Application of biomarker in preparation of products for treating and predicting myasthenia gravis

Abstract

The invention discloses an application of a biomarker in preparation of a product for treating and predicting myasthenia gravis. The biomarker is one or more of OSM, CASP-8, TRAIL, TGF-alpha, FLT3L, IFN-gamma, IL-7 and AXIN1. According to the application, through multi-omics conjoint analysis, the dynamic change and molecular mechanism of immune inflammation related protein after complement inhibition are explored from double dimensions of gene transcription and protein expression, and a comprehensive biomarker capable of systematically evaluating symptom improvement dynamics and disease stability degree is found; the method has the advantages of accurately guiding treatment adjustment and predicting disease outcome.

Description

Technical Field

[0001] The present invention relates to the field of biological medicine technology. Specifically, it relates to the application of a biomarker in the preparation of products for the treatment and prediction of myasthenia gravis. Background Art

[0002] Myasthenia Gravis (MG) is an acquired autoimmune disease mediated by cell immunity and antibodies. The core pathological mechanism of MG with positive anti-acetylcholine receptor (AChR) antibodies is that autoantibodies activate the classical complement cascade reaction, resulting in the destruction of the postsynaptic membrane structure of the neuromuscular junction (NMJ). Although complement C5 inhibitors represented by eculizumab can effectively block this pathological process, complement-related hematological indicators such as anaphylatoxin C5a, membrane attack complex (MAC), and serum complement hemolytic activity (CH50) are difficult to accurately reflect the degree of damage at the NMJ and the disease activity, resulting in the current clinical decision-making still relying on symptom scores and antibody titers, and there is a lag in the evaluation of prognosis and treatment plan decision-making.

[0003] However, the asynchrony between complement inhibition therapy and symptom improvement, as well as the individual differences in symptom fluctuations after drug withdrawal, indicate that the current evaluation system has significant limitations.

[0004] It should be noted that the information disclosed in the above background art section is only used to enhance the understanding of the background of the present disclosure, and thus may include information that does not constitute the prior art known to those of ordinary skill in the art. Summary of the Invention

[0005] The technical task of this application is to address the above deficiencies and provide the application of a biomarker in the preparation of products for the treatment and prediction of myasthenia gravis. Through multi-omics combined analysis, this application explores the dynamic changes and molecular mechanisms of immune-inflammatory related proteins after complement inhibition from the dual dimensions of gene transcription and protein expression, and finds a comprehensive biomarker that can systematically evaluate the dynamic changes of symptom improvement and the degree of disease stability, which has the advantages of accurately guiding treatment adjustment and predicting disease prognosis.

[0006] To achieve the above object, this application provides the following technical solutions:

[0007] According to one aspect of the present application, there is provided an application of a biomarker in the preparation of products for treating and predicting myasthenia gravis, and the biomarker is one or more of OSM, CASP-8, TRAIL, TGF-α, FLT3L, IFN-γ, IL-7, and AXIN1.

[0008] In some embodiments, the biomarker is OSM or IFN-γ.

[0009] In some embodiments, the biomarker includes OSM and IL-7.

[0010] According to another aspect of the present application, there is also provided a biomarker combination, which is composed of serum total complement hemolytic activity, OSM expression level, and IL-7 concentration.

[0011] According to another aspect of the present application, there is also provided an application of the biomarker combination in a kit for predicting the therapeutic effect evaluation of myasthenia gravis.

[0012] In some embodiments, the kit includes reagents for detecting serum total complement hemolytic activity, OSM expression level, and IL-7 concentration.

[0013] In some embodiments, the kit uses ELISA technology or a microfluidic chip to detect serum total complement hemolytic activity, OSM expression level, and IL-7 concentration.

[0014] According to another aspect of the present application, there is also provided an application of the biomarker combination in constructing a clinical score prediction model for myasthenia gravis.

[0015] In some embodiments, the prediction model includes a quantitative myasthenia gravis score prediction model and a daily living ability score prediction model, and the prediction model is constructed using binary Logistic regression analysis.

[0016] In some embodiments, the equation of the quantitative myasthenia gravis score prediction model is: the probability that the quantitative myasthenia gravis score is higher than the median = -5.1975 + 0.0700 × serum total complement hemolytic activity + 0.6866 × OSM expression level - 0.3134 × IL-7 concentration; the equation of the daily living ability score prediction model is: the probability that the daily living ability score is higher than the median = -10.2096 + 0.0446 × serum total complement hemolytic activity + 1.6621 × OSM expression level - 0.4761 × IL-7 concentration.

[0017] Compared with the prior art, the advantages and positive effects of the present application are as follows: Through multi-omics joint analysis, the present application screens the dynamic changes and molecular mechanisms of immune inflammation-related proteins after complement inhibition from the dual dimensions of gene transcription and protein expression, finds comprehensive biomarkers capable of systematically evaluating the dynamic improvement of symptoms and the degree of disease stability, and has the advantages of accurately guiding treatment adjustment and predicting disease prognosis.

[0018] Furthermore, the prediction model combining CH50, OSM and IL-7 proposed for the first time in the present application shows highly reliable prediction ability (C-index = 0.889) and cost / benefit decision-making ability (net benefit close to 1.0). BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required in the embodiments. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0020] Figure 1 Shows the volcano plot of differentially expressed proteins in the serum proteomics of MG patients before and after eculizumab treatment in the embodiments of the present application.

[0021] Figure 2 Shows the paired scatter plot of differentially expressed proteins in the serum proteomics of MG patients before and after eculizumab treatment in the embodiments of the present application.

[0022] Figure 3 Shows the volcano plot and bar chart of differentially expressed genes in the transcriptomics of MG patients before and after eculizumab treatment in the embodiments of the present application.

[0023] Figure 4 Shows the Venn diagram of the correlation analysis between proteomics and transcriptomics in the embodiments of the present application.

[0024] Figure 5 Shows the scatter plot of the correlation analysis between proteomics and transcriptomics in the embodiments of the present application.

[0025] Figure 6 Shows the scatter plot of the correlation analysis between differentially expressed proteins in the serum proteomics before and after eculizumab treatment and the MG clinical score in the embodiments of the present application.

[0026] Figure 7 Shows the ROC curve of the prediction efficacy of the differentially expressed protein level in serum proteomics on the QMG score affected by eculizumab treatment in the embodiments of the present application.

[0027] Figure 8Show the correlation between serum CH50 levels and clinical scores before and after eculizumab treatment in the embodiments of the present application.

[0028] Figure 9 Show the verification of the clinical decision-making value of the QMG score combined prediction model in the embodiments of the present application.

[0029] Figure 10 Show the verification of the clinical decision-making value of the ADL score combined prediction model in the embodiments of the present application. Detailed implementation manners

[0030] In order to more clearly understand the above objects, features and advantages of the present application, the present application will be further described below with reference to the accompanying drawings and embodiments. It should be noted that, without conflict, the embodiments of the present application and the features in the embodiments may be combined with each other.

[0031] Corresponding abbreviations of main biomarkers:

[0032] Oncostatin M (OSM); Caspase 8 (CASP-8); TNF-related apoptosis-inducing ligand (TRAIL);

[0033] Transforming growth factor α (TGF-α); FMS-like tyrosine kinase 3 ligand (FLT3L); Interferon γ (IFN-γ); Interleukin 7 (IL-7); Axin 1 (AXIN1); Quantitative myasthenia gravis (QMG); Activities of daily living (ADL).

[0034] The present application will be further described below with reference to the accompanying drawings and specific embodiments.

[0035] In the embodiments of the present application, 10 patients with anti-AChR antibody-positive generalized myasthenia gravis (gMG) who received standard eculizumab treatment were enrolled. Under the premise of ethical approval and informed consent, peripheral blood samples were collected before treatment, after single induction treatment (week 1), and after completing 4 induction treatments (week 5). After the single induction treatment, the serum total complement hemolytic activity (CH50) was measured by liposome immunoassay; after completing 4 induction treatments, the proximity extension assay (PEA) (Olink Proteomics, Target 96 Inflammation Panel) was used to quantitatively detect serum inflammation-related proteins, and the RNA-seq transcriptome sequencing was combined to analyze the gene expression profile of peripheral blood leukocytes, and the cross-omics data was integrated to describe the complement inhibition-driven immune regulation network.

[0036] 1. Screening of Biomarkers.

[0037] Through the cross-omics joint analysis of Olink ultra-low abundance proteomics (PEA technology) and Bulk RNA-seq transcriptomics, combined with a longitudinal study design, as Figure 1-5 shown, for the first time, 18 serum differentially expressed proteins (P < 0.05) related to the clinical score of myasthenia gravis (MG) were screened from 92 inflammation-related proteins. Among them, OSM and IFN-γ were significant at both the gene transcription and protein expression levels.

[0038] (1) Screening of differentially expressed proteins by Olink ultra-low abundance proteomics.

[0039] Figure 1 The volcano plot of differentially expressed proteins in the serum proteomics of MG patients before and after eculizumab treatment in the embodiments of the present application is shown. Specifically, the volcano plot shows the change trends of all proteins in the differential expression analysis. The abscissa ( FC value) represents the differential change of proteins in different treatments (positive values indicate upregulation, and negative values indicate downregulation), and the ordinate (Pvalue) represents the statistical significance of the change in protein abundance value. Gray represents non-differentially expressed proteins, red represents significantly upregulated differentially expressed proteins, including Ftl3L (FMS-like Tyrosine Kinase 3 ligand), CSF-1 (Colony Stimulating Factor 1), FGF-19 (Fibroblast Growth Factor 19), CD8A (CD8 Alpha Chain), TRAIL (TNF-Related Apoptosis-Inducing Ligand), IFN-gamma (Interferon-gamma), CCL19 (C-C Motif Chemokine Ligand 19), IL7 (Interleukin-7), OPG (Osteoprotegerin), CCL28 (C-C Motif Chemokine Ligand 28). Blue represents significantly downregulated differentially expressed proteins, including CASP-8 (Caspase-8), AXIN1 (Axis Inhibition Protein 1), TGF-alpha (Transforming Growth Factor-alpha), SIRT2 (Sirtuin 2), ST1A1 (Sulfotransferase Family 1A Member 1), STAMBP (Signal Transducing Adaptor Molecule Binding Protein), OSM (Oncostatin M), IL-22 RA1 (Interleukin-22 Receptor Alpha 1).

[0040] Figure 2 The paired scatter plot of differentially expressed proteins in the serum proteomics of MG patients before and after eculizumab treatment in the examples of this application is shown. Among them, the paired scatter plot takes the NPX value as the vertical coordinate, and red and blue respectively represent the differential protein expression levels before and after eculizumab treatment. * indicates the statistical significance of differential expression, *P < 0.05, **P < 0.01.

[0041] (2)Bulk RNA-seq transcriptomics screening for differentially expressed genes:

[0042] Figure 3 Shows the volcano plot and bar chart of differentially expressed genes in the transcriptomics of MG patients before and after eculizumab treatment in the embodiments of the present application. Among them, the abscissa of the volcano plot ( Fold Change) represents the fold change of gene expression (upregulation for positive values, downregulation for negative values), and the ordinate ( Adjust P value) represents the significance of the difference. Red dots (UP) mark significantly upregulated genes, blue dots (DOWN) mark significantly downregulated genes, and gray dots (STABLE) are genes with no significant difference; the bar chart on the right shows the distribution of the number of differentially expressed genes (red bar (Up): upregulation, dark blue bar (Down): downregulation, light blue bar (Total): total).

[0043] (3)Integrated analysis of proteomics and transcriptomics:

[0044] Figure 4 Shows the Venn diagram of the correlation analysis between proteomics and transcriptomics in the embodiments of the present application. All_protein: represents all proteins that can be quantified in the proteome; among them, All_mRNA: represents all quantifiable genes obtained from the transcriptome; DE_Protein: represents differentially expressed proteins identified in the proteome; DE_Gene: represents differentially expressed genes identified in the transcriptome. The color scale bar on the right intuitively maps the element quantity scale of each region through color gradients.

[0045] Figure 5 Shows the scatter plot of the correlation analysis between proteomics and transcriptomics in the embodiments of the present application. Among them, the abscissa represents (Ratio)represents the fold change of protein expression (upregulation for positive values, downregulation for negative values), and the ordinate (FC)represents the fold change of differentially expressed gene expression (upregulation for positive values, downregulation for negative values). Data point color classification: purple dots (NDEPs_NDEGs): the overlapping part of non-differentially expressed proteins and non-differentially expressed genes; green dots (DEPs_NDEGs): the overlapping part of differentially expressed proteins and non-differentially expressed genes; blue dots (NDEPs_DEGs): the overlapping part of non-differentially expressed proteins and differentially expressed genes; red dots (DEPs_DEGs): the overlapping part of differentially expressed proteins and differentially expressed genes.

[0046] 2. Correlation verification between biomarkers and clinical scores of myasthenia gravis (MG) (QMG score and ADL score).

[0047] Figure 6The scatter plot showing the correlation analysis between differentially expressed proteins in serum proteomics before and after eculizumab treatment and the MG clinical score in the embodiments of the present application is shown. In the figure, the QMG score is the quantitative myasthenia gravis score, and the ADL is the activity of daily living score. As Figure 6 shown, through Spearman rank correlation analysis, it was found that the dynamic changes in the expression of OSM, CASP-8, TRAIL, TGF-α, FLT3L, IFN-γ, IL-7 and AXIN1 were significantly correlated with the improvement of the MG clinical score (P<0.05). The results of univariate linear regression analysis further found that among them, OSM (β=3.233, 95%CI 0.3279 to 6.138, P<0.05), CASP-8 (β=3.058, 95%CI 0.6723 to 5.443, P<0.05), TGF-α (β=3.962, 95%CI 0.1077 to 7.816, P<0.05) and TRAIL (β=-8.719, 95%CI -16.57 to -0.8645, P<0.05) had a potential causal association with the improvement of the MG-QMG score. OSM (β=1.905, 95%CI 0.4230 to 3.387, P<0.05), CASP-8 (β=1.950, 95%CI 0.8168 to 3.082, p<0.01) and TGF-α (β=2.441, 95%CI 0.4997 to 4.383, P<0.05) had a potential causal association with the improvement of the MG-ADL score.

[0048] 3. Analysis of the predictive efficacy of the levels of differentially expressed proteins in serum proteomics on the QMG score affected by eculizumab treatment.

[0049] Figure 7 The ROC curve showing the predictive efficacy of the levels of differentially expressed proteins in serum proteomics on the QMG score affected by eculizumab treatment in the embodiments of the present application is shown. Among them, the horizontal axis is 1-specificity, the vertical axis is sensitivity, each protein is represented by a curve of a different color, and the AUC value is marked beside it. Figure 7 a are the differentially expressed proteins with significant Spearman correlation analysis. Figure 7 b are the differentially expressed proteins with significant Spearman correlation analysis and univariate linear regression analysis. As Figure 7As shown, by plotting the ROC curve, it was found that IL-7 (AUC = 0.99, cut-off value > 3.643 pg / mL, sensitivity 95%, specificity 100%) had both high sensitivity and high specificity, demonstrating good predictive ability for treatment response. The other 7 serum differentially expressed proteins all had good predictive efficacy: AXIN1 (AUC = 0.9875, cut-off value > 3.810 pg / mL, sensitivity 45%), TGF-α (AUC = 0.89, < 2.442 pg / mL, sensitivity 50%), IFN-γ (AUC = 0.7325, cut-off value < 4.046 pg / mL, sensitivity 60%), TRAIL (AUC = 0.732, cut-off value < 4.046 pg / mL, sensitivity 65%), CASP-8 (AUC = 0.7994, cut-off value > 7.790 pg / mL, sensitivity 67%, specificity 94%), OSM (AUC = 0.7513, cut-off value > 7.915 pg / mL, sensitivity 67%, specificity 94%), FLT3L (AUC = 0.5802, cut-off value < 6.024 pg / mL).

[0050] 4. Construct a combined prediction model based on potential therapeutic targets to optimize the complement inhibition treatment cycle.

[0051] Figure 8 Shows the correlation between serum CH50 levels and clinical scores before and after eculizumab treatment in the embodiments of the present application, where **P < 0.01. As Figure 8 shown, IL-7 demonstrated excellent response classification efficacy (AUC = 0.99). However, its protein level increased significantly after complement inhibition treatment (P < 0.05), but the change in mRNA expression was not statistically significant (FDR = 0.078). Complement inhibition mediates the downregulation of OSM protein expression at the gene transcription level, which is more suitable for tracking the evolution of chronic immune homeostasis. Based on this, constructing a prediction model combining CH50 (complement activity) + IL-7 + OSM can break through the limitations of the time effect and predictive efficacy of single markers, providing a full-cycle management strategy for complement inhibition treatment from acute response to chronic homeostasis.

[0052] Figure 9 Shows the verification of the clinical decision-making value of the QMG score combined prediction model (calibration curve and decision curve analysis) in the embodiments of the present application, Figure 10 Shows the verification of the clinical decision-making value of the ADL score combined prediction model (calibration curve and decision curve analysis) in the embodiments of the present application, where Figure 9 a and Figure 10For a (calibration curve), the abscissa represents the probability value that the clinical score predicted by the model is higher than the median, and the ordinate represents the true proportion of the actually observed clinical score higher than the median. The three curves in the figure represent respectively: the Ideal line (black dashed line) represents the ideal state where the predicted probability is exactly equal to the actual probability; the Apparent line (gray dashed line) represents the prediction performance of the uncorrected model on the training data; the Bias-corrected line represents the calibration result after overfitting correction by 40 times of Bootstrap resampling (B = 40). Figure 9 b and Figure 10 b (Decision Curve Analysis): The horizontal axis (threshold probability) represents the minimum risk probability that needs to be intervened as defined by the clinical decision maker, and the cost-benefit ratio (Cost:Benefit Ratio) marked above it converts the economic consideration into the clinical decision weight (1:100 means that for every 1 unit of cost paid, 100 units of health loss can be avoided). The vertical axis represents the standardized net benefit, which quantifies the clinical benefit efficiency of different strategies. The red curve represents the net benefit curve model of the CH50 + IL-7 + OSM combined biomarker model, the gray curve represents the net benefit assuming intervention for all patients, and the black curve represents the net benefit assuming no intervention for all patients.

[0053] As Figure 9 shown, its prediction probability calibration degree and clinical decision value are evaluated through Decision Curve Analysis (DCA) of the calibration curve. Based on the fluctuations of serum CH50, OSM, and IL-7 levels before and after treatment, a QMG score prediction model (QMG_binary) is constructed through binary Logistic regression analysis: logit(p QMG ) = -5.1975 + 0.0700 × serum CH50 + 0.6866 × OSM - 0.3134 × IL-7, where p QMGIndicates the probability that the QMG score is higher than the median. The calibration curve results show that the model has a systematic overestimation tendency in the predicted probability range of 0.3 - 0.7 (the Apparent curve deviates from the ideal line by up to 0.25). However, after B = 40 times of Bootstrap bias correction, the calibration accuracy is significantly improved, and the mean absolute error (MAE) decreases from 0.138 to 0.098. Especially in the high-risk threshold area > 0.6, the error between the predicted probability and the actual observed probability is controlled within ±0.1 after correction, confirming the risk stratification ability of the model for critically ill patients (C-index = 0.889). The DCA curve further quantifies the clinical decision-making value of the model: when the high-risk threshold set by the clinical decision-maker is between 17% and 73% (corresponding to the cost-benefit ratio of 1:5 to 3:2), the net benefit value of the individualized treatment plan guided by the combined biomarker stabilizes in the range of 0.778 - 0.95, significantly superior to the empirical full-intervention strategy (gray line) and the conservative treatment strategy (black line). It is worth noting that the net benefit peak of 0.97 (sensitivity 95%, specificity 100%) is reached at the threshold probability of 21%, indicating that the model has the optimal cost-benefit balance in the medium-risk interval, and preliminarily confirming the clinical translation potential of the combined prediction model.

[0054] As Figure 10 shown, after establishing the predictive efficacy of the biomarkers related to the QMG (Quantitative Myasthenia Gravis) score, an independent validation model based on the ADL (Activities of Daily Living) score system was constructed simultaneously to comprehensively evaluate the predictive efficacy of the combined prediction model for the patient's muscle function (QMG) and quality of life (ADL). Based on the pre- and post-treatment level fluctuations of serum CH50, OSM, and IL-7, a predictive system for the ADL score (ADL_binary) was constructed through binary Logistic regression analysis: logit(p ADL ) = -10.2096 + 0.0446 × serum CH50 + 1.6621 × OSM - 0.4761 × IL-7, where p ADLIndicates the probability that the QMG score is higher than the median. The calibration curve results show that there is a systematic overestimation in the predicted probability range of 0.4 - 0.8 (the Apparent curve deviates from the ideal line by up to 0.23). After B = 40 times of Bootstrap bias correction, the mean absolute error (MAE) decreases from 0.132 to 0.093. The error between the actual probability and the predicted value in the high-risk threshold > 0.5 area is stable within ±0.08, verifying the identification efficacy of the model for patients with severe functional impairment (C-index = 0.938). The DCA curve further quantifies its clinical utility: when the risk threshold is set at 19% - 73% (corresponding cost-benefit ratio of 1:4 to 3:2), the net benefit value of the intervention strategy guided by the combined biomarker remains in the range of 0.795 - 0.951, significantly superior to the empirical full-intervention strategy (gray line) and the non-intervention strategy (black line), and reaches the net benefit peak of 0.96 (sensitivity 93%, specificity 92%) at the threshold probability of 22%, highlighting the clinical decision-making advantage of this model in patients with moderate risk and revealing the precise assessment value of this prediction model for the risk of impaired activities of daily living. It is worth noting that the robustness characteristics of this combined prediction model in the moderate-risk interval highly coincide with the "moderate-risk priority optimization" strategy proposed in previous studies. This theory means that in medical decision-making, precise intervention strategies are preferentially formulated for the moderate-risk patient group to achieve the balance of maximizing clinical benefits and minimizing resource consumption and avoid the interference of the extreme-risk area on clinical translation. In addition, this combined prediction model breaks through the fragmented interpretation of the pathological process by the traditional single biomarker system and realizes the closed-loop verification from the value of health economics to the mechanism interpretability under the dual background of the lag in the development of biomarkers for myasthenia gravis and the pressure of optimizing medical cost-benefit.

[0055] The application in the clinical decision-making and efficacy evaluation of myasthenia gravis (MG) is mainly reflected in: Based on the CH50 + IL-7 + OSM combined prediction model (C-index = 0.889 - 0.938, net benefit peak > 0.95), developing a rapid detection kit:

[0056] (1) Technical carrier: Using ELISA technology or microfluidic chips to achieve synchronous detection of multiple targets (quantification of IL-7 / OSM + determination of CH50 activity).

[0057] (2) Application scenarios:

[0058] A. Stratification before treatment: Screening high-responding patients through the baseline level of IL-7 (critical value > 3.643 pg / mL) (AUC = 0.99) to guide the selection of applicable populations for complement inhibitors.

[0059] B. Dynamic monitoring of therapeutic efficacy: Evaluate the steady-state repair process of ECM by combining the dynamic changes of OSM (a decrease of > 30% in the 5th week of treatment), and optimize the medication cycle (such as extending the maintenance interval when CH50 < 10 U / mL).

[0060] Through the above specific embodiments, those skilled in the technical field can easily implement this application. However, it should be understood that this application is not limited to the above specific embodiments. Based on the disclosed embodiments, those skilled in the technical field can arbitrarily combine different technical features to achieve different technical solutions.

Claims

1. Use of a biomarker in the preparation of a product for treating and predicting myasthenia gravis, characterized in that, The biomarkers include OSM and IL-7.

2. A biomarker combination, characterized in that, The biomarker combination consists of serum total complement hemolytic activity, OSM expression level, and IL-7 concentration.

3. Use of the biomarker combination according to claim 2 in a kit for predicting the therapeutic effect evaluation of myasthenia gravis.

4. The application according to claim 3, wherein The kit includes reagents for detecting serum total complement hemolytic activity, OSM expression level, and IL-7 concentration.

5. The application according to claim 4, characterized in that, The kit uses ELISA technology or a microfluidic chip to detect serum total complement hemolytic activity, OSM expression level, and IL-7 concentration.

6. Use of the biomarker combination according to claim 2 in constructing a prediction model for the clinical score of myasthenia gravis.

7. The application according to claim 6, wherein The prediction model includes a quantitative myasthenia gravis score prediction model and a daily living ability score prediction model, and the prediction model is constructed using binary Logistic regression analysis.

8. The application according to claim 7, characterized in that The equation of the quantitative myasthenia gravis score prediction model is: probability of quantitative myasthenia gravis score higher than the median = -5.1975 + 0.0700 × serum total complement hemolytic activity + 0.6866 × OSM expression level - 0.3134 × IL-7 concentration; the equation of the daily living ability score prediction model is: probability of daily living ability score higher than the median = -10.2096 + 0.0446 × serum total complement hemolytic activity + 1.6621 × OSM expression level - 0.4761 × IL-7 concentration.

Images