Method for identifying skin mucosal lesion of ALL patient after HD-MTX treatment based on column diagram

Abstract

The invention belongs to the technical field of medicine, and particularly relates to a method for identifying skin mucosal lesions of ALL patients after HD-MTX treatment based on a column diagram, and the column diagram for predicting the skin mucosal lesions after acute lymphocytic leukemia patients receive high-dose methotrexate scheme treatment is established based on a multi-element logistic regression model, a decision tree and a neural network result. The method comprises the following steps: (1) determining risk factors of treating the skin mucosal lesion of an acute lymphocytic leukemia (ALL) patient by a large dose of methotrexate through a multivariate logistic regression model; (2) establishing the occurrence conditions of the skin mucosal lesion of the acute lymphocytic leukemia patient by using a decision tree model; and (3) the primary and secondary relationship between the 48-hour methotrexate blood concentration and the acute lymphocytic leukemia patient is found by using a neural network. The method provided by the invention has an accurate prediction capability, finds a lower and more accurate critical concentration for the first time, and can timely and effectively send out an early warning signal of skin mucosa injury of the acute lymphocytic leukemia patient.

Description

Technical Field

[0001] The present invention belongs to the field of medical technology, and particularly relates to a nomogram-based method for identifying skin and mucosal lesions in ALL patients after HD-MTX treatment. Background Art

[0002] High-dose methotrexate (HD-MTX) was defined as MTX dose ≥ 0.5 g m -2 When HD-MTX is administered, the blood concentration of MTX increases significantly, and it can not only penetrate the blood-testosterone barrier and the blood-brain barrier, but also reach solid tumors with poor blood circulation.

[0003] High-dose methotrexate (HD-MTX) is an important drug for the treatment of children with acute lymphoblastic leukemia (ALL). However, while HD-MTX enhances cytotoxicity and antitumor efficacy, it can also cause multiple adverse drug reactions (ADRs), including hepatotoxicity, nephrotoxicity, hematotoxicity, gastrointestinal toxicity, neurotoxicity, and mucocutaneous adverse reactions. High-risk children, in particular, experience varying degrees of mucocutaneous adverse reactions, such as oral mucositis and perianal inflammation. Mild manifestations include mucosal congestion, edema, and erythema, while severe manifestations include extreme mucosal congestion and erosion, confluent white membranes, extensive mucosal ulcers, and severe pain. Severe oral mucositis and perianal inflammation can lead to nosocomial infections. Therefore, establishing early and rapid prediction and identification of mucocutaneous adverse reactions facilitates the timely development and adjustment of mucocutaneous injury repair and nursing plans.

[0004] Currently, the HD-MTX regimen is used clinically to treat children with ALL. MTX blood concentrations are tested 24 hours after HD-MTX administration, and a CF rescue regimen is used. Over time, the CF rescue regimen and MTX metabolism in the body change. MTX concentrations are retested at 48 and 72 hours, and the CF rescue regimen is adjusted until the MTX blood concentration is <0.2 μmol / L. However, mucocutaneous and oral mucositis are dose-limiting toxicities that can be caused by lower MTX concentrations. In clinical practice, even when the 48-hour MTX blood concentration is <0.2 μmol / L, some patients develop varying degrees of mucocutaneous and skin damage. Therefore, identifying the critical concentration for the occurrence of mucocutaneous and skin damage is crucial.

[0005] In view of this, the present invention establishes a method based on machine learning to identify skin and mucosal damage in children with high-risk acute lymphoblastic leukemia after receiving high-dose methotrexate treatment, which is of great significance for the prevention and treatment of skin and mucosal damage. Summary of the Invention

[0006] In view of the shortcomings of the existing technology, the present invention proposes a nomogram-based method for identifying skin and mucosal lesions in ALL patients after HD-MTX treatment.

[0007] This is achieved specifically through the following technical solutions:

[0008] A nomogram-based method for identifying skin and mucosal lesions in ALL patients after HD-MTX treatment includes the following steps:

[0009] The first step is to collect raw data and obtain continuous variables;

[0010] The second step was to construct a logistic regression model: Continuous variables were converted into binary variables using the ROC curve. These variables were then incorporated into the univariate and multivariate logistic regression equations to screen for risk factors for skin and mucosal injuries. The binary variables were classified based on the actual situation of skin and mucosal injuries. If the input variable indicated skin and mucosal injuries, the variable was assigned a 1; if the input variable did not indicate skin and mucosal injuries, the variable was assigned a 0.

[0011] The third step is to construct a CART decision tree model: The CART algorithm is used to analyze risk factors and establish a CART decision tree model to determine the conditions for the occurrence of skin and mucosal damage. The CART decision tree model has a split node and significance test level of 0.05, a maximum tree depth of 5 layers, a minimum sample size of 20 for parent nodes, and a minimum sample size of 15 for child nodes. The conditions for the occurrence of skin and mucosal damage are high-risk ALL patients with a 48-hour MTX blood concentration ≥ 0.14 μmol / L and receiving HD-MTX treatment.

[0012] Step 4: Neural network analysis: Risk factors were analyzed using a neural network to obtain the independent variable importance value and normalized importance value of each risk factor, and the primary and secondary relationships of the risk factors were determined. That is, high-risk ALL patients receiving HD-MTX treatment were the primary factor, and 48-hour MTX blood concentration ≥ 0.14 μmol / L was the secondary factor. The independent variable importance value of high-risk ALL patients receiving HD-MTX treatment was 0.613, and the normalized importance value was 100.0%. The independent variable importance value of 48-hour MTX blood concentration ≥ 0.14 μmol / L was 0.387, and the normalized importance value was 63.2%.

[0013] The fifth step is to construct a nomogram module: use the nomogram to visualize the results of the Logistic regression model, CART decision tree model and neural network analysis to obtain the nomogram module;

[0014] Step 6 Nomogram Warning:

[0015] 6.1 Define the patient's treatment type and blood drug concentration type, including:

[0016] The disease treatment type is divided into two groups: when the high-risk ALL patients receive HD-MTX treatment, the disease treatment type is 1; when the high-risk ALL patients do not receive HD-MTX treatment, or the intermediate-risk ALL patients receive HD-MTX treatment, or the low-risk patients receive HD-MTX treatment, the disease treatment type is 0;

[0017] The blood drug concentration type is divided into two groups: when the MTX blood drug concentration is ≥0.14 μmol / L at 48 hours, the blood drug concentration type is 1; when the MTX blood drug concentration is <0.14 μmol / L at 48 hours, the blood drug concentration type is 0;

[0018] 6.2 Input the treatment type and blood drug concentration type into the nomogram module. The nomogram module first scores the current patient's condition treatment type and blood drug concentration type separately, then sums them. Finally, the risk probability value is calculated based on the sum and the data results of the neural network multi-layer perception domain decision tree in SPSS26.0 software;

[0019] The 0.14 μmol / L was determined by the Youden index of the ROC curve; the ROC curve was based on the CCLG-ALL 2018 protocol, and 100 or more high-risk ALL patients meeting the symbol requirements were included. Their 48-hour MTX blood concentrations and skin and mucosal damage were statistically analyzed, and the Youden index was generated in SPSS26.0 software;

[0020] The ALL patients are minors aged ≤ 18 years old.

[0021] The continuous variables include 24-hour methotrexate blood concentration, 48-hour methotrexate blood concentration, 72-hour methotrexate blood concentration, body weight, age, methotrexate dosage, number of methotrexate administrations, folinate dosage, number of methotrexate administrations, HD-MTX treatment regimen, and disease type.

[0022] The risk factors are 48-hour MTX serum concentration, HD-MTX treatment regimen, and disease type.

[0023] The disease type is divided into high risk, intermediate risk and low risk according to the diagnosis results of the Diagnosis and Treatment Program for Children's Acute Lymphoblastic Leukemia (CCLG-ALL2018).

[0024] The disease treatment type score includes a score of 100 for high-risk ALL patients receiving HD-MTX treatment; a score of 0 for high-risk ALL patients not receiving HD-MTX treatment or children with non-high-risk ALL receiving HD-MTX treatment.

[0025] The blood drug concentration type score includes: when the MTX blood drug concentration in 48 hours is ≥0.14 μmol / L, the blood drug concentration score is 70; when the MTX blood drug concentration in 48 hours is <0.14 μmol / L, the blood drug concentration score is 0.

[0026] HD-MTX treatment for patients with acute lymphoblastic leukemia refers to MTX dose ≥ 0.5 g·m -2 .

[0027] The HD-MTX treatment regimens for ALL patients include: (1) HD-MTX+6-MP: HD-MTX combined with mercaptopurine; (2) HR-1′: HD-MTX combined with dexamethasone, vincristine, cyclophosphamide, cytarabine, and aspartase; (3) HR-2′: HD-MTX combined with dexamethasone, vindesine, ifosfamide, daunorubicin, and aspartase.

[0028] The method for detecting the MTX blood concentration is any one of high performance liquid chromatography (HPLC), high performance liquid chromatography-tandem mass spectrometry (HPLC-MS / MS), fluorescence polarization immunoassay (FPIA), enzyme amplified immunoassay (EMIT), and chemiluminescent microparticle immunoassay (CMIA).

[0029] The detection method of the MTX blood concentration is any one of HPLC, HPLC-MS / MS, FPIA, EMIT, and CMIA.

[0030] The risk probability includes: 48 hours after ALL patients are given HD-MTX, if the MTX blood concentration is less than the critical value, the risk probability of skin and mucosal damage is less than 6%; if the MTX blood concentration is greater than or equal to the critical value, the risk probability of skin and mucosal damage is greater than or equal to 85%.

[0031] Beneficial effects:

[0032] The method of the present invention has accurate prediction capabilities and has discovered a lower and more accurate critical concentration for the first time. It can promptly and effectively issue early warning signals of skin and mucous membrane damage in patients with acute lymphoblastic leukemia, and can provide a powerful reference for medical staff to adjust rescue plans as early as possible.

[0033] This paper establishes a nomogram for predicting skin and mucosal damage in patients with acute lymphoblastic leukemia treated with high-dose methotrexate based on the results of a multivariate logistic regression model, a decision tree, and a neural network. The paper identifies the risk factors for skin and mucosal damage in patients with acute lymphoblastic leukemia (ALL) treated with high-dose methotrexate using a multivariate logistic regression model. Furthermore, a decision tree model is used to establish the conditions for the occurrence of skin and mucosal damage in patients with ALL. Finally, a neural network is used to identify the primary and secondary relationships between 48-hour methotrexate blood concentrations and ALL patients. This demonstrates the scientific and feasible nature of the present invention.

[0034] The method of the present invention is suitable for identifying and predicting the risk probability of skin and mucous membrane damage in children with high-risk acute lymphoblastic leukemia, and is of great significance for proposing corresponding rescue plans for high-risk children in a timely manner. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 Regression tree prediction model for skin and mucosal injury;

[0036] Figure 2 Schematic diagram of the importance of independent variables for neural network analysis;

[0037] Figure 3 This is a nomogram for predicting skin and mucosal damage in children with high-risk acute lymphoblastic leukemia. In the figure, A represents children with high-risk acute lymphoblastic leukemia; B represents a 48-hour serum methotrexate concentration ≥ 0.14 μmol / L; 1 represents yes, and 0 represents no;

[0038] Figure 4 Comparison of ROC curves between the Logistic regression model and the CART decision tree model;

[0039] Figure 5 ROC curves for the training set and validation set;

[0040] Figure 6 is the calibration curve for the training set;

[0041] Figure 7 DCA curves for the training set and validation set. DETAILED DESCRIPTION

[0042] The specific embodiments of the present invention are further described in detail below, but the present invention is not limited to these embodiments. Any improvement or replacement based on the basic spirit of the present embodiment still falls within the scope of protection required by the claims of the present invention.

[0043] Example 1 Construction of Nomogram Early Warning Module

[0044] In the first part, 142 patients meeting the symbol requirements were enrolled according to the CCLG-ALL 2018 protocol, and their clinical data are shown in Table 1 .

[0045] Table 1 Clinical data of 142 ALL patients

[0046]

[0047] Note: HD-MTX treatment regimens: (1) 6-MP+HD-MTX: HD-MTX combined with mercaptopurine; (2) VD+HD-MTX: HD-MTX combined with vincristine and dexamethasone; (3) HR-1′: HD-MTX combined with dexamethasone, vincristine, cyclophosphamide, cytarabine, and aspartase; (4) HR-2′: HD-MTX combined with dexamethasone, vindesine, ifosfamide, daunorubicin, and aspartase.

[0048] The skin and mucosal injury conditions of the 142 patients in the second part are shown in Table 2.

[0049] Table 2 Description of skin and mucous membrane damage in 142 patients treated with high-dose methotrexate

[0050]

[0051] Note: a) Three patients had oral mucositis and perianal inflammation.

[0052] Part III Univariate and multivariate analysis of skin and mucosal injury

[0053] The ROC curve was used to convert continuous variables (24-hour methotrexate blood concentration, 48-hour methotrexate blood concentration, 72-hour methotrexate blood concentration, body weight, age, methotrexate dosage, number of methotrexate administrations, folinate dosage, number of methotrexate administrations) into binary variables (see Table 3).

[0054] Table 3 Univariate logistic regression analysis and multivariate logistic regression analysis of risk factors

[0055]

[0056]

[0057] The results of univariate analysis showed that 24h CMTX (MTX blood concentration after 24h), 48h CMTX (MTX blood concentration after 48h), gender, weight, age, risk level, albumin, MTX dose, MTX frequency, CF dose, CF frequency, and HD-MTX regimen were associated with a higher incidence of skin and mucosal injuries (P<0.05).

[0058] The forward stepwise regression method was used to include single factors with p < 0.10 and clinically significant factors such as 24h CMTX, 48h CMTX, gender, weight, age, albumin, MTX dose, CF dose, and HD-MTX treatment regimen into multivariate logistic regression analysis. The results showed that 48h CMTX and high-risk ALL patients receiving HD-MTX treatment regimen were independent risk factors for skin and mucosal damage (p < 0.05).

[0059] Part IV: Classification and Regression Tree (Decision Tree) Analysis (CART) for Predicting HD-MTX-Related Skin and Mucosal Injuries

[0060] A decision tree model was constructed using the CART algorithm. The splitting and significance testing levels were set at 0.05, the maximum tree depth was 5 layers, the minimum sample size for parent nodes was 20, and the minimum sample size for child nodes was 15.

[0061] Figure 1 As shown, the mucocutaneous injury model is divided into three layers, with a total of seven nodes, four of which are terminal nodes. The identified influencing factors primarily include the 48-hour CMTX and HD-MTX regimens. Among them, 48-hour CMTX was most closely associated with mucocutaneous injury, followed by the HD-MTX regimen. The model showed that when the 48-hour CMTX concentration was <0.1395 μmol / L, the incidence of mucocutaneous injury was 5.60%. When the 48-hour CMTX exposure was ≥0.1395 μmol / L, the incidence was 31.3%. When the 48-hour CMTX concentration was ≥0.1395 μmol / L, the incidence was 85.70% when patients chose regimens 3 or 4. When the 48-hour CMTX concentration was ≥0.1395 μmol / L, the incidence was only 9.90% when patients received regimens 1 or 2. When 48-hour CMTX ≥ 0.1395 μmol / L, the incidence of mucocutaneous and skin damage was 13.20% in patients receiving regimen 1. When 48-hour CMTX ≥ 0.1395 μmol / L was maintained in patients receiving regimen 2, no mucocutaneous and skin damage occurred. Regimens 3 and 4 are recommended for high-risk patients. This suggests that this model, based on 48-hour methotrexate plasma concentrations, can be used to identify high-dose methotrexate regimens for treating high-risk children with acute lymphoblastic leukemia and associated mucocutaneous damage.

[0062] Part V: Prediction of HD-MTX-related skin and mucosal damage using CART neural network analysis

[0063] Neural network results are shown in Figure 2As shown in Table 4, the results showed that the incidence of skin and mucosal injury was 86%, the normalization importance of the independent variable high-risk acute lymphoblastic leukemia children was 100%, and the normalization importance of 48hMTX serum concentration was 63.2%. It can be seen that high-risk acute lymphoblastic leukemia children was the primary factor for acute lymphoblastic leukemia children, followed by 48hMTX serum concentration.

[0064] Table 4 Importance of independent variables of neural network

[0065]

[0066]

[0067] Part VI: Nomogram for predicting skin and mucosal damage in children with high-risk acute lymphoblastic leukemia treated with high-dose methotrexate

[0068] Figure 3 This is a nomogram for predicting skin and mucosal lesions in children with high-risk acute lymphoblastic leukemia. Nomogram Description: Each point on the variable axis corresponds to a different score. The scores for each predictor are summed to obtain a total score. Different total scores correspond to different risks of developing SADR in hospitalized patients. For example, a child with high-risk acute lymphoblastic leukemia (100 points) and a 48-hour methotrexate serum concentration (≥0.14 μmol / L) (70 points) would have a corresponding score of 100 + 70 = 170, resulting in a Disease Risk of SADR > 0.80 (0.842).

[0069] Part VII Comparison of Logistic Regression Model and CART Decision Tree Model Based on ROC Curve

[0070] Draw the ROC curves of the two models ( Figure 4 ), the results showed that the AUC of the logistic regression model was greater than that of the CART decision tree model, with p < 0.001. The accuracy of the logistic regression model and the decision tree model was similar.

[0071] Part 8 Nomogram Validation

[0072] The ORC curve, C-index, calibration curve, and DCA curve show certain guiding significance in clinical practice.

[0073] The data set was divided into a training set (70%) and a validation set (30%). The performance of the nomogram prediction model was evaluated using the training set, and the receiver operating characteristic (ROC) curve (AUC) of the training set was 0.874 (95% confidence interval: 0.796–0.953) ( Figure 5The C index is 0.875, indicating good discrimination ability. The calibration curve of the training set shows good consistency between the predicted probability and the actual observation results, as shown in the calibration curve ( Figure 6 ) The calibration curve of the validation set is the same as that of the training set. Clinical decision curve analysis (DCA) is used to evaluate the clinical application value of the nomogram, such as Figure 7 In summary, the nomogram showed a significant positive net benefit in predicting SADR, highlighting its important clinical value in SADR prediction.

[0074] Example 2

[0075] A nomogram-based method for identifying skin and mucosal lesions in ALL patients after HD-MTX treatment includes the following steps:

[0076] The first step is to collect raw data and obtain continuous variables;

[0077] The second step was to construct a logistic regression model: using the ROC curve in SPSS 26.0 software, continuous variables were converted into binary variables, which were then incorporated into the univariate logistic regression equation and the multivariate logistic regression equation to screen out the risk factors for skin and mucosal damage. The binary variables were classified according to the actual situation of skin and mucosal damage. When the input variable had skin and mucosal damage, it was classified as 1; when the input variable did not have skin and mucosal damage, it was classified as 0.

[0078] The third step was to construct a CART decision tree model: the CART algorithm was used in SPSS 26.0 software to analyze risk factors and establish a CART decision tree model to determine the conditions for the occurrence of skin and mucosal damage. The split node and significance test level in the CART decision tree model was 0.05, the maximum tree depth was 5 layers, the minimum sample size for parent nodes was 20, and the minimum sample size for child nodes was 15. The conditions for the occurrence of skin and mucosal damage were high-risk ALL patients with a 48-hour MTX blood concentration ≥ 0.14 μmol / L and receiving HD-MTX treatment. High-risk ALL patients were diagnosed according to the Childhood Acute Lymphoblastic Leukemia Diagnosis and Treatment Program (CCLG-ALL2018).

[0079] Step 4: Neural network analysis: SPSS 26.0 software was used to analyze the risk factors using a neural network. The importance value of the independent variable and the normalized importance value of each risk factor were obtained to determine the primary and secondary relationship of the risk factors. That is, high-risk ALL patients receiving HD-MTX treatment were the primary factor, and 48-hour MTX blood concentration ≥ 0.14 μmol / L was the secondary factor. The independent variable importance value of high-risk ALL patients receiving HD-MTX treatment was 0.613, and the normalized importance value was 100.0%. The independent variable importance value of 48-hour MTX blood concentration ≥ 0.14 μmol / L was 0.387, and the normalized importance value was 63.2%.

[0080] The fifth step is to construct a nomogram module: In R software, the nomogram is used to visualize the results of the Logistic regression model, CART decision tree model, and neural network analysis to obtain a nomogram module;

[0081] Step 6 Nomogram Warning:

[0082] 6.1 Define the patient's treatment type and blood drug concentration type, including:

[0083] The disease treatment type is divided into two groups: when the high-risk ALL patients receive HD-MTX treatment, the disease treatment type is 1; when the high-risk ALL patients do not receive HD-MTX treatment, or the intermediate-risk ALL patients receive HD-MTX treatment, or the low-risk patients receive HD-MTX treatment, the disease treatment type is 0;

[0084] The blood drug concentration type is divided into two groups: when the MTX blood drug concentration is ≥0.14 μmol / L at 48 hours, the blood drug concentration type is 1; when the MTX blood drug concentration is <0.14 μmol / L at 48 hours, the blood drug concentration type is 0;

[0085] 6.2 Input the treatment type and blood drug concentration type into the nomogram module. The nomogram module first scores the current patient's condition treatment type and blood drug concentration type separately, then sums them. Finally, the risk probability value is calculated based on the sum and the data results of the neural network multi-layer perception domain decision tree in SPSS26.0 software;

[0086] The 0.14 μmol / L was determined by the Youden index of the ROC curve; the ROC curve was based on the CCLG-ALL 2018 protocol, and 100 or more high-risk ALL patients meeting the symbol requirements were included. Their 48-hour MTX blood concentrations and skin and mucosal damage were statistically analyzed, and the Youden index was generated in SPSS26.0 software;

[0087] The ALL patients are minors aged ≤18 years old;

[0088] The continuous variables include 24-hour methotrexate blood concentration, 48-hour methotrexate blood concentration, 72-hour methotrexate blood concentration, body weight, age, methotrexate dosage, number of methotrexate administrations, folinate dosage, number of methotrexate administrations, and HD-MTX treatment regimen.

[0089] The risk factors were 48-hour MTX serum concentration and HD-MTX treatment regimen.

[0090] The disease treatment type score includes a score of 100 for high-risk ALL patients receiving HD-MTX treatment; a score of 0 for high-risk ALL patients not receiving HD-MTX treatment or children with non-high-risk ALL receiving HD-MTX treatment.

[0091] The blood drug concentration type score includes: when the MTX blood drug concentration in 48 hours is ≥0.14 μmol / L, the blood drug concentration score is 70; when the MTX blood drug concentration in 48 hours is <0.14 μmol / L, the blood drug concentration score is 0.

[0092] HD-MTX treatment for patients with acute lymphoblastic leukemia refers to MTX dose ≥ 0.5 g·m -2 .

[0093] The HD-MTX treatment regimens for ALL patients include: (1) HD-MTX+6-MP: HD-MTX combined with mercaptopurine; (2) HR-1′: HD-MTX combined with dexamethasone, vincristine, cyclophosphamide, cytarabine, and aspartase; (3) HR-2′: HD-MTX combined with dexamethasone, vindesine, ifosfamide, daunorubicin, and aspartase.

[0094] The method for detecting the MTX blood concentration is any one of high performance liquid chromatography (HPLC), high performance liquid chromatography-tandem mass spectrometry (HPLC-MS / MS), fluorescence polarization immunoassay (FPIA), enzyme amplified immunoassay (EMIT), and chemiluminescent microparticle immunoassay (CMIA).

[0095] The detection method of the MTX blood concentration is any one of HPLC, HPLC-MS / MS, FPIA, EMIT, and CMIA.

[0096] The risk probability includes: 48 hours after ALL patients are given HD-MTX, if the MTX blood concentration is less than the critical value, the risk probability of skin and mucosal damage is less than 6%; if the MTX blood concentration is greater than or equal to the critical value, the risk probability of skin and mucosal damage is greater than or equal to 85%.

[0097] The method for detecting the MTX serum concentration also includes any one of HPLC, HPLC-MS / MS, FPIA, and CMIA. The EMIT value is converted according to the method of Zhang Yuanyuan, Xu Kangkang. Comparative analysis of three detection methods for determining methotrexate blood concentration. Anhui Medicine, 2014, 18(2): 357-360.

[0098] The high-risk ALL patients or non-high-risk ALL patients are determined according to the Diagnosis and Treatment Program for Children with Acute Lymphoblastic Leukemia (CCLG-ALL2018).

[0099] Example 3

[0100] Methods for using a nomogram to predict the incidence of skin and mucosal lesions in children with acute lymphoblastic leukemia aged ≤18 years include:

[0101] (1) Diagnose the child's condition according to clinical standards and determine whether it is a high-risk condition;

[0102] (2) According to the CCLG-ALL 2018 protocol (e.g., HR-1′ and HR-2′) and the revised protocol, high-dose methotrexate was used, and half-dose hydration (1500-2000 ml / m2) was performed 1 day before HD-MTX treatment. 2 ) and alkalinization to make urine pH ≥7.

[0103] (3) High-risk children were given a dose of 5 g / m 2 ;

[0104] (4) At 24 hours after infusion, the serum MTX concentration was monitored using enzyme-enhanced immunosorbent assay; the CF rescue plan, hydration plan, and alkalinization plan were adjusted according to the MTX blood concentration;

[0105] (5) MTX serum concentration was monitored by enzyme-enhanced immunosorbent assay at 48 h after HD-MTX infusion;

[0106] (6) Nomogram module prediction: The obtained patient's condition and MTX serum concentration value 48 hours after HD-MTX administration are classified according to the condition treatment type and blood drug concentration type. If the condition treatment type is divided into two groups: high-risk ALL patients who receive HD-MTX treatment, the treatment type is 1; high-risk ALL patients who do not receive HD-MTX treatment or non-high-risk ALL patients who receive HD-MTX treatment, the treatment type is 0; if the blood drug concentration type is divided into two groups: when the 48-hour MTX blood drug concentration is ≥0.14μmol / L, the blood drug concentration type is 1; when the 48-hour MTX blood drug concentration is <0.14μmol / L, the blood drug concentration type is 0;

[0107] Then, the treatment type and blood drug concentration type are input into the nomogram module generated by R software. The nomogram module first scores the current patient's condition treatment type and blood drug concentration type separately, then sums them. Finally, based on the sum, the risk probability value is calculated using the data results obtained by the neural network multi-layer perception domain decision tree of SPSS26.0 software;

[0108] The risk probability statistics are as follows: if the serum MTX concentration of high-risk children is less than 0.14 μmol / L 48 hours after receiving HD-MTX treatment, the skin and mucosal injury rate is less than 6%; if the serum MTX concentration of high-risk children is ≥0.14 μmol / L 48 hours after receiving HD-MTX treatment, the skin and mucosal injury rate is ≥85%;

[0109] (7) Clinical recommendations: In traditional rescue plans, protective measures should be taken for high-risk children whose serum MTX concentration is ≥0.14 μmol / L 48 hours after receiving HD-MTX treatment. Specifically, immediate gargle with an appropriate amount of calcium folinate to prevent oral ulcers; oral therapeutic dose (10 ml, tid) of Kangfuxin solution to prevent gastrointestinal ulcers; and use of light iodine for perianal ulcers.

[0110] The traditional rescue plan is: when the 48h MTX blood concentration is >0.2μmol / L, continue to implement the CF rescue plan: when the 48h MTX blood concentration is >0.2μmol / L and less than or equal to 1.0μmol / L, give 15mg / m 2 Rescue dose for CF. When the 48h MTX blood concentration is >1.0μmol / L and ≤2.0μmol / L, give 30mg / m 2 Rescue dose for CF. When the 48h MTX blood concentration is 2.0μmol / L or ≤3.0μmol / L, give 45mg / m 2 Similarly, for every 1.0 μmol / L increase in MTX blood concentration, the CF rescue dose increases by 15 mg / m 2 The rescue dose of CF should not exceed 20 mg / kg, and the rescue should be stopped when the MTX blood concentration of the child is <0.2 μmol / L after 48 hours.

[0111] (8) Observe the child for skin and mucous membrane damage to prevent the occurrence of skin and mucous membrane damage.

[0112] Clinical Case 1

[0113] A 15-year-old boy weighing 46.5 kg was diagnosed with acute lymphoblastic leukemia for more than 3 months and returned to the hospital for chemotherapy. The preliminary diagnosis was acute lymphoblastic leukemia (B-lineage high-risk KMT2A-ITPR2, SETD2, FLT3 mutations); right subclavian vein thrombosis. After admission, the patient was given the first round of HR-1' chemotherapy according to the CCLG-ALL 2018 protocol, excluding contraindications to medication. A total of 6.862 g of MTX was administered in 7 doses (1 intrathecal injection of 12 mg, 1 intravenous infusion of 0.5 g, and the remaining 5 intravenous infusions of 1.27 g / time). Other drugs of the HR-1' regimen were also administered. On the second day of chemotherapy with the HR-1' regimen, the child had no nausea, vomiting, abdominal pain, bloating, diarrhea, rash, mouth pain, fever, cough, runny nose, numbness of the limbs, active bleeding, and was in good spirits and had a good appetite. , urination and defecation are normal. Physical examination: vital signs are stable, the patient is conscious, responsive, ruddy in complexion, no bleeding spots on the skin, no congestion or flushing in the pharynx. The MTX blood concentration (60.6 μmol / L) was measured 24 hours after administration, and CF rescue was immediately given (0.02 g / time, 6 hours between each time, for a total of 3 times). On the third day of HR-1' chemotherapy, the patient had no gross hematuria, no frequent urination, no urgency, no pain when urinating, no fever, no rash, no mouth pain, no nausea, vomiting, no abdominal pain, no cough, runny nose, sneezing, and no quadruple Numbness and paresthesia in the limbs, no active bleeding, good spirits, good diet, and normal bowel movements. Physical examination: stable vital signs, clear mind, good reaction, ruddy complexion, no bleeding spots on the skin, no congestion or flushing of the throat. 48-hour MTX blood concentration was measured (0.185μmol / L). The child's 48-hour methotrexate blood concentration was within the safe range. Calcium folinate was given for rescue 3 times and then discontinued. The amount of hydration and alkalinization fluids was reduced, and the methotrexate concentration was monitored for 72 hours. The treatment plan was adjusted as appropriate. HR-1' scheme On the fourth day of treatment, the MTX blood concentration was measured at 72 hours and was 0.07 μmol / L. On the fifth day of HR-1' chemotherapy, skin and mucosal lesions (flushing, blistering, and fissures at the corners of the mouth) occurred, but otherwise normal findings were observed. On the seventh day of HR-1' chemotherapy (the first day of rest and the ninth day of hospitalization), the patient developed oral mucosal ulceration. Kangfuxin solution was added to repair the oral mucosal ulcer, and oral vitamin B2 tablets were continued for vitamin supplementation. Subsequently, the patient's oral mucosal lesions improved, and he was discharged in stable condition (a total of 19 days of hospitalization).

[0114] That is, the patient had high-risk ALL. The MTX blood concentration was >0.14 (0.185 μmol / L) 48 hours after HD-MTX administration, and oral ulcers occurred on the 9th day of hospitalization.

[0115] Clinical Case 2

[0116] A 5-year-old boy weighing 16.5 kg returned to the hospital for chemotherapy after being diagnosed with acute lymphoblastic leukemia for more than 4 months. The initial diagnosis was acute lymphoblastic leukemia (B-lineage high-risk TCF3 / ZNF384, EVI1 positive) and post-chemotherapy bone marrow suppression. Following remission, medication contraindications were eliminated and, according to the CCLG-ALL 2018 protocol, the first round of HR-1' regimen consisted of 3.362 g of MTX administered in five divided doses (one intrathecal injection of 12.0 mg, one intravenous infusion of 0.35 g, and three subsequent intravenous infusions of 1.20 g). On the second day of HR-2' chemotherapy, the child had no gross hematuria, frequency, urgency, or dysuria, fever, rash, or oral pain, and his appetite was slightly poor. This morning, he complained of nausea, but no vomiting, abdominal pain, cough, runny nose, numbness or paresthesia in the extremities, and no active bleeding. He was in good spirits and had normal bowel movements. Physical examination: The patient was clear-headed, mentally alert, and responsive. There was no rash or bleeding spots on the skin of the whole body. The complexion was ruddy, and there was no ulceration or flushing of the oral mucosa. The 24-hour TXM concentration was 62.94 μmol / L. CF rescue was immediately given (0.01 g / time, 6 hours apart, for a total of 3 times). On the third day of HR-2' chemotherapy, the patient's 48-hour TXM concentration was 0.159 μmol / L. The patient's methotrexate concentration was within the safe range. Calcium folinate rescue was given 3 times and then discontinued. Calcium folinate rescue was given 3 times and then discontinued. Hydration and alkalinization were continued. On the fourth day of R-2' chemotherapy (0 6.17), the child complained of mouth pain, poor diet, no dysphagia; no rash; no abdominal pain, vomiting or diarrhea, no frequent urination, urgency, or pain during urination; normal urine volume and color; no fever, cough, numbness or paresthesia in the limbs, no bleeding, good spirits, and normal bowel movements. Physical examination: clear consciousness, good spirits and reactions, no rash or hemorrhages on the skin, ruddy complexion, redness of the oral mucosa, slight ulceration on the left oral mucosa, and no pharyngeal congestion. 72h TXM concentration (0. 032μmol / L). Treatment plan: Discontinue cyclophosphamide chemotherapy and continue hydration and alkalinization. After high-dose methotrexate chemotherapy, the patient developed mouth pain, redness, and ulceration of the oral mucosa. Diagnosis: oral mucosal ulcer. Calcium folinate gargles were administered to mitigate the toxic side effects of methotrexate. Oral gold powder was administered for antibacterial purposes, and Kangfuxin solution was administered for mucosal repair. Other treatment remained the same, with careful observation for adverse chemotherapy reactions. On day 6 (June 19) of R-2' chemotherapy, the patient still had mouth pain and sore throat, accompanied by poor appetite and fatigue. Nocturnal restlessness improved compared to the previous period. There was no fever, rash, redness or pain in the eyes, nausea, vomiting, abdominal pain, bloating, or diarrhea. There was no cough, shortness of breath, or palpitations. Urine color and volume were normal. There was no active bleeding, dizziness, or headache. His spirits and appetite were poor, and his bowel movements were normal. Physical examination revealed stable vital signs, clear consciousness, good mental performance, and no rash or petechiae. His complexion was slightly pale. There was oral mucosal ulceration, and he continued with current treatment.On the 14th day after chemotherapy (June 27), the child had no chills, convulsions, or chills; however, he continued to experience abdominal pain, primarily around the umbilicus, with paroxysmal exacerbations. He did not experience abdominal distension or vomiting. He passed two small, brown, loose stools with tenesmus and a small volume, without mucus, pseudomembranes, or bloody pus. Urine output was normal. His mouth pain had improved, and he had no cough or shortness of breath, headache, or convulsions. He was in poor spirits and had a poor appetite. Physical examination revealed stable vital signs, clear consciousness, good mental performance, no bleeding spots or rashes, and a slightly pale complexion. The oral mucosal ulcer had decreased in size, and his pharynx was congested. On the 16th hospital day, the oral mucosal ulcer had improved, and his condition was stable, and he was discharged.

[0117] That is, the patient had high-risk ALL. The MTX blood concentration was >0.14 (0.159 μmol / L) 48 hours after HD-MTX administration, and oral ulcers occurred on the fifth day of hospitalization.

[0118] Clinical Case 3

[0119] A 12-year-old female patient, weighing 25.0 kg, returned to the hospital for chemotherapy after being diagnosed with acute lymphoblastic leukemia for more than 8 months. The initial diagnosis was acute lymphoblastic leukemia (B-lineage high-risk remission); post-chemotherapy bone marrow suppression. Following remission, medication contraindications were eliminated and a second round of treatment, the HR-2' regimen, according to the CCLG-ALL 2018 protocol, was administered: a total of 4.877 g of MTX in six divided doses (one intrathecal injection of 12.0 mg, one intravenous infusion of 0.48 g, three intravenous infusions of 1.46 g each, and the remaining 5 mg intravenous infusion). On the second day of HR-2' chemotherapy, the patient had no nausea, vomiting, abdominal pain, bloating, diarrhea, rash, mouth pain, fever, cough, runny nose, numbness in the limbs, or active bleeding. She was in good spirits, had a slightly poor appetite, and had normal bowel movements. Physical examination: clear consciousness, good spirit and reaction, no rash or bleeding spots on the skin, pale face, moderate anemia, no ulceration of oral mucosa, no congestion or flushing of pharynx. 24h TXM concentration (74.88μmol / L, CF rescue was immediately given (0.02g / time, 6 hours between each time, for a total of 2 times. On the third day of HR-2' chemotherapy, the child's urine volume was normal, no gross hematuria, no frequent urination, urgency, or pain, no palpitations, shortness of breath, no nausea, vomiting, abdominal pain, diarrhea, no fever, rash, no mouth pain, slightly poor diet, no cough, runny nose, no numbness or abnormal sensation in the limbs, no active bleeding, good spirits, normal bowel movements. Physical examination: clear consciousness, good spirits , the reaction was acceptable, there was no rash or bleeding spots on the skin of the whole body, the face was pale, and there was a moderate anemia, no ulceration of the oral mucosa, and no congestion in the pharynx. 48h TXM concentration (0.183μmol / L), the child's methotrexate concentration was within the safe range, and calcium folinate was rescued twice and then discontinued, and hydration and alkalinization were continued. On the fourth day of HR-2' chemotherapy, the child did not complain of mouth pain, fever, cough, gross hematuria, frequent urination, urgency, or pain, and the urine volume was normal, there was no rash, no abdominal pain, abdominal distension, vomiting or diarrhea, and no activity. The patient had bleeding, good spirits, poor diet, and normal bowel movements. Physical examination: stable vital signs, clear consciousness, good spirits and reactions, no rash or bleeding spots on the skin, ruddy complexion, ulceration of the left oral mucosa, no congestion in the pharynx, large tonsils, clear and symmetrical breath sounds in both lungs, no dry or wet rales, the heart examination was the same as before, the abdomen was flat and soft, no tenderness, the liver and spleen were not palpable under the ribs, bowel sounds were normal, no ulceration of the perianal mucosa, and the rest of the physical examination was the same as before. 72h TXM concentration (0.049μmol / L), for children with oral mucosal rupture Ulcers were diagnosed as stomatitis. Kangfuxin solution and human epidermal growth factor were administered for mucosal repair, and oral care was advised. On the fifth day of HR-2 chemotherapy (June 19), the patient underwent the same physical examination as the previous day. Treatment included discontinuation of dexamethasone and daunorubicin chemotherapy and hydration and alkalinization. For oral mucosal ulcers, vitamin C supplementation was administered to promote mucosal repair. On the fifteenth day of HR-2 chemotherapy (June 29), the patient's temperature returned to normal, the mouth pain improved, and there was no active bleeding. Other symptoms were normal, the patient was in good spirits, had a good diet, and had normal bowel movements.Physical examination: Vital signs are stable, mental state is good, no rash or bleeding spots, complexion is rosier than before, oral mucosal ulcers are healed, and there is no congestion in the pharynx. On the 24th day of hospitalization, the child's body temperature was normal, there was no cough, runny nose, dizziness, fatigue, active bleeding, abdominal pain, vomiting or diarrhea, mental state and appetite were good, and bowel movements were normal. Physical examination: Vital signs are stable, mental state is clear, mental state is good, there are no rashes or bleeding spots on the skin, no yellowing of the skin and sclera, no congestion in the pharynx, large tonsils, no pus spots or herpes, oral mucosal ulcers have basically healed, bilateral lung breath sounds are clear and symmetrical, and no dry or wet rales are heard. The heart rhythm is regular, the heart sounds are strong, the abdomen is flat and soft, there is no tenderness, the bowel sounds are normal, there are no ulcers in the skin around the anus, the extremities are warm, and the pulse is strong. The condition is stable and he was discharged after improvement.

[0120] That is, the patient had high-risk ALL. The MTX blood concentration was >0.14 (0.183 μmol / L) 48 hours after HD-MTX administration, and oral mucosal ulcers occurred on the 6th day of hospitalization.

[0121] Clinical Case 4

[0122] A 10-year-old female patient, weighing 36.0 kg, returned to the hospital for chemotherapy after being diagnosed with acute lymphoblastic leukemia for more than 5 months. The initial diagnosis was acute lymphoblastic leukemia (T-lineage high-risk remission DEK:CAN). After admission, medication contraindications were eliminated and the first round of chemotherapy, HR-2', was administered according to the CCLG-ALL 2018 protocol: a total of 5.832 g of MTX was administered in six divided doses (one intrathecal injection of 12.0 mg, one intravenous infusion of 0.50 g, and the remaining four intravenous infusions of 1.33 g). On the second day of HR-2' chemotherapy, the patient had no nausea, vomiting, fever, cough, abdominal pain, or diarrhea. Her spirits and diet were good, and her bowel movements were normal. Physical examination revealed stable vital signs, clear consciousness, good reflexes, a rosy complexion, no petechiae on the skin, and no congestion or flushing of the throat. 24h TXM concentration (99.66μmol / L, CF rescue was immediately given (0.02g / time, 6h interval each time, for a total of 3 times. On the third day of HR-2' chemotherapy, the child had no nausea, vomiting, fever, cough, abdominal pain, diarrhea, mental state and diet, and normal bowel movements. Physical examination: vital signs were stable, clear-headed, responsive, ruddy complexion, no bleeding spots on the skin, no congestion or flushing of the throat. 48h TXM concentration (0.282μmol / L), the child's methotrexate concentration was acceptable, and subcutaneous injection was given. After 3 rescues of calcium carbonate, it was discontinued, and the methotrexate concentration was tracked for 72 hours, and chemotherapy was continued as planned. On the 4th day of HR-2' chemotherapy (11.18), the child did not complain of special discomfort, no nausea, vomiting, fever, cough, abdominal pain, diarrhea, mental state and diet were good, and bowel movements were normal. Physical examination: vital signs were stable, the patient was clear-headed, responsive, with ruddy complexion, and no bleeding spots on the skin. The rest of the physical examination was the same as before. 72h TXM concentration (0.00μmol / L). The child was on the 6th day of rest (10th day after chemotherapy; 11.2 4), the child complained of sore throat, no nausea, vomiting, no epistaxis, no gingival bleeding, no fever, cough, no abdominal pain, diarrhea, good spirits and appetite, normal bowel movements. Physical examination: stable vital signs, clear mind, good reaction, ruddy complexion, no bleeding spots on the skin. The pharynx was slightly congested, and a white film was seen in the oral mucosa. The superior physician instructed: considering the white film in the child's mouth, oral mucositis should be considered, and symptomatic treatment should be given. On the 8th day of rest (the 12th day after chemotherapy; 11.26), the child had no sore throat, no nausea, vomiting, no epistaxis, no gingival bleeding, no fever, cough, abdominal pain, diarrhea, no fever, no abdominal pain, no diarrhea, no fever, no abdominal pain, no diarrhea, no abdominal pain ... Fever, cough, no abdominal pain, diarrhea, good spirits and appetite, normal bowel movements. Physical examination: stable vital signs, clear mind, good reaction, slightly pale complexion, no bleeding spots on the skin. No congestion in the pharynx, and improved oral white film. On the 20th day of hospitalization, the child had no nausea, vomiting, fever, cough, epistaxis, gingival bleeding, abdominal pain, diarrhea, good spirits and appetite, and normal bowel movements. Physical examination: stable vital signs, clear mind, good reaction, ruddy complexion, no bleeding spots on the skin. No congestion in the pharynx, and no white film in the mouth. The condition was stable and he was discharged after improvement.

[0123] That is, the patient had high-risk ALL. The MTX blood concentration was >0.14 (0.282 μmol / L) 48 hours after HD-MTX administration, and oral mucosal injury occurred on the 12th day of hospitalization.

[0124] Clinical Case 5

[0125] A 9-year-old boy, weighing 20.0 kg, returned to the hospital for chemotherapy after being diagnosed with acute lymphoblastic leukemia for more than 4 months. The initial diagnosis was acute lymphoblastic leukemia (B-lineage high-risk remission). After admission, medication contraindications were eliminated and the first HR-2' chemotherapy regimen was administered according to the CCLG-ALL 2018 protocol: a total of 4.012 g of MTX was administered in five divided doses (one intrathecal injection of 12 mg, one intravenous infusion of 0.4 g, and three subsequent intravenous infusions of 1.20 g). Other HR-2' regimen drugs were also administered. On the second day of HR-2' chemotherapy, the child had no oral pain, nausea, vomiting, fever, cough, dizziness, headache, abdominal pain, or diarrhea. His spirits and appetite were normal, and his bowel movements were normal. Physical examination revealed stable vital signs, clear consciousness, good reflexes, a rosy complexion, no petechiae on the skin, and no congestion or flushing of the throat. The MTX blood concentration (37.20 μmol / L) was measured 24 hours after administration, and CF rescue was immediately given (0.01 g / time, 6 hours between each time, for a total of 3 times). On the third day of HR-2' chemotherapy, the child's 48-hour TXM concentration (0.136 μmol / L) and the child's methotrexate concentration were within the safe range. Calcium folinate rescue was given 3 times and then discontinued. The 72-hour methotrexate concentration and liver function test results were tracked today, and chemotherapy was continued as planned. On the fourth day of HR-2' chemotherapy, the 72-hour methotrexate blood concentration was 0.07 μmol / L. The child had no mouth pain, no nausea, vomiting, no fever, no cough, no dizziness, no headache, no abdominal pain, no diarrhea, and was in good spirits and appetite. His bowel movements were normal. Physical examination: vital signs were stable, he was alert, his reactions were good, his complexion was still rosy, there were no bleeding spots on the skin, and there was no congestion or flushing in the throat. Children undergoing HR-2' chemotherapy On the 6th day, the child had no nausea, vomiting, abdominal pain, diarrhea, fever, cough, dizziness, or headache. His spirit and appetite were slightly worse, and his bowel movements were normal. Physical examination: no ulceration of the oral mucosa, no congestion of the pharynx, large tonsils, no pus spots or white film attached. On the first day of rest and treatment (the 7th day), the child was in good general condition, with no nausea, vomiting, abdominal pain, diarrhea, fever, or cough. His spirit and appetite were slightly better than before, and his bowel movements were normal. Physical examination: vital signs The patient's physical signs were stable, with clear consciousness and good reflexes. His face, conjunctiva, nail beds, and auricles were slightly pale. There were no petechiae or ecchymoses on the skin. There was no palpable enlargement of superficial lymph nodes throughout the body. There was no ulceration of the oral mucosa, no pharyngeal congestion, and no large tonsils. There were no pus spots or white film attached. His neck was soft, his respiratory activity was weak, and his cardiopulmonary and abdominal examinations were normal. He rested for the third day (9th day) and was discharged on the same day as on the first day. On the 18th day of hospitalization, the patient had no skin or mucosal lesions and had improved.

[0126] That is, the patient had high-risk ALL, and the MTX blood concentration was <0.14 (0.136 μmol / L) 48 hours after HD-MTX administration, and there was no skin or mucosal damage during hospitalization.

[0127] Clinical Case 6

[0128] A 14-year-old boy weighing 53.0 kg was returned to the hospital for treatment due to a diagnosis of leukemia more than 8 years prior and a testicular relapse more than 1 year prior. The initial diagnosis was acute lymphoblastic leukemia (B-lineage relapse, high-risk NRAS, CREBBP-positive, in remission); testicular leukemia. After admission, medication contraindications were eliminated and the second HR-2' chemotherapy regimen was administered according to the CCLG-ALL 2018 protocol: a total of 7.5625 g of MTX was administered in 7 divided doses (1 intrathecal injection of 12.5 mg, 1 intravenous infusion of 0.5 g, 4 intravenous infusions of 1.20 g / dose, and the remaining 1 dose of 1.45 g / dose). Other drugs of the HR-2' regimen were also administered. On the second day of HR-2' chemotherapy, the 24-hour TXM concentration was (89.82 μmol / L), and CF rescue (0.02 g / dose, 6 hours apart, for a total of 3 doses) was immediately administered. On the third day of HR-2' chemotherapy, the child There was no gross hematuria, no urinary frequency, urgency, or pain, no fever, rash, nausea, vomiting, abdominal pain, cough, runny nose, or sneezing, no numbness or paresthesia in the limbs, no active bleeding, and a clear mind and a good diet. Bowel movements were normal. Physical examination: clear consciousness, good spirits, and reactions, no rash or hemorrhages on the skin, rosy complexion, and no ulceration or flushing of the oral mucosa. The 48-hour TXM concentration (0.114 μmol / L) was within the safe range for methotrexate blood concentrations over 48 hours, so calcium folinate was administered three times before the rescue was discontinued and the dose was reduced with hydration and alkalinization. On the 4th day of chemotherapy with R-2' regimen, the child had no urinary frequency, urgency, or pain, and the urine volume and color were normal. There was no fever, cough, rash, or mouth pain. There was no ulceration of the oral mucosa, no congestion of the pharynx, no abdominal pain, vomiting, or diarrhea, no numbness or paresthesia in the limbs, no bleeding, and the child's spirit and diet were normal. The physical examination showed that the child was clear-headed, with good spirits and reactions. There was no rash or bleeding spots on the skin of the whole body. The complexion was slightly pale, there was no ulceration of the oral mucosa, and no congestion in the pharynx. 72h TXM concentration was (0.000μmol / L). On the 6th day of chemotherapy, the child had no nausea, vomiting, or abdominal pain. The patient had no pain, diarrhea, or abdominal distension, but no fever or rash, no cough or runny nose, no mouth pain, no ulceration of the oral mucosa, no sore throat, no active bleeding, no dizziness, headache, or fatigue, normal urine volume and color, and good spirits and appetite. Physical examination: The patient was clear-headed, mentally alert, and had good reactions. There were no rashes or bleeding spots on the skin of the whole body. The face, conjunctiva, nail beds, and auricles were pale, showing signs of moderate anemia. There were no ulceration of the oral mucosa, and no congestion in the pharynx. On the second day of rest (the eighth day after chemotherapy), the physical examination was the same as after chemotherapy (the sixth day after chemotherapy). On the 17th day of hospitalization, the child had no skin or mucosal damage and was discharged after improvement.

[0129] That is, the patient had high-risk ALL, and the MTX blood concentration was <0.14 (0.114 μmol / L) 48 hours after HD-MTX administration, and there was no skin or mucosal damage during hospitalization.

[0130] The above clinical cases 1-6 have verified the feasibility and accuracy of the present invention.

Claims

1. A nomogram-based method for identifying skin and mucosal lesions in ALL patients after HD-MTX treatment, characterized in that: The steps include: The first step is to collect raw data and obtain continuous variables; The second step was to construct a logistic regression model: Continuous variables were converted into binary variables using the ROC curve. These variables were then incorporated into the univariate and multivariate logistic regression equations to screen for risk factors for skin and mucosal injuries. The binary variables were classified based on the actual situation of skin and mucosal injuries. If the input variable indicated skin and mucosal injuries, the variable was assigned a 1; if the input variable did not indicate skin and mucosal injuries, the variable was assigned a 0. The third step is to construct a CART decision tree model: Using the CART algorithm to analyze risk factors, a CART decision tree model is established to determine the conditions for the occurrence of skin and mucosal damage. The CART decision tree model has a split node and significance test level of 0.05, a maximum tree depth of 5 layers, a minimum sample size of 20 for parent nodes, and a minimum sample size of 15 for child nodes. The conditions for the occurrence of skin and mucosal damage are high-risk ALL patients with a 48-hour MTX blood concentration ≥ 0.14 μmol / L who are receiving HD-MTX treatment. Step 4: Neural network analysis: Risk factors were analyzed using a neural network to obtain the independent variable importance value and normalized importance value of each risk factor, and the primary and secondary relationships of the risk factors were determined. That is, high-risk ALL patients receiving HD-MTX treatment were the primary factor, and 48-hour MTX blood concentration ≥ 0.14 μmol / L was the secondary factor. The independent variable importance value of high-risk ALL patients receiving HD-MTX treatment was 0.613, and the normalized importance value was 100.0%. The independent variable importance value of 48-hour MTX blood concentration ≥ 0.14 μmol / L was 0.387, and the normalized importance value was 63.2%. The fifth step is to construct a nomogram module: In R software, the nomogram is used to visualize the results of the Logistic regression model, CART decision tree model, and neural network analysis to obtain a nomogram module; Step 6 Nomogram Warning: 6.1 Define the patient's treatment type and blood drug concentration type, including: The disease treatment type is divided into two groups: when the high-risk ALL patients receive HD-MTX treatment, the disease treatment type is 1; when the high-risk ALL patients do not receive HD-MTX treatment, or the intermediate-risk ALL patients receive HD-MTX treatment, or the low-risk patients receive HD-MTX treatment, the disease treatment type is 0; The blood drug concentration type is divided into two groups: when the MTX blood drug concentration is ≥0.14 μmol / L at 48 hours, the blood drug concentration type is 1; when the MTX blood drug concentration is <0.14 μmol / L at 48 hours, the blood drug concentration type is 0; 6.2 Input the treatment type and blood drug concentration type into the nomogram module. The nomogram module first scores the current patient's condition treatment type and blood drug concentration type separately, then sums them. Finally, the risk probability value is calculated based on the sum and the data results of the neural network multi-layer perception domain decision tree in SPSS26.0 software; The 0.14 μmol / L was determined by the Youden index of the ROC curve; the ROC curve was based on the CCLG-ALL2018 protocol, and 100 or more high-risk ALL patients meeting the symbol requirements were included. Their 48-hour MTX blood concentrations and skin and mucosal damage were statistically analyzed, and the Youden index was generated in SPSS26.0 software; The ALL patients are minors aged ≤ 18 years old.

2. The method for identifying skin and mucous membrane lesions in ALL patients after HD-MTX treatment based on a nomogram according to claim 1, wherein: The continuous variables include 24-hour methotrexate blood concentration, 48-hour methotrexate blood concentration, 72-hour methotrexate blood concentration, body weight, age, methotrexate dosage, number of methotrexate administrations, folinate dosage, number of methotrexate administrations, HD-MTX treatment regimen, and disease type.

3. The method for identifying skin and mucous membrane lesions in ALL patients after HD-MTX treatment based on a nomogram according to claim 1, wherein: The risk factors are 48-hour MTX serum concentration, HD-MTX treatment regimen, and disease type.

4. The method for identifying skin and mucosal lesions in ALL patients after HD-MTX treatment based on a nomogram according to claim 1, wherein: The disease treatment type score includes a score of 100 for high-risk ALL patients receiving HD-MTX treatment; a score of 0 for high-risk ALL patients not receiving HD-MTX treatment or children with non-high-risk ALL receiving HD-MTX treatment.

5. The method for identifying skin and mucous membrane lesions in ALL patients after HD-MTX treatment based on a nomogram according to claim 1, wherein: The blood drug concentration type score includes: when the MTX blood drug concentration in 48 hours is ≥0.14 μmol / L, the blood drug concentration score is 70; when the MTX blood drug concentration in 48 hours is <0.14 μmol / L, the blood drug concentration score is 0.

6. The method for identifying skin and mucosal lesions in ALL patients after HD-MTX treatment based on a nomogram according to claim 1, wherein: HD-MTX treatment for patients with acute lymphoblastic leukemia refers to MTX dose ≥ 0.5 g·m -2 .

7. The method for identifying skin and mucous membrane lesions in ALL patients after HD-MTX treatment based on a nomogram according to claim 6, wherein: The detection method of the MTX blood concentration is any one of HPLC, HPLC-MS / MS, FPIA, EMIT, and CMIA.

8. The method for identifying skin and mucous membrane lesions in ALL patients after HD-MTX treatment based on a nomogram according to claim 1, wherein: The risk probability includes: 48 hours after ALL patients are given HD-MTX, if the MTX blood concentration is less than the critical value, the risk probability of skin and mucosal damage is less than 6%; if the MTX blood concentration is greater than or equal to the critical value, the risk probability of skin and mucosal damage is greater than or equal to 85%.

Images