expression trends of PARVB, JADE3, and G-3-P among the three groups. (BE) ROC curves and confusion matrix of the training set and test set. (F) Mean NRI of different models in the implantation / penetration group, adhesion group, and control group. PI: implantation / penetration group, PA: adhesion group, CON: control group.

[0046] Figure 6 Example 2 shows the application of the three-marker panel kit. (A) Validation of plasma levels of JADE3, PARVB, and G-3-P in the cohort. Significance thresholds were defined by p-values ​​as: *, <0.05; **, <0.01; ***, <0.001; ***, <0.001. (BD) Training set, ROC curves of the testing machine, and confusion matrix. PI: implantation / penetration group, PA: adhesion group, CON: control group.

[0047] Figure 7 Expression of PARVB and JADE3 in placental tissue. Representative images and quantitative analysis of PARVB-positive areas (A) and JADE3 expression levels (B) at the maternal-fetal interface in patients with implantation / penetration (n=5) and individuals in the control group (n=6). Scale bar = 50 μm. Significance thresholds defined by p-value were: *, <0.05; **, <0.01; ***, <0.001; ***, <0.001. Detailed Implementation

[0048] Example 1: Screening of biomarkers for early diagnosis of PAS

[0049] Sample Source: From April 2020 to December 2022, plasma samples were collected from 60 pregnant women at 12-18 weeks of gestation at Changzhou Maternal and Child Health Hospital and Nanjing Maternal and Child Health Hospital. These samples were divided into three groups: control group (n=20, no PAS, history of uterine surgery, placental spontaneous detachment with intact decidual structure), adhesion type (n=20, requiring mechanical / surgical placental removal), and implantation / penetration type (n=20, characterized by placental bulging, neovascularization, etc.). Basic characteristics are shown in Table 1.

[0050] Table 1 Basic characteristics of the included population

[0051]

[0052] Sample processing: Collect 5 mL of peripheral venous blood, centrifuge at 4℃ and 1600×g for 10 minutes within 1 hour to separate the plasma, aliquot and freeze at -80℃ for later use.

[0053] Non-targeted proteomics detection: 100 µL of plasma was incubated with pre-washed magnetic beads at 37 °C for 1 hour (1200 rpm), and trypsin was added for overnight digestion at 37 °C; 5 mM DTT was used for reduction at 56 °C for 30 minutes, followed by alkylation with 11 mM iodoacetamide at room temperature in the dark for 15 minutes; after desalting with C18 Zip Tips, the solution was concentrated with SpeedVac and reconstituted in an aqueous phase containing 0.1% formic acid (mobile phase A); UHPLC-MS / MS analysis was performed (mobile phase B: 80% acetonitrile + 0.1% formic acid, flow rate 300 nL / min; NSI ion source voltage 1900 V, DIA mode, precursor ion scan range 480-780 m / z, resolution 240000).

[0054] Non-targeted metabolomics assay: 100 µL of plasma was added to 300 µL of acetonitrile-methanol (1:4, including internal standard), vortexed for 3 minutes, centrifuged at 12000 rpm for 10 minutes at 4°C, incubated at -20°C for 30 minutes, and then centrifuged again; the supernatant was separated using a Waters ACQUITY Premier HSS T3 column (1.8 µm, 2.1 mm × 100 mm) (column temperature 40°C, flow rate 0.4 mL / min, injection volume 4 µL); Q Exactive HF-X mass spectrometry detection (positive and negative ion modes, spray voltage 3.5kV / 3.2kV, ion transmission tube temperature 320℃, vaporization temperature 300℃, collision energy 30 / 40 / 50eV); data were converted to mzXML format by ProteoWizard, processed by XCMS software, and matched with HMDB / KEGG / mzCloud databases to identify metabolites (identification score >0.5, QC sample CV <0.3, Level 1 / 2 endogenous metabolites).

[0055] Proteomics identified 1368 (control), 1450 (adhesion type), and 1610 (implantation / penetration type) proteins in the three groups, respectively. 1356 proteins were quantified in ≥50% of the samples. After batch effect correction using the ComBat algorithm, PCA showed significant differences in the proteomic profiles among the implantation / penetration, adhesion, and control groups (p<0.001). Figure 1 A); Compared with the control, 121 proteins were upregulated and 21 were downregulated in the implanted / penetrating type, and 46 proteins were upregulated and 16 were downregulated in the adhesion type. Figure 1 B). Bioinformatics analysis showed that differentially expressed proteins were enriched in cytoskeleton remodeling, ECM metabolism, cell migration, endothelial cell development, and immune responses. Figure 1 CH).

[0056] A total of 532 metabolites were identified by metabolomics. PLS-DA showed a segregation trend among the three groups (displacement test p=0.005). Figure 2 (AB) For implanted / penetrating types, 50 metabolites were upregulated and 30 were downregulated; for adhesive types, 15 metabolites were upregulated and 9 were downregulated. Figure 2 CE), differential metabolites are enriched in carbon metabolism, amino acid metabolism, and lipid metabolism pathways ( Figure 2 F).

[0057] The 60 samples were re-divided into a training set (n=39) and a validation set (n=21) according to their data sources. Biomarker screening: A triple machine learning algorithm was used to screen features: ① LASSO to screen features with non-zero coefficients; ② RFE-SVM to screen features that maximize AUC; ③ Boruta combined with random forest algorithm to screen stable features. The intersection of these three algorithms was used to determine candidate biomarkers. The intersection of the three algorithms selected 2 proteins and 8 metabolites respectively. Figure 3 (See Table 2).

[0058] Table 2 List of candidate small molecule substances

[0059]

[0060] To develop a cost-effective, practical, and efficient early diagnosis model for PAS, we further conducted a comprehensive combinatorial search of the aforementioned candidate small molecule substances with 2 to 5 variables, as shown in the appendix. Figure 4Among these combinations, the preferred combination is PARVB, JADE3, and G-3-P. PARVB expression levels were significantly higher in adhesive and implantation / penetration plasma than in the control (p<0.001), and increased with increasing disease severity; JADE3 expression levels were significantly lower in adhesive and implantation / penetration plasma than in the control (p<0.001), with lower expression levels in more severe cases; G-3-P expression levels were significantly higher in implantation / penetration plasma than in the control and adhesive types (p<0.001), with no significant difference between the adhesive and control types (p=0.162) (see...). Figure 5 A).

[0061] Subject operating curves for five machine learning classifiers are as follows: Figure 5 As shown in BC, the Logistic Regression (LR) model is preferred, and its performance is shown in [the table]. Figure 5 DE and Table 3. In the training set (n=39), the mean AUC was 0.900 (95% CI: 0.826–0.963), the mean sensitivity was 0.769 (95% CI: 0.615–0.897), and the mean specificity was 0.885 (95% CI: 0.808–0.949). Adding G-3-P to the two-protein model significantly improved performance, especially in distinguishing between implantation and penetration. The overall net reclassification improvement (NRI) of the LR model was 1.000 (95% CI: 0.437–1.541, p<0.001), with significant improvement in the classification of implantation / penetration cases (NRI = 1.615, 95% CI: 1.171–1.929, p<0.001) (see Table 3). Figure 5 F). In the validation set (n=21), the three-marker model maintained strong discriminative power, with a mean AUC of 0.951 (95% CI: 0.847–1.000), a mean sensitivity of 0.957 (95% CI: 0.714–1.000), and a mean specificity of 0.929 (95% CI: 0.857–1.000).

[0062] Table 3 Performance of the LR model on the training and validation sets.

[0063]

[0064] Example 2: PARVB, JADE3, G-3-P Combined Detection Kit and Its Application in Early Risk Assessment of Placental Accretion Disease

[0065] To verify the stability and clinical application value of the three candidate biomarkers screened in Example 1 in an independent population, this example established a combined detection kit based on PARVB, JADE3 and G-3-P, and evaluated its application in a prospectively collected cohort of pregnant women in the first and second trimesters.

[0066] The study participants were drawn from the prenatal care clinics of Changzhou Maternal and Child Health Hospital and Nanjing Maternal and Child Health Hospital, with the inclusion period from January 2023 to December 2024. Based on the FIGO clinical and imaging classification criteria for placenta accreta, combined with perinatal surgical records and pathological results, a nested case-control cohort was established, including 134 pregnant women who met the inclusion and exclusion criteria, including patients with placenta accreta / penetrating placenta accreta, patients with placental adhesions, and non-PAS controls.

[0067] Inclusion criteria included: singleton pregnancy, routine prenatal examinations performed between 12 and 18 weeks of gestation with signed informed consent; and pregnancy outcome could be clearly determined according to FIGO criteria, including the presence and subtype of PAS. Exclusion criteria included: severe pregnancy complications (such as severe gestational hypertension, active autoimmune diseases, etc.), malignant tumors, severe hemolysis of peripheral blood samples from pregnancy, or failure to comply with collection / preservation standards.

[0068] The basic characteristics of the included population are shown in Table 4:

[0069] Table 4 Basic characteristics of the included population

[0070]

[0071] The PARVB, JADE3, and G-3-P combined detection kit provided in this embodiment includes reagent components for the quantitative detection of the above three biomarkers and common components, including but not limited to the following:

[0072] 1. PARVB detection unit:

[0073] Pre-coated plates: 96-well polystyrene microplates coated with capture antibodies against human PARVB;

[0074] PARVB Standards: Quantitatively calibrated recombinant human PARVB protein is prepared into a series of standard solutions with different mass concentrations for plotting standard curves;

[0075] Enzyme-labeled antibody: a monoclonal antibody solution that binds to the non-competitive epitope of PARVB and is conjugated with horseradish peroxidase (HRP) or other enzyme labels;

[0076] Diluent: Used for diluting plasma samples and reagents;

[0077] Washing solution (×10 or ×20 concentrated solution);

[0078] Colorimetric substrate solution (such as TMB solution);

[0079] Termination solution (e.g., 1–2 mol / L sulfuric acid).

[0080] 2. JADE3 Detection Unit:

[0081] Pre-coated plates: Microplates coated with capture antibodies against human JADE3, available as co-plates with PARVB or as separate plates;

[0082] JADE3 Standards: A series of quantitatively calibrated recombinant human JADE3 protein standard solutions;

[0083] HRP-conjugated JADE3 detection antibody;

[0084] The corresponding sample diluent, washing solution, chromogenic substrate, and stop solution.

[0085] 3. G-3-P Detection Unit:

[0086] Reaction buffer: Tris or phosphate buffer with a suitable pH;

[0087] Enzyme working solution: Contains glycerol-3-phosphate oxidase, peroxidase, etc., used to catalyze G-3-P to generate products that can be detected by chromogenic substrates;

[0088] G-3-P standard: Sn-glycerol-3-phosphate sodium salt standard solution, prepared as a series of concentration gradients for plotting standard curves;

[0089] Colorimetric substrate solution;

[0090] Termination solution (such as dilute acid solution).

[0091] 4. Common components:

[0092] Sample diluent, negative control, and positive control;

[0093] Instructions: Specify the sample type, testing steps, result calculation method, and risk interpretation threshold.

[0094] The proportions of the above components can be adjusted within the conventional range in the field to ensure that the linear range, sensitivity, and repeatability meet the requirements of clinical testing. For example, the concentration of the capture antibody coating can be 1–10 μg / mL, and the coverage range of the standard curve can be set within 0.1–100 ng / mL or other suitable ranges depending on the level of the population to be tested.

[0095] In this embodiment, the combined detection kit is used to quantitatively detect the levels of PARVB, JADE3, and G-3-P in plasma samples from pregnant women at 12–18 weeks of gestation. Specific steps include:

[0096] 1. Sample processing:

[0097] Collect 5 mL of blood from the pregnant woman's elbow vein and place it in an EDTA anticoagulant tube. Centrifuge at 1600×g for 10 minutes at 4 ℃ within 1 hour after blood collection. Separate the supernatant plasma, aliquot it, and store it at -80 ℃, avoiding repeated freeze-thaw cycles. Thaw the plasma sample at room temperature and mix thoroughly before testing.

[0098] 2. ELISA detection of PARVB and JADE3:

[0099] (1) Dilute the plasma with sample diluent as required by the instructions (e.g., 1:10–1:1000 times).

[0100] (2) Add standards, negative / positive controls and diluted test samples to the microplate coated with capture antibodies, and incubate at 37 °C for 30–60 minutes;

[0101] (3) After washing the plate wells 3–5 times, add the HRP-conjugated detection antibody and incubate at 37 °C for 30–60 minutes;

[0102] (4) Wash again, add chromogenic substrate solution, incubate at room temperature in the dark for 10–20 minutes, and read the optical density (OD) value of each well at 450 nm after the color development is terminated.

[0103] (5) Plot a four-parameter or linear regression standard curve based on the OD value of the standard, and calculate the mass concentration of PARVB and JADE3 in each sample.

[0104] 3. G-3-P Enzyme-Catalyzed Colorimetric Detection:

[0105] (1) Use the 3KD ultrafiltration tube to ultrafilter the plasma sample according to the instructions;

[0106] (2) Add the G-3-P standard, blank control and sample to be tested at the preset concentration to the reaction well, and add enzyme working solution;

[0107] (3) Add the chromogenic substrate and incubate at 37 °C for 10–30 minutes. Use a fluorescence microplate reader to detect the fluorescence value of each well at an excitation wavelength of 535 nm and an emission wavelength of 587 nm, and calculate the G-3-P concentration of each sample.

[0108] Data from 134 plasma samples showed that, compared with non-PAS controls, PARVB and JADE3 were abnormally expressed in the plasma of pregnant women with placenta accreta, and their trends were consistent with the multi-omics screening results in Example 1; G-3-P levels were significantly higher in pregnant women with accreta / penetrating PAS than in controls and adhesion-type PAS, while there was no significant difference between adhesion-type and controls (see...). Figure 6 A).

[0109] The 134 samples were divided into a training set (n=81) and a validation set (n=52) based on the data source center. Based on the quantitative results of PARVB, JADE3, and G-3-P, this embodiment employs five machine learning supervised learning methods, including logistic regression, to construct an early risk assessment model for placental implantation disease (see...). Figure 6 B). Preferably, the LR model performs better.

[0110] Specifically, the concentrations of PARVB, JADE3, and G-3-P in the plasma of each pregnant woman were logarithmically transformed using log10 and denoted as X1, X2, and X3. The presence and severity of placenta accreta were used as dependent variables to establish a logistic regression equation in the following form:

[0111] logit(P) = β0+ β1·X1+ β2·X2+ β3·X3

[0112] Where P represents the predicted probability, and β0–β3 are the model parameters obtained by fitting the training set samples mentioned above. Examples of the prediction equations for the three groups are shown in Table 5:

[0113] Table 5. Coefficients of the Logistic Regression Prediction Model

[0114]

[0115] When testing samples, the concentrations of PARVB, JADE3, and G-3-P in the plasma of pregnant women were logarithmically transformed using log10 and denoted as X1, X2, and X3. These values ​​were then substituted into three sets of logistic regression equations. Based on the highest predicted probability for each group, the pregnant woman was classified as low, medium, or high risk for PAS. Specifically, the probability of each pregnant woman being in the control group (low risk), adhesion group (medium risk), or implantation / penetration group (high risk) was calculated, and the group with the highest probability was identified as the affected group.

[0116] Using the combined detection kit and accompanying risk assessment model described in this embodiment, plasma samples from 134 pregnant women were tested and analyzed. The confusion matrix is ​​attached. Figure 6The specific performance results for CD are shown in Table 6. In the training set (n=82), the average AUC was 0.964 (95% CI: 0.923–0.991), the average sensitivity was 0.837 (95% CI: 0.714–0.947), and the average specificity was 0.948 (95% CI: 0.863–1.000). In the validation set (n=52), the three-marker model maintained strong discriminative power, with an average AUC of 0.821 (95% CI: 0.683–0.931), an average sensitivity of 0.769 (95% CI: 0.592–0.916), and an average specificity of 0.769 (95% CI: 0.600–0.916).

[0117] Table 6 Performance of the LR model on the training and validation sets.

[0118]

[0119] When using "whether placenta accreta has occurred" as the endpoint, the sensitivity and specificity of the three-marker combined model were superior to those of the single-marker model and the model based solely on clinical risk factors. Furthermore, the model's AUC remained at a high level in distinguishing between accreta / penetration-type severe PAS and non-severe cases, which can provide a basis for the early identification and stratified management of high-risk groups.

[0120] In summary, this embodiment constructs a combined detection kit for PARVB, JADE3, and G-3-P, and combines it with a risk assessment model based on the above biomarkers. This allows for an objective and quantitative early assessment of the risk and severity of placental implantation disease based on a single blood sample taken in the early to mid-pregnancy. This overcomes the shortcomings of existing methods that rely solely on imaging examinations, such as strong subjectivity and insufficient sensitivity in early pregnancy, and has promising clinical application prospects.

[0121] Example 3 Immunohistochemical expression of PARVB and JADE3 in placental tissue

[0122] Based on Examples 1 and 2, this embodiment verifies the expression of PARVB and JADE3 in placental tissue at the histological level to further illustrate the correlation between the biomarkers described in this invention and placental implantation diseases.

[0123] This embodiment selected several cases of maternal-fetal interface tissue from pregnant women with placenta accreta / penetrating placenta accreta confirmed by perinatal surgery and pathological examination, as well as several cases of maternal-fetal interface tissue from non-PAS pregnant women during the same period as controls. All specimens were obtained from cases of cesarean section or curettage during the early to mid-pregnancy. Immediately after the surgery, samples were taken from the placental accreta site and the corresponding control area, fixed in 4% paraformaldehyde solution for 24–48 hours, then routinely dehydrated, embedded in paraffin, and prepared into paraffin sections with a thickness of approximately 4 μm for later use.

[0124] Immunohistochemical staining methods:

[0125] 1. Dewaxing and hydration: Paraffin sections are baked in a 70℃ oven for about 1 hour, dewaxed with xylene, and then hydrated in stages with 100%, 95%, 85%, and 75% ethanol. Finally, they are placed in distilled water for later use.

[0126] 2. Antigen retrieval and blocking: Place the slides in Tris-EDTA buffer (pH 9.0) and perform high-temperature antigen retrieval for 10 min using microwave or water bath. After natural cooling, wash three times with PBS for 3–5 min each time. Then add 3% hydrogen peroxide working solution and incubate at room temperature for about 10 min to block endogenous peroxidase. After washing with PBS, block with 5%–10% normal goat serum at room temperature for 15–30 min.

[0127] 3. Primary and Secondary Antibody Incubation: After discarding the blocking solution, without washing the slides, add either the working solution of the primary antibody against human PARVB or the working solution of the primary antibody against human JADE3 (prepared according to the recommended dilution ratio in the manufacturer's instructions), and incubate overnight at 4°C. The next day, after thawing, wash three times with PBS for 3–5 minutes each time. Then add the working solution of the secondary antibody conjugated with HRP, incubate at room temperature for 20–30 minutes, and wash three more times with PBS.

[0128] 4. Staining and Counterstaining: Add DAB staining solution to the slides. Observe the staining process in real time under an optical microscope. When the appropriate staining intensity is reached, terminate the reaction with distilled water. Afterwards, lightly counterstain the cell nuclei with hematoxylin, rinse with running water, allow differentiation, and return to blue. Then, dehydrate with graded ethanol, clear with xylene, and mount with neutral resin. For the negative control, use isotype IgG at the same concentration instead of the primary antibody; the remaining steps are the same.

[0129] As attached Figure 7As shown in Figure A, in the control group placenta, PARVB staining was mostly negative or weakly positive at the junction of a few decidual areas and free villi, with only a few areas showing moderate-intensity staining. In the implantation / penetration group placenta, enhanced PARVB staining was observed in the placental villi and decidual / muscular junction areas. Statistical analysis showed that the area of ​​PARVB-positive staining in the implantation / penetration group was higher than that in the control group, and the difference was statistically significant (p < 0.05). This result indicates that PARVB expression shows an increasing trend at the placental-uterine interface.

[0130] As attached Figure 7 As shown in Figure B, in the control group placenta, relatively obvious JADE3 staining was observed in the extravillous trophoblastic cell region, with moderate positivity mainly in the nucleus and some cytoplasm, and the corresponding H-score was at a moderate to high level. In the implantation / penetration group placenta, JADE3 staining was weaker in the extravillous trophoblastic cell region associated with placental implantation compared to the control group: some EVT cells were only weakly positive or showed no obvious staining; the proportion of positive cells decreased, and the overall H-score was significantly lower than that of the control group. Statistical analysis showed that the JADE3 expression score in the extravillous trophoblastic cell region of the implantation / penetration group was lower than that of the control group, and the difference was statistically significant (p < 0.05). This result suggests that JADE3 expression is decreasing in the extravillous trophoblastic cell population associated with placental invasion.

[0131] The immunohistochemical detection results in this embodiment show that the placental tissue of patients with placenta accreta disease is as follows:

[0132] PARVB expression levels were elevated at the villi and decidual / muscular junction.

[0133] JADE3 expression levels were reduced in the extravillous trophoblast region.

[0134] The above-mentioned histological expression characteristics are consistent with the multi-omics screening results in Example 1 and the plasma biomarker detection results in Example 2 in terms of the direction of change, which provides further support at the tissue level for PARVB and JADE3 as biomarkers related to placental implantation disease.

Claims

1. A biomarker for early assessment of the risk and / or severity of placenta accreta, characterized in that, The biomarkers mentioned are selected from any one of the following: (1) β-Paven protein; (2) PHD in the jade family refers to protein 3; (3) β-Pavenin PARVB and Emerald family PHD finger protein 3; (4) One or two of β-Paven protein PARVB, Emerald family PHD finger protein 3, and Sn-glycerol-3-phosphate G-3-P; Among them, β-Paven protein and Emerald family PHD finger protein 3 are selected from their natural forms, fragments with the same or equivalent detection properties, variants or functional equivalents.

2. The use of the reagent for detecting the biomarker of claim 1 in the preparation of a kit or system for early assessment of the risk and / or severity of placental implantation disease.

3. The application according to claim 2, characterized in that, The detection method for the biomarkers is selected from immunological detection methods, enzymatic detection methods, mass spectrometry detection methods or equivalent techniques. The β-Paven protein or Emerald family PHD finger protein 3 is preferably detected by enzyme-linked immunosorbent assay (ELISA), and the Sn-glycerol-3-phosphate G-3-P is preferably quantitatively detected by enzymatic or mass spectrometry methods.

4. A kit for early assessment of the risk and / or severity of placental implantation disease, characterized in that, The kit includes reagents for detecting the biomarkers of claim 1 in blood.

5. The early assessment kit according to claim 4, characterized in that, Reagents for detecting the biomarkers in blood include antibodies, enzymes, substrates, buffer solutions, markers, or their functional equivalents for detecting the biomarkers.

6. The early assessment kit according to claim 5, characterized in that, The kit also includes calibrators, controls, and / or instructions for use, which are used to guide the assessment of the risk of placental implantation and / or the severity of the disease based on the test results.

7. An early assessment system for the risk and / or severity of placenta accreta, characterized in that, Include: (1) Sample collection unit: containing reagents or devices for collecting blood samples; (2) Detection unit: containing reagents and apparatus for detecting the biomarkers described in claim 1 in blood; (3) Data processing unit: Data processing software and equipment that processes the raw data obtained by the detection unit and obtains the results; (4) Result output unit: Outputs visual results.

8. The early assessment system for the risk and / or severity of placenta accreta disease according to claim 7, characterized in that, Reagents for detecting the biomarkers in blood include antibodies, enzymes, substrates, buffer solutions, markers, or their functional equivalents for detecting the biomarkers.

9. The early assessment system for the risk and / or severity of placenta accreta disease according to claim 7, characterized in that, The data processing unit performed a logarithmic transformation (log10) on the concentrations of β-Pavenin, Emerald family PHD finger protein 3, and Sn-glycerol-3-phosphate G-3-P in the plasma of each pregnant woman, denoted as X1, X2, and X3. Using the presence and severity of placenta accreta as the dependent variable, the predicted probabilities were calculated for healthy controls, placental adhesion type, and placenta accreta / penetration type according to the following logistic regression equation. Based on the highest predicted probability calculated for each group, pregnant women were classified as low, medium, or high-risk for PAS. Specifically, the highest predicted probability calculated using the logistic regression equation for healthy controls indicated a low-risk group; the highest predicted probability calculated using the logistic regression equation for placental adhesion type indicated a medium-risk group; and the highest predicted probability calculated using the logistic regression equation for placenta accreta / penetration type indicated a high-risk group. logit(P) =β0+ β1·X1+ β2·X2+ β3·X3 Where P represents the prediction probability, and β0–β3 are the model coefficients obtained by fitting known samples for healthy controls, placental adhesion type, and placental implantation / penetration type, respectively.

10. The early assessment system for the risk and / or severity of placenta accreta according to claim 7, characterized in that, The coefficients of each group in the logistic regression equation logit(P) = β0 + β1·X1 + β2·X2 + β3·X3 are shown in the table below: 。

Images