A prediction model for diagnosis of infectious mononucleosis in children and a method for constructing the same

Abstract

The application discloses a prediction model for diagnosing infectious mononucleosis in children and a construction method thereof. The application adopts single factor analysis and logistic stepwise regression to screen independent influencing factors by using confirmed IM inpatient cases and suspected IM but EBV-DNA negative child cases, and adopts multiple factor regression to screen six independent influencing factors of age, ALB, GLB, PLT, LYMPH# and LYMPH% to construct a diagnosis model. Researches show that in the confirmed diagnosis scene, the multi-index combined model with blood routine and liver function has excellent diagnosis efficiency (AUC=0.986), realizes the ideal balance of high sensitivity and high specificity, provides a hierarchical tool for the precise diagnosis of IM in different clinical scenes, and is suitable for the precise diagnosis of inpatient children.

Description

Technical Field

[0001] This invention belongs to the field of biomedical technology and relates to a predictive model for the diagnosis of infectious mononucleosis in children and its construction method. Background Technology

[0002] Epstein-Barr virus (EBV) is a human lymphocyte-tropic gamma herpesvirus with an extremely high infection rate in the population. Globally, the seropositivity rate in adults exceeds 90%, and the infection rate is high among children as young as age 10 in my country. EBV infection can lead to a variety of diseases, mainly falling into three categories: The first category consists of non-malignant lymphoproliferative disorders, including infectious mononucleosis (IM), chronic active Epstein-Barr virus infection (CAEBV), and Epstein-Barr virus-related hemophagocytic lymphohistiocytosis (EBV-HLH). The latter two have poor prognoses, with mortality rates exceeding 26.2% and 50%, respectively. The second category includes autoimmune diseases such as systemic lupus erythematosus and multiple sclerosis. The third category is malignant tumors; EBV was the first confirmed carcinogenic virus and is closely associated with nasopharyngeal carcinoma, lymphoma, and gastric cancer. The complex clinical manifestations of EBV infection suggest that its pathogenesis involves multiple factors such as immune dysregulation and metabolic changes, and the relevant mechanisms need further in-depth research.

[0003] Infectious mononucleosis (IM) is the main clinical type of EBV infection in children. IM is an acute clinical syndrome caused by primary EBV infection, with typical clinical manifestations including fever, pharyngitis, and cervical lymphadenopathy (the clinical "triad"), which may be accompanied by hepatosplenomegaly. Peripheral blood is characterized by an increased proportion of lymphocytes and atypical lymphocytes (i.e., reactive lymphocytes). IM is a benign, self-limiting disease with a mostly good prognosis, but complications such as upper airway obstruction, encephalitis, hemolytic anemia, and thrombocytopenic purpura can occur. In rare cases, serious complications such as hemophagocytic lymphohistiocytosis (HLH) and splenic rupture may occur, requiring timely identification and intervention. Studies have shown that the epidemiological characteristics of IM in my country differ significantly from those in Western countries. The diagnosis of IM in Chinese children needs to be combined with the country's own epidemiological characteristics and cannot simply apply Western experience.

[0004] The current status of IM diagnosis and treatment: the dilemma of "dual differentiation" between specialist and primary care. The current status of IM diagnosis and treatment in my country shows a significant "dual differentiation" characteristic. There are huge differences in diagnostic accuracy, diagnosis time and treatment cost among different levels of medical institutions: (1) Difference in diagnostic accuracy: In primary care institutions, due to the lack of specific testing methods, the initial diagnosis accuracy of IM is generally less than 50%. Most cases are misdiagnosed as bacterial tonsillitis or common upper respiratory tract infection in the early stage. This misdiagnosis not only leads to unnecessary use of antibiotics (such as amoxicillin or ampicillin), but can also induce drug rash risk of up to 90%. In tertiary specialist hospitals, thanks to the complete laboratory testing conditions, the diagnostic accuracy can be increased to 80%-90%, mainly relying on serological testing (such as EBV specific antibodies) and the standardized count of atypical lymphocytes (reactive lymphocytes) in blood routine. (2) Differences in diagnosis time: From the onset of illness to diagnosis, primary healthcare institutions usually require 7-14 days, mainly because initial diagnosis often involves empirical anti-infective treatment, and referral or specialized testing is only performed after the treatment is ineffective. Specialized hospitals can shorten this to 3-7 days, relying on faster EBV-specific antibodies (such as VCA-IgM) and nucleic acid testing (PCR) technology. However, serological testing still requires one week after the onset of illness to reach reliable sensitivity, resulting in some cases still requiring observation and waiting. (3) Differences in treatment costs: The diagnostic cost of IM mainly comes from laboratory testing items.

[0005] The 2021 Expert Consensus on Diagnosis and Treatment Principles of EBV Infection-Related Diseases in Children clarified the diagnostic criteria for IM. The diagnostic criteria recommended in the 2021 consensus have good application value in specialized hospitals, but face the following challenges in primary hospitals: (1) Insufficient accessibility of specific antibody testing: EBV specific antibody and nucleic acid testing are often not carried out in primary hospitals or need to be sent out, with long testing cycles and high costs, resulting in a heavy economic burden on patients. This makes it difficult for primary doctors to obtain laboratory evidence of primary EBV infection and to confirm suspected cases. (2) The diagnostic efficacy of blood routine indicators is scenario-dependent: Although the non-specific laboratory tests recommended in the consensus (atypical lymphocyte ratio ≥10%, lymphocyte ratio >50%) can be carried out at the primary level, their diagnostic efficacy may be significantly scenario-dependent. In the early screening scenario in outpatient clinics, most children are in the early stage of the disease, and lymphocytes and atypical lymphocytes have not yet reached their peak. At this time, the above indicators may have extremely low sensitivity, resulting in a large number of IM patients being missed. In hospitalized diagnosis scenarios, children with long disease courses and typical symptoms may show good diagnostic value for the above indicators. This potential difference has not yet been fully studied and verified. (3) The practical needs of primary care: Due to the atypical early symptoms of IM, most children are first diagnosed in primary healthcare institutions. In cases where serological or PCR test results cannot be obtained immediately, primary care physicians urgently need a diagnostic tool based on routine examinations, which can be implemented quickly and is suitable for early screening scenarios to help identify suspected IM children early and guide reasonable triage and referral. At the same time, for hospitalized or more seriously ill children, there is also a diagnostic tool with excellent diagnostic efficacy based on routine indicators to reduce reliance on expensive specific tests. Currently, there is an urgent need for an IM diagnostic tool that can be promoted and applied in primary hospitals, is based on routine examinations, is economical and convenient, and is applicable to different clinical scenarios. Summary of the Invention

[0006] The purpose of this invention is to address the above-mentioned problems by providing a predictive model for the diagnosis of infectious mononucleosis in children and a method for constructing it.

[0007] To achieve its objective, the present invention employs the following technical solution: A first aspect of the present invention provides biomarkers for diagnosing and / or predicting infectious mononucleosis in children, said biomarkers including age, ALB, GLB, PLT, and LYMPH.

[0008] A second aspect of the invention provides the use of the above-described biomarkers for diagnosing and / or predicting childhood infectious mononucleosis in the preparation of reagents or kits for diagnosing and / or predicting childhood infectious mononucleosis.

[0009] A third aspect of the present invention provides a method for constructing a predictive model for the diagnosis of infectious mononucleosis (IM) in children, comprising the following steps: S1. Data collection and preprocessing: Inpatients diagnosed with IM were collected as the case group, and children with clinical manifestations suspected of having IM but negative EBV-DNA test results were randomly selected as the control group. Clinical data were selected according to the inclusion and exclusion criteria and preprocessed. The clinical data included: (1) Basic information: age and gender; (2) Routine laboratory indicators: blood routine: white blood cell count and classification, platelet count and related parameters and red blood cell related parameters, liver function: ALT, AST, TP, ALB and GLB, where ALT is alanine aminotransferase, AST is aspartate aminotransferase, TP is total protein, ALB is albumin and GLB is globulin; S2. Screening independent influencing factors: First, univariate analysis was used to compare the differences of various indicators between the case group and the control group to screen candidate indicators that may be related to the diagnosis of IM; then, multivariate binary logistic stepwise regression was used to screen independent influencing factors of IM. S3. Multivariate Logistic Regression Prediction Model: A simple random sampling method is used to divide all research subjects into training and validation sets. Based on independent influencing factors, the regression coefficients and corresponding scores of each variable are determined, and a prediction model is constructed. The calculation formula for the prediction model is as follows: Logit(P) = 0.623 × age (years) - 0.301 × ALB (g / L) + 0.308 × GLB (g / L) - 9.195 × PLT (10 9 / L) + 0.763 × LYMPH# (10 9 / L) + 0.066 × LYMPH%-5.292; Multivariate logistic regression analysis showed that age, ALB, GLB, PLT, LYMPH#, and LYMPH% were independent risk factors for IM diagnosis, while ALB and PLT were protective factors. ALB is albumin, GLB is globulin, PLT is platelet count, and LYMPH# is the absolute lymphocyte count, representing the number of these cells in units of 10-1. 9 g / L, LYMPH% is the percentage of lymphocytes, indicating the proportion / composition of this type of cell in the five differential serum tests, expressed as a percentage. S4. Internal validation and evaluation of model performance: After building the model based on the training set, the model is evaluated in the training set and validation set from three dimensions: discrimination (ROC curve), calibration (calibration curve) and clinical applicability (decision curve). The stability of the model is judged by comparing the consistency of the results of the two sets.

[0010] In the aforementioned construction method, the prediction probability P is:

[0011] If P > 0.5, IM is diagnosed; if P ≤ 0.5, IM is basically ruled out.

[0012] In the calculation formula of the prediction model in step S3: Age: OR 1.838, 95% CI: 1.482–2.280, P < 0.001; ALB: OR value 0.748, 95% CI: 0.644-0.868, P<0.001; GLB: OR value 1.333, 95% CI: 1.185-1.500, P<0.001; PLT: OR value 0.993, 95% CI: 0.987-0.999, P=0.004; LYMPH#: OR value 2.075, 95% CI: 1.511-2.849, P<0.001; LYMPH%: OR value 1.069, 95% CI: 1.037-1.103, P<0.001.

[0013] A fourth aspect of the present invention provides a predictive model for the diagnosis of infectious mononucleosis (IM) in children, obtained by the construction method according to any one of the preceding claims.

[0014] A fifth aspect of the present invention provides a predictive system for the diagnosis of infectious mononucleosis (IM) in children, the system comprising an input module, a calculation module, and an output module; (1) Input module: used to transmit the following information of the subject to the calculation module: age, ALB, GLB, PLT, LYMPH#, LYMPH% (2) Calculation module: It has a built-in predictive model for the diagnosis of infectious mononucleosis (IM) in children. The calculation formula for the prediction model is as follows: Logit(P) = 0.623 × age (years) - 0.301 × ALB (g / L) + 0.308 × GLB (g / L) - 9.195 × PLT (10 9 / L) + 0.763 × LYMPH# (10 9 / L) + 0.066 × LYMPH%-5.292; Among them, ALB is albumin, GLB is globulin, PLT is platelet count, LYMPH# is absolute lymphocyte count, and LYMPH% is percentage of lymphocytes. (3) Output module: Used to output the IM risk probability value obtained by the calculation module:

[0015] If P > 0.5, IM is diagnosed; if P ≤ 0.5, IM is basically ruled out.

[0016] Preferably, in the prediction system, the input module and the calculation module are connected via a wired and / or wireless means.

[0017] Preferably, in the prediction system, the computing module includes a computer host, a central processing unit, or a network server; the output module is a display, a printer, or an audio output device.

[0018] The beneficial effects of this invention are as follows: This invention uses univariate analysis and stepwise logistic regression to screen independent influencing factors in hospitalized children diagnosed with intraepithelial neoplasia (IM) and suspected IM patients with negative EBV-DNA. Multivariate regression identifies six independent influencing factors: age, ALB, GLB, PLT, LYMPH#, and LYMPH% of the IM patient. The constructed joint model achieved an AUC of 0.986 (95% CI: 0.973-0.995) with a sensitivity of 93.7% and a specificity of 98.7% in the training set and 0.986 (95% CI: 0.956-1.000) with a sensitivity of 90.0% and a specificity of 90.6% in the validation set. Calibration curves show a good model fit, and decision curve analysis indicates a significant net clinical benefit. This invention demonstrates that in hospitalized diagnosis scenarios, the multi-indicator joint model incorporating liver function exhibits excellent diagnostic efficacy (AUC=0.986), achieving an ideal balance between high sensitivity and high specificity. It provides a stratified tool for accurate diagnosis of IM in different clinical scenarios and is suitable for the accurate diagnosis of hospitalized children.

[0019] (1) In hospital settings, the combined diagnostic efficacy of multiple indicators is superior: The fundamental reason lies in the temporal evolution of various indicators after EBV infection. Significant increases in lymphocytes and atypical lymphocytes (reactive lymphocytes) usually occur 5-7 days after the onset of the disease, and liver function changes such as increased globulin and decreased albumin gradually appear in the middle and late stages of the disease. Outpatients are mostly in the early stages of the disease, and these indicators have not yet been fully expressed, so it is difficult to achieve high-precision diagnosis based solely on blood routine tests; while hospitalized children have a longer disease course, and typical manifestations such as increased lymphocytes and increased globulin have been fully manifested. At this time, including liver function indicators can comprehensively capture the multidimensional impact of EBV infection on the immune system, liver, and hematopoietic system, thereby achieving a leapfrog improvement in diagnostic efficacy.

[0020] (2) Advantages of the predictive model of this invention in its applicability to primary care: Primary hospitals in my country generally lack EBV-specific testing, but routine blood tests and liver function tests are widely available. The two models constructed in this study require indicators that are routine items at the primary care level, with a total cost of less than 100 yuan, which is a light burden on patients. Doctors can quickly obtain the predicted probability by inputting their age and test values, realizing "same-day consultation and same-day judgment", shortening the diagnosis time from the routine 7-14 days at the primary care level to within 1 day.

[0021] (3) The predictive model of this invention can effectively reduce misdiagnosis: Currently, the initial diagnosis accuracy of IM in primary hospitals is less than 50%, and the proportion of misdiagnosis as bacterial tonsillitis is high, leading to the neglect of serious complications such as antibiotic abuse (incidence of amoxicillin-related drug eruption >90%) and splenic rupture (mortality rate 30%). After the predictive model of this invention is widely applied, the accuracy of inpatient IM diagnosis can be increased to over 90%, while the misdiagnosis rate can be controlled within 10%, which has significant clinical benefits.

[0022] (4) Good prospects for promotion and application: All indicators are routine testing items at the grassroots level, and no additional testing costs are required, so the prospects for promotion and application at the grassroots level are good.

[0023] This invention provides hospitals (especially primary hospitals) with a set of IM diagnostic tools that are based on routine examinations, are economical and convenient, and are applicable to different clinical scenarios. This can improve diagnostic accuracy, ultimately optimize the allocation of medical resources, and improve the quality of diagnosis and treatment of EBV infection-related diseases (especially IM) in children. Attached Figure Description

[0024] Figure 1 This is the ROC curve of the IM diagnostic prediction model of the present invention.

[0025] Figure 2 This is the calibration curve of the IM diagnostic prediction model of the present invention.

[0026] Figure 3 This is the decision curve analysis of the IM diagnostic prediction model of the present invention. Detailed Implementation

[0027] The present invention will be further described below with reference to embodiments, but these embodiments are not intended to limit the scope of the invention.

[0028] Unless otherwise specified, the experimental methods described in the following examples are conventional methods.

[0029] Example 1: Early Screening Scenario in Outpatient Clinics: EB Virus Infection Screening Model Based on Routine Blood Parameters 1. Research Subjects This part of the study focuses on early screening in outpatient settings, using children suspected of having EBV infection who visited the outpatient department of the Affiliated Nanchuan Hospital of Chongqing Medical University between April 17, 2023 and May 2, 2024 as the study subjects. Cases were screened according to inclusion and exclusion criteria, and patients who met the inclusion criteria were identified as study subjects.

[0030] 1.1 Inclusion Criteria (1) The patient seeks medical attention due to suspected EBV infection with clinical manifestations such as fever, pharyngitis, and cervical lymphadenopathy; (2) Complete blood routine and whole blood EBV-DNA PCR tests are performed; (3) Complete clinical data are available.

[0031] The diagnostic criteria for IM refer to the 2021 Expert Consensus on Diagnosis and Treatment Principles of EBV Infection-Related Diseases in Children, which requires meeting any 3 of the following clinical manifestations and any laboratory evidence of primary EBV infection.

[0032] I. Clinical manifestations: ① Fever; ② Pharyngitis; ③ Cervical lymphadenopathy; ④ Hepatomegaly; ⑤ Splenomegaly; ⑥ Eyelid edema.

[0033] II. Laboratory evidence of primary EBV infection: ① Positive for anti-EBV-CA-IgM and anti-EBV-CA-IgG antibodies, and negative for anti-EBV-NA-IgG; ② Positive for a single anti-EBV-CA-IgG antibody, and the EBV-CA-IgG is a low-affinity antibody.

[0034] III. Non-specific laboratory tests: ① Peripheral blood atypical lymphocyte ratio > 0.10; ② Peripheral blood lymphocyte ratio > 0.50 or absolute lymphocyte count > 5.0 × 10⁻⁶ in children over 6 years of age. 9 / L.

[0035] 1.2 Exclusion Criteria (1) Those with missing clinical data or laboratory tests; (2) Those with other pathogen infections; (3) Those with immune system diseases, tumors, or other diseases that seriously affect their own immunity; (4) Those who are pregnant or still in the postpartum period.

[0036] 1.3 Medical Ethics and Privacy Protection This study has been approved by the Medical Ethics Committee of the Affiliated Nanchuan Hospital of Chongqing Medical University (Ethics Approval No.: Cx202322). The study was conducted strictly in accordance with the provisions of the Declaration of Helsinki.

[0037] 2. Research Methods 2.1 Data Acquisition and Processing Clinical data of suspected EBV-infected patients who visited the outpatient clinic of the Affiliated Nanchuan Hospital of Chongqing Medical University from April 17, 2023 to May 2, 2024 were retrospectively collected, including: (1) basic information: age and gender; (2) laboratory tests: whole blood EBV-DNA PCR test and first blood routine test results completed at the time of visit or within one week thereafter.

[0038] A total of 2,044 suspected EBV-infected patients were collected. 535 adults (aged ≥18 years) were excluded, as were 155 patients for whom blood routine test parameters could not be collected and 32 patients for whom data were missing. Finally, the clinical data of 1,322 children were included for analysis.

[0039] The following blood cell ratios were calculated based on complete blood count results: neutrophil-to-lymphocyte ratio (NLR), platelet-to-lymphocyte ratio (PLR), monocyte-to-lymphocyte ratio (MLR), derived neutrophil-to-lymphocyte ratio (dNLR), neutrophil-to-monocyte-plus-lymphocyte ratio (NMLR), systemic inflammation response index (SIRI), and systemic inflammatory index (SII). High fluorescence intensity lymphocytes (HFLCs) were included in the analysis as an instrumental indicator for the detection of atypical lymphocytes (reactive lymphocytes).

[0040] 2.2 Statistical Analysis Statistical analysis was performed using SPSS 26.0 software, and statistical graphs were generated using R language version 4.4.1. The Shapiro-Wilk test was used to test the normality of continuous variables. Normally distributed continuous data were expressed as mean ± standard deviation. Comparisons between two groups were performed using the t-test, and comparisons among multiple groups were performed using one-way ANOVA and the SNK-q test. Non-normally distributed continuous data were expressed as median (P < 0.05). 25 P 75 The numbers indicate that comparisons between two groups were performed using the Mann-Whitney U test. Count data are expressed as percentages (%), and comparisons between groups were performed using the chi-square test (χ²). 2 The diagnostic test or Fisher's exact test was used. Binary logistic stepwise regression analysis was employed to screen independent influencing factors of EBV infection and construct a diagnostic model. Receiver operating characteristic (ROC) curves were used to evaluate the diagnostic efficacy of each indicator and the combined model, calculating the area under the curve (AUC), sensitivity, and specificity. A p-value < 0.05 was considered statistically significant.

[0041] 3. Research Results 3.1 Comparison of General Data Of the 1322 children ultimately included in the analysis, 366 (27.69%) were EBV-DNA positive and 956 (72.31%) were negative. There was no statistically significant difference in gender composition between the two groups (χ²). 2 =0.006, P=0.938), but the difference in age distribution was statistically significant (χ² = 0.006, P=0.938). 2 =384.062, P<0.001). Stratified by age, the EBV-DNA positivity rate increased with age: 23.00% (267 / 1161) for 0-5 years, 42.86% (102 / 238) for 6-10 years, 51.81% (43 / 83) for 11-15 years, 75.61% (31 / 41) for 16-20 years, and 71.59% (373 / 521) for >20 years. This trend is consistent with the characteristics of EBV infection.

[0042] 3.2 Validation of the diagnostic efficacy of the 2021 expert consensus Using EBV-DNA PCR detection results as the gold standard, a four-fold table was used to analyze the diagnostic efficacy of three lymphocyte-related markers recommended in the 2021 expert consensus in adolescents aged 6-18 years suspected of having intraepithelial neoplasia (IM). HFLC, as an instrumental indicator for atypical lymphocytes (reactive lymphocytes), was analyzed using an HFLC ratio ≥0.10 instead of an atypical lymphocyte (reactive lymphocyte) ratio ≥0.10.

[0043] The diagnostic efficacy of the three indicators is as follows: Lymphocyte percentage > 0.50: Among 192 children aged 6-18 years, there were 13 true positives, 4 false positives, 106 true negatives, and 69 false negatives. The sensitivity was 15.85% (13 / 82), and the specificity was 96.36% (106 / 110). χ² 2 =8.689, P=0.003, the difference was statistically significant, the positive predictive value was 76.47%, and the negative predictive value was 60.57%.

[0044] Absolute lymphocyte count > 5.0 × 10⁻⁶ 9 / L: 10 true positives, 0 false positives, 110 true negatives, and 72 false negatives. Sensitivity was 12.20% (10 / 82), specificity was 100% (110 / 110), χ² 2 =11.789, P=0.001, positive predictive value 100%, negative predictive value 60.44%.

[0045] HFLC proportion ≥0.10: Among all 1322 children, there was 1 true positive, 1 false positive, 955 true negatives, and 365 false negatives. The sensitivity was 0.27% (1 / 366), the specificity was 99.90% (955 / 956), and Fisher's exact test P=0.477, indicating no statistically significant difference.

[0046] The above results indicate that all three indicators exhibit high specificity (96.36%-100%) but low sensitivity (0.27%-15.85%), suggesting that the diagnostic criteria for lymphocyte-related blood routine tests in the 2021 expert consensus have an excessively high rate of missed diagnoses when screening febrile children for IM in outpatient settings.

[0047] 3.3 Comparison of blood routine indicators between EBV-infected and non-infected groups The differences in various blood routine indicators between the EBV-infected group and the negative group were compared. The results showed that there were no statistically significant differences between the two groups in terms of gender, RBC, MCHC#, RDW_CV, P_LCR, LYMPH#, EO#, BASO#, NEUT%, EO%, PLR, NLR, dNLR, NMLR, SIRI, and SII (P>0.05).

[0048] The indicators that showed statistically significant differences between the two groups included (all P < 0.05): Red blood cell-related indicators: The EBV-infected group had higher levels of HGB (122.34±11.17 vs. 120.29±11.14), HCT (36.15±3.03 vs. 35.41±2.98), MCV (79.80±5.56 vs. 78.72±4.68), MCH (27.02±2.41 vs. 26.74±1.90), and RDW_SD (37.55±2.86 vs. 37.19±2.68) than the non-infected group, indicating that the number and size of red blood cells were increased in the infected group.

[0049] Platelet-related indicators: PLT (260.87±79.95 vs. 292.92±100.86) and PCT (0.25±0.07 vs. 0.28±0.09) were significantly lower in the infected group than in the non-infected group, while PDW (10.84±2.36 vs. 10.28±2.13) and MPV (9.86±0.99 vs. 9.74±0.99) were significantly higher in the infected group than in the non-infected group, suggesting that the number of platelets was reduced but the platelet volume was increased in the infected group.

[0050] White blood cell-related parameters: WBC (10.07±5.14 vs. 10.74±5.35), NEUT# (5.96±4.40 vs. 6.66±4.59), and MONO# (0.79±0.47 vs. 0.95±0.57) were significantly lower in the infected group than in the non-infected group, while LYMPH% (33.58±19.39 vs. 31.15±17.58), BASO% (0.29±0.21 vs. 0.26±0.16), HFLC (0.13±0.23 vs. 0.09±0.08), and HFLC% (1.23±1.51 vs. 0.96±1.09) were significantly higher in the infected group than in the non-infected group, and MONO% (8.26±3.29 vs. 0.95±0.57) was significantly higher. The 9.20±3.54 (9.20±3.54) was significantly lower than that in the non-infected group, indicating a decrease in neutrophils and monocytes in the infected group. Although the total number of lymphocytes did not change significantly, it showed an increasing trend. In particular, the absolute and relative numbers of atypical lymphocytes (reactive lymphocytes) reflected by HFLC were significantly increased, which was statistically significant between the two groups.

[0051] Derived inflammatory markers: The ratio of monocytes to lymphocytes (MLR) in the infected group (0.36±0.28 vs. 0.40±0.28) was significantly lower than that in the non-infected group (P=0.009), while other derived markers showed no significant differences.

[0052] 3.4 Comparison of blood routine indicators between EBV-infected and non-infected groups The differences in various blood routine indicators between the EBV-infected group and the negative group were compared. The results showed that there were no statistically significant differences between the two groups in terms of gender, RBC, MCHC, RDW_CV, P_LCR, LYMPH#, EO#, BASO#, NEUT%, EO%, PLR, NLR, dNLR, NMLR, SIRI, and SII (P>0.05).

[0053] The indicators that showed statistically significant differences between the two groups included (all P < 0.05): Red blood cell-related indicators: The EBV-infected group had higher levels of HGB (122.34±11.17 vs. 120.29±11.14), HCT (36.15±3.03 vs. 35.41±2.98), MCV (79.80±5.56 vs. 78.72±4.68), MCH (27.02±2.41 vs. 26.74±1.90), and RDW_SD (37.55±2.86 vs. 37.19±2.68) than the non-infected group, indicating that the number and size of red blood cells were increased in the infected group.

[0054] Platelet-related indicators: PLT (260.87±79.95 vs. 292.92±100.86) and PCT (0.25±0.07 vs. 0.28±0.09) were significantly lower in the infected group than in the non-infected group, while PDW (10.84±2.36 vs. 10.28±2.13) and MPV (9.86±0.99 vs. 9.74±0.99) were significantly higher in the infected group than in the non-infected group, suggesting that the number of platelets was reduced but the platelet volume was increased in the infected group.

[0055] White blood cell-related parameters: WBC (10.07±5.14 vs. 10.74±5.35), NEUT# (5.96±4.40 vs. 6.66±4.59), and MONO# (0.79±0.47 vs. 0.95±0.57) were significantly lower in the infected group than in the non-infected group, while LYMPH% (33.58±19.39 vs. 31.15±17.58), BASO% (0.29±0.21 vs. 0.26±0.16), HFLC (0.13±0.23 vs. 0.09±0.08), and HFLC% (1.23±1.51 vs. 0.96±1.09) were significantly higher in the infected group than in the non-infected group, and MONO% (8.26±3.29 vs. 0.95±0.57) was significantly higher. The 9.20±3.54 (9.20±3.54) was significantly lower than that in the non-infected group, indicating a decrease in neutrophils and monocytes in the infected group. Although there was no statistical difference in the total number of lymphocytes, the absolute and relative number of atypical lymphocytes (reactive lymphocytes) reflected by HFLC were significantly increased.

[0056] Derived inflammatory markers: The ratio of monocytes to lymphocytes (MLR) in the infected group (0.36±0.28 vs. 0.40±0.28) was significantly lower than that in the non-infected group (P=0.009), while other derived markers showed no significant differences.

[0057] The above results indicate that hematological changes in IM patients are not limited to the traditionally recognized lymphocyte lineage, but also involve multiple lineages such as neutrophils, monocytes, platelets, and erythrocytes. Moreover, compared to the qualitative and quantitative changes in the lymphocyte lineage, the changes in neutrophils, monocytes, platelets, and erythrocytes are more pronounced in outpatient IM patients. These findings not only provide a basis for constructing a multi-indicator combined diagnostic model, but also support the view that lymphocyte changes often occur in the middle and late stages of the disease.

[0058] 3.5 Multivariate Logistic Regression Analysis and Diagnostic Model Construction Using EBV-DNA positivity as the dependent variable (positive = 1, negative = 0), a multivariate logistic regression analysis was performed, incorporating all relevant variables. Stepwise regression was used to screen variables, and the results are shown in Table 1.

[0059] Table 1

[0060] Table 1 shows that age and HFLC are independent risk factors for EBV-DNA positivity, while MCHC, PCT, and MONO% are protective factors. Specifically: For every year of age increase, the risk of EBV-DNA positivity increased by 22.3% (OR=1.223, 95%CI: 1.168-1.281, P<0.001). For every 1×10 of HFLC 9 / L, the positive risk increased 19.25 times (OR=20.251, 95%CI: 6.152-66.633, P<0.001). For every 1 g / L increase in MCHC, the risk of a positive result decreased by 1.3% (OR=0.987, 95% CI: 0.977-0.997, P=0.015). For every 1% increase in PCT, the risk of a positive result decreased by 96.0% (OR=0.040, 95%CI: 0.007-0.218, P<0.001). For every 1% increase in MONO%, the risk of a positive result decreased by 5.5% (OR=0.945, 95%CI: 0.907-0.984, P=0.006).

[0061] Based on the above five indicators, a Logistic regression diagnostic model is constructed: Logit(P) = 0.201 × Age (years) - 0.013 × MCHC (g / L) - 3.220 × PCT (%) - 0.057 × MONO (%) + 3.008 × HFLC (10 9 / L) + 3.685

[0062] 3.6 Performance Evaluation of the Diagnostic Model The diagnostic efficacy of the combined model was analyzed using ROC curves. The results (Table 2) show that the AUC of the model's predicted probabilities was 0.708 (95% CI: 0.675-0.741), the sensitivity corresponding to the optimal cutoff value was 71.1%, the specificity was 58.7%, and the Youden index was 0.23. Compared with the 2021 version of the expert consensus index validated in Section 3.2, the sensitivity of this model (71.1%) was significantly higher than that of the consensus index (0.3%-15.9%), but the specificity (58.7%) was lower than that of the consensus index (94.7%-100%).

[0063] Table 2

[0064] The results showed that the diagnostic model based on multiple indicators of routine blood tests significantly improved the sensitivity of early screening in outpatient clinics and could more effectively identify suspected IM patients in primary care settings. However, the specificity and overall discrimination of the model still need to be further improved.

[0065] 4. Analysis and Summary This study focuses on early screening in outpatient settings and systematically analyzes the clinical data of 1322 children suspected of having EBV infection in outpatient settings. The main results are as follows: (1) The actual diagnostic efficacy of lymphocyte-related indicators recommended in the 2021 expert consensus in early outpatient screening scenarios. Data analysis clearly shows that the three indicators recommended by the consensus (lymphocyte percentage > 0.50, absolute lymphocyte count > 5.0 × 10⁻⁶) are effective. 9 In adolescents aged 6-18 years with suspected IM, the sensitivities of lymphocytes ( / L, HFLC ratio ≥0.10) were only 15.85%, 12.20%, and 0.27%, respectively, while the specificities all exceeded 94%, with absolute lymphocyte counts >5.0×10⁻⁶. 9The specificity of / L even reached 100%. This result suggests that in early outpatient screening scenarios, the core value of the consensus standard lies in diagnosis—once the indicator is positive, IM can almost be diagnosed; however, its value as a screening tool is very limited, with over 84% of IM children being missed because they do not meet these indicators. In primary care outpatient clinics, when children are first seen based solely on blood routine results, if they do not meet the above indicators, they are easily misdiagnosed as having an upper respiratory tract infection or bacterial tonsillitis, leading to delayed diagnosis and unnecessary antibiotic use. This finding confirms its limitation of "strong diagnostic power but weak screening power" and clarifies the specific manifestations of this limitation in early outpatient scenarios.

[0066] (2) Feasibility of constructing a better screening model based on multiple indicators of blood routine in early outpatient screening scenarios. Univariate analysis revealed multi-lineage hematological changes in IM patients: the EBV infection group not only had an elevated percentage of lymphocytes (33.58% vs. 31.15%), but also showed an increased percentage of neutrophils (5.96 × 10⁻⁶). 9 / L vs. 6.66×10 9 / L) and monocytes (0.79×10 9 / L vs. 0.95×10 9 The platelet count decreased ( / L), and the platelet count decreased (260.87×10). 9 / L vs. 292.92×10 9 While the EBV count was 9.86 fL vs. 9.74 fL, the volume of erythrocytes increased (HGB 122.34 g / L vs. 120.29 g / L) and the volume of erythrocytes increased (MCV 79.80 fL vs. 78.72 fL). These systemic changes suggest that even in the early outpatient stage, the immune response induced by EBV infection has already affected multiple blood cell lineages, providing a biological basis for constructing a multi-indicator combined model.

[0067] Based on this, we constructed a joint diagnostic model using logistic regression, incorporating five indicators: age, MCHC, PCT, MONO%, and HFLC. The results showed that HFLC was the strongest independent risk factor (OR=20.251), followed by age (OR=1.223), while MCHC, PCT, and MONO% were protective factors. The ROC analysis in Table 2 shows that the model's AUC was 0.708, and its sensitivity was improved to 71.1%, significantly better than the consensus indicators. This means that this model can identify over 70% of children with IM in early outpatient screening, greatly reducing the risk of missed diagnoses and playing a significant role in improving the initial identification capabilities of primary care clinics.

[0068] However, this model also has significant limitations. The model's specificity is only 58.7%, meaning that 41.3% of non-IM children were still misdiagnosed as positive, potentially leading to unnecessary referrals or further examinations. The AUC of 0.708 indicates moderate discrimination, suggesting that relying solely on routine blood tests may be insufficient to fully capture the pathophysiological characteristics of IM in the early outpatient stage. Furthermore, while HFLC, as an instrumental alternative for atypical lymphocytes (reactive lymphocytes), can reduce the subjectivity of manual interpretation, its detection is not yet widespread in some primary care hospitals, which may hinder the model's wider application.

[0069] The following preliminary conclusions were drawn from this study: The lymphocyte markers recommended in the 2021 expert consensus report show a strong diagnostic effect but a weak screening effect in early outpatient screening scenarios, and are not suitable as screening tools for primary care clinics. 2. EBV infection can cause changes in multiple lineages of blood cells in the early stages of outpatient treatment, providing a basis for constructing a multi-index model; 3. The combined model based on multiple blood routine indicators can significantly improve the sensitivity of outpatient screening to 71.1%, but the specificity (58.7%) and overall discrimination (AUC=0.708) are still not ideal, and cannot simultaneously meet the clinical needs of high sensitivity and high specificity.

[0070] The limitations mentioned above suggest that early screening in outpatient settings requires a more comprehensive combination of indicators to improve diagnostic efficiency.

[0071] Example 2: Inpatient Diagnosis Scenario: Optimization of an EBV Infection Screening Model Based on Combined Blood Routine and Liver Function Tests 1. Research Subjects This study focuses on inpatient diagnosis, collecting data from 109 hospitalized children diagnosed with intraepithelial neoplasia (IM) admitted to the Affiliated Nanchuan Hospital of Chongqing Medical University from January 2020 to December 2024 as the case group. During the same period, 465 children with clinically suspected IM but negative EBV-DNA test results were randomly selected as the control group. The diagnostic criteria for IM and the inclusion and exclusion criteria were the same as in Example 1. All patients had complete clinical data.

[0072] 2. Research Methods 2.1 Data Collection and Indicator Selection Clinical data of the above-mentioned subjects were retrospectively collected, including: (1) basic information: age and gender; (2) routine laboratory indicators: complete blood count (white blood cell count and differential, platelet count and related parameters, red blood cell related parameters), and liver function (ALT, AST, TP, ALB, GLB). Univariate analysis was used to compare the differences of each indicator between the case group and the control group to screen candidate indicators that may be related to the diagnosis of IM.

[0073] 2.2 Model Construction and Validation A simple random sampling method was used to divide all research subjects into a training set (for model construction) and a validation set (for internal model validation) at an 8:2 ratio. Based on the training set data, statistically significant indicators from the univariate analysis were included in a multivariate binary logistic regression analysis. Stepwise regression was used to screen independent influencing factors of IM, and a diagnostic model was constructed.

[0074] 2.3 Model Performance Evaluation Discrimination assessment: The discrimination ability of the model was evaluated using receiver operating characteristic (ROC) curves. The area under the curve (AUC) and its 95% confidence interval (CI) were calculated, as well as the sensitivity and specificity corresponding to the optimal cutoff value.

[0075] Calibration evaluation: The Hosmer-Lemeshow goodness-of-fit test was used to evaluate the model's calibration. P > 0.05 indicates that there is no significant difference between the model's predicted probability and the actual probability. Simultaneously, the Bootstrap method (1000 repeated samplings) was used to plot a calibration curve, visually demonstrating the consistency between the model's predicted probability and the actual probability.

[0076] Clinical applicability evaluation: Decision curve analysis (DCA) was used to evaluate the clinical net benefit of the model under different threshold probabilities to determine its clinical application value.

[0077] Statistical analysis was performed using SPSS 26.0 and R 4.4.1. A p-value < 0.05 was considered statistically significant.

[0078] 3. Research Results 3.1 General information of the cases and univariate analysis A total of 574 patients were included, of which 109 (19.0%) were in the IM group and 465 (81.0%) were in the non-IM group. There was no statistically significant difference in gender composition between the two groups (P=0.721), but there were statistically significant differences in age and several laboratory indicators (P<0.05), as detailed in Table 3.

[0079] Table 3 Comparison of routine indicators between children in the IM group and the non-IM group

[0080] 3.2 Validation of the diagnostic efficacy of the 2021 expert consensus in hospitalized diagnosis scenarios Using EBV-DNA PCR test results as the gold standard, a four-fold table was used to analyze the diagnostic efficacy of two lymphocyte-related indicators recommended in the 2021 expert consensus in hospitalized children aged 6-18 years suspected of having IM. The results are shown in Table 4.

[0081] Table 4. Diagnostic Performance Analysis of Lymphocyte-Related Indicators in Hospitalized Children with Fever and Impatient Disease (IM) Based on the 2021 Expert Consensus.

[0082] The above results indicate that, in hospital-based diagnosis scenarios, the diagnostic efficacy of lymphocyte-related indicators recommended in the 2021 expert consensus is significantly improved compared to early outpatient scenarios: absolute lymphocyte count > 5.0 × 10⁻⁶. 9 The sensitivity of / L was 75.61% and the specificity was 95.12%; the sensitivity of lymphocyte percentage >0.50 was 46.34% and the specificity was 100%. This suggests that the consensus indicators are more suitable for diagnosis in hospitalized or late-stage patients, but have limited value in early outpatient screening (as confirmed in Part II). However, even in hospital settings, the sensitivity of both indicators is still not ideal (maximum 75.61%), with the possibility of missed diagnoses, and they lack a comprehensive reflection of multi-system involvement of EBV infection (such as changes in liver function), requiring further optimization.

[0083] 3.3 Multivariate Logistic Regression Analysis Indicators showing statistically significant differences in univariate analysis were included in multivariate logistic regression, and stepwise regression was used to screen for independent influencing factors. The results showed that age, ALB, GLB, PLT, LYMPH#, and LYMPH% were independent factors associated with the diagnosis of IM (all P < 0.05), as detailed in Table 5.

[0084] Table 5 Results of IM multivariate logistic regression stepwise screening method analysis

[0085] According to the regression results, age, GLB, LYMPH, and LYMPH% were independent risk factors for IM, while ALB and PLT were protective factors. Specifically: for every year of age increase, the risk of IM increased by 83.8%; for every 1 g / L increase in globulin, the risk of IM increased by 33.3%; for every 1 × 10⁻⁶ increase in absolute lymphocyte count, the risk of IM increased by [missing information]. 9 For every 1 g / L increase in albumin, the risk of IM increases by 107.5%; for every 1% increase in lymphocyte percentage, the risk of IM increases by 6.9%; for every 1 g / L increase in albumin, the risk of IM decreases by 25.2%; for every 1 × 10⁶ g / L increase in platelet count, the risk of IM decreases by 25.2%. 9 / L, IM risk decreased by 0.7%.

[0086] 3.4 Construction of the Diagnostic Model Based on the above six independent influencing factors, a Logistic regression diagnostic model is constructed: Logit(P) = 0.623 × age (years) - 0.301 × ALB (g / L) + 0.308 × GLB (g / L) - 9.195 × PLT (10 9 / L) + 0.763 × LYMPH# (10 9 / L) + 0.066 × LYMPH% - 5.292

[0087] 3.5 Performance Evaluation of the Diagnostic Model (1) Discrimination evaluation ROC curve analysis results show that ( Figure 1 Training set (460 cases): The model's AUC for predicting IM was 0.986 (95% CI: 0.973-0.995), with a sensitivity of 93.67% and a specificity of 98.68% corresponding to the optimal cutoff value. Validation set (114 cases): The model's AUC for predicting IM was 0.986 (95% CI: 0.956-1.000), with a sensitivity of 90.00% and a specificity of 90.59%. This indicates that the model exhibits excellent discriminative ability on both the training and validation sets.

[0088] (2) Calibration evaluation Hosmer-Lemeshow goodness-of-fit test results show: Training set: χ² 2 =2.34, P=0.992; Validation set: χ² 2 =10.12, P=0.259; both P values ​​are >0.05, indicating that there is no significant difference between the model's predicted probability and the actual probability, and the model fits well.

[0089] The calibration curve was plotted using the Bootstrap method (1000 repeated samplings). Figure 2 The results showed that the calibration curves of both the training and validation sets were close to the ideal diagonal, further confirming that the model's prediction accuracy was good and there was no obvious overfitting.

[0090] (3) Evaluation of clinical applicability Decision curve analysis (DCA) results show that ( Figure 3 Within a relatively wide range of risk thresholds, the net benefit (red line) of clinical decision-making based on this model is higher than that of the "all intervention" (green line) or "no intervention" (black line) strategies, indicating that the model has good clinical application value.

[0091] 3.6 Model Performance Comparison The diagnostic efficacy of this model was compared with the indicators of the 2021 expert consensus, and the results are shown in Table 6.

[0092] Table 6. Comparison of the efficacy of different diagnostic models

[0093] As can be seen, the multi-indicator joint model constructed in this section significantly outperforms the outpatient model and consensus indicators in terms of AUC, sensitivity, and specificity in the inpatient diagnosis scenario, achieving comprehensive optimization of diagnostic efficacy.

[0094] 4. Summary and Analysis This embodiment focuses on the inpatient diagnosis scenario. Based on the outpatient model in Embodiment 1, it incorporates more routine test indicators such as liver function and successfully constructs a joint diagnostic model based on six indicators: age, albumin (ALB), globulin (GLB), platelet (PLT), absolute lymphocyte count (LYMPH#), and lymphocyte percentage (LYMPH%).

[0095] First, in hospitalized diagnosis scenarios, the diagnostic efficacy of the 2021 expert consensus indicators was significantly improved compared to outpatient settings. Table 6 shows that an absolute lymphocyte count > 5.0 × 10⁻⁶... 9 The sensitivity of / L reached 75.61%, and the sensitivity of lymphocyte proportion >0.50 was 46.34%, both significantly higher than the 0.3%-15.9% of outpatients in Example 1. This confirms the "scenario-dependent" consensus indicators proposed in the introduction—these indicators are more suitable for hospitalized children in the later stages of the disease with typical symptoms, and have high value as a diagnostic tool. However, their sensitivity still did not reach the ideal level (maximum 75.61%), meaning that about a quarter of hospitalized IM patients may be missed because they do not meet the consensus criteria; and the consensus indicators fail to reflect the impact of EBV infection on the liver and other systems, resulting in information blind spots.

[0096] Second, the multi-indicator joint model constructed in this study demonstrated excellent diagnostic efficacy in inpatient diagnosis scenarios. The model achieved an AUC of 0.986 on both the training and validation sets, with sensitivity and specificity exceeding 90%, significantly outperforming the outpatient model in Example 1 (AUC=0.708, sensitivity 71.1%, specificity 58.7%) and the 2021 consensus indicators. This means that the model can not only identify over 90% of inpatient IM patients but also accurately exclude over 90% of non-IM patients, achieving an ideal balance between high sensitivity and high specificity, thus simultaneously meeting the dual needs of screening and diagnosis.

[0097] Third, the newly included liver function indicators (ALB and GLB) have significant pathophysiological implications. A decrease in ALB may reflect a negative nitrogen balance and relative suppression of liver synthetic function during the acute infection phase; an increase in GLB directly reflects B cell polyclonal activation and the intensity of humoral immune responses. The combination of decreased ALB and increased GLB constitutes a unique serum protein profile in hospitalized patients with IM, information that cannot be captured by routine blood tests alone. In addition, the decrease in platelets (PLT) confirms the immune platelet destruction mechanism associated with EBV infection, suggesting that the diagnostic perspective needs to be expanded from "lymphocytes" to "platelet lineage".

[0098] Fourth, the complementary relationship with the outpatient model in Example 1. The outpatient model in Example 1 (AUC=0.708, sensitivity 71.1%, specificity 58.7%) is suitable for early screening in primary care clinics, prioritizing high sensitivity to minimize missed diagnoses. This inpatient model, with its advantages of high sensitivity and high specificity, is suitable for accurate diagnosis of hospitalized children. Together, they constitute a complete IM diagnosis and treatment pathway: those who test positive in the initial outpatient screening are referred to the hospital, confirmed by this model after hospitalization, and difficult cases are further supplemented with EBV-specific antibody or nucleic acid testing.

[0099] In summary, this embodiment successfully constructed a multi-indicator joint diagnostic model for IM with excellent diagnostic efficacy, easy-to-obtain indicators, and suitability for inpatient diagnosis scenarios. It complements the outpatient model in Embodiment 1 and provides a hierarchical tool for IM identification in different clinical scenarios in primary hospitals.

Claims

1. A biomarker for diagnosing and / or predicting infectious mononucleosis in children, characterized in that, The markers include age, ALB, GLB, PLT, and LYMPH.

2. The use of the biomarker for diagnosing and / or predicting childhood infectious mononucleosis as described in claim 1 in the preparation of reagents or kits for diagnosing and / or predicting childhood infectious mononucleosis.

3. A method for constructing a predictive model for the diagnosis of infectious mononucleosis (IM) in children, characterized in that, Includes the following steps: S1. Data collection and preprocessing: Inpatients diagnosed with IM were collected as the case group, and children with clinical manifestations suspected of having IM but negative EBV-DNA test results were randomly selected as the control group. Clinical data were selected according to the inclusion and exclusion criteria and preprocessed. The clinical data included: (1) Basic information: age and gender; (2) Routine laboratory indicators: blood routine: white blood cell count and classification, platelet count and related parameters and red blood cell related parameters, liver function: ALT, AST, TP, ALB and GLB, where ALT is alanine aminotransferase, AST is aspartate aminotransferase, TP is total protein, ALB is albumin and GLB is globulin; S2. Screening independent influencing factors: First, univariate analysis was used to compare the differences of various indicators between the case group and the control group to screen candidate indicators that may be related to the diagnosis of IM; then, multivariate binary logistic stepwise regression was used to screen independent influencing factors of IM. S3. Multivariate Logistic Regression Prediction Model: A simple random sampling method is used to divide all research subjects into training and validation sets. Based on independent influencing factors, the regression coefficients and corresponding scores of each variable are determined, and a prediction model is constructed. The calculation formula for the prediction model is as follows: Logit(P) = 0.623 × age (years) - 0.301 × ALB (g / L) + 0.308 × GLB (g / L) - 9.195 × PLT (10 9 / L) + 0.763 × LYMPH# (10 9 / L) + 0.066 × LYMPH%-5.292; Multivariate logistic regression analysis showed that age, ALB, GLB, PLT, LYMPH#, and LYMPH% were independent associated factors for the diagnosis of IM, and that age, GLB, LYMPH#, and LYMPH% were independent risk factors for IM, while ALB and PLT were protective factors; among them, ALB is albumin, GLB is globulin, PLT is platelet count, LYMPH# is absolute lymphocyte count, and LYMPH% is percentage of lymphocytes. S4. Internal validation and evaluation of model performance: After building the model based on the training set, the model is evaluated in the training set and validation set from three dimensions: discrimination, calibration and clinical applicability. The stability of the model is judged by comparing the consistency of the results of the two sets.

4. The construction method according to claim 3, characterized in that: Predicted probability P: ， If P > 0.5, IM is diagnosed; if P ≤ 0.5, IM is basically ruled out.

5. The construction method according to claim 3, characterized in that: In the calculation formula of the prediction model in step S3: Age: OR 1.838, 95% CI: 1.482–2.280, P < 0.001; ALB: OR value 0.748, 95% CI: 0.644-0.868, P<0.001; GLB: OR value 1.333, 95% CI: 1.185-1.500, P<0.001; PLT: OR value 0.993, 95% CI: 0.987-0.999, P=0.004; LYMPH#: OR value 2.075, 95% CI: 1.511-2.849, P<0.001; LYMPH%: OR value 1.069, 95% CI: 1.037-1.103, P<0.

001.

6. A predictive model for the diagnosis of infectious mononucleosis (IM) in children, obtained by the construction method according to any one of claims 3 to 5.

7. A predictive system for the diagnosis of infectious mononucleosis (IM) in children, characterized in that, The system includes an input module, a calculation module, and an output module; (1) Input module: used to transmit the following information of the subject to the calculation module: age, ALB, GLB, PLT, LYMPH#, LYMPH% (2) Calculation module: It has a built-in predictive model for the diagnosis of infectious mononucleosis (IM) in children. The calculation formula for the prediction model is as follows: Logit(P) = 0.623 × age (years) - 0.301 × ALB (g / L) + 0.308 × GLB (g / L) - 9.195 × PLT (10 9 / L) + 0.763 × LYMPH# (10 9 / L) + 0.066 × LYMPH%-5.292; Among them, ALB is albumin, GLB is globulin, PLT is platelet count, LYMPH# is absolute lymphocyte count, and LYMPH% is percentage of lymphocytes. (3) Output module: Used to output the IM risk probability value obtained by the calculation module: ， If P > 0.5, IM is diagnosed; if P ≤ 0.5, IM is basically ruled out.

8. The prediction system according to claim 7, characterized in that, The input module and the computing module are connected via wired and / or wireless means.

9. The prediction system according to claim 7, characterized in that, The computing module includes a computer host, a central processing unit, or a network server; the output module is a monitor, a printer, or an audio output device.

Images