Prediction model of glaucoma and training method and related application thereof

Abstract

The invention discloses a prediction model of glaucoma and a training method and related application thereof, and relates to the field of biomedicine, myeloperoxidase is used as a marker for predicting glaucoma, a machine learning algorithm is combined, a new prediction model is constructed, and the prediction model can predict the glaucoma based on the marker content in the plasma of a subject. The method can be used for predicting the disease risk and / or disease progress of the glaucoma, has the advantages of convenience, simplicity and rapidness, provides a new means for diagnosis or auxiliary diagnosis of the glaucoma and severity judgment of the glaucoma, and has better clinical application value.

Description

Technical Field

[0001] The present invention relates to the field of biomedicine, and in particular, to a prediction model for glaucoma, a training method thereof, and related applications. Background Art

[0002] Glaucoma is a chronic progressive optic neuropathy characterized by the gradual damage and loss of retinal ganglion cells (RGCs) and corresponding visual field defects. As the main cause of irreversible vision loss globally, the number of glaucoma patients will increase to approximately 111.8 million by 2040. Although the exact mechanism of glaucoma is still unclear, elevated intraocular pressure is the main risk factor leading to apoptosis of retinal ganglion cells (RGC) and axonal loss. In addition to intraocular pressure, more and more evidence indicates that oxidative stress and immune factors play important roles in the development of glaucoma, including immune mediators such as bilirubin and lactoferrin, and immune cells such as Th1 and B cells, which can lead to the progression of glaucoma. However, these findings have not fully elucidated the roles of oxidative stress and immune factors in glaucoma.

[0003] In view of this, the present invention is specifically proposed. Summary of the Invention

[0004] The purpose of the present invention is to provide a prediction model for glaucoma, a training method thereof, and related applications.

[0005] The present invention is implemented as follows:

[0006] In a first aspect, an embodiment of the present invention provides an application of a reagent for detecting the content of a biomarker in the preparation of a product for predicting glaucoma, wherein the biomarker includes myeloperoxidase.

[0007] In a second aspect, an embodiment of the present invention provides a training method for a prediction model of glaucoma, which includes: obtaining the detection results of the biomarker content in the training samples and their annotation results; wherein, the biomarker includes myeloperoxidase, and the annotation results include labels representing the disease risk and / or disease progression of the sample glaucoma; inputting the detection results of the training samples into a pre-constructed machine learning model to obtain prediction results; and updating the parameters of the machine learning model according to the prediction results and the annotation results to obtain a prediction model.

[0008] In a third aspect, an embodiment of the present invention provides a prediction device for glaucoma, which includes:

[0009] An acquisition module for obtaining the detection results of the biomarker content in a sample to be tested; the biomarker includes myeloperoxidase;

[0010] A prediction module is configured to input the detection result of a sample to be detected into the prediction model trained by the training method described in the foregoing embodiments to obtain a prediction result.

[0011] In a fourth aspect, an embodiment of the present invention provides an electronic device, which includes a processor and a memory. The memory is used to store a program. When the program is executed by the processor, the processor implements the prediction method for glaucoma. Wherein, the prediction method includes: obtaining the detection result of the content of a biomarker in a sample to be detected; inputting the detection result of the sample to be detected into the prediction model trained by the training method described in the foregoing embodiments to obtain a prediction result.

[0012] In a fifth aspect, an embodiment of the present invention provides a computer-readable medium, on which a computer program is stored. When the computer program is executed by a processor, it implements the prediction method for glaucoma described in the foregoing embodiments.

[0013] The present invention has the following beneficial effects:

[0014] In the embodiments of the present invention, myeloperoxidase is used as a biomarker for predicting glaucoma, and combined with a machine learning algorithm, a new prediction model is constructed. The prediction model can predict the risk of glaucoma and / or the disease progression based on the content of the biomarker in the plasma of the subject, and has the advantages of being convenient, simple, and fast, providing a new means for the diagnosis or auxiliary diagnosis of glaucoma and the judgment of the severity of glaucoma, and having good clinical application value. Description of the Drawings

[0015] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required for the embodiments. It should be understood that the following drawings only show some embodiments of the present invention, and therefore should not be regarded as limiting the scope. For those of ordinary skill in the art, other related drawings can be obtained based on these drawings without creative efforts.

[0016] Figure 1 is the plasma MPO level of glaucoma patients; (A) Plasma MPO in glaucoma patients is elevated, (B) Comparing the plasma MPO levels of healthy controls (HC), primary angle-closure glaucoma (PACG), and primary open-angle glaucoma (POAG) subgroups; Mann-Whitney U test was used for between-group comparison, and Kruskal-Wallis test was used to determine the differences between primary angle-closure glaucoma (PACG), primary open-angle glaucoma (POAG), and healthy controls, ***p<0.001;

[0017] Figure 2The correlation between plasma MPO levels, optic nerve damage, and the severity of glaucoma; (A) shows the evaluation of plasma MPO levels in each group stratified by vertical cup-to-disc ratio (C / D); (B) shows the plasma MPO levels in different groups of glaucoma patients, stratified by disease severity using the Hodapp-Parrish-Anderson (H-P-A) classification system; the differences in all parameters were evaluated using the unpaired Mann-Whitney U test, and the significance levels were expressed as *p < 0.05, **p < 0.01, ***p < 0.001, and NS (not significant) indicates no statistical significance;

[0018] Figure 3 To perform receiver operating characteristic (ROC) curve analysis to determine the ability to distinguish between the healthy control group and patients with different disease severities according to the H-P-A (A) classification system, and to determine the ability to distinguish between early glaucoma patients and severe glaucoma patients (B); where AUC is the area under the ROC curve. Detailed implementation manners

[0019] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be described clearly and completely below. Those not specified in the embodiments are carried out according to conventional conditions or conditions recommended by the manufacturer. The reagents or instruments not specified by the manufacturer are all conventional products that can be obtained through commercial purchase.

[0020] Myeloperoxidase (MPO) is a heme-containing peroxidase mainly expressed in neutrophils. Currently, the relationship between MPO and glaucoma is unclear. The inventors of the present application have found that by using myeloperoxidase as a marker and combining it with a machine learning algorithm model, a prediction model for predicting glaucoma can be constructed. This prediction model can predict the risk of glaucoma or the disease progression by detecting the MPO content in the plasma of subjects (minimally invasive and easily obtainable), providing an effective supplement to the current diagnostic methods and severity judgment of glaucoma.

[0021] On the one hand, the embodiments of the present invention provide the application of a reagent for detecting the content of a marker in the preparation of a product for predicting glaucoma, and the marker includes myeloperoxidase (MPO).

[0022] In some embodiments, the prediction of glaucoma includes: predicting any one or more of the risk of glaucoma and / or the disease progression.

[0023] In some embodiments, the prediction of the disease progression of glaucoma includes: predicting whether glaucoma is in the early, middle, or late stage.

[0024] In some embodiments, the method for detecting the content of the detection marker includes at least one of the BCA method, Bradford method, Western blotting, immunohistochemistry, ELISA detection, flow cytometry, protein chip, and two-dimensional electrophoresis.

[0025] In some embodiments, the product includes a kit.

[0026] On the other hand, an embodiment of the present invention provides a method for training a prediction model for glaucoma, which includes:

[0027] Obtaining the detection results of the marker content in the training samples and their annotation results; wherein, the marker includes myeloperoxidase, and the annotation results include labels representing the disease risk and / or disease progression of the glaucoma in the samples;

[0028] Inputting the detection results of the training samples into a pre-constructed machine learning model to obtain prediction results;

[0029] Updating the parameters of the machine learning model according to the prediction results and the annotation results to obtain a prediction model.

[0030] It can be understood that the categories and quantities of the training samples are conventionally selectable by those skilled in the art. The total number of training samples and the number of various categories of training samples (such as healthy people, patients, and patients can include early, middle, and / or late stages) can be any value among ≥10, 50, 100, 200, 300, 400, and 500 or the range between any two of them.

[0031] In some embodiments, the label can be a character or a string. The content of the prediction result corresponds to the content of the annotation result.

[0032] The present invention places no special restrictions on the machine learning model, and a model commonly used in the art for predicting the disease risk or disease progression of diseases can be selected. In some embodiments, the machine learning model includes any one of a decision tree, a logistic regression model, a support vector machine, KNN, a naive Bayes, and a random forest.

[0033] On the other hand, an embodiment of the present invention provides a prediction device for glaucoma, which includes:

[0034] An acquisition module, configured to acquire the detection results of the marker content in the sample to be tested; the marker includes myeloperoxidase;

[0035] A prediction module, configured to input the detection results of the sample to be tested into the prediction model trained by the training method described in any of the foregoing embodiments to obtain a prediction result.

[0036] The modules described in the embodiments of the present invention may be stored in a memory in the form of software or firmware or solidified in the operating system (OS) of the electronic device provided by the present application, and may be executed by a processor in the electronic device. At the same time, the data and program codes required to execute the above modules may be stored in the memory.

[0037] On the other hand, an embodiment of the present invention provides an electronic device, comprising a processor and a memory, wherein the memory is used to store a program, and when the program is executed by the processor, the processor implements a prediction method for glaucoma; wherein the prediction method comprises: obtaining the detection result of the marker content in the sample to be tested; inputting the detection result of the sample to be tested into the prediction model trained by the training method described in any of the aforementioned embodiments to obtain a prediction result.

[0038] An electronic device may include a memory, a processor, a bus, and a communication interface, wherein the memory, the processor, and the communication interface are electrically connected to each other directly or indirectly to achieve data transmission or interaction. For example, these components may be electrically connected to each other via one or more buses or signal lines.

[0039] The memory can be, but is not limited to, random access memory (RAM), read only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable read-only memory (EEPROM), etc.

[0040] The processor can be an integrated circuit chip with signal processing capabilities. The processor 120 can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0041] The electronic device may be a server, a cloud platform, a mobile phone, a tablet computer, a laptop computer, an ultra-mobile personal computer (UMPC), a handheld computer, a netbook, a personal digital assistant (PDA), a wearable electronic device, a virtual reality device, etc. Therefore, the embodiments of the present application do not limit the types of electronic devices.

[0042] In addition, an embodiment of the present invention further provides a computer-readable medium, on which a computer program is stored. When the computer program is executed by a processor, the prediction method of glaucoma described in any of the foregoing embodiments is implemented.

[0043] In some embodiments, the computer-readable medium may be a general storage medium, such as a removable disk, a hard disk, etc.

[0044] The features and performance of the present invention will be further described in detail below in conjunction with embodiments.

[0045] Example 1

[0046] In this embodiment, glaucoma patients and matched healthy controls were recruited, and blood samples were collected, as follows.

[0047] 1. 127 patients were recruited. All participants in the study received relevant information and signed an informed consent form. Ophthalmologists diagnosed glaucoma based on comprehensive ophthalmic examinations such as optic nerve damage and visual field loss. The following ophthalmic examinations were performed on the participants: intraocular pressure (IOP), mean deviation (MD), vertical cup-to-disc ratio (VCDR), retinal nerve fiber layer thickness (RNFL), and visual field (VF). IOP was measured using a Goldmann tonometer, MD, RFNL, and VCDR were measured using an optical coherence tomography scanner (DRIOCT Triton, TOPCON), and VF was selected in the 24-2 mode using a projection perimeter (SK-950, Shangbang). Glaucoma patients had to meet the following criteria to be included in the study: they had to be clinically diagnosed with the disease and had no autoimmune, inflammatory, or neurodegenerative diseases (such as AD or PD).

[0048] 2. 106 healthy controls matched for age and gender were recruited. Exclusion criteria included: having glaucoma or a family history of glaucoma, having eye pain, elevated IOP (>21 mmHg), having had surgery recently, or having any other neurological diseases.

[0049] Table 1. Clinical information table of cases

[0050]

[0051] Note: Age is expressed as mean ± SD; Abbreviations: HPA: Hodapp, Parish, and Anderson; POAG: primary open-angle glaucoma; PAGG: primary angle-closure glaucoma; The age differences between glaucoma patients and healthy control groups were analyzed by two-tailed unpaired Student's t-test, and gender differences were analyzed by chi-square test; P < 0.05 was considered statistically significant.

[0052] 3. After each subject fasted for 8 hours in the early morning, venous blood was collected into an EDTA anticoagulant vacuum container. The blood collection tube was centrifuged at a speed of 3000 revolutions per minute, and the supernatant liquid aspirated was immediately frozen in an ultra-low temperature refrigerator at -80 °C until analysis.

[0053] Example 2

[0054] A method for detecting the plasma myeloperoxidase concentration levels in patients and healthy control populations, comprising the following steps.

[0055] 1 Preparation of washing working solution

[0056] Take 480 ml of ultrapure water in a 500 ml narrow-mouth bottle, pour 20 ml of concentrated washing solution (diluted 25-fold) into the narrow-mouth bottle, and mix well to obtain the 1× washing solution required for the experiment. Let it stand for use.

[0057] 2 Preparation of standards

[0058] 2.1 Vortex the standard tube (product number: CSB-E08721h, Huamei Biotech) thoroughly and dilute it 1:100 times with sample diluent. The specific operation is as follows: Take 10 μl of the standard (100x) and add it to 990 μl of sample diluent and mix well. The obtained standard is labeled as S7 (10 ng / ml).

[0059] 2.2 Take 7 sterile EP tubes, label them as S0, S1, S2, S3, S4, S5, S6 respectively, and add 250 μl of sample diluent to each EP tube. Aspirate 250 μl of standard S7 into the first centrifuge tube (S6), and gently pipette and mix well. Aspirate 250 μl from S6 into the second EP tube (S5) and gently pipette and mix well. And so on for serial dilution of the standard. S0 is the sample diluent.

[0060] 3 Preparation of biotin antibody working solution

[0061] Dilute the biotin-labeled antibody solution with biotin-labeled antibody diluent at a ratio of 1:100. Take 10 μl of biotin-labeled antibody and add 990 μl of biotin-labeled antibody diluent, mix gently, and prepare within 10 minutes before use.

[0062] 4. Preparation of horseradish peroxidase-labeled avidin working solution

[0063] Dilute horseradish peroxidase-labeled avidin with horseradish peroxidase-labeled avidin diluent at a ratio of 1:100. Take 10 μl of horseradish peroxidase-labeled avidin and add 990 μl of horseradish peroxidase-labeled avidin diluent, mix gently, and prepare within 10 minutes before use.

[0064] 5. Determination steps

[0065] 5.1 Move all reagents to room temperature (18-25°C) to equilibrate for at least 30 minutes, prepare reagents according to the above method and set aside.

[0066] 5.2 Sample addition: Set up standard wells and sample wells. Add 100μl of standard or sample to be tested to each well, shake gently to mix, cover with plate sticker, and incubate at 37℃ for 2 hours. Avoid creating bubbles and touching the well wall with the gun tip when adding samples.

[0067] 5.3 Add biotin antibody: Prepare biotin antibody working solution 10 minutes in advance. When the time is up, remove the film, remove the liquid in the well, spin dry, and tap 2-3 times with medium force on absorbent paper without washing. Then, add 100μl of the prepared biotin antibody working solution to each well, apply the film, and incubate in a 37℃ constant temperature box for 1 hour.

[0068] 5.4 Wash the plate and add horseradish peroxidase-labeled avidin working solution: Prepare the horseradish peroxidase-labeled avidin working solution 10 minutes in advance. When the time is up, remove the film, discard the liquid in the well, spin dry, use a spray gun to absorb 200μl of washing solution and add it to the well, soak for 2 minutes each time, spin off the washing solution in the well, and then add 200μl of washing solution. Wash the plate 3 times according to this step. After the last wash, spin dry and tap 2-3 times with medium force on absorbent paper. Finally, add 100μl of horseradish peroxidase-labeled avidin working solution to the enzyme-labeled well, cover with film, and incubate in a 37℃ constant temperature box for 1 hour.

[0069] 5.5 Wash the plate and add TMB: Remove the film and wash the plate 5 times according to step 5.4. Then add 90μl TMB substrate solution to each well, apply the film, and incubate in a 37℃ constant temperature box in the dark for 15-30min. The reaction time can be shortened or extended according to the actual color development. When the standard shows a good blue gradient, stop the reaction.

[0070] 5.6 Add reaction stop solution: Add 50μl reaction stop solution, the color changes from blue to yellow. Note that the order of adding the stop solution should be the same as the order of adding the TMB substrate solution.

[0071] 5.7 OD value determination: Preheat the microplate reader 15 minutes in advance. Immediately after termination, measure the absorbance at 450nm using a microplate reader and read the OD value.

[0072] 5.8 Concentration calculation: According to the concentration and OD value of the standard, use Curve Expert software to calculate the standard curve, and then use the OD value of the sample to calculate the standard curve. 450 The value is used to calculate the concentration.

[0073] The plasma myeloperoxidase concentration level of patients and healthy controls can be measured according to the above steps.

[0074] Example 3

[0075] The myeloperoxidase content in the plasma sample was obtained by testing based on the sample of Example 1 and the method of Example 2. The measured plasma myeloperoxidase was used to perform difference analysis, correlation analysis and receiver operating characteristic curve analysis (ROC).

[0076] 1. Analysis of the difference in expression levels of plasma MPO in healthy controls and glaucoma patients

[0077] Plasma MPO was measured using a commercially available MPO ELISA kit (Cat. No.: CSB-E08721h, Huamei Bio). Statistical analysis was performed using R language (version number: 4.1.3). The differences in MPO expression between 127 glaucoma patients and 106 healthy controls were compared. Compared with the healthy control group, the plasma MPO level in glaucoma patients was statistically significantly increased ( Figure 1 A).

[0078] Among the glaucoma subtypes, primary open-angle glaucoma (POAG) and primary angle-closure glaucoma (PACG) are the most common. Therefore, the plasma MPO levels of the PACG group and the POAG group were compared with the healthy control group. Compared with the control group, the median plasma MPO levels of the PACG group and the POAG group were significantly increased. However, there was no significant difference in plasma MPO levels between the POAG group and the PACG group ( Figure 1 B).

[0079] The above results confirm that plasma MPO has significant differences in glaucoma, glaucoma subtypes and healthy controls.

[0080] 2. Logistic regression analysis to correct for confounding factors

[0081] To better understand the association between glaucoma and plasma MPO, a logistic regression analysis was performed to control for confounding factors. In the univariate logistic regression analysis, plasma MPO levels were significantly associated with glaucoma (OR = 1.05, 95% CI: 1.03 - 1.07, p < 0.01). In the multivariate logistic regression, after adjusting for confounding factors such as age, gender, hypertension, and diabetes, plasma MPO levels remained independently associated with glaucoma (OR = 1.05, 95% CI: 1.04 - 1.07, p < 0.001), indicating that plasma MPO is an independent risk factor for glaucoma (Table 2).

[0082] Table 2. Adjustment of confounding factors in logistic regression analysis

[0083]

[0084]

[0085] 3. Relationship between plasma MPO levels and glaucoma nerve damage and glaucoma severity

[0086] To investigate whether plasma MPO levels are associated with optic nerve damage, it was evaluated through retinal structural changes (VCDR) and functional loss (H - P - A). Patients were stratified according to their VCDR, and an increase in VCDR was associated with an increase in plasma MPO levels. The plasma MPO levels in the group with a C / D ratio between 0.80 and 1.0 were significantly higher than those in the group with a C / D ratio of 0.6 or lower (p < 0.05) ( Figure 2 A in

[0087] Next, it aimed to compare the plasma MPO levels of glaucoma patients with different disease severities classified using the H - P - A classification system. Patients were divided into three subgroups: early, moderate (mid - stage), and severe (late - stage). According to the H - P - A classification, the severity of glaucoma was associated with an increase in plasma MPO levels. There were significant differences in plasma MPO levels between the early and severe stages (p < 0.05) and between the moderate and severe stages (p < 0.05), while there was no significant difference between the early and moderate stages (p > 0.05) ( Figure 2 B in

[0088] In addition, Spearman analysis showed that plasma MPO levels were positively correlated with the H - P - A grade (rho = 0.557, p < 0.001) and were not affected by IOP (rho = 0.078, p = 0.397).

[0089] 4. Receiver operating characteristic (ROC) curve analysis

[0090] To prove whether plasma MPO can distinguish between the healthy control group and patients with different disease severities, ROC curves were plotted and the area under the ROC curve (AUC) was calculated.

[0091] The staging of glaucoma samples was based on the H-P-A (Hodapp Parrish Anderson) staging system. The AUC values of plasma MPO for differentiating 106 healthy controls from 29 early-stage, 28 moderate-stage, and 67 severe-stage glaucoma patients were 0.743, 0.776, and 0.880, respectively ( Figure 3 A in []). The results showed that the discriminatory ability of plasma MPO increased with the severity of glaucoma.

[0092] Next, it was verified whether the plasma MPO level was sufficient to distinguish between early-stage glaucoma patients and late-stage glaucoma patients. The results were as Figure 3 shown in B of []. MPO showed relatively good accuracy in differentiating between early-stage and late-stage glaucoma. The AUC values for early-stage and moderate-stage glaucoma were 0.516 and 0.633, respectively (evaluated by the H-P-A system). This indicates that MPO may be a valuable indicator for identifying the disease severity.

[0093] The above are only the preferred embodiments of the present invention and are not used to limit the present invention. For those skilled in the art, the present invention can have various changes and modifications. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. Use of a reagent for detecting the content of a biomarker in the preparation of a product for predicting glaucoma, characterized in that, The biomarker includes myeloperoxidase.

2. The application according to claim 1, characterized in that, The prediction of glaucoma includes any one or more of: predicting the risk of developing glaucoma and / or the disease progression.

3. The application according to claim 2, characterized in that, The prediction of the disease progression of glaucoma includes: predicting whether glaucoma is in the early, middle or late stage.

4. The application according to any one of claims 1 to 3, characterized in that, The method for detecting the content of the biomarker includes at least one of the BCA method, Bradford method, Western blotting, immunohistochemistry, ELISA assay, flow cytometry, protein chip and two-dimensional electrophoresis.

5. The application according to any one of claims 1 to 3, characterized in that, The product includes a kit.

6. A training method for a prediction model of glaucoma, characterized in that, It includes: Obtaining the detection result of the biomarker content in the training sample and its annotation result; wherein, the biomarker includes myeloperoxidase, and the annotation result includes a label representing the risk of developing glaucoma and / or the disease progression of the sample; Inputting the detection result of the training sample into a pre-constructed machine learning model to obtain a prediction result; Updating the parameters of the machine learning model according to the prediction result and the annotation result to obtain a prediction model.

7. The training method according to claim 6, wherein The machine learning model includes any one of decision tree, logistic regression model, support vector machine, KNN, naive Bayes and random forest.

8. A prediction device for glaucoma, characterized in that, It includes: An acquisition module for obtaining the detection result of the biomarker content in the sample to be tested; the biomarker includes myeloperoxidase; A prediction module for inputting the detection result of the sample to be tested into the prediction model trained by the training method according to claim 6 or 7 to obtain a prediction result.

9. An electronic device, characterized in that, It includes a processor and a memory, and the memory is used to store a program, which when executed by the processor enables the processor to implement the method for predicting glaucoma; wherein, the prediction method includes: obtaining the detection result of the biomarker content in the sample to be tested; inputting the detection result of the sample to be tested into the prediction model trained by the training method according to claim 6 or 7 to obtain a prediction result.

10. A computer-readable medium, characterized in that, A computer program is stored on the computer-readable medium, and when the computer program is executed by a processor, it implements the method for predicting glaucoma according to claim 9.

Images