A quantitative prediction model for the response of multiple sclerosis patients to steroid pulse therapy

By constructing a quantitative predictive model for the response to hormone therapy in patients with multiple sclerosis, and utilizing factors such as the number of lesions enhanced by MRI, the total protein content of cerebrospinal fluid, and the side effects of hormone therapy, the problem of unpredictable response to hormone therapy in patients with multiple sclerosis has been solved, enabling the formulation of precise treatment plans and reducing side effects and costs.

CN117649936BActive Publication Date: 2024-11-26THE SECOND AFFILIATED HOSPITAL ARMY MEDICAL UNIV
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202311687908.X
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2023-12-11
Publication Date
2024-11-26
Estimated Expiration
2043-12-11

AI Technical Summary

Technical Problem

Currently, there is no simple, quick, and quantitative method to predict the response of multiple sclerosis patients to hormone therapy, which may lead to side effects or increased medical costs due to blind treatment.

Method used

A quantitative predictive model for the response of multiple sclerosis patients to hormone pulse therapy was constructed. The probability of insensitivity was calculated using a multivariate logistic regression analysis model based on the number of lesions on enhanced MRI, total protein content in cerebrospinal fluid, waiting period from onset to treatment, and hormone therapy side effect scores.

Benefits of technology

It provides a simple, fast, and accurate prediction method to guide clinicians in developing precise treatment plans, reducing side effects and medical costs, and improving treatment outcomes.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure SMS_1
    Figure SMS_1
  • Figure SMS_2
    Figure SMS_2
  • Figure SMS_3
    Figure SMS_3
Patent Text Reader

Abstract

The present invention discloses a quantitative prediction model and construction method for the response effect of multiple sclerosis patients to hormone shock therapy. The method constructs a statistical model for predicting the effect of hormone therapy by using statistical methods by analyzing the baseline clinical data, clinical symptoms and signs, magnetic resonance imaging results, cerebrospinal fluid test parameters, and side effects that occur within 5 days of hormone therapy of multiple sclerosis patients who are sensitive and insensitive to hormone shock therapy before treatment, and converts the model into an easy-to-operate scoring scale by statistical methods, which is convenient for clinical promotion and application. The present invention helps to quantitatively predict the response effect of multiple sclerosis patients to hormone shock therapy, thereby guiding clinicians to formulate accurate and effective treatment plans.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention belongs to the technical field of medical diagnosis, and in particular relates to a quantitative prediction model for the response effect of multiple sclerosis patients to hormone shock therapy. Background Art

[0002] Multiple sclerosis is an immune-mediated disease characterized by inflammatory demyelinating lesions in the central nervous system, which mainly affect the white matter. The disease is characterized by diverse clinical manifestations, which can cause a variety of symptoms such as decreased vision and limb sensory and motor disorders. The course of the disease is mostly a relapse-remission-progression process. Delays in diagnosis and treatment may lead to severe disability.

[0003] Treatment of the acute phase of multiple sclerosis is crucial for improving prognosis, shortening the course of the disease, and reducing disability. Hormone shock therapy (methylprednisolone, starting at 1g / d) is the first choice for the acute phase of multiple sclerosis, but it has great side effects (infection, arrhythmia, gastrointestinal bleeding, electrolyte imbalance, osteoporosis, hyperglycemia, etc.). More importantly, hormone use is ineffective for some patients. At this time, plasma exchange and intravenous high-dose immunoglobulin need to be used or combined to improve the prognosis. However, plasma exchange and immunoglobulin are extremely expensive and cannot be carried out in some primary medical institutions. At the same time, there are also side effects such as invasiveness and allergies. Therefore, early prediction of whether the patient is sensitive to hormone therapy is conducive to the formulation of a reasonable treatment plan to significantly reduce the disability and mortality rates, avoid the side effects caused by blind high-dose hormone shocks, or blindly add plasma exchange and immunoglobulin to cause high medical expenses and related side effects. However, there is currently no simple, fast and quantitative method to predict the response of multiple sclerosis patients to hormone therapy. Summary of the invention

[0004] The purpose of the present invention is to provide a quantitative prediction model for the response of multiple sclerosis patients to hormone pulse therapy. In view of the lack of effective means for predicting the response of multiple sclerosis patients to hormone therapy, the present invention provides a simple, fast and quantitative prediction method for guiding the precise treatment of patients and improving their prognosis.

[0005] To achieve the above-mentioned object of the invention, the present invention provides the following implementation scheme:

[0006] In one embodiment, the present invention provides a quantitative prediction model for the response of multiple sclerosis patients to hormone pulse therapy, and the model is the following scoring quantification table, see Table 1 below.

[0007] Table 1. Quantification table of risk score for hormone therapy insensitivity

[0008]

[0009] The total score is the sum of the MRI enhanced lesion number score, the cerebrospinal fluid total protein score, the onset to treatment waiting period score, and the hormone therapy side effect score.

[0010] The quantitative prediction model of the present invention is as follows: the number of MRI enhanced lesions (n) is divided into four levels: n<2, 2≤n<4, 4≤n<6, n≥6, and is assigned 0, 1, 2, and 3 points respectively;

[0011] The total protein content score of cerebrospinal fluid, the total protein content (n, g / L) is divided into five levels: n < 0.2, 0.2 ≤ n < 0.4, 0.4 ≤ n < 0.6, 0.6 ≤ n < 0.8, n ≥, 0.8, and assigned 0, 0, 1, 2, 2 points respectively;

[0012] The waiting period from onset to treatment score, the waiting period (n, months) is divided into four levels: n<3, 3≤n<6, 6≤n<12, n≥12, and is assigned 0, 0, 1, and 2 points respectively;

[0013] The side effects during the hormone treatment were divided into none or two levels, with 0 and 1 points respectively.

[0014] In the quantitative prediction model of the present invention, the multiple sclerosis patients are patients diagnosed with multiple sclerosis for the first time. The hormone shock therapy is administered by intravenous drip starting from 1 g / d for 3 to 4 hours for a total of 3 to 5 days.

[0015] In another embodiment, a method for constructing a quantitative prediction model for the response of multiple sclerosis patients to hormone pulse therapy comprises the following steps:

[0016] 1) Collect cases of multiple sclerosis patients who received hormone pulse therapy for the first time, determine and classify them into hormone-sensitive group and insensitive group, with no less than 25 cases in each group, and analyze the index parameters of each patient before treatment, including baseline clinical data, clinical symptoms and signs, magnetic resonance imaging results, cerebrospinal fluid test parameters, and side effects data within 5 days of hormone treatment;

[0017] 2) The index parameters collected in the previous step were screened through a variety of statistical methods to identify possible predictive factors for the effect of hormone therapy, and finally four predictive factors were determined: the number of MRI-enhanced lesions, the total protein content of cerebrospinal fluid, the waiting period from onset to treatment, and the side effects of hormone therapy;

[0018] 3) Divide the indicator parameters of each predictive factor into multiple levels, assign corresponding scores to each level, and calculate the distribution range of the total scores of the four predictive factors;

[0019] 4) Based on the multi-factor logistics regression analysis model, the insensitive probability corresponding to each score within the total score distribution range is calculated and converted into a simple scoring scale.

[0020] Preferably, in the method for constructing the quantitative prediction model of the present invention, the levels and the corresponding scores in step 3) are as follows:

[0021] The number of MRI-enhanced lesions (n) was divided into four levels: n < 2, 2 ≤ n < 4, 4 ≤ n < 6, and n ≥ 6, and was assigned 0, 1, 2, and 3 points, respectively;

[0022] The total protein content of cerebrospinal fluid (n, g / L) was divided into five levels: n < 0.2, 0.2 ≤ n < 0.4, 0.4 ≤ n < 0.6, 0.6 ≤ n < 0.8, and n ≥ 0.8, and was assigned 0, 0, 1, 2, and 2 points, respectively;

[0023] The waiting period from onset to treatment (n, months) was divided into four levels: n < 3, 3 ≤ n < 6, 6 ≤ n < 12, and n ≥ 12, and was assigned 0, 0, 1, and 2 points, respectively;

[0024] Side effects during hormone therapy were divided into none or two levels, assigned 0 and 1 points respectively.

[0025] As described above, the total score distribution range of the four predictive factors is: 0-15 points. Preferably, the method for constructing the quantitative prediction model of the present invention, the multi-factor logistics regression analysis model, its statistical model equation is: LogitP = -4.344 + 0.931 × number of MRI enhanced lesions + 5.302 × total cerebrospinal fluid protein + 0.199 × waiting period from onset to treatment + 2.312 × side effects of hormone treatment.

[0026] In another aspect, the present invention also provides the use of the quantitative prediction model of the response effect of multiple sclerosis patients to hormone pulse therapy of the present invention for predicting the response effect of multiple sclerosis patients to hormone pulse therapy. The use comprises taking diagnostic index parameters for multiple sclerosis patients diagnosed for the first time, confirming the score values ​​of the number of MRI enhanced lesions, the total protein content of cerebrospinal fluid, the waiting period from onset to treatment and the side effects during hormone treatment, summing the scores, calculating the total score, checking the quantitative score table of hormone treatment insensitivity risk, and directly obtaining the probability of hormone treatment insensitivity.

[0027] The above-mentioned related diseases of the present invention are patients diagnosed with multiple sclerosis for the first time.

[0028] The dosage regimen of the hormone shock therapy described in the above invention is to start with 1g / d, intravenously drip for 3 to 4 hours, for a total of 3 to 5 days.

[0029] In a specific embodiment, the method for constructing a quantitative prediction model of the response effect of multiple sclerosis patients to hormone pulse therapy of the present invention comprises the following steps:

[0030] 1) After receiving hormone shock therapy, patients with multiple sclerosis who were first diagnosed were judged to belong to the hormone-sensitive group or the hormone-insensitive group, with more than 25 cases in each group. Then the index parameters of the patients before receiving treatment were analyzed;

[0031] 2) Analyze the index parameters, including baseline clinical data, clinical symptoms and signs, magnetic resonance imaging results, and cerebrospinal fluid test parameters. Combined with the side effects data within 5 days of hormone treatment, use a variety of statistical methods to screen out possible predictive factors for the effect of hormone treatment, and construct a multivariate logistics regression analysis model to determine the significant predictive factors;

[0032] 3) identify four significant predictive factors: the number of MRI-enhanced lesions, the amount of total cerebrospinal fluid protein, the waiting period from onset to treatment, and the side effects of hormonal treatment;

[0033] 4) The four identified predictive factors were scored, and the probability of hormone insensitivity was further calculated using a statistical regression model, which was further converted into a simple and easy-to-use scoring scale. The four predictive factors were stratified and assigned scores. Finally, the total score was calculated and the corresponding probability table was used to predict the probability of hormone insensitivity, as follows:

[0034] 1. The number of MRI enhanced lesions (n) was divided into four levels: n<2, 2≤n<4, 4≤n<6, and n≥6, and was assigned 0, 1, 2, and 3 points respectively;

[0035] 2. The total protein content of cerebrospinal fluid (n, g / L) is divided into five levels: n < 0.2, 0.2 ≤ n < 0.4, 0.4 ≤ n < 0.6, 0.6 ≤ n < 0.8, n ≥ 0.8, and is assigned 0, 0, 1, 2, and 2 points respectively;

[0036] 3. The waiting period from onset to treatment (n, months) was divided into four levels: n<3, 3≤n<6, 6≤n<12, and n≥12, and was assigned 0, 0, 1, and 2 points respectively;

[0037] 4. Side effects during hormone therapy were divided into none or two levels, with 0 and 1 points respectively;

[0038] 5. Sum the scores of the above four factors to calculate the total score, and then use the converted probability scale to find the probability of predicting insensitivity to hormone therapy.

[0039] Terminology: Hormone therapy side effects refer to side effects or adverse reactions occurring within 5 days of hormone therapy. When using the quantitative prediction model of the present invention, patients diagnosed with multiple sclerosis for the first time are treated with hormone therapy for 3 to 5 days to obtain hormone therapy side effect parameter data (factors).

[0040] The quantitative prediction model construction method of the present invention establishes a model for predicting the response of multiple sclerosis patients to hormone pulse therapy from the perspective of simple, fast and quantitative evaluation. The model is simple, intuitive, and has high prediction accuracy. It can guide clinicians to formulate accurate and effective treatment plans and has good prospects for clinical application and promotion. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 Flowchart for the use of a quantitative prediction model for the response of patients with multiple sclerosis to steroid pulse therapy. DETAILED DESCRIPTION

[0042] The following is a specific description of a quantitative prediction model of the response effect of multiple sclerosis patients to hormone pulse therapy in the present invention through examples. It is necessary to point out that the following examples are only used to further illustrate the present invention to help understand the spirit of the present invention, and cannot be understood as limiting the scope of protection of the present invention. Technical personnel in this field can make some non-essential improvements and modifications based on the spirit of the present invention mentioned above, which also fall within the scope of the present invention.

[0043] Example 1 Construction of a statistical model for predicting the response of multiple sclerosis patients to hormone pulse therapy

[0044] 1. After receiving hormone pulse therapy, 55 patients with multiple lesions who were first diagnosed had their treatment effects determined by three senior neurologists. The patients were divided into a hormone-sensitive group (26 patients) and an insensitive group (29 patients).

[0045] 2. The baseline clinical data, clinical symptoms and signs, magnetic resonance imaging results, cerebrospinal fluid test parameters, and side effects within 5 days of hormone treatment were collected, and possible predictive factors for the effect of hormone treatment were screened out through various statistical methods such as t-test, chi-square test, and univariate logistics regression analysis.

[0046] 2.1 The baseline clinical data before treatment were analyzed by t-test and chi-square test, and it was found that there were differences in "BMI" and "waiting period from onset to treatment" between the hormone-sensitive group and the insensitive group, which may be predictive factors. See Table 2 for details.

[0047] Table 2. Baseline clinical data of patients before hormone therapy

[0048]

[0049] 2.2 The MRI imaging data of patients before treatment were analyzed by t-test, and it was found that there were differences in the "number of MRI lesions" and "number of MRI enhanced lesions" between the hormone-sensitive group and the insensitive group, which may be predictive factors. See Table 3 for details.

[0050] Table 3. MRI imaging analysis of patients before hormone therapy

[0051]

[0052] 2.3 The t-test and chi-square test were used to analyze the cerebrospinal fluid test results before treatment. It was found that there were differences in the "cerebrospinal fluid cell count" and "cerebrospinal fluid total protein level" between the hormone-sensitive group and the insensitive group, which may be predictive factors. See Table 4 for details.

[0053] Table 4. Analysis of cerebrospinal fluid test results before patients received hormone treatment

[0054]

[0055] 2.4 The relationship between the clinical symptoms / signs of patients before treatment and the effect of hormone treatment was analyzed by univariate logistic regression, and it was found that "Barbina sign / active tendon reflex" may be a predictive factor (see Table 5 for details).

[0056] Table 5. Univariate logistic regression analysis of the relationship between symptoms and signs before treatment and the effect of hormone therapy

[0057] factor OR 95% confidence interval for OR P Babinski sign / brisk tendon reflexes 8.531 2.533to28.729 0.001

[0058] 2.5 The relationship between hormone-related side effects and the efficacy of hormone therapy was analyzed by univariate logistic regression analysis, and it was found that "the presence or absence of hormone side effects" may be a predictive factor (see Table 6 for details).

[0059] Table 6. Univariate logistic regression analysis of the relationship between hormone-related side effects and hormone therapy efficacy in patients receiving hormone therapy

[0060] factor OR 95% confidence interval for OR P Side effects of hormone therapy 17.000 3.362to85.968 0.001

[0061] 2.6 Through multivariate logistic regression analysis, the "BMI", "waiting period from onset to treatment", "number of MRI lesions", "number of MRI enhanced lesions", "number of cerebrospinal fluid cells", "total cerebrospinal fluid protein", "Babinski sign / active tendon reflexes" and "side effects during hormone treatment" mentioned in 2.1-2.5 were all included in the analysis, and finally the four predictive factors of "number of MRI enhanced lesions", "total cerebrospinal fluid protein", "waiting period from onset to treatment" and "side effects during hormone treatment" were determined, as shown in Table 7. The statistical model equation was constructed as follows:

[0062] LogitP=-4.344+0.931×number of MRI enhanced lesions+5.302×total cerebrospinal fluid protein+0.199×waiting period from onset to treatment+2.312×side effects of hormone treatment.

[0063] Table 7. Relationship between relevant factors and hormone therapy effects in multivariate logistic regression analysis

[0064]

[0065] Example 2 Converting the statistical prediction model into a scoring scale to facilitate clinical application

[0066] 1. Classify each risk factor and specify the reference value W for each group ij , see columns 1, 2, and 3 of Table 8 for details;

[0067] 2. Determine the basic risk reference value W for each risk factor iREF , see column 3 of Table 8 for details;

[0068] 3. Calculate the distance D between each risk factor group and the basic risk reference value, formula D = (W ij -W iREF )*β i , see column 5 of Table 8 for details;

[0069] 4. Set the constant B corresponding to 1 point in the scoring tool, B = 2 × β MRI增强病灶数目 =2×0.931=1.862, that is, the number of MRI-enhanced lesions was used as a reference;

[0070] 5. Calculate the points corresponding to each stratification of risk factors ij , the calculation formula is Points ij =D / B=(W ij -W iREF )*β i / B, and finally the calculated value is rounded to the nearest integer, which is the score corresponding to the group, see the sixth column of Table 8 for details;

[0071] 6. Calculate the correspondence table between the total score and the risk prediction probability. The formula is:

[0072] in,

[0073] The corresponding table of the calculated total score risk prediction probability is shown in Figure 1 .

[0074] Table 8. Construction of a scoring scale for predicting the effect of hormone therapy using a statistical prediction model

[0075]

[0076] β i : Regression coefficient; B = 2 × β MRI增强病灶数目 =2×0.931=1.862.

[0077] Example 3: Example application of the scoring scale and comparison with the probability of hormone therapy insensitivity calculated by the statistical prediction model

[0078] The process of using the quantitative prediction model of the present invention is shown in Figure 1 .

[0079] When the patient was admitted to the hospital, MRI revealed three enhancing lesions ( Figure 1 , step 1: 1 point), the total protein content of cerebrospinal fluid is 0.45g / L ( Figure 1 , Step 2: 1 point), the waiting period from onset to treatment is 4 months ( Figure 1 , Step 3: 0 points), no side effects occurred after 5 days of hormone treatment ( Figure 1 , Step 4: 0 points), the total score is 2 points ( Figure 1 , step 5), query the corresponding table to find out that the probability of the patient being insensitive to hormone treatment is 81.16% ( Figure 1 , step 6).

[0080] The probability is calculated using a statistical prediction model, the formula is:

[0081]

[0082] in,

[0083] Finally calculated

[0084] It can be seen that there is only a 2.47% difference between the prediction results of the scoring scale and the Logistic regression model, which is sufficient to meet the requirements of disease risk prediction. It is also more intuitive and convenient to use, and has good prospects for clinical promotion and application.

Claims

1. A method for constructing a quantitative prediction model for the response of multiple sclerosis patients to hormone pulse therapy. The following steps are involved: 1) Collect cases of multiple sclerosis patients who received hormone shock therapy for the first time, determine and classify them into hormone-sensitive group and insensitive group, with more than 25 cases in each group, and analyze the index parameters of each patient before treatment, including baseline clinical data, clinical symptoms and signs, magnetic resonance imaging results, cerebrospinal fluid test parameters, and side effects data index parameters that occurred within 5 days of hormone treatment; 2) The index parameters collected in the previous step were statistically screened for possible predictive factors for the effect of hormone therapy, and finally four predictive factors were determined by multivariate logistics regression analysis: the number of MRI-enhanced lesions, the total protein content of cerebrospinal fluid, the waiting period from onset to treatment, and the side effects of hormone therapy; 3) Divide the indicator parameters of each predictive factor into multiple levels, assign corresponding scores to each level, and calculate the distribution range of the total scores of the four predictive factors; 4) Based on the multi-factor logistics regression analysis model, the insensitive probability corresponding to each score within the total score distribution range is calculated and converted into a simple scoring scale. Among them, the model has the following scoring quantization table, The total score is the sum of the MRI enhanced lesion number score, the cerebrospinal fluid total protein score, the onset to treatment waiting period score, and the hormone therapy side effect score. The number of MRI enhanced lesions (n) is divided into four levels: n<2, 2≤n<4, 4≤n<6, and n≥6, and is assigned 0, 1, 2, and 3 points respectively; The total protein content of cerebrospinal fluid (n, g / L) is divided into five levels: n < 0.2, 0.2 ≤ n < 0.4, 0.4 ≤ n < 0.6, 0.6 ≤ n < 0.8, and n ≥ 0.8, and is assigned 0, 0, 1, 2, and 2 points respectively; The waiting period from onset to treatment (n, months) was divided into four levels: n<3, 3≤n<6, 6≤n<12, and n≥12, and was assigned 0, 0, 1, and 2 points respectively; The side effects of hormone therapy were divided into none or two levels, with 0 and 1 points respectively. The total score distribution range of the above four predictive factors was: 0-15 points. The statistical model equation of the multi-factor logistics regression analysis model is: LogitP=-4.344+0.931×number of MRI enhanced lesions+5.302×total cerebrospinal fluid protein+0.199×waiting period from onset to treatment+2.312×side effects of hormone treatment.

2. The method for constructing a quantitative prediction model according to claim 1, wherein the multiple sclerosis patient is a patient diagnosed with multiple sclerosis for the first time.

3. The method for constructing a quantitative prediction model as described in claim 1, wherein the hormone shock therapy is started with 1 g / d, intravenously dripped for 3 to 4 hours, for a total of 3 to 5 days.

4. Use of the quantitative prediction model constructed by the construction method of claim 1 for predicting the response of multiple sclerosis patients to hormone pulse therapy.

5. The use as claimed in claim 4, comprising taking diagnostic index parameters for patients with multiple sclerosis diagnosed for the first time, confirming the score values ​​of the number of MRI enhanced lesions, the total protein content of cerebrospinal fluid, the waiting period from onset to treatment, and the side effects of hormone treatment, summing the scores, calculating the total score, checking the quantitative score table for the risk of insensitivity to hormone treatment, and directly obtaining the probability of insensitivity to hormone treatment.