Hierarchical prediction method and system for risk of malignant tumors related to dermatomyositis

Abstract

The invention relates to a layered prediction method and system for the risk of malignant tumors related to dermatomyositis. The method comprises the following steps: collecting clinical index data of a patient suffering from dermatomyositis, the clinical index data comprising an anti-TIF1-gamma antibody detection result, an interstitial lung disease existence state, a dermatomyosis existence state, an anemia state and a dermatomyositis type; constructing a comprehensive scoring model to generate a risk stratification prediction result according to the clinical index data; the comprehensive scoring model is used for performing binary assignment on each index to obtain a score of a corresponding item; calculating the sum of scores of all indexes of the patient to obtain a total risk score; and layering the risk of the malignant tumor accompanied by the patient based on the total risk score, and outputting a risk layering result. According to the method, the cancer risk of the patient with dermatomyositis can be effectively predicted, and the reliability, convenience and practicability of layered prediction of the cancer risk are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent medical treatment, in particular to a dermatomyositis related malignant tumor risk stratification prediction method and system. BACKGROUND

[0002] Dermatomyositis (DM) is a representative disease of idiopathic inflammatory myopathy (IIM), characterized by skin and / or muscle inflammation, often involving multiple organ systems and seriously threatening patient survival. In addition to interstitial lung disease (ILD), malignant tumor is one of the most common comorbidities of DM, significantly increasing patient mortality. The potential association between DM and malignant tumor was first proposed by Stertz in 1916, and subsequent studies showed that the standardized incidence of DM patients with malignant tumor was between 2.2%-7.7%. Adult DM patients have a 4.66-fold higher risk of developing malignant tumors than the general population, and the risk continues to rise within three years before and after the diagnosis of DM.

[0003] Due to the high heterogeneity of malignant tumor types, the clinical and laboratory manifestations of DM associated with malignant tumor (CRDM) are significantly different. Although previous studies have explored the potential risk factors for DM associated with malignant tumors, there is a lack of large-scale multicenter studies that integrate these factors to assess the predictive value. Notably, the heterogeneity of DM further increases the difficulty of clinical assessment: patients with skin symptoms are more likely to be first diagnosed in dermatology, while those with muscle, joint or respiratory symptoms are more likely to be diagnosed in rheumatology, which may lead to differences in the observed clinical manifestations of DM between the two departments.

[0004] Therefore, a prediction method is needed that can overcome the above problems and be suitable for the risk prediction of malignant tumors in all subgroups of DM. SUMMARY

[0005] The technical problem to be solved by the present application is to provide a dermatomyositis related malignant tumor risk stratification prediction method and system, which can effectively predict the risk of dermatomyositis patients developing malignant tumors and improve the reliability, convenience and practicality of risk stratification prediction.

[0006] The technical solution adopted by the present application to solve its technical problem is to provide a dermatomyositis related malignant tumor risk stratification prediction method, comprising the following steps:

[0007] Collecting clinical indicator data of dermatomyositis patients, the clinical indicator data including anti-TIF1-gamma antibody detection results, interstitial lung disease presence status, skin discoloration presence status, anemia status, and dermatomyositis typing;

[0008] Constructing a comprehensive scoring model to generate risk stratification prediction results according to the clinical indicator data; the comprehensive scoring model is used to:

[0009] The binary assignment is performed on each index to obtain the score of the corresponding index;

[0010] The sum of the scores of all indexes of the patient is calculated to obtain a total risk score;

[0011] Based on the total risk score, the risk of the patient with malignant tumor is stratified, and the risk stratification prediction result is output.

[0012] Further, the binary assignment is performed on each index to obtain the score of the corresponding index, comprising:

[0013] If the anti-TIF1-γ antibody detection result is positive, the corresponding score is 1, otherwise 0;

[0014] If there is interstitial lung disease, the corresponding score is 0, otherwise 1;

[0015] If there is skin heterochromia, the corresponding score is 1, otherwise 0;

[0016] If hemoglobin is reduced, the corresponding score is 1, otherwise 0;

[0017] If the clinical diagnosis is dermatomyositis, the corresponding score is 1, otherwise if the clinical diagnosis is clinical non-muscular dermatomyositis, the corresponding score is 0.

[0018] Further, the comprehensive scoring model is constructed by the following method:

[0019] The dermatomyositis patients are divided into a dermatomyositis patient group with malignant tumor and a dermatomyositis patient group without malignant tumor, and the candidate index data of the two groups of patients are obtained, including demographic characteristics, dermatomyositis classification, blood routine, liver and kidney function, tumor markers, and clinical symptoms;

[0020] The candidate index data of the dermatomyositis patient group with malignant tumor and the dermatomyositis patient group without malignant tumor are compared, and the candidate index with a significant difference greater than a set threshold is extracted as a first associated index;

[0021] The LASSO regression algorithm is used to screen a second associated index with statistical significance from the first associated index;

[0022] The LASSO regression coefficients of the second associated index are standardized, and a comprehensive scoring model is constructed based on the processed second associated index.

[0023] Further, the comparison of the candidate index data of the dermatomyositis patient group with malignant tumor and the dermatomyositis patient group without malignant tumor comprises:

[0024] The continuous variable data in the candidate index data is converted into the form of mean and standard deviation;

[0025] Converting the categorical variable data into percentage form;

[0026] Respectively comparing the converted two groups of continuous variable data and two groups of categorical variable data.

[0027] Further, the respective comparison of the converted two groups of continuous variable data and two groups of categorical variable data comprises:

[0028] Using Fisher's test to compare the converted two groups of continuous variable data;

[0029] Using Mann-Whitney U test to compare the converted two groups of categorical variable data.

[0030] Further, when comparing the converted two groups of continuous variable data, the cutoff value of the continuous variable is calculated by the ROC curve to set the significant difference of the corresponding index.

[0031] Further, the anti-TIF1-γ antibody detection result is obtained by the following method:

[0032] The suspicious sample is preliminarily screened out by the Euroblot linear immunoblotting method;

[0033] The sample with missing myositis-specific antibody (MSA) data or weak positive sample is retested by ELISA method, and if the retest result is greater than the set threshold, it is determined as positive.

[0034] Further, it further comprises the step of formulating individualized screening and treatment recommendations for the prediction results.

[0035] Further, the individualized screening and treatment recommendations for the prediction results comprise:

[0036] For high-risk patients, the following recommendations are generated:

[0037] Systematic screening according to the IMACS "basic" and "intensive" tumor screening standards at the first hospitalization;

[0038] CT examination of neck, chest, abdomen and pelvis every 3 years;

[0039] Gastroscopy and colonoscopy every 3 years;

[0040] Nasopharyngoscopy every 3 years;

[0041] Molybdenum target of breast, transvaginal ultrasound and CA125 detection every year for women;

[0042] For medium-risk patients, the following recommendations are generated:

[0043] Systematic screening according to the IMACS "basic" and "intensive" tumor screening standards at the first hospitalization;

[0044] Patients over 45 years old should have CT examination of neck, chest, abdomen and pelvic every 3 years;

[0045] Gastroscopy and colonoscopy should be performed every 5 years;

[0046] Patients with EB virus infection should have nasopharyngoscopy every 3 years;

[0047] Women over 40 years old should have molybdenum target mammography, transvaginal ultrasound and CA125 detection every year;

[0048] For low-risk patients, the following recommendations are generated:

[0049] When first diagnosed with dermatomyositis, a single system examination is performed according to the "basic" tumor screening standard recommended by IMACS;

[0050] Subsequent follow-up examinations should be determined according to the clinical manifestations and laboratory characteristics of the patient at the time of follow-up, or refer to the general population standard screening program.

[0051] The present application also provides a dermatomyositis-related malignant tumor risk stratification prediction system, comprising:

[0052] A collection module for collecting clinical index data of dermatomyositis patients, the clinical index data including anti-TIF1-γ antibody detection results, interstitial lung disease status, skin heterochromia status, anemia status, and dermatomyositis typing;

[0053] A prediction module for constructing a comprehensive scoring model to generate a risk stratification prediction result according to the clinical index data; the comprehensive scoring model comprises:

[0054] A scoring module for binary assignment of each index to obtain the score of the corresponding index;

[0055] A comprehensive evaluation module for calculating the sum of scores of all indexes of the patient to obtain a total risk score;

[0056] A stratification module for stratifying the risk of the patient with malignant tumors based on the total risk score and outputting a risk stratification prediction result.

[0057] Advantages

[0058] Compared with the prior art, the present application has the following advantages and positive effects: the present application fully considers data diversity and complexity by fusing traditional statistical methods and machine learning technology, and ensures that the model is robust and reliable; the TIP-CA model constructed based on multi-center queue data and containing 5 risk factors has good prediction accuracy for the risk of malignant tumors of dermatomyositis / clinical amyopathic dermatomyositis (DM / CADM) patients; the TIP-CA model constructed by the present application adopts a comprehensive scoring system to predict risk stratification, which is more simple and practical than other schemes, and has a broad application prospect. BRIEF DESCRIPTION OF DRAWINGS

[0059] Figure 1 is a flowchart of the first embodiment of the present application;

[0060] Figure 2 is a flowchart of constructing the TIP-CA model in the first embodiment of the present application;

[0061] Figure 3 is a coefficient path diagram of the regularization parameter λ of the LASSO algorithm in the first embodiment of the present application;

[0062] Figure 4 is a cross-validation diagram of the regularization parameter λ of the LASSO algorithm in the first embodiment of the present application;

[0063] Figure 5 is an ROC curve diagram of the training queue in the first embodiment of the present application;

[0064] Figure 6 is an ROC curve diagram of the verification queue in the first embodiment of the present application;

[0065] Figure 7 is a system structure diagram of the second embodiment of the present application. DETAILED DESCRIPTION

[0066] The present application will be further described below in conjunction with specific embodiments. It should be understood that these embodiments are only used to illustrate the present application and not to limit the scope of the present application. In addition, it should be understood that those skilled in the art can make various modifications or changes to the present application after reading the content taught by the present application, and these equivalent forms also fall within the scope defined by the appended claims of the present application.

[0067] The first embodiment of the present application relates to a dermatomyositis-related malignant tumor risk stratification prediction method, as shown in Figure 1 , comprising the following steps:

[0068] Collecting clinical index data of dermatomyositis patients, specifically including anti-TIF1-γ antibody detection results, interstitial lung disease presence status, skin heterochromia presence status, anemia status, and dermatomyositis typing;

[0069] A comprehensive scoring model TIP-CA is constructed to generate a prediction result of risk stratification according to the clinical index data described above; the comprehensive scoring model is used for:

[0070] Each index is assigned a binary value to obtain the score of the corresponding item;

[0071] The sum of the scores of all items of the patient is calculated to obtain a total risk score;

[0072] Based on the total risk score, the risk of the patient developing a malignant tumor is stratified, and the risk stratification result is output.

[0073] The model scoring factors and scores are shown in Table 1.

[0074] Table 1

[0075]

[0076] More specifically, the total score obtained by the patient after scoring by the TIP-CA model can be divided into low risk (0-1 points), medium risk (2-3 points), and high risk (4-5 points). In some preferred embodiments, individualized screening and treatment recommendations can also be made based on the prediction results.

[0077] The following further illustrates the TIP-CA model and its construction method by combining the data of 300 DM / clinical amyopathic dermatomyositis (CADM) patients admitted to the Department of Dermatology of Ruijin Hospital Affiliated to Shanghai Jiaotong University School of Medicine from 2015 to 2022, and the data of 263 patients admitted to the Department of Rheumatology of Renji Hospital Affiliated to Shanghai Jiaotong University School of Medicine from 2018 to 2023.

[0078] Among them, the inclusion criteria are DM / CADM patients aged 18-80 years old and meeting the 2017 EULAR / ACR criteria or the 2002 Sontheimer criteria, while excluding patients with other connective tissue diseases (including anti-synthetase syndrome) and those with a history of malignant tumor more than 5 years before the onset of DM / CADM. Interstitial lung disease (ILD) is diagnosed based on respiratory symptoms, pulmonary function tests, and high-resolution CT (HRCT), and dysphagia and skin discoloration data are obtained through admission physical examination.

[0079] In the training cohort of Ruijin Hospital, 70 patients had a history of concurrent or past malignant tumors (including 54 patients with non-remitting malignant tumors), and 230 patients were found to have no malignant tumors upon admission screening and had no progression during at least 1 year of follow-up. Considering the differences in treatment status, 16 patients who were confirmed to have complete remission by imaging were excluded, and finally 284 patients were included in the training cohort.

[0080] The Renji Hospital validation cohort contained 263 DM / CADM patients, of whom 53 had a malignancy within 5 years before or after DM diagnosis, and 210 had no malignancy at admission and remained free of progression during 1-5 years of follow-up.

[0081] The TIP-CA model construction method is as shown in Figure 2 The specific steps include:

[0082] The dermatomyositis patients are divided into a dermatomyositis patient group with a malignancy and a dermatomyositis patient group without a malignancy, and the candidate index data of the two groups of patients are obtained, wherein the candidate index data includes demographic characteristics, dermatomyositis classification, blood routine, liver and kidney function, tumor markers, and clinical symptom manifestations.

[0083] The candidate index data of the dermatomyositis patient group with a malignancy and the dermatomyositis patient group without a malignancy are compared, and candidate indexes with a difference significance greater than a set threshold are extracted as first associated indexes.

[0084] The LASSO regression algorithm is used to screen second associated indexes with statistical significance from the first associated indexes.

[0085] The LASSO regression coefficients of the second associated indexes are standardized, and a comprehensive score model is constructed based on the processed second associated indexes.

[0086] More specifically, the collected indexes include:

[0087] Gender, age, DM / CADM clinical classification;

[0088] Blood routine, liver and kidney function, tumor markers, anti-TIF1-γ antibody, and other laboratory parameters;

[0089] Dysphagia, skin discoloration, and other clinical manifestations.

[0090] Among them, the laboratory data are obtained from the detection results during the first hospitalization. The anti-TIF1-γ antibody detection mainly adopts the EUROMUN linear immunoblot method (EUROIMMUN, Germany), and then the missing myositis-specific antibody (MSA) data or weak positive samples are retested by ELISA method (Beijing Medical Biotechnology Institute), and >30 IU / L is determined as positive.

[0091] When performing single factor analysis, first represent the distribution of continuous variables in the data with mean ± standard deviation, present the classification variables in the data in percentage form, then use Fisher's exact test and Mann-Whitney U test to compare the classification variables and continuous variables of the CRDM group and the non-CRDM group, and set the significance level of all statistical tests to P<0.05.

[0092] Single factor analysis found 16 indicators significantly different between CRDM and non-CRDM patients, as shown in Table 2.

[0093] Table 2

[0094]

[0095]

[0096] The average age of the training cohort was 52.7 ± 14.5 years, with the non-CRDM group being 51.6 ± 14.7 years and the CRDM group being 57.5 ± 13.1 years. The proportion of female patients was higher than that of male patients in both groups, and the proportion of male patients in the CRDM group (29.1%) was significantly lower than that in the non-CRDM group (46.3%). Among the 284 patients, 115 (40.5%) had interstitial lung disease (ILD), and the incidence of ILD in the non-CRDM group (45.7%) was significantly higher than that in the CRDM group (18.5%) (P < 0.001), which was consistent with clinical observations, suggesting that DM patients without ILD were more likely to be associated with malignant tumors. In addition, the incidence of anemia in the CRDM group (47.2%) was significantly higher than that in the non-CRDM group (26.9%) (P = 0.005).

[0097] The CRDM group had higher levels of neutrophil / lymphocyte ratio (NLR), erythrocyte sedimentation rate (ESR), C-reactive protein (CRP), neuron-specific enolase (NSE), squamous cell carcinoma antigen (SCC), and D-dimer, while having lower lymphocyte counts and ALT / AST ratios. The positive rate of anti-TIF1-γ antibody was 63.7% (100 cases), of which 64.8% (35 cases) were in the CRDM group and 28.3% (65 cases) were in the non-CRDM group. Dysphagia (35.2% vs. 18.7%, P = 0.016) and skin discoloration (53.7% vs. 18.7%, P < 0.001) were more common in CRDM patients.

[0098] Next, LASSO regression combined with multivariate logistic regression was used to further optimize variable selection. As shown in Figure 4 the cross-validation results and visualization process of the regularization parameter (λ) are shown. The λ value corresponding to the minimum mean square error (MSE) is selected to reduce the independent variables from 16 to 8, including: gender, age, clinical diagnosis (DM / CADM), ILD, low hemoglobin, ESR, anti-TIF1-γ antibody, and skin discoloration. Based on sample size considerations, the 8 factors corresponding to the λ.min value were finally selected for subsequent analysis, of which 5 factors were statistically significant: clinical diagnosis, skin discoloration, ILD, anti-TIF1-γ antibody, and anemia, as shown in Table 3 below.

[0099] Table 3

[0100]

[0101] After calculating the standardized regression coefficient (β) for the five predictive factors, a 1-point scoring system was established: 1 point each for positive anti-TIF1-γ antibodies (T), absence of ILD (I), presence of poikiloderma (P), clinical diagnosis of DM (C), and anemia (A), and 0 point for all other conditions. The model was named TIP-CA based on the initials of each factor. The specific statistical data are shown in Table 4.

[0102] Table 4

[0103]

[0104] Multicollinearity diagnosis showed that the model had no significant collinearity problem, as shown in Table 5 below.

[0105] Table 5

[0106]

[0107] A tolerance greater than 0.1 and a variance inflation factor (VIF) less than 3 indicate that the TIP-CA model does not have multicollinearity among the independent variables. An eigenvalue greater than 0, a conditional index less than 10, and a variance ratio less than 1 indicate that the model does not have multicollinearity. Multicollinearity (also known as collinearity) refers to the phenomenon in linear regression where highly correlated independent variables lead to distorted predictions or difficulty interpreting results. Severe collinearity can inflate standard errors, cause unstable coefficient estimates, and lead to falsely significant results, effectively reducing the model's predictive power.

[0108] The TIP-CA score range is 0-5 points. The CRDM rate in both cohorts increased with the increase of the score, and the 4-5 group was significantly higher than the 0-3 group. Based on this, the 0-1, 2-3 and 4-5 points were defined as low, medium and high risk groups, respectively. The ROC curve showed that the training cohort AUC = 0.809 (sensitivity 0.743, specificity 0.722), the validation cohort AUC = 0.808 (sensitivity 0.823, specificity 0.660), and the optimal cutoff value was 2.5 points. Figure 5 and Figure 6 Considering the simplicity of the model and the heterogeneity of CRDM tumors, the TIP-CA model achieved satisfactory results.

[0109] The TIP-CA model divides DM patients into three risk levels: 0-1 points (low risk), 2-3 points (intermediate risk), and 4-5 points (high risk). High-risk patients should follow the screening protocols for high-risk groups of various tumors. Based on the 2023 International Myositis Assessment and Clinical Studies Group (IMACS) guidelines and the latest advances in oncology research, screening recommendations are formulated for lung, nasopharyngeal, breast, gastrointestinal, and ovarian cancers. Specifically, they include:

[0110] (1) High-risk patients (score 4-5 points):

[0111] Systematic screening according to IMACS "basic" and "intensive" tumor screening standards at the first hospitalization;

[0112] CT examination of the neck, chest, abdomen and pelvis every 3 years;

[0113] Gastroscopy and colonoscopy every 3 years;

[0114] Nasopharyngoscopy every 3 years;

[0115] Mammography, transvaginal ultrasound and CA125 detection every year for women;

[0116] (2) Medium-risk patients (score 2-3 points):

[0117] Systematic screening according to IMACS "basic" and "intensive" tumor screening standards at the first hospitalization;

[0118] CT examination of the neck, chest, abdomen and pelvis every 3 years for patients over 45 years old;

[0119] Gastroscopy and colonoscopy every 5 years;

[0120] Nasopharyngoscopy every 3 years for patients with EB virus infection;

[0121] Mammography, transvaginal ultrasound and CA125 detection every year for women over 40 years old;

[0122] (3) Low-risk patients (score 0-1 points):

[0123] Single systematic examination according to the "basic" tumor screening standard recommended by IMACS at the first diagnosis of dermatomyositis;

[0124] Follow-up examination should be determined according to the clinical manifestations and laboratory characteristics of the patient at the time of follow-up, or refer to the general population standard screening program.

[0125] The second embodiment of the present application relates to a dermatomyositis-related malignant tumor risk stratification prediction system for realizing the system as described above, as shown in Figure 7 , comprising:

[0126] The acquisition module is used for acquiring the clinical index data of the dermatomyositis patient, and specifically includes the anti-TIF1-γ antibody detection result, the interstitial lung disease existence state, the skin heterochromia existence state, the anemia state, and the dermatomyositis type;

[0127] A prediction module is configured to construct a comprehensive scoring model to generate a risk stratification prediction result according to the clinical index data; the comprehensive scoring model comprises:

[0128] A scoring module is configured to perform binary assignment on each index to obtain a score of the corresponding index.

[0129] A comprehensive evaluation module is configured to calculate a sum of scores of all indexes of the patient to obtain a total risk score.

[0130] A stratification module is configured to stratify the risk of the patient suffering from a malignant tumor based on the total risk score and output a risk stratification prediction result.

[0131] In some preferred embodiments, a suggestion generation module is further included to formulate individualized screening and treatment suggestions for the prediction result. Specifically, the suggestion generation module comprises:

[0132] A first suggestion generation submodule is configured to generate the following suggestions for a high-risk patient:

[0133] Perform systematic screening according to the IMACS "basic" and "intensive" tumor screening standards at the first hospitalization;

[0134] Perform neck, chest, abdominal and pelvic CT examination every 3 years;

[0135] Perform gastroscopy and colonoscopy every 3 years;

[0136] Perform nasopharyngoscopy every 3 years;

[0137] Female patients perform breast molybdenum target, transvaginal ultrasound and CA125 detection every year;

[0138] A second suggestion generation submodule is configured to generate the following suggestions for a medium-risk patient:

[0139] Perform systematic screening according to the IMACS "basic" and "intensive" tumor screening standards at the first hospitalization;

[0140] Perform neck, chest, abdominal and pelvic CT examination every 3 years for patients over 45 years old;

[0141] Perform gastroscopy and colonoscopy every 5 years;

[0142] Perform nasopharyngoscopy every 3 years for patients with EB virus infection;

[0143] Female patients over 40 years old perform breast molybdenum target, transvaginal ultrasound and CA125 detection every year;

[0144] A third suggestion generation submodule is configured to generate the following suggestions for a low-risk patient:

[0145] A single system check was performed at initial diagnosis of dermatomyositis according to the "base" oncology screening criteria recommended by IMACS;

[0146] Subsequent follow-up examinations should be determined based on the clinical manifestations and laboratory characteristics of the patient at the time of follow-up, or refer to the general population standard screening program.

Claims

1. A method for predicting the risk stratification of dermatomyositis-related malignant tumors, characterized in that: The following steps are involved: Collecting clinical indicator data of dermatomyositis patients, including anti-TIF1-γ antibody test results, presence of interstitial lung disease, presence of poikiloderma, anemia status, and dermatomyositis classification; A comprehensive scoring model is constructed to generate risk stratification prediction results based on the clinical indicator data; the comprehensive scoring model is used to: Perform binary assignment on each indicator to obtain the score of the corresponding item; Calculate the sum of the scores of all indicators of the patient to obtain the total risk score; The patient's risk of developing malignant tumors is stratified based on the total risk score, and the risk stratification prediction results are output.

2. The method according to claim 1, characterized in that The binary assignment of each indicator to obtain the score of the corresponding item includes: If the anti-TIF1-γ antibody test result is positive, the corresponding score is 1, otherwise it is 0; If interstitial lung disease is present, the corresponding score is 0, otherwise it is 1; If poikiloderma is present, the corresponding score is 1, otherwise 0; If hemoglobin is decreased, the corresponding score is 1, otherwise it is 0; If the clinical diagnosis was dermatomyositis, the corresponding score was 1; otherwise, if the clinical diagnosis was clinically amyopathic dermatomyositis, the corresponding score was 0.

3. The method according to claim 1, characterized in that The comprehensive scoring model is constructed by the following method: Patients with dermatomyositis were divided into a group with dermatomyositis associated with malignant tumors and a group without dermatomyositis associated with malignant tumors. Data on candidate indicators were obtained for the two groups, including demographic characteristics, dermatomyositis classification, blood routine, liver and kidney function, tumor markers, myositis-specific antibodies, and clinical symptoms. Compare the candidate index data of the dermatomyositis patients with malignant tumors and the dermatomyositis patients without malignant tumors, and extract the candidate index with a significant difference greater than the set threshold as the first correlation index; The LASSO regression algorithm is used to screen out statistically significant second correlation indicators from the first correlation indicators; the LASSO regression coefficient of the second correlation indicator is standardized, and a comprehensive scoring model is constructed based on the processed second correlation indicator.

4. The method according to claim 3, characterized in that The comparison of the candidate indicator data between the dermatomyositis patient group with malignant tumor and the dermatomyositis patient group without malignant tumor is performed by comparing the continuous variable data and the categorical variable data respectively.

5. The method according to claim 4, characterized in that The continuous variable data and categorical variable data are compared separately, including: If the variable was a continuous variable, Fisher's test was used to compare the two groups of continuous variable data; If the variable is categorical, the Mann-Whitney U test was used to compare the two groups of categorical variable data.

6. The method according to claim 1, characterized in that The anti-TIF1-γ antibody detection result is obtained by the following method: Suspicious samples were initially screened using the O'Meng line immunoblotting method; The ELISA method was used to retest samples lacking myositis-specific antibodies or weakly positive samples. If the retest result was greater than the set threshold, it was judged as positive.

7. The method according to claim 1, characterized in that It also includes steps to develop individualized screening and treatment recommendations based on the predicted results.

8. The method according to claim 7, characterized in that The personalized screening and treatment recommendations based on the predicted results include: For high-risk patients, the following recommendations are generated: Systematic screening was performed according to the IMACS "basic" and "enhanced" cancer screening criteria during the first hospitalization; CT scan of the neck, chest, abdomen, and pelvis every 3 years; Gastroscopy and colonoscopy every 3 years; Nasopharyngoscopy every 3 years; Women should undergo annual mammography, transvaginal ultrasound, and CA125 testing; For moderate-risk patients, the following recommendations are generated: Systematic screening was performed according to the IMACS "basic" and "enhanced" cancer screening criteria during the first hospitalization; Patients over 45 years of age should undergo CT scans of the neck, chest, abdomen, and pelvis every 3 years; Gastroscopy and colonoscopy every 5 years; Patients with EBV infection should undergo nasopharyngeal endoscopy every 3 years; Women over 40 years old should undergo annual mammography, transvaginal ultrasound, and CA125 testing; For low-risk patients, the following recommendations are generated: When dermatomyositis is first diagnosed, a single systematic examination should be performed according to the "basic" tumor screening criteria recommended by IMACS; Subsequent follow-up examinations should be determined based on the patient's clinical manifestations and laboratory characteristics at the time of follow-up, or refer to the standard screening program for the general population.

9. A dermatomyositis-related malignant tumor risk stratification prediction system, characterized in that: include: An acquisition module is used to collect clinical indicator data of dermatomyositis patients, wherein the clinical indicator data include anti-TIF1-γ antibody test results, the presence of interstitial lung disease, the presence of poikiloderma, anemia status, and dermatomyositis classification; A prediction module, configured to construct a comprehensive scoring model to generate risk stratification prediction results based on the clinical indicator data; The comprehensive scoring model includes: The scoring module is used to assign binary values ​​to each indicator and obtain the score of the corresponding item; Comprehensive evaluation module, used to calculate the sum of scores of all indicators of the patient to obtain the total risk score; The stratification module is used to stratify the patient's risk of developing concomitant malignant tumors based on the total risk score and output the risk stratification prediction results.

Images