Platelet trajectory calculator and application thereof in construction of acute graft versus host disease risk assessment model

Abstract

The invention discloses application of a platelet trajectory calculator or a platelet trajectory model in construction of an acute graft versus host disease (aGVHD) risk assessment model. According to the method, a two-stage GBTM model is utilized to find that a platelet mode after transplantation has three different change trajectories, the platelet trajectories are found to be independent risk factors for predicting aGVHD occurrence, especially severe aGVHD occurrence, and patients in a low platelet trajectory group have a higher severe aGVHD accumulative occurrence rate, so that the aGVHD occurrence rate is improved. An earlier time window and a simpler clinical index are provided for clinical prediction of aGVHD, and a new thought is provided for prediction of aGVHD and other diseases. Meanwhile, based on the research result of the invention, a platelet trajectory calculator is developed, and a unique aGVHD risk assessment model is constructed, so that the technical scheme of the invention has higher clinical practical application value.

Description

Technical Field

[0001] The present disclosure relates to the field of medical informatics, and in particular, to a platelet trajectory calculator / platelet trajectory model and use thereof in constructing an acute graft-versus-host disease (aGVHD) risk assessment model. Background Art

[0002] Hematopoietic stem cell transplantation (HSCT) is an important means to cure blood diseases such as acute myeloid leukemia (AML), especially for those high-risk (chemotherapy ineffective) or relapsed patients, HSCT provides the only possible cure.

[0003] Acute graft-versus-host disease (aGVHD) is the most common short-term complication after transplantation. Its severity is closely related to the outcome of transplantation. The one-year survival rate of patients with severe aGVHD is less than 50%, and the one-year non-relapse mortality (NRM) of patients with aGVHD who have failed treatment can even reach 60%. Early prediction of the occurrence of aGVHD and drug prevention can reduce the incidence of aGVHD, thereby improving the long-term survival of patients. At present, many research results on the prediction of aGVHD have been published internationally, and most of its predictive factors are biomarkers including some cytokines. In addition, machine prediction models based on large-scale clinical data have been reported more and more, but the existing prediction models have not yet reached the ideal state. How to identify high-risk aGVHD patients earlier and more accurately after transplantation remains a challenge. At present, there is an urgent need to explore risk factors that can predict aGVHD earlier after transplantation, and such factors should have clinical applicability. Summary of the invention

[0004] Technical issues solved:

[0005] One aspect of the present disclosure is to address the lack of new prediction methods and models for acute graft-versus-host disease in the prior art, and to provide a platelet trajectory calculator / platelet trajectory model and its use in constructing an acute graft-versus-host disease (aGVHD) risk assessment model.

[0006] Specifically, the inventors creatively discovered that the platelet number (count) trajectory can be used as an independent risk factor for predicting / evaluating the risk of acute graft-versus-host disease in patients after allogeneic hematopoietic stem cell transplantation. Applying the calculator / platelet trajectory model for the platelet number (count) trajectory to the construction of an aGVHD risk assessment model can provide an earlier time window and simpler clinical indicators for clinical prediction of the occurrence of aGVHD, and provide new ideas for the prediction of diseases such as aGVHD, thereby solving the above technical problems.

[0007] The technical solution provided by this disclosure:

[0008] Use of a platelet trajectory calculator or platelet trajectory model in constructing a risk assessment model for acute graft-versus-host disease (aGVHD).

[0009] In some embodiments of the present disclosure, the inventors further studied the early platelet trajectory of patients after allogeneic hematopoietic stem cell transplantation and found that patients with early platelet transfusion requirements are more likely to develop aGVHD. Therefore, in some embodiments of the present disclosure, the above-mentioned platelet trajectory is the platelet trajectory of the patient from the 0th day to the onset of acute graft-versus-host disease after allogeneic hematopoietic stem cell transplantation. Further, in other embodiments of the present disclosure, the above-mentioned platelet trajectory is the platelet trajectory of the patient from the 0th day to the 10th day after allogeneic hematopoietic stem cell transplantation.

[0010] In the present disclosure, the acute graft-versus-host disease (aGVHD) is a common complication after allogeneic hematopoietic stem cell transplantation and one of the important complications of early death after transplantation. Diseases that can be treated with allogeneic hematopoietic stem cell transplantation include, for example, leukemia (e.g., acute myeloid leukemia, acute lymphocytic leukemia, chronic myeloid leukemia, chronic lymphocytic leukemia), lymphoma, multiple myeloma, severe aplastic anemia, immunodeficiency disease, paroxysmal nocturnal hemoglobinuria, autoimmune diseases (e.g., lupus erythematosus), malignant non-hematological tumors, etc. The above diseases are all likely to develop acute graft-versus-host disease (aGVHD) complications after treatment with haploidentical / haploidentical transplantation (haploidentical HSCT, HID-HSCT). In one embodiment of the present disclosure, the above acute graft-versus-host disease is produced by patients with acute myeloid leukemia after treatment.

[0011] In some embodiments of the present disclosure, the acute graft-versus-host disease is severe acute graft-versus-host disease, and the severe acute graft-versus-host disease is grade 3-4 according to the Glucksberg criteria.

[0012] In the present disclosure, the inventors performed GBTM modeling on the platelet counts of 165 AML-CR patients in the training cohort 0-10 days after HID-HSCT, and determined three platelet trajectories: low platelet count trajectory, medium platelet count trajectory, and high platelet count trajectory, and found that the low platelet count trajectory group had a higher proportion of severe aGVHD than the high platelet count trajectory group (28.6% vs 4.9%, P<0.001). Therefore, in some embodiments of the present disclosure, the platelet trajectory is the low platelet count trajectory.

[0013] In some embodiments of the present disclosure, the method for constructing the platelet trajectory calculator comprises the following steps:

[0014] S1: The GBTM model (Group-Based Trajectory Modeling) is used to divide the platelet trajectories into three groups, namely, a low platelet count trajectory group, a medium platelet count trajectory group, and a high platelet count trajectory group. The posterior probability of the class is as follows:

[0015] ,

[0016] in, =3,

[0017] y = (y0, y1, y2, . . . , yT ) represents the trajectory observed from time 0 to T,

[0018] m k is a quadratic function of time t, with different sets of coefficients for different groups: ,

[0019] are all the parameters that have;

[0020] S2: Reconstruct the model obtained in S1 using a two-stage decision tree method:

[0021] First, in the initial stage, set = 2 and use formula (2) in S1 to classify the observed trajectories into one of the following two groups: non-group 3 ( =1) or Group 3 ( =2),

[0022] Then in the second stage, for the trajectories assigned to the non-group 3 categories in the initial stage, they are further divided into two subgroups using formula (2) in S1: Group 1 ( =1) or Group 2 ( =2);

[0023] S3: The probability of obtaining the three groups of observation trajectories is calculated as follows:

[0024]

[0025] Further, in one embodiment of the present disclosure, the model parameters in the initial stage in S2 are:

[0026]

[0027]

[0028] ;

[0029] The model parameters of the second stage in S2 above are:

[0030]

[0031]

[0032] .

[0033] Utilizing the research ideas and results in this disclosure, another aspect of this disclosure is to provide a method for constructing an acute graft-versus-host disease (aGVHD) risk assessment model, comprising:

[0034] Step 1) Collect the patient’s clinical data;

[0035] Step 2) determining predictive variables of the risk assessment model, wherein the predictive variables include platelet trajectory;

[0036] Step 3) establishing a GVHD prediction risk nomogram using the predictor variables described in step 2);

[0037] Step 4) constructing the acute graft-versus-host disease risk assessment model using the aGVHD prediction risk nomogram obtained in step 3).

[0038] Further, in some embodiments of the present disclosure, the above-mentioned platelet trajectory is an independent risk factor in the predicted risk nomogram.

[0039] In order to achieve better results, in some embodiments of the present disclosure, the predictor variables described in the above step 2) also include known aGVHD risk factors, wherein the known aGVHD risk factors include those selected from patient age, transplant type, donor-recipient gender pair, aGVHD prevention regimen that the patient has been administered, total mononuclear cell count or CD34 + One or more of a number of cells.

[0040] Furthermore, in some embodiments of the present disclosure, the acute graft-versus-host disease is caused by acute myeloid leukemia patients after allogeneic hematopoietic stem cell transplantation.

[0041] Utilizing the research ideas and results in the present disclosure, another aspect of the present disclosure is to provide a use of a platelet trajectory calculator or a platelet trajectory model in the preparation of a product for evaluating the prognosis of treatment of acute myeloid leukemia.

[0042] Using the research ideas and results in the present disclosure, another aspect of the present disclosure is to provide a device for assessing the risk of acute graft-versus-host disease, the device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor;

[0043] When the computer program is executed by the processor, the acute graft-versus-host disease (aGVHD) risk assessment model constructed by the above construction method is run.

[0044] Utilizing the research ideas and results in the present disclosure, another aspect of the present disclosure is to provide a computer-readable storage medium, on which is stored an acute graft-versus-host disease (aGVHD) risk assessment model program constructed by the above-mentioned construction method, and when the program is executed by a processor, the functions of the above-mentioned device are realized.

[0045] Beneficial effects:

[0046] The present invention uses the GBTM model to find that there are three different change trajectories in the platelet pattern after transplantation, and finds that the platelet trajectory is an independent risk factor for predicting the occurrence of aGVHD, especially severe aGVHD. Patients in the low platelet trajectory group have a higher cumulative incidence of severe aGVHD, which provides an earlier time window and simpler clinical indicators for clinical prediction of the occurrence of aGVHD. At the same time, a risk prediction model is constructed in combination with other clinical factors, which provides new ideas for the prediction of diseases such as aGVHD. At the same time, based on the research results of the present invention, a platelet trajectory calculator is developed, and a unique aGVHD risk assessment model is constructed, which makes the technical solution of the present invention more clinically practical. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 A nomogram for predicting the risk of acute graft-versus-host disease (aGVHD) constructed in an embodiment of the present disclosure;

[0048] Figure 2 This is a result diagram of the classification of hemoglobin and platelet trajectory groups when the vanilla GBTM algorithm is directly applied in the embodiment of the present disclosure;

[0049] Figure 3 It is a result diagram of the occurrence of aGVHD in different hemoglobin trajectory groups when the vanilla GBTM algorithm is directly applied in the embodiment of the present disclosure;

[0050] Figure 4 This is a result diagram of the occurrence of aGVHD in different platelet trajectory groups when the vanilla GBTM algorithm is directly applied in the embodiment of the present disclosure;

[0051] Figure 5 It is a result diagram of the occurrence of severe aGVHD in different platelet trajectory groups when the vanilla GBTM algorithm is directly applied in the embodiment of the present disclosure;

[0052] Figure 6 This is a result diagram of the classification of platelet trajectory groups when the two-stage GBTM algorithm is applied in the embodiment of the present disclosure;

[0053] Figure 7 This is a result diagram of the occurrence of severe aGVHD in two cohorts with different platelet trajectory groups when the two-stage GBTM algorithm is applied in the embodiment of the present disclosure;

[0054] Figure 8 This is a result diagram of the occurrence of severe aGVHD in different risk groups of two cohorts when the acute graft-versus-host disease (aGVHD) prediction risk model constructed according to the embodiment of the present disclosure is applied. DETAILED DESCRIPTION

[0055] The present invention discloses a use of a platelet trajectory calculator in constructing an acute graft-versus-host disease (aGVHD) risk assessment model. Those skilled in the art can refer to the content of this article and appropriately improve the process parameters to achieve it. It should be particularly pointed out that all similar substitutions and modifications are obvious to those skilled in the art, and they are all considered to be included in the present invention, and relevant personnel can obviously modify or appropriately change and combine the contents described herein without departing from the content, spirit and scope of the present invention to implement and apply the technology of the present invention.

[0056] In the present invention, unless otherwise specified, the scientific and technical terms used herein have the meanings commonly understood by those skilled in the art. Unless otherwise expressly indicated, throughout the specification and claims, the term "including" or its variations such as "comprising" or "including" etc. will be understood to include the elements or components stated, without excluding other elements or other components. The term "a", "an" and "the" include plural indicators. The term "multiple" refers to two or more. The terms "such as", "for example", etc. are intended to refer to exemplary embodiments, and are not intended to limit the scope of the present disclosure.

[0057] In the present disclosure, when a range of values ​​is provided, it should be understood that the endpoints are included in the range and each intervening value between the upper and lower limits of the range and any other specified value or intervening value in the specified range and any value in the smaller range between the specified values ​​are included unless the context clearly dictates otherwise.

[0058] In this disclosure, the term "about" generally refers to variations within a range of 0.5%-10% above or below a specified value, for example, 0.5%, 1%, 1.5%, 2%, 2.5%, 3%, 3.5%, 4%, 4.5%, 5%, 5.5%, 6%, 6.5%, 7%, 7.5%, 8%, 8.5%, 9%, 9.5%, or 10% above or below a specified value.

[0059] In the present disclosure, unless otherwise specified, the scientific and technical terms used herein have the meanings commonly understood by those skilled in the art. The definition of common terms in molecular biology can be found in Lewin's GENES, Twelfth Edition, Jocelyn E. Krebs, Elliott S. Goldstein, Stephen T. Kilpatrick, Publisher: Jones & Bartlett Learning. The definition of common terms in biochemistry can be found in Lehninger Principles of Biochemistry, Eighth Edition, David L. Nelson, Michael M. Cox, Publisher: WH Freeman. The definition of common terms in cell biology can be found in Molecular Biology of the Cell, Sixth Edition, Bruce Alberts, Alexander Johnson, Julian Lewis, David Morgan, Martin Raff, Keith Roberts, Peter Walter, Publisher: Garland Science. The definition of common terms in genetics can be found in Genetics: Analysis of Genes and Genomes, Eighth Edition, Daniel L. Hartl, Maryellen Ruvolo, Publisher: Jones & Bartlett Learning.

[0060] Unless otherwise specified, the experimental techniques herein employ conventional techniques of immunology, biochemistry, chemistry, molecular biology, microbiology, cell biology, genomics and recombinant DNA, which can be found in standard books such as the following: Molecular Cloning: A Laboratory Manual; Cell Biology: A Laboratory Handbook, etc.

[0061] definition:

[0062] The term "platelet" in the present disclosure refers to a daughter cell with hemostatic function that is produced by bone marrow megakaryocytes and released into the peripheral blood. Normal engraftment and reconstruction of platelets are important indicators for judging the success of hematopoietic reconstruction after transplantation. There are different ways of platelet reconstruction in patients after HSCT, among which poor platelet reconstruction will significantly reduce the patient's overall survival after transplantation. In the case of aGVHD, patients usually have lower platelet counts and higher platelet transfusion requirements, and even slower platelet recovery and reconstruction speed, resulting in a shortened overall survival after transplantation.

[0063] The terms "platelet count trajectory" and "platelet number trajectory" in the present disclosure have the same or very similar meanings as those in the present disclosure and can be used interchangeably. Generally, platelet count or platelet number refers to the absolute number of platelets. The platelet count / number trajectory is the change in the number of platelets over a period of time and can be characterized and analyzed using an appropriate trajectory model. For example, a joint latent class model (JLCM), etc. In some embodiments of the present disclosure, a group-based trajectory model (GBTM) is used to characterize the platelet trajectories in groups.

[0064] Group-Based Trajectory Modeling (GBTM), also known as latent class growth model (LCGM), is an effective method for analyzing longitudinal data. GBTM was proposed by DANIELS.NAGIN and described in the article "Analyzing Developmental Trajectories: A Semiparametric, Group-Based Approach" (Psychol Methods, 1999, 4 (2): 139-157.). It is a semi-parametric finite mixture model that can classify the trajectories of a group of people and generate several representative motion trajectory models, and then analyze each trajectory model to understand the person's motion characteristics, physiological level and risk level. The core idea of ​​GBTM is to divide the population into several groups, each of which has similar motion patterns or development trajectories. These patterns are used to describe and predict their future trajectories, thereby providing a more scientific basis for the health management of individuals and groups. GBTM is widely used in the biomedical field to study the trajectory of biomarkers over time, evaluate the heterogeneity of clinical interventions, and predict the relationship between biomarkers and clinical events in different fields. At the same time, by mapping the development trajectory of the disease, GBTM helps evaluate the effectiveness of clinical interventions and identify high-risk groups for the disease.

[0065] The term "platelet trajectory calculator" in the present disclosure may also be referred to as a "platelet trajectory model", which is a mathematical or computational model used to describe and analyze the trajectory of platelet movement in blood or blood vessels. Common modeling methods include, for example, fluid mechanics simulation methods, cell mechanics models, image processing techniques, and the like. Those skilled in the art can construct it according to the methods in the prior art. For example, Chen J, Gao X, Shen S, et al. disclosed a platelet trajectory model in Association of longitudinal platelet count trajectory with ICU mortality: A multi-cohort study. Front Immunol. 2022;13:936662. For another example, Ye Q, Wang X, Xu X, Chen J, Christiani DC, Chen F, et al. disclosed a platelet trajectory model in Serial platelet count as a dynamic prediction marker of hospital mortality among septic patients. Burns Trauma.2024;12:tkae016. In one embodiment of the present disclosure, the inventor converts the constructed platelet trajectory model into the form of a web calculator for easy use. The construction of the web calculator can be achieved using known methods in the prior art.

[0066] The term "acute graft-versus-host disease" in this disclosure is used interchangeably with "aGVHD", "acute GVHD", "classic acute graft-versus-host disease", and "classic acute GVHD" in this disclosure and refers to graft-versus-host disease that develops in a transplant recipient within about 100 days after transplantation and exhibits clinical features typically associated with acute graft-versus-host disease, including but not limited to inflammation and tissue damage in the skin, oral and genital mucosa, eyes, intestines, liver, lungs, joints, and muscles. GVDH is discussed in Nassereddine, Anticancer Research, 37(4): 1547-1555 (2017).

[0067] The term "Glucksberg criteria" in the present disclosure is a commonly used international clinical acute GVHD severity grading standard (Przepiorka D, Weisdorf D, Martin P, et al. 1994 Consensus Conference on Acute GVHD Grading [J]. Bone Marrow Transplant, 1995, 15 (6): 825-828.), which is formulated based on the impact of acute GVHD on non-relapse-related death after transplantation, and the acute GVHD of the skin, gastrointestinal tract and liver is integrated to form a total grading. The existing modified Glucksberg criteria are shown in Table 1 below.

[0068] Table 1 Modified Glucksberg grading criteria for acute graft-versus-host disease (GVHD)

[0069] The severe acute graft-versus-host disease (GVHD) described in the present disclosure refers to grades III-IV according to the Glucksberg grading standard.

[0070] The term "prediction risk nomogram" (Nomogram) in the present disclosure, also known as the nomogram or alignment diagram, is a quantitative analysis diagram that uses scaled line segments in plane coordinates to represent the functional relationship between multiple variables. The basic principle of the prediction risk nomogram is to construct a multi-factor regression model (such as Cox regression, Logistic regression, etc.), and assign points to each value level of each influencing factor according to the contribution of each influencing factor in the model to the outcome variable (that is, the size of the regression coefficient). Then, these scores are plotted on the same plane according to a certain proportion to form a scaled line segment. When in use, the total score can be obtained by looking up the scores of each variable and adding them together. Finally, the predicted value of the individual outcome event can be calculated through the functional conversion relationship between the total score and the probability of occurrence of the outcome event. In one embodiment of the present disclosure, the obtained aGVHD prediction risk nomogram is as follows: Figure 1 As shown, the prediction risk nomogram alone or in combination with other prediction models constitutes the acute graft-versus-host disease (aGVHD) risk assessment model described in the present disclosure.

[0071] Example:

[0072] In order to enable those skilled in the art to better understand the technical solution of the present invention, the present invention is further described in detail below in conjunction with specific embodiments.

[0073] The research objects in this embodiment are:

[0074] The research subjects in this embodiment are AML patients who achieved complete remission (AML-CR). These patients received HID-HSCT at the Hematology Hospital of the Chinese Academy of Medical Sciences from January 2017 to March 2023. A total of 276 patients were included and divided into a training cohort (165 cases) and a validation cohort (111 cases).

[0075] The primary endpoint of the study in this example is the occurrence of aGVHD, and the secondary endpoint is the cumulative incidence of aGVHD within 100 days. The grading of aGVHD refers to the Glucksberg standard.

[0076] 1. The GBTM model was used to divide the early platelet and hemoglobin trajectories after transplantation into three groups.

[0077] The absolute values ​​of neutrophils, hemoglobin, and platelets were collected within the time range of 0-10 days after transplantation to explore whether patients showed obvious dynamic changes in the early post-transplant period. Population-based trajectory modeling (GBTM) is an unsupervised clustering machine learning model implemented through the "traj" package. We used this approach to characterize the evolution of the above blood cell counts over time and evaluated the association between trajectory group membership and the occurrence of aGVHD. The modeling method is as follows:

[0078] Let y = (y0, y1, y2, . . . , yT ) denote the trajectory observed in time from 0 to T.

[0079] In the context of GBTM, y is modeled as a mixture of multivariate Gaussian distributions as follows:

[0080]

[0081]

[0082] Here, the (T + 1)-dimensional mean vector mk of the kth group is a quadratic function of time t, with different sets of coefficients for different groups:

[0083]

[0084] If you have all the parameters:

[0085]

[0086] Then given an observed sequence Y, the posterior probability that Y belongs to class K can be calculated as follows:

[0087]

[0088] The denominator in (1) is a normalization constant and does not need to be computed explicitly. Instead, one can compute (2) for k = 1, . . . , K and then normalize the result. In (2), wk is incorporated into the exponential component because, in practice, all computations are usually performed within the exponential component.

[0089] result:

[0090] When the vanilla GBTM algorithm was applied directly, the results revealed a model with K = 3 classes, indicating that hemoglobin and platelets can be divided into high, medium, and low trajectory groups (see Figure 2 When the association with the development of aGVHD was examined, no significant differences were observed between the different hemoglobin trajectory groups (see Figure 3 However, although the overall incidence of aGVHD was slightly different between the three platelet trajectory groups (see Figure 4 ), but they showed a significant cumulative incidence of severe aGVHD. Of note, the cumulative incidence of severe aGVHD in the low, medium, and high platelet trajectory groups was 21.4%, 10.9%, and 6.7%, respectively. The platelet trajectory showed a clear trend (see Figure 5 ).

[0091] 2. A two-stage decision tree method revealed the association between platelet trajectory and the occurrence of severe aGVHD.

[0092] Furthermore, this example developed a two-stage decision tree approach to mathematically reconstruct platelet trajectories. This approach produced a better Bayesian Information Criterion (BIC), indicating improved model fit. Based on this new model, it was possible to identify distinct dynamic changes in patients' platelets early after transplantation (see Figure 6 Importantly, statistically significant differences in the incidence of severe aGVHD among the three platelet trajectory groups were observed, which was not evident in the original GBTM model. The modeling approach was as follows:

[0093] Two-Stage group trajectory modeling:

[0094] Step 1: In this initial stage, set K = 2 and use (2) to classify the observed trajectories into one of two groups: not group 3 (k = 1) or group 3 (k = 2).

[0095] The model parameters at this stage are as follows:

[0096]

[0097]

[0098]

[0099] Step 2: For the trajectories assigned to the “non-Group 3” category in Step 1, they are further divided into two subgroups using (2): Group 1 (k = 1) or Group 2 (k = 2).

[0100] The model parameters at this stage are as follows:

[0101]

[0102]

[0103]

[0104] The probability of an observed trajectory belonging to these three groups can be calculated as follows:

[0105]

[0106] result:

[0107] To validate the above findings, the patient cohort was divided into a training cohort and a testing cohort. In the training cohort consisting of 165 patients, 30 patients developed severe aGVHD. The incidence of individual severe aGVHD was significantly higher in the low platelet trajectory group compared with the high platelet trajectory group (28.6% and 4.9%, respectively, both P values ​​< 0.001), while no significant differences were observed in the other aGVHD groups. In addition, among the 30 patients with severe aGVHD, 18 patients (60%) showed a low platelet trajectory in the early post-transplant period, while 2 patients (6.7%) showed a high platelet trajectory. To further confirm these findings, patients from the validation cohort were included in the two-stage trajectory modeling. The analysis revealed three different platelet trajectories in the early post-transplant period. Consistently, patients in the low platelet trajectory group showed a higher incidence and proportion of severe aGVHD compared with patients in the high platelet trajectory group. These findings highlight the utility of a two-stage decision tree approach to identify meaningful platelet trajectories and their association with the risk of severe aGVHD after HSCT.

[0108] 3. Develop a network calculator based on the above method.

[0109] 4. Acute graft-versus-host disease (aGVHD) risk assessment model.

[0110] 1) Demonstrate that platelet trajectory is an independent risk factor for aGVHD.

[0111] The unique clinical results associated with platelet trajectory suggest that platelet trajectory has potential utility as a risk factor for the development of aGVHD. To assess this possibility, a COX proportional hazard regression model was established in this example. Considering competing risks, the cumulative incidence of aGVHD was estimated using the Fine-Gray test, with relapse and death within 100 days as competing events. Statistical significance was defined at the 0.05 level.

[0112] In the factor regression analysis of this example, factors that were previously reported to be clinically relevant or to show univariate association with the outcome were included in the multivariate Cox proportional hazards regression model. Given the limited number of available events, great caution was exercised in variable selection to ensure the simplicity of the final model. The results of univariate analysis showed that platelet trajectory, rather than other factors, was a risk factor for the development of aGVHD, especially severe aGVHD (P=0.001, HR [95%CI]=0.403 [0.231-0.702]). When platelet trajectory and previously reported factors were included in multivariate analysis, platelet trajectory remained an independent risk factor for aGVHD, especially severe aGVHD, even after adjusting for other factors (P=0.004, HR [95%CI] = 0.454 [0.257−0.771]). HR [95%CI] = 0.312 [0.130−0.752]). Further analysis of the cumulative incidence of aGVHD, especially severe aGVHD, in the training and validation groups showed that patients in the low platelet trajectory group exhibited a higher incidence of severe aGVHD at 100 days (28.6% vs 25.8%), while patients in the high trajectory group had significantly lower incidences, 4.9% vs 4.5%, respectively, in both cohorts (see Figure 7 ), the incidence of severe aGVHD was significantly different between the two trajectory groups (P=0.009, P=0.011). Interestingly, the platelet trajectory was found to be unrelated to the occurrence of overall aGVHD. Therefore, the low platelet trajectory in the early post-transplant period is a risk factor for the development of aGVHD, especially severe aGVHD, but not for overall aGVHD.

[0113] 2) Demonstrate that platelet trajectory combined with other factors can better predict the occurrence of severe aGVHD.

[0114] This example attempts to determine whether combining platelet trajectory with other factors associated with aGVHD can enhance the ability to predict the onset of aGVHD, especially severe aGVHD. The analysis in this example involves evaluating the cumulative incidence of aGVHD by combining these comprehensive factors. In this example, risk factors for the occurrence of aGVHD reported in the previous literature were selected, multifactorial nomogram modeling was performed, and weights were assigned to re-define the risk population. The results of the study in this example produced a noteworthy observation: when combined with a low platelet trajectory, the incidence of severe aGVHD can be significantly increased ( Figure 8 ). This highlights the complexity of predicting aGVHD, particularly severe aGVHD, and suggests that the predictive value derived from platelet trajectories may itself be robust.

[0115] 3) How to establish a risk prediction nomogram.

[0116] See Figure 1 .in,

[0117] The first line is the score scale, with a range of 0-100; the second line is age, with a range of 5-65 years old;

[0118] The third row was transplant type, including PB+BM (peripheral blood + bone marrow), PB (peripheral blood);

[0119] The fourth line is the gender of the donor and recipient, including other (non-female donor to male), female for male (female donor to male);

[0120] The fifth line is the prevention regimen for acute aGVDH: including FK506 (tacrolimus) and other (non-tacrolimus);

[0121] The sixth row is the total number of mononuclear cells, with a value range of 2-30;

[0122] The seventh row is the number of CD34+ cells, with a value range of 0-18;

[0123] The eighth row is the platelet trajectory, including 1 (low trajectory group), 2 (middle trajectory group), and 3 (high trajectory group);

[0124] The ninth line is the total score, which is the sum of the scores from the second to the eighth line, ranging from 0 to 350;

[0125] The tenth row is the risk prediction value of severe aGVHD, which corresponds to the ninth row.

[0126] 4) Construct a risk assessment model for acute graft-versus-host disease (aGVHD).

[0127] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. Use of a platelet trajectory calculator or platelet trajectory model in constructing an acute graft-versus-host disease risk assessment model.

2. The use according to claim 1, characterized in that The platelet trajectory is the platelet trajectory of the patient from day 0 to before the occurrence of acute graft-versus-host disease after allogeneic hematopoietic stem cell transplantation.

3. The use according to claim 2, characterized in that: The platelet trajectory is the platelet count trajectory of the patient from day 0 to day 10 after allogeneic hematopoietic stem cell transplantation.

4. The use according to any one of claims 1 to 3, characterized in that: The acute graft-versus-host disease is produced after patients with acute myeloid leukemia are treated with allogeneic hematopoietic stem cell transplantation.

5. The use according to claim 1, characterized in that: The acute graft-versus-host disease is severe acute graft-versus-host disease, and the severe acute graft-versus-host disease is grade 3-4 according to the Glucksberg standard.

6. The use according to claim 1, characterized in that: The method for constructing the platelet trajectory calculator or the platelet trajectory model comprises the following steps: S1: The platelet trajectories are divided into three groups using the GBTM model, which are low platelet count trajectory group, medium platelet count trajectory group and high platelet count trajectory group. The calculation of y belongs to The posterior probability of the class is as follows: ， in, =3, y = (y0, y1, y2, . . . , yT ) represents the trajectory observed from time 0 to T, m k is a quadratic function of time t, with different sets of coefficients for different groups: , are all the parameters that have; S2: Reconstruct the model obtained in S1 using a two-stage decision tree method: First, in the initial stage, set = 2 and use formula (2) in S1 to classify the observed trajectories into one of the following two groups: =1 or Group 3 =2, Then in the second stage, for the trajectories assigned to the non-group 3 categories in the initial stage, they are further divided into two subgroups using formula (2) in S1: Group 1 =1 or Group 2 =2; S3: The probability of obtaining the three groups of observation trajectories is calculated as follows:

7. The use according to claim 6, characterized in that The model parameters in the initial stage of S2 are: ； ； ； The model parameters of the second stage in S2 are: ； ； 。 8. A method for constructing an acute graft-versus-host disease risk assessment model, characterized in that: include: Step 1) Collect the patient’s clinical data; Step 2) determining predictive variables of the risk assessment model, wherein the predictive variables include platelet trajectory; Step 3) establishing a GVHD prediction risk nomogram using the predictor variables described in step 2); Step 4) constructing the acute graft-versus-host disease risk assessment model using the aGVHD prediction risk nomogram obtained in step 3).

9. The construction method according to claim 8, characterized in that: The platelet trajectory is an independent risk factor in the predicted risk nomogram.

10. The construction method according to claim 8, characterized in that: The predictor variables in step 2) also include known aGVHD risk factors, which are selected from patient age, transplant type, donor-recipient gender pair, aGVHD prevention regimen that the patient has been administered, total mononuclear cell count or CD34 + One or more of a number of cells.

11. The construction method according to claim 8, 9 or 10, characterized in that: The acute graft-versus-host disease is produced after patients with acute myeloid leukemia are treated with allogeneic hematopoietic stem cell transplantation.

12. Use of a platelet trajectory calculator or a platelet trajectory model in the preparation of a product for evaluating the prognosis of treatment of acute myeloid leukemia.

13. A device for assessing the risk of acute graft-versus-host disease, characterized in that: The device comprises: a memory, a processor, and a computer program stored in the memory and executable on the processor; When the computer program is executed by the processor, the acute graft-versus-host disease risk assessment model constructed by the construction method described in any one of claims 8 to 11 is run.

14. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores an acute graft-versus-host disease risk assessment model program constructed by the construction method as described in any one of claims 8 to 11, and when the program is executed by the processor, the function of the device as described in claim 13 is realized.

Images