Kit and system for identifying the pathogenic cause of diarrhea after allogeneic hematopoietic stem cell transplantation

Abstract

The application provides a method for differentiating pathogenic causes of diarrhea after allogeneic hematopoietic stem cell transplantation, and specifically provides application of a reagent for detecting a combination of biological indexes of an individual in preparation of a product for differentiating pathogenic causes of diarrhea after allogeneic hematopoietic stem cell transplantation, wherein the biological indexes include: absolute count of NK cells, CD4 + CD28 + T cells account for CD4 + T lymphocyte percentage; and the detection time point of the biological indexes is within seven days before the diarrhea symptom appears. The method for differentiating pathogenic causes of diarrhea after allogeneic hematopoietic stem cell transplantation provided by the application realizes minimally invasive, low-risk, early and accurate pathogenic cause differentiation by detecting peripheral blood samples and constructing a detection model in combination with multi-time point data.

Description

Technical Field

[0001] This invention relates to the field of medical testing and clinical diagnostic technology, and to a method for differential diagnosis of the causes of diarrhea after allogeneic hematopoietic stem cell transplantation, specifically a method for differential diagnosis of the causes of diarrhea after allogeneic hematopoietic stem cell transplantation based on dynamic monitoring of immune cell subsets and cytokines. Background Technology

[0002] Allogeneic hematopoietic stem cell transplantation (allo-HSCT) is a key treatment for various high-risk hematological malignancies (such as acute myeloid leukemia (AML), acute lymphoblastic leukemia (ALL), and myelodysplastic syndromes (MDS)) and non-malignant hematological diseases (such as aplastic anemia (AA)). However, post-HSCT diarrhea is an extremely common complication after allo-HSCT, with an incidence rate as high as 40%-91%.

[0003] The pathogenesis of post-HSCT diarrhea is complex, with gastrointestinal graft-versus-host disease (GI-GVHD) and enterovirus infections (such as cytomegalovirus (CMV), Epstein-Barr virus (EBV), human herpesvirus 6 (HHV-6), and adenovirus infections) being the two main causes. Although GI-GVHD and viral infections share highly similar clinical symptoms (such as watery stools, bloody stools, abdominal pain, nausea, and vomiting), and there are no differences between the two groups in terms of diarrhea onset time, duration, frequency, total amount of diarrhea, and bloody stools, their treatment strategies are completely opposite: GI-GVHD requires intensive immunosuppressive therapy to suppress the graft's attack on host tissues, while viral infections require reduced immunosuppressant dosage and targeted antiviral therapy. Therefore, early and accurate identification of the cause of diarrhea is crucial for guiding clinical treatment and improving patient prognosis.

[0004] Currently, the "gold standard" for clinically differentiating GI-GVHD from viral infection-related diarrhea is intestinal endoscopic biopsy combined with pathological examination. However, this method has significant limitations: First, obtaining intestinal tissue samples is difficult, especially for patients who are weak after transplantation and cannot tolerate endoscopy, making secondary sampling difficult; second, the pathological features of early or mild GI-GVHD (such as epithelial cell apoptosis and crypt destruction) are atypical, easily leading to misdiagnosis; third, both present with ulcers and erosions endoscopically, and viral infections (such as CMV enteritis) can also cause intestinal epithelial apoptosis, overlapping with the pathological features of GI-GVHD, easily leading to misdiagnosis; fourth, the pathological diagnosis is highly dependent on the pathologist's experience, highly subjective, and difficult to standardize. In clinical practice, the initial clinical thought for post-transplant diarrhea is often GVHD, with viral infections often misdiagnosed as GVHD, and in rare cases, GVHD misdiagnosed as an infection. These reasons lead to incorrect differential diagnoses of post-transplant diarrhea.

[0005] Therefore, there is an urgent need in this field for a less invasive, lower-risk, earlier, more accurate, and standardized method for the differential diagnosis of the causes of post-transplant diarrhea, in order to make up for the shortcomings of existing pathological biopsy techniques and meet the clinical demand for precision diagnosis and treatment. Summary of the Invention

[0006] To address the aforementioned problems, the present invention aims to provide a method for differential diagnosis of the causes of diarrhea after allogeneic hematopoietic stem cell transplantation based on dynamic monitoring of immune cell subsets and cytokines. This method utilizes peripheral blood sample testing and multi-timepoint dynamic monitoring data of immune cell subsets and cytokines to construct a detection model, enabling minimally invasive, low-risk, early, and accurate identification of the causes of diarrhea following allogeneic hematopoietic stem cell transplantation (allo-HSCT) caused by gastrointestinal graft-versus-host disease (GI-GVHD) and enterovirus infection.

[0007] Specifically, on the one hand, the present invention provides the application of a reagent for detecting individual biomarkers in the preparation of products for identifying the pathogenic causes of diarrhea after allogeneic hematopoietic stem cell transplantation, wherein the biomarkers include:

[0008] absolute NK cell count, CD4 + CD28 + T cells account for CD4+ + Percentage of T lymphocytes;

[0009] The biomarkers were detected within seven days before the onset of diarrhea symptoms.

[0010] According to a specific embodiment of the present invention, the biomarker further includes the following grouping biomarkers:

[0011] Group 1: CD3 levels detected within 7 days prior to the onset of diarrhea symptoms + CD8 + Percentage of T cells among lymphocytes, CD4 + Type 2 helper T cells account for a significant portion of CD4+. + Percentage of T lymphocytes;

[0012] Second group: TNFR1 levels and absolute NK cell counts were measured within 4 weeks after allo-HSCT;

[0013] Group 3: GM-CSF levels, absolute NK cell count, and CD3 count measured within 5 weeks after allo-HSCT. + CD38 + T occupies CD3 + T lymphocyte percentage; This invention involves a series of monitoring sessions for patients according to a pre-set, fixed time schedule, rather than single tests performed only when symptoms appear. Specifically, immunophenotypic analysis is conducted after leukocyte engraftment, following our center's standardized follow-up protocol: weekly monitoring within 3 months post-transplantation; bi-weekly monitoring from 3 to 6 months post-transplantation; and monthly monitoring beyond 6 months post-transplantation. This prospective, fixed-interval, intensive monitoring strategy allows us to capture the dynamic process of post-transplant immune reconstitution and provides a systematic data foundation for analyzing changes in immune markers before and after the onset of diarrhea.

[0014] Specifically, the testing time point is every Monday. If diarrhea occurs on any day of the following week, the results of the previous Monday will be reviewed (i.e., the test results at the time point 7 days before the onset of diarrhea symptoms).

[0015] Since diarrhea develops post-transplant, the 7-day timeframe before the onset of diarrhea symptoms may overlap with the 4th or 5th week after allo-HSCT. In such cases, the data is normally included in the statistics; that is, the data for "7 days before the onset of diarrhea symptoms" and "4th or 5th week after allo-HSCT" are analyzed using the same duplicate dataset. Furthermore, in actual cases, the number of cases where the 7-day timeframe before the onset of diarrhea symptoms overlaps with the 4th or 5th week after allo-HSCT is extremely small, and its impact on the statistical analysis is negligible.

[0016] The early time points selected in this invention accurately reflect the patient's immune status. In the later stages of transplantation, especially after the onset of diarrhea, the use of anti-GVHD, anti-infective, or hormonal drugs can significantly interfere with the results of immune cell subset analysis. In the early stages of transplantation, such as before the third week, immune reconstitution has just begun, and the number of cells observed is relatively small, making the data unreliable to reflect the true cellular status. Therefore, this invention selects three key time points—7 days before the onset of diarrhea, 4 weeks after allo-HSCT, and 5 weeks after allo-HSCT—to achieve dynamic monitoring.

[0017] According to a specific embodiment of the present invention, the causes of diarrhea include gastrointestinal graft-versus-host disease and / or enterovirus infection.

[0018] According to a specific embodiment of the present invention, preferably, the enterovirus infection includes one or a combination of two or more of cytomegalovirus (CMV) infection, Epstein-Barr virus (EBV) infection, human herpesvirus 6 (HHV-6) infection, and adenovirus infection.

[0019] According to a specific embodiment of the present invention, the test sample of the individual being tested includes a peripheral blood sample.

[0020] According to a specific embodiment of the present invention, the pathogenic factors of diarrhea after allogeneic hematopoietic stem cell transplantation are identified based on individual biomarkers;

[0021] The detection data of the aforementioned biomarkers are calculated according to the following steps:

[0022] Step 1: Calculate the linear detection value (logit value):

[0023] z1 = 0.6162 + 0.0141 × A - 0.0270 × B

[0024] Where A represents the absolute NK cell count measured within seven days prior to the onset of diarrhea symptoms, and B represents the CD4 count measured within seven days prior to the onset of diarrhea symptoms. + CD28 + T cells account for CD4+ + Percentage of T lymphocytes;

[0025] Step 2: Convert the logit value to a detection probability, where e is a constant, e = 2.71828.

[0026]

[0027] When the detection probability P1 > 0.408, the cause of diarrhea is determined to be enterovirus infection; when the detection probability P1 ≤ 0.408, the cause of diarrhea is determined to be gastrointestinal graft-versus-host disease.

[0028] According to a specific embodiment of the present invention, the equation relating the detection data of the biomarker to P1 is as follows:

[0029]

[0030] According to a specific embodiment of the present invention, the pathogenic factors of diarrhea after allogeneic hematopoietic stem cell transplantation are identified based on the detection data of individual biomarkers;

[0031] The detection data of the aforementioned biomarkers are calculated according to the following steps:

[0032] The test score is calculated as follows: 2.6714 + A×(-0.0116) + B×0.0049 + C×(-0.0217) + D×(-0.0256) + E×(-0.0005) + F×(0.0006) + G×(-1.6570) + H×(0.0038) + I×0.0243;

[0033] Where A represents the absolute NK cell count measured within seven days prior to the onset of diarrhea symptoms, and B represents the CD4 count measured within seven days prior to the onset of diarrhea symptoms. + CD28 + T cells account for CD4+ + T lymphocyte percentage; C represents CD4 count measured within 7 days prior to the onset of diarrhea symptoms. + Type 2 helper T cells account for a significant portion of CD4+. + T lymphocyte percentage, and D represents CD3 count measured within 7 days prior to the onset of diarrhea symptoms. + CD8 + The percentage of T cells among lymphocytes; E represents the TNFR1 level measured within 4 weeks after allo-HSCT; F represents the absolute NK cell count measured within 4 weeks after allo-HSCT; G represents the GM-CSF level measured within 5 weeks after allo-HSCT; H represents the absolute NK cell count measured within 5 weeks after allo-HSCT; and I represents the CD3+ level measured within 5 weeks after allo-HSCT. + CD38 + T occupies CD3 + Percentage of T lymphocytes.

[0034] Step 2: Calculate the linear detection value (logit value), where e is a constant, e = 2.71828;

[0035]

[0036] Step 3: Convert the logit value into a detection probability.

[0037]

[0038] When the detection probability P2 > 0.379, the cause of diarrhea is determined to be enterovirus infection; when the detection probability P2 ≤ 0.379, the cause of diarrhea is determined to be gastrointestinal graft-versus-host disease.

[0039] According to a specific embodiment of the present invention, the equation relating the detection data of the biomarker to P2 is as follows:

[0040]

[0041] Examples of units for each biomarker in the above model formula are as follows: NK absolute count (cells / μL), TNFR1 (pg / mL), GM-CSF (pg / mL).

[0042] On the other hand, the present invention provides a kit comprising reagents for detecting a combination of biomarkers of an individual in the above applications.

[0043] According to a specific embodiment of the present invention, the reagent includes an antibody and a reagent for detecting cytokine levels;

[0044] The antibodies comprise four groups of antibodies:

[0045] The first group of antibodies includes fluorescently labeled CD3, CD56, CD16, CD45, CD4, CD19, and CD8 antibodies. The fluorescent labeling order of each antibody is FITC, PE, PE, PerCP-Cy5.5, PE-Cy7, APC, and APC-Cy7. These antibodies are added to the flow cytometry absolute counting tube in which the sample to be tested is in a single-cell suspension state.

[0046] The second group of antibodies includes fluorescently labeled CD183 antibody, CD3 antibody, CD4 antibody, CD196 antibody and CD8 antibody. The fluorescent labeling order of each antibody is PE, PerCP-Cy5.5, PE-Cy7, APC and APC-Cy7. They are used to add to the flow cytometry tube 2 where the sample to be tested is in a single-cell suspension state.

[0047] The third group of antibodies includes fluorescently labeled CD8, CD38, CD3, CD4, and CD28 antibodies. The fluorescent labeling order of each antibody is FITC, PE, PerCP-Cy5.5, PE-Cy7, and APC. These antibodies are added to flow cytometry tubes in which the test sample is in a single-cell suspension state.

[0048] The reagents used to detect cytokine levels include:

[0049] A reagent for detecting TNFR1 and GM-CSF levels at the protein level.

[0050] According to a specific embodiment of the present invention, the first group of antibodies is used to detect CD8+T%, absolute NK cell count.

[0051] According to a specific embodiment of the present invention, the second group of antibodies is used to detect CD4. + Th2 / CD4 + T%, CD4 + CD28 + / CD4 + T%.

[0052] According to a specific embodiment of the present invention, the second group of antibodies is used to detect CD4. + CD28 + / CD4 + T%, CD3 + CD38 + T%.

[0053] According to a specific embodiment of the present invention, the kit further includes one or more of the following: cell lysis buffer, buffer solution, and flow cytometry tubes for use with a flow cytometer.

[0054] According to a specific embodiment of the present invention, the reagent includes a cell preservation solution for preserving individual peripheral blood samples to obtain a cell suspension suitable for the detection of the biomarkers; each 100 ml of the cell preservation solution contains 1-10 mL of compound electrolyte glucose injection, 1-10 mL of human serum albumin injection, 1-10 mL of compound amino acid injection, 10-20 μM salvianolic acid B, 20-60 μM rhodioloside, 1-10 μM ginkgo biloba extract, and the compound electrolyte injection is used to make up the difference.

[0055] According to a specific embodiment of the present invention, each 100ml of the cell protection solution contains 3-8mL of compound electrolyte glucose injection, 3-8mL of human serum albumin injection, 3-8mL of compound amino acid injection, 10-20μM salvianolic acid B, 20-60μM rhodioloside, 1-10μM ginkgo biloba extract, and the compound electrolyte injection is used to make up the difference.

[0056] According to a specific embodiment of the present invention, each 100ml of the cell protection solution contains 5mL of compound electrolyte glucose injection, 5mL of human serum albumin injection, 5mL of compound amino acid injection, 10-20μM salvianolic acid B, 20-60μM rhodioloside, 1-10μM ginkgo biloba extract, and the compound electrolyte injection is used to make up the difference.

[0057] According to a specific embodiment of the present invention, each 100ml of the cell protection solution contains 5mL of compound electrolyte glucose injection, 5mL of human serum albumin injection, 5mL of compound amino acid injection, 15-20μM salvianolic acid B, 40-60μM rhodioloside, 5-10μM ginkgo biloba extract, and the compound electrolyte injection is used to make up the difference.

[0058] According to a specific embodiment of the present invention, preferably, the cell protection solution further contains 10-50 μM hesperidin and 20-50 μM total saponins of Panax notoginseng.

[0059] According to a specific embodiment of the present invention, preferably, the cell protection solution further contains 30-50 μM hesperidin and 30-50 μM total saponins of Panax notoginseng.

[0060] According to a specific embodiment of the present invention, preferably, the compound amino acid injection is compound amino acid injection 18AA-II.

[0061] According to a specific embodiment of the present invention, preferably, the pH of the cell protection solution is 6-8, and the osmolar concentration is 270-400 Osmol, more preferably 300-380 Osmol.

[0062] In a specific embodiment of the present invention, the method for preparing the cell protection solution includes:

[0063] The salvianolic acid B, rhodioloside, and ginkgo biflavonoids were each prepared into solutions. The resulting solutions were then mixed with compound electrolyte injection, compound electrolyte glucose injection, human serum albumin injection, and compound amino acid injection in proportion to the final concentration.

[0064] On the other hand, the present invention provides a method for processing peripheral blood samples, the method comprising the following steps:

[0065] The peripheral blood sample and cell preservation solution are thoroughly mixed, with the volume ratio of peripheral blood sample to cell preservation solution being 1:4 to 1:9.

[0066] This invention's cell preservation solution addresses the excessive free radicals (such as reactive oxygen species (ROS) and reactive nitrogen species (RNS) during cryopreservation. It establishes an antioxidant defense system at its core, using free radical removal and antioxidation as entry points. The solution adds salvianolic acid B as an antioxidant to rapidly scavenge free radicals in the cytoplasm, stabilize mitochondria, directly inhibit the Caspase cascade, and quickly block apoptosis. Ginkgo biloba extract is added as a membrane stabilizer, utilizing its lipid solubility to specifically scavenge lipid peroxidation free radicals while activating the Nrf2 / ARE pathway, upregulating the cell's own antioxidant enzyme (SOD), and enhancing cellular oxidative stress capacity. Rhodiola rosea extract is added as a cell homeostasis agent to activate the PI3K / Akt and ERK survival pathways, improving cell viability, and regulating the AMPK pathway to maintain cell homeostasis. These three components synergistically enhance cellular antioxidant capacity, maintain biological activity, and extend cell preservation time. The addition of tangeretin as a RORα / γ agonist maintains cellular homeostasis, improves mitochondrial energy supply, and enhances stress resistance during cell preservation. The addition of total saponins from Panax notoginseng acts as a calcium homeostasis regulator, maintaining intracellular calcium ion homeostasis. Simultaneously, the saponin components bind to membrane cholesterol, further enhancing membrane stability, ensuring cell membrane integrity, and regulating osmotic pressure. Both tangeretin and total saponins from Panax notoginseng work synergistically to reduce the risk of thrombosis during clinical use. Cells preserved in the cell preservation solution of this invention exhibit stable pH and osmotic pressure, high cell viability, no decrease in cell surface marker expression, intact cell function, and unaffected cell quality.

[0067] The herbal extracts selected in this invention have the following advantages compared to synthetic or other natural antioxidants: ① Complex composition: This invention uses the main components of multiple herbal extracts, which work synergistically through different mechanisms (tanshinone B scavenging cytoplasmic free radicals, ginkgo biloba flavonoids scavenging lipid free radicals, and rhodioloside regulating cellular homeostasis) to form a more comprehensive antioxidant network; ② Multi-faceted combined effect: The antithrombotic ability of hesperidin combined with the vasodilatory effect of Panax notoginseng saponins in this invention results in a stronger antithrombotic ability; ③ Low side effects: Herbal antioxidants have high safety with long-term use, while some synthetic antioxidants (such as BHA / BHT) may have potential hepatotoxicity or carcinogenic controversies. Moreover, the combination of components in this invention focuses more on regulating metabolism and restoring homeostasis, which is more in line with the biological laws of such active substances in cells.

[0068] On the other hand, the present invention provides a method for identifying the pathogenic cause of diarrhea after allogeneic hematopoietic stem cell transplantation, comprising the following steps:

[0069] (a) Obtain peripheral blood samples from individuals at multiple time points after transplantation;

[0070] (b) Detect a set of biological indicators in the sample to obtain detection data of the biological indicators; the biological indicators are the combination of the above-mentioned biological indicators;

[0071] (c) Input the detection data of the biomarkers obtained in step (b) into the detection model to obtain the detection probability of the individual developing gastrointestinal graft-versus-host disease or enterovirus infection, thereby realizing the identification of the cause of diarrhea.

[0072] According to a specific embodiment of the present invention, preferably, the biomarkers are detected by flow cytometry in step (b).

[0073] According to a specific embodiment of the present invention, preferably, the detection model is constructed in the following manner:

[0074] Training set sample data were collected from patients with pathologically confirmed gastrointestinal graft-versus-host disease and enterovirus infection.

[0075] The above combination of bioindicators was obtained by screening key detection variables from candidate detection variables through Lasso regression combined with logistic regression.

[0076] A logistic regression detection model was constructed based on the combination of the key biological indicators.

[0077] On the other hand, the present invention provides a system for identifying the pathogenic factors of diarrhea after allogeneic hematopoietic stem cell transplantation, wherein the system comprises:

[0078] The data acquisition module is used to acquire detection data of a set of biological indicators in peripheral blood samples of an individual at multiple time points after allogeneic hematopoietic stem cell transplantation, wherein the biological indicators are the combination of the above-mentioned biological indicators.

[0079] The processing and analysis module stores a detection model, which is used to input the detection data of the biomarkers into the detection model and output the detection probability of the individual developing gastrointestinal graft-versus-host disease or enterovirus infection.

[0080] On the other hand, the present invention provides a non-transitory computer-readable storage medium storing a computer program that, when executed by a processor, enables the above-described system for identifying the cause of diarrhea following allogeneic hematopoietic stem cell transplantation.

[0081] On the other hand, the present invention provides a computer program product comprising computer instructions that, when executed by a processor, enable the system described above to identify the pathogenic causes of diarrhea following allogeneic hematopoietic stem cell transplantation.

[0082] Compared with the prior art, the present invention has the following beneficial effects:

[0083] 1. Minimally invasive and low-risk: Only peripheral blood samples need to be collected, and there is no need to perform intestinal endoscopic biopsy, which avoids the trauma of endoscopic examination to weak patients after transplantation and improves patient compliance;

[0084] 2. Early diagnosis time: Monitoring and pathogen detection can begin immediately after transplantation. Compared with the traditional method that relies on pathological biopsy after diarrhea, this method can make a diagnosis earlier, clarify the pathogenic cause, and provide a window of opportunity for early intervention of the disease.

[0085] 3. High accuracy: The multi-time point model has an AUC of 0.908, a sensitivity of 87.5%, and a specificity of 79.4%; the single-time point model has an AUC of 0.799 and a specificity of 88.2%, which are significantly better than traditional pathological biopsy (misdiagnosis rate of 30%-40%). Moreover, the results can be standardized and are not affected by the doctor's experience.

[0086] 4. Advantages of dynamic monitoring: By constructing a model based on dynamic data from three time points—7 days before the onset of diarrhea symptoms, 4 weeks after allo-HSCT, and 5 weeks after allo-HSCT—the model can better reflect the dynamic changes in the patient's immune status compared to single-time-point detection, and the identification results are more stable and reliable.

[0087] 5. High clinical guidance value: Clear identification of pathogenic causes can directly guide the selection of clinical treatment plans (e.g., GI-GVHD requires enhanced immunosuppression, viral infections require antiviral treatment and reduction of immunosuppressants). The correct treatment direction can reduce patient complications and improve patient prognosis. Attached Figure Description

[0088] Figure 1 Box plots showing the visualization results of univariate differences between the GVHD group and the viral infection group 7 days before the onset of diarrhea symptoms.

[0089] Figure 2 The effect size forest plot is a visualization of the univariate differences between the GVHD group and the viral infection group 7 days before the onset of diarrhea symptoms.

[0090] Figure 3 Box plot of the visualization results of univariate differences between the GVHD group and the virus infection group at week 4 after allo-HSCT.

[0091] Figure 4 The effect size forest plot is the visualization result of the univariate difference index between the GVHD group and the virus infection group at week 4 after allo-HSCT.

[0092] Figure 5 Box plots showing the visualization results of univariate differences between the GVHD group and the virus infection group at week 5 after allo-HSCT.

[0093] Figure 6 The effect size forest plot is a visualization of the univariate difference indices between the GVHD group and the virus infection group at week 5 after allo-HSCT.

[0094] Figure 7A The coefficient path diagram is shown in the Lasso regression screening results for indicators in the 7 days before the onset of diarrhea symptoms.

[0095] Figure 7B The cross-validation error curve is shown in the Lasso regression screening results for indicators in the 7 days before the onset of diarrhea symptoms.

[0096] Figure 8A This is a nomogram showing the performance of a bivariate detection model at a single time point (7 days before the onset of diarrhea symptoms).

[0097] Figure 8B This is the calibration curve for the performance of the bivariate detection model at a single time point (7 days before the onset of diarrhea symptoms).

[0098] Figure 8C The decision curve represents the performance of the bivariate detection model at a single time point (7 days before the onset of diarrhea symptoms).

[0099] Figure 8D The ROC curve represents the performance of the bivariate detection model at a single time point (7 days before the onset of diarrhea symptoms).

[0100] Figure 9A This is a graph showing the coefficient path in the Lasso regression results for variables detected at multiple time points.

[0101] Figure 9B This is the cross-validation error curve in the Lasso regression results for variables detected at multiple time points.

[0102] Figure 10A This is a nomogram showing the performance of the multi-time-point detection model.

[0103] Figure 10B This is the calibration curve for the performance of the multi-time-point detection model.

[0104] Figure 10C This is the decision curve in the performance of the multi-time-point detection model.

[0105] Figure 10D The ROC curve is used to measure the performance of the multi-time point detection model.

[0106] Figure 11 To verify the ROC curve in the performance of the centralized multi-time point detection model. Detailed Implementation

[0107] In order to provide a clearer understanding of the technical features, objectives and beneficial effects of the present invention, the technical solution of the present invention will now be described in detail below, but it should not be construed as limiting the scope of implementation of the present invention.

[0108] In some embodiments of the present invention, a model for identifying the pathogenic factors of diarrhea after allogeneic hematopoietic stem cell transplantation and a method for constructing the same are provided, as detailed below:

[0109] 1. Inclusion and grouping of study subjects

[0110] Patients who developed diarrhea after allo-HSCT and had a confirmed pathological diagnosis via intestinal endoscopic biopsy were retrospectively or prospectively included. Based on the pathological results, they were divided into two groups:

[0111] GVHD group: Patients whose pathological diagnosis met the 2015 National Institutes of Health (NIH) consensus guidelines on gastrointestinal GVHD histology and the Lerner classification criteria and were diagnosed with GVHD.

[0112] Viral infection group: Patients whose pathological diagnosis or laboratory tests (such as immunohistochemistry, in situ hybridization, PCR) confirm enterovirus infection, including CMV enteritis, EBV enteritis, CMV combined with EBV enteritis, HHV-6 enteritis, adenovirus enteritis, etc.

[0113] Diarrhea is defined as: having more than 3 bowel movements per day, with a stool volume of more than 200 grams per day, and the stool is loose and has a water content of >85%; bloody stool is defined as stool containing blood, which is bright red, dark red or black in color.

[0114] Sample collection and testing time points

[0115] Peripheral blood samples were collected from patients, and the testing was conducted at three key time points to achieve dynamic monitoring:

[0116] Time point 1: 7 days before the onset of diarrhea symptoms; Time point 2: 4 weeks after allo-HSCT; Time point 3: 5 weeks after allo-HSCT.

[0117] This embodiment involves a series of monitoring sessions for patients according to a pre-set, fixed time schedule, rather than performing single tests only when symptoms appear. Specifically, immunophenotypic analysis is conducted after leukocyte engraftment, following the center's standardized follow-up protocol: weekly monitoring within 3 months post-transplantation; bi-weekly monitoring from 3 to 6 months post-transplantation; and monthly monitoring after 6 months post-transplantation. This prospective, fixed-interval, intensive monitoring strategy allows us to capture the dynamic process of post-transplant immune reconstitution and provides a systematic data foundation for analyzing changes in immune markers before and after the onset of diarrhea.

[0118] Specifically, the testing time point is every Monday. If diarrhea occurs on any day of the following week, the results of the previous Monday will be reviewed (i.e., the test results at the time point 7 days before the onset of diarrhea symptoms).

[0119] Since leukocyte engraftment after stem cell transplantation usually occurs more than 14 days post-transplantation, diarrhea requiring differential diagnosis typically develops after 14 days. Therefore, the 7-day window before the onset of diarrhea symptoms may overlap with the 4th or 5th week post-allo-HSCT. In such cases, the data is normally included in the statistics; that is, the data for "7 days before the onset of diarrhea symptoms" and "4th or 5th week post-allo-HSCT" are analyzed using the same duplicate dataset. Furthermore, in actual cases, the number of cases where the 7-day window before the onset of diarrhea symptoms overlaps with the 4th or 5th week post-allo-HSCT is extremely small, and its impact on the statistical analysis is negligible.

[0120] The early time points selected in this embodiment accurately reflect the patient's immune status. Later in the transplantation process, especially after the onset of diarrhea, the use of anti-GVHD, anti-infective, or hormonal drugs can significantly interfere with the results of immune cell subset analysis. In the early stages of transplantation, such as before the third week, immune reconstitution has just begun, and the number of cells observed is relatively small, making the data unreliable to reflect the true cellular status. Therefore, this embodiment selects three key time points—7 days before the onset of diarrhea, 4 weeks after allo-HSCT, and 5 weeks after allo-HSCT—to achieve dynamic monitoring.

[0121] Detection indicators and detection methods

[0122] (1) Detection of immune cell subsets

[0123] The following immune cell subsets in peripheral blood were detected by flow cytometry using a BD Cantoplus flow cytometer (BD Biosciences, Inc., USA) with the accompanying Diva 8.0.2 analysis software. Antibody reagents, hemolysin (10x concentrated, catalog number 349202), and absolute counting tubes (catalog number 340334) were all products of BD Biosciences, Inc., USA.

[0124] T cell-related subsets: CD3 + The percentage of T cells among lymphocytes (CD3) + T% and CD3 + CD8 + Percentage of T cells among lymphocytes (CD8) + T%), initial CD4 + T cells account for CD4+ + T lymphocyte percentage (Naive CD4)+ / CD4 + T% and CD4 + Type 2 helper T cells account for a significant portion of CD4+. + Percentage of T lymphocytes (CD4) + Th2 / CD4 + T% and CD4 + CD28 + T cells account for CD4+ + Percentage of T lymphocytes (CD4) + CD28 + / CD4 + T% and CD3 + CD38 + absolute T cell count, CD8 + CD28 + CD38 + T cells account for a significant portion of CD8. + CD28 + Percentage of T lymphocytes (CD8) + CD28 + CD38 + / CD8 + CD28 + T% and CD3 + CD38 + T % (CD3) + CD38 + T occupies CD3 + (Percentage of T lymphocytes)

[0125] Regulatory T cells (Treg): Absolute count of Treg cells;

[0126] Natural killer (NK) cells: absolute NK cell count, percentage of NK cells among lymphocytes (NK%).

[0127] Flow cytometry reagents, lymphocyte subset combination schemes, and detection reagents:

[0128] Tube 1 (Absolute Count Tube): CD3FITC / CD56+CD16 PE / CD45PerCP Cy5.5 / CD4 PE CY7 / CD19 APC / CD8 APC CY7; Lymphocyte Subset Detection Reagent (Flow Cytometry - 6 Colors), Catalog No. 662967;

[0129] Tube 2: CD45RA FITC / CD127 PE / CD3 PerCP Cy5.5 / CD4 PE CY7 / CD25 APC / CD8APC CY7 / CD197 BV421; CD45RA assay kit, catalog number 662840; CD127 assay kit (flow cytometry method), catalog number 664400; CD3 assay kit (flow cytometry method), catalog number 665748; CD4 assay kit, catalog number 663493; CD25 assay kit (flow cytometry method), catalog number 666484; CD8 assay kit, catalog number 663521; BV421 conjugated with anti-human CCR7 (CD197) antibody, catalog number 566743;

[0130] Tube 3: CD183 PE / CD3 PerCP Cy5.5 / CD4 PE CY7 / CD196 APC / CD8 APC CY7;

[0131] PE-conjugated human CD183 antibody, catalog number 557185; APC-conjugated human CD196 (CCR6) antibody, catalog number 560619;

[0132] Tube 4: CD8 FITC / CD38 PE / CD3 PerCP Cy5.5 / CD4 PE CY7 / CD28 APC; CD8 FITC fluorescent monoclonal antibody reagent, catalog number 347313; CD38 detection reagent (flow cytometry method), catalog number 665746; APC conjugated with antibody against human antigen CD28, catalog number 559770.

[0133] Flow cytometry specimen processing steps:

[0134] Absolute counting tube processing: Add 100 μL of peripheral blood and corresponding fluorescent antibody to each tube, mix thoroughly, and incubate at room temperature in the dark for 15 min; add 500 μL of diluted 1× hemolysin, mix well, and incubate at room temperature in the dark for 10 min, then directly perform the test.

[0135] Standard flow cytometry processing: The white blood cell count in peripheral blood (unit: cells / µL) was obtained using a cell counter. The required peripheral blood volume for adding 1×10^6 cells was calculated. 3-20 μL of antibodies labeled with different fluorescein were added, thoroughly mixed, and incubated at room temperature in the dark for 15 min. 3 mL of diluted 1× hemolysin was added, mixed, and incubated at room temperature in the dark for 10 min. The mixture was centrifuged at 300×g for 5 min, the supernatant was discarded, 3 mL of phosphate-buffered saline (PBS) was added, mixed, centrifuged, washed, and resuspended in 0.5 mL of PBS before analysis. In the above results, the absolute count is the peripheral blood sample content converted from absolute counting tubes, and the percentage of T-cell-related subsets is also the percentage content in the peripheral blood sample.

[0136] (2) Detection of cytokines and chemokines

[0137] Using liquid-phase chip multifactor flow cytometry, a QBPlex flow cytometer (Changzhou Bidaco) and two reagent kits (factor 14 and factor 10, Beijing Kuangbo Biotechnology Co., Ltd.) were employed. This technology captured multiple soluble proteins, including interleukin (IL)-1β, IL-2, IL-4, IL-5, IL-6, IL-8, IL-10, IL-12p70, IL-17A, IL-17F, IL-22, and tumor necrosis factor-α (TNF), through microspheres of known size and fluorescence intensity. The study included the following cytokines and chemokines: α-α, tumor necrosis factor-β (TNF-β), interferon-gamma (IFN-γ), soluble CD25 (sCD5), macrophage inflammatory protein 1α (MIP-1α), monocyte chemoattractant protein-1 (MCP-1), granulocyte-macrophage colony-stimulating factor (GM-CSF), IL-15, regenerating islet-derived protein 3α (Reg3α), Elafin, growth-stimulating gene 2 protein (ST2), granzyme B, and tumor necrosis factor receptor 1 (TNFR1). The corresponding fluorescence was detected using flow cytometry, and the levels of these cytokines and chemokines in the sample were determined by the fluorescence intensity.

[0138] This includes the following valuable indicators from screened peripheral blood serum:

[0139] Inflammation-related factors: tumor necrosis factor receptor 1 (TNFR1), granulocyte-macrophage colony-stimulating factor (GM-CSF).

[0140] Intestinal injury-related factor: Regenerating islet-derived protein 3α (Reg3α).

[0141] Cytokine detection steps:

[0142] Serum preparation: Peripheral blood samples were collected using coagulation tubes, coagulated at room temperature (20-25℃) for 30 min, centrifuged at 1200×g for 15 min, and the clear layer serum was collected for testing;

[0143] Preparation of standard: Take 9 EP tubes and dilute them serially to prepare the standard;

[0144] Sample testing: Two EP tubes were used as parallel test tubes for each sample. Capture microspheres, detection antibody, and diluent were added to each tube and mixed. After mixing, the mixture was incubated at room temperature in the dark. Then, PE-labeled streptavidin was added, and the mixture was mixed again and incubated at room temperature in the dark. After washing, the samples were analyzed, and the levels of corresponding cytokines and chemokines in the test samples were calculated based on the fluorescence intensity. In the above test results, the cytokine levels have been converted to the levels found in peripheral blood samples.

[0145] Data analysis and detection model construction

[0146] The data analysis was performed using R (version 4.4.3). The specific steps are as follows:

[0147] (1) Data preprocessing

[0148] The Multiple Imputation (MI) method is used to handle missing values ​​in the data to ensure data integrity.

[0149] (2) Univariate analysis

[0150] Categorical variables were tested using chi-square or exact tests, while continuous variables were tested using t-tests (for normal distribution) or Mann-Whitney U tests (for non-normal distribution). Differences in immune cell subsets and cytokine indicators between the GVHD group and the virus infection group at each time point were compared, and indicators with statistically significant differences between groups (P < 0.05) were selected.

[0151] (3) Screening of key detection variables

[0152] Key detection variables were selected using a combination of Lasso regression and logistic regression.

[0153] Lasso regression: The optimal regularization parameter (λ=0.06) is determined through 10-fold cross-validation, and indicators with non-zero coefficients are retained as candidate variables. The variable selection process is visualized through coefficient path graphs.

[0154] Logistic Regression: Logistic regression analysis is performed on the candidate variables selected by Lasso. Based on the significance of the variables (P-value), model parsimony, and sample size, the key detection variables are finally determined.

[0155] (4) Detection model construction

[0156] Two detection models were built for different clinical scenarios:

[0157] Single-timepoint model: Constructed based on key detection variables in the 7 days prior to the onset of diarrhea symptoms. The final identified key biomarkers were absolute NK cell count (positive detection factor) and CD4. + CD28 + / CD4 + T% (negative detection factor), the model formula is:

[0158] logit(P1 (infection)) = 0.6162 + 0.0141 × absolute NK cell count - 0.0270 × CD4 + CD28 + / CD4 + T%

[0159] The model's cut-off threshold is 0.408. When the detection probability P1 > 0.408, it is considered a high risk of infection; otherwise, it is considered a high risk of GVHD.

[0160] Multi-time point model: Integrating key detection variables (a total of 9, including CD8) at 7 days before the onset of diarrhea symptoms, 4 weeks after allo-HSCT, and 5 weeks after allo-HSCT. + T%, absolute NK cell count, CD4 + CD28 + / CD4 + T%, CD4+Th2 / CD4+T%, TNFR1 at week 4 after allo-HSCT, absolute NK cell count at week 4 after allo-HSCT, GM-CSF at week 5 after allo-HSCT, absolute NK cell count at week 5 after allo-HSCT, CD3 at week 5 after allo-HSCT + CD38 + T %), calculate the test score, test score = 2.6714 + (CD8 % 7 days before the onset of diarrhea symptoms) + T%) × (-0.0116) + Absolute NK cell count 7 days before the onset of diarrhea symptoms × 0.0049 + (CD4+ 7 days before the onset of diarrhea symptoms) + CD28 + / CD4 + %)×(-0.0217)+(CD4 count 7 days before the onset of diarrhea symptoms) + Th2 / CD4 +T%)×(-0.0256)+(TNFR1 content at week 4 after allo-HSCT)×(-0.0005)+(absolute NK cell count at week 4 after allo-HSCT)×(0.0006)+(GM-CSF content at week 5 after allo-HSCT)×(-1.6570)+(absolute NK cell count at week 5 after allo-HSCT)×(0.0038)+(CD3+ at week 5 after allo-HSCT)×(-0.0006)+(GM-CSF content at week 5 after allo-HSCT)×(-1.6570)+(absolute NK cell count at week 5 after allo-HSCT)×(0.0038)+(CD3+ at week 5 after allo-HSCT)×(-0.0006)×(-0.0006)×(-0.0006)×(GM-CSF content at week 5 after allo-HSCT ...GM-CSF content at week 5 after allo- + CD38 + T%)×(0.0243);

[0161] The model was then built using a "detection score," and the model formula is as follows:

[0162] logit(P2 (infection)) = 0.238 + 1.632 × (test score)

[0163] The model's cut-off threshold is 0.379. When the detection probability P2 > 0.379, it is considered a high risk of infection; otherwise, it is considered a high risk of GVHD.

[0164] (5) Model validation

[0165] The model performance was verified using the following methods, with r language's rms package used to construct nodal plots and ggplot2 and ggpubr packages used to create charts:

[0166] Discriminant power: Plot the receiver operating characteristic (ROC) curve and calculate the area under the curve (AUC).

[0167] Calibration: Plot calibration curves to evaluate the consistency between the model's detection probability and the actual pathological diagnosis results;

[0168] Clinical utility: Decision curve analysis (DCA) plots were generated to assess the benefits of model-guided clinical interventions.

[0169] Differential Diagnostic Applications

[0170] By substituting the immune cell subsets and cytokine detection data of the patients to be diagnosed into the above single-timepoint or multi-timepoint models, and based on the detection probability / score and corresponding threshold output by the model, it can be determined whether the cause of the patient's diarrhea is GI-GVHD or enterovirus infection, providing a basis for the selection of clinical treatment plans.

[0171] In some specific embodiments of the present invention, the collected peripheral blood samples need to be preserved using cell preservation solution before being used for immune cell subset detection.

[0172] This invention provides a cell protection solution, wherein each 100ml of the cell protection solution contains 5mL of compound electrolyte glucose injection, 5mL of human serum albumin injection, 5mL of compound amino acid injection, 10-20μM salvianolic acid B, 20-60μM rhodioloside, 1-10μM ginkgo biloba extract, and the compound electrolyte injection is used to make up the difference. As a preferred embodiment, the cell protection solution of this invention also contains 10-50μM hesperidin and 20-50μM total saponins of Panax notoginseng.

[0173] In some specific embodiments of the present invention, the preparation method of the cell protection solution of the present invention includes preparing the salvianolic acid B, the rhodioloside, the total saponins of Panax notoginseng, the hesperidin and the ginkgo biloba flavonoids into solutions respectively, and mixing the resulting solutions with compound electrolyte injection, compound electrolyte glucose injection, human serum albumin injection and compound amino acid injection in proportion to the final concentration.

[0174] Example 1

[0175] This embodiment provides a method for processing peripheral blood samples with cell preservation solution and verifies its effectiveness:

[0176] Preparation of cell preservation solution:

[0177] In the following examples, each traditional Chinese medicine component was sourced from Chengdu Mansite Biotechnology Co., Ltd., and was prepared and stored as follows: each component was prepared as a 1mL 10mM solution.

[0178]

[0179] The manufacturers of Compound Electrolyte Glucose Injection and Compound Electrolyte Injection are Sichuan Kelun Pharmaceutical Co., Ltd., with product numbers H20063443 and H20113476 respectively. The manufacturer of Compound Amino Acid Injection (18AA-II) is Fresenius Kabi (China) Investment Co., Ltd., with product number H10980029. The manufacturer of Human Serum Albumin Injection is Octapharma AG (Austria), with product number SJ20160037.

[0180] The salvianolic acid B, rhodioloside, total saponins of Panax notoginseng, hesperidin and ginkgo biflavonoids were each prepared into solutions. The resulting solutions were then mixed with compound electrolyte injection, compound electrolyte glucose injection, human serum albumin injection and compound amino acid injection in proportion to the final concentration.

[0181] The cell preservation solution used in this embodiment contains the following raw material components per 100 ml: 84.5 mL compound electrolyte injection, 5 mL compound electrolyte glucose injection, 5 mL human serum albumin injection, 5 mL compound amino acid injection (18AA-II), 20 μM salvianolic acid B, 60 μM rhodioloside, 10 μM ginkgo biloba extract, 50 μM hesperidin, and 50 μM total saponins of Panax notoginseng. The pH of this cell preservation solution is 6.89, and the osmolality is 376 Osmol.

[0182] Sample collection and processing

[0183] After collecting peripheral blood from patients using EDTA anticoagulant tubes, the peripheral blood is thoroughly mixed with cell preservation solution within the shortest possible time (within 2 hours) to obtain a peripheral blood mixed suspension, with a volume ratio of peripheral blood to cell preservation solution of 1:4. In this invention, patients are monitored serially according to a pre-set, fixed time schedule, rather than undergoing a single test only when symptoms appear. Therefore, if immediate testing is not possible after collecting peripheral blood samples, this invention provides a cell preservation solution for processing peripheral blood samples to immediately stabilize cell state and absolute cell count, prevent cell apoptosis or activation in vitro, and thus maintain their original phenotypic characteristics. Blood samples mixed with cell preservation solution can be stored and transported at 4°C, significantly extending the sample's processable window without causing a significant decrease in cell count or changes in biomarker expression.

[0184] Experimental Group 1: After collecting peripheral blood from patients using EDTA anticoagulant tubes, the peripheral blood was thoroughly mixed with cell preservation solution within the shortest possible time (within 2 hours) to obtain a peripheral blood mixed suspension with a volume ratio of peripheral blood to cell preservation solution of 1:4. The suspension was stored at 2-8℃ for 120 hours and then analyzed by flow cytometry.

[0185] Control group 1: Peripheral blood was directly analyzed by flow cytometry after collection;

[0186] Control group 2: Peripheral blood was collected without adding cell preservation solution, stored at 2-8℃ for 120 hours, and then analyzed by flow cytometry;

[0187] Control group 3: After collecting peripheral blood from patients using EDTA anticoagulant tubes, the peripheral blood and cell preservation solution were thoroughly mixed within the shortest possible time (within 2 hours) to obtain a peripheral blood mixed suspension with a volume ratio of peripheral blood to cell preservation solution of 1:4, which was then analyzed by flow cytometry.

[0188] The flow cytometry indicators used in this embodiment are: absolute NK cell count and CD4+. + CD28 + T cells account for CD4+ + T lymphocyte percentage, CD3 + CD8+ Percentage of T cells among lymphocytes, CD4 + Type 2 helper T cells account for a significant portion of CD4+. + Percentage of T lymphocytes.

[0189] The flow cytometry reagent combination scheme and detection reagents for lymphocyte subsets are carried out in accordance with the above-described embodiments.

[0190] Flow cytometry specimen processing steps:

[0191] Absolute counting tube processing: Add 100 μL of peripheral blood suspension to each tube, add the corresponding fluorescent antibody, mix thoroughly, and incubate at room temperature in the dark for 15 min; add 500 μL of diluted 1× hemolysin, mix well, and incubate at room temperature in the dark for 10 min, then directly perform the test.

[0192] Standard flow cytometry processing: The white blood cell count (cells / µL) in the peripheral blood suspension was obtained using a cell counter. The required volume of peripheral blood suspension for adding 1×10^6 cells was calculated. The suspension was centrifuged at 300×g for 5 min, the supernatant was discarded, and 200 μL of phosphate-buffered saline (PBS) was added and mixed thoroughly. Then, 3-20 μL of antibodies labeled with different fluorescein were added, mixed thoroughly, and incubated at room temperature in the dark for 15 min. Next, 3 mL of diluted 1× hemolysin was added, mixed thoroughly, and incubated at room temperature in the dark for 10 min. The suspension was centrifuged at 300×g for 5 min, the supernatant was discarded, and 3 mL of phosphate-buffered saline (PBS) was added and mixed. After centrifugation and washing, the suspension was resuspended in 0.5 mL of phosphate-buffered saline before flow cytometry analysis. In the above results, since the sample used for flow cytometry analysis is a peripheral blood suspension, the absolute count of NK cells and the percentage of T cell-related subsets in the original peripheral blood sample need to be calculated based on the ratio of peripheral blood to cell preservation fluid volume.

[0193] Test results statistics:

[0194]

[0195] Results Analysis: ① In the group without preservation solution, cell counts plummeted to 31% of baseline due to massive apoptosis and necrosis. Experimental group 1 successfully maintained 88% of cells, with minimal loss, effectively ensuring cell viability; ② NK cells are sensitive to in vitro stress, with only 25% remaining in the group without preservation solution. Experimental group 1 protected 87%, improving NK cell stress response and ensuring NK cell viability; ③ In the group without preservation solution, the more vulnerable effector CD8+ cells... + T cells preferentially die, artificially reducing their percentage (from 28% to 20.5%), creating a false impression of "CD8 cell reduction." Experimental group 1 maintained the original proportion; ④ Severe oxidative and metabolic stress in the group without preservation solution led to downregulation or internalization of CD28 molecules, resulting in CD4... + CD28 +A significant decrease in cell proportion can lead to misjudgments in the evaluation of CD4+ T lymphocyte function. Experimental group 1 effectively maintained stable CD28 expression. ⑤ Cells in the group without preservation solution were in poor condition, with unstable or lost expression of chemokine receptors used to define Th2 cells (such as CRTH2), resulting in blurred cell boundaries, difficulty in gating, and extremely low detection rates. Cells in experimental group 1 were healthy with clear surface markers, allowing for accurate gating and analysis. ⑥ The group without preservation solution produced a large amount of cell debris and dead cells, severely interfering with flow cytometry analysis, affecting gating accuracy, and potentially causing result bias. Samples from experimental group 1 were clean, resulting in high-quality data.

[0196] The cell preservation solution of this invention maintains cell numbers and avoids pseudocytosis through its core anti-apoptotic and anti-oxidative mechanisms, and can preserve total lymphocytes and various subsets (especially NK and CD8) in peripheral blood. + The absolute count of T cells remained near baseline (>85%) after 120 hours, while traditional methods could only retain less than 1 / 3. In addition, the cell preservation solution of the present invention can maintain phenotypic stability during sample preservation, which is crucial for clinical detection of immune function status and prevents clinical misjudgment due to sample test results distortion. Furthermore, samples using the cell preservation solution of the present invention have less debris and a lower proportion of dead cells during flow cytometry detection, with clear and accurate gating and high data quality, ensuring the reliability and reproducibility of experimental results.

[0197] Example 2

[0198] This embodiment provides a model for identifying the pathogenic factors of diarrhea after allogeneic hematopoietic stem cell transplantation and its construction method:

[0199] 1. Study subjects and baseline characteristics

[0200] A retrospective study included 58 patients who underwent allo-HSCT at Beijing Lu Daopei Hospital between April 2021 and December 2024, developed diarrhea, and underwent intestinal endoscopic biopsy. Patients were grouped according to pathological results:

[0201] GVHD group: 34 cases, including 5 cases of Lerner grade I, 10 cases of grade II, 14 cases of grade III, and 5 cases of grade IV;

[0202] Viral infection group: 24 cases, including 8 cases of CMV enteritis, 9 cases of EBV enteritis, 5 cases of CMV combined with EBV enteritis, 1 case of HHV-6 enteritis, and 1 case of adenovirus enteritis.

[0203] There were no statistically significant differences in baseline characteristics between the two groups of patients (P > 0.05), as detailed in Table 1.

[0204] Table 1

[0205]

[0206] 2. Results of immune cell subsets and cytokine detection

[0207] Intergroup differences in the 7 days prior to the onset of diarrhea symptoms (time point 1)

[0208] The GVHD group and the virus infection group showed significant differences in the following indicators (P < 0.05), as shown in Table 2.

[0209] Table 2

[0210]

[0211] Note: CD8 + T% is CD3 + CD8 + The percentage of T lymphocytes among all lymphocytes; data that conform to a normal distribution are expressed as mean ± SD, and data that do not conform to a normal distribution are expressed as median (Q1, Q3).

[0212] Differences between groups at week 4 after allo-HSCT (time point 2)

[0213] The GVHD group and the virus infection group showed significant differences in the following indicators (P < 0.05), as shown in Table 3.

[0214] Table 3

[0215]

[0216] Differences between groups at week 5 after allo-HSCT (time point 3)

[0217] The GVHD group and the virus infection group showed significant differences in the following indicators (P < 0.05), as shown in Table 4.

[0218] Table 4

[0219]

[0220] in, Figure 1 This is a box plot showing the visualization results of univariate differences in Reg3α and CD3 between the GVHD group and the viral infection group 7 days before the onset of diarrhea symptoms. + T%, CD8 + T%, NaiveCD4 + / CD4 + T%, CD4 + Th2 / CD4 + Differences in the distribution of T%, NK%, and absolute NK cell counts between groups.

[0221] Figure 2The effect size forest plot is a visualization of the univariate differences between the GVHD group and the viral infection group 7 days before the onset of diarrhea symptoms. It shows the effect size and 95% confidence interval of each difference indicator (** indicates P < 0.01, * indicates P < 0.05).

[0222] Figure 3 Box plots showing the visualization results of univariate differences between the GVHD group and the virus infection group at week 4 after allo-HSCT (indicators: TNFR1, NaiveCD4). + / CD4 + T%, absolute NK cell count, and absolute Treg count.

[0223] Figure 4 Forest plot of effect size at week 4 after allo-HSCT, showing the visualization results of univariate differences between the GVHD group and the virus infection group at week 4 after allo-HSCT.

[0224] Figure 5 Box plots showing the visualization results of univariate differences between the GVHD group and the virus infection group at week 5 after allo-HSCT (indicators: GM-CSF, CD3). + CD38 + absolute T count, absolute NK cell count, CD8 + CD28 + CD38 + / CD8 + CD28 + T%).

[0225] Figure 6 The effect size forest plot is a visualization of the univariate difference indices between the GVHD group and the virus infection group at week 5 after allo-HSCT.

[0226] Key findings: The absolute NK cell count was significantly lower in the GVHD group than in the virus-infected group at all three time points, and the effect size was moderate in all cases, making it the core identification indicator throughout the entire process.

[0227] 3. Detection Model Construction and Validation

[0228] Single-point model (7 days before the onset of diarrhea symptoms)

[0229] (1) Screening of key variables

[0230] Lasso regression initially identified 5 candidate variables (IL-4, IL-22, CD3). + T%, absolute NK cell count, CD4 + CD28 + / CD4 + T%); Figure 7AThe coefficient path diagram in the Lasso regression screening results of indicators for the 7 days before the onset of diarrhea symptoms shows the changing trend of the coefficients of each indicator as the regularization parameter changes.

[0231] Figure 7B The optimal regularization parameter λ=0.06 was determined by the cross-validation error curve in the Lasso regression screening results of indicators 7 days before the onset of diarrhea symptoms.

[0232] After further screening using logistic regression, two key variables were finally identified, as shown in Table 5.

[0233] Table 5

[0234]

[0235] Absolute NK cell count: a positive detection factor, P < 0.1, with an increase of 1 unit in the probability of infection.

[0236] CD4 + CD28 + / CD4 + T%: Negative detection factor, P=0.0043 (95% CI: 0.950, 0.991). For every 1 unit increase, the probability of infection decreases by approximately 3%.

[0237] A single-timepoint model was constructed based on key detection variables from the 7 days prior to the onset of diarrhea symptoms. The final identified key biomarkers were absolute NK cell count (a positive detection factor) and CD4. + CD28 + / CD4 + T% (negative detection factor), the model formula is:

[0238] logit(P1 (infection)) = 0.6162 + 0.0141 × absolute NK cell count - 0.0270 × CD4 + CD28 + / CD4 + T%

[0239] The model's cut-off threshold is 0.408. When the detection probability P1 > 0.408, it is considered a high risk of infection; otherwise, it is considered a high risk of GVHD.

[0240] (2) Model performance

[0241] Figure 8A This is a nomogram of the performance of a bivariate detection model at a single time point (7 days before the onset of diarrhea symptoms), used to calculate the patient's infection risk score;

[0242] Figure 8BThe calibration curve is shown in the bivariate detection model performance at a single time point (7 days before the onset of diarrhea symptoms), demonstrating the consistency between the model's detection probability and the actual probability.

[0243] Figure 8C The decision curves for the bivariate detection model at a single time point (7 days before the onset of diarrhea symptoms) are used to evaluate the benefits of the model in guiding clinical intervention.

[0244] Figure 8D The ROC curve for the bivariate detection model at a single time point (7 days before the onset of diarrhea symptoms) is shown. The AUC is 0.799 (CUT-OFF = 0.408, sensitivity 70.8%, specificity 88.2%).

[0245] Discrimination efficacy: ROC curve AUC = 0.799 (95% CI 0.66-0.98);

[0246] Threshold and diagnostic efficacy: Optimal cut-off = 0.408, sensitivity 70.8%, specificity 88.2%;

[0247] Calibration accuracy: The calibration curve shows good consistency between the detection probability and the actual pathological results;

[0248] Clinical utility: The decision curve shows that when the threshold probability is in the range of 0.1-0.8, the clinical intervention benefit based on this model is higher than the "all intervention" or "no intervention" strategies.

[0249] Multi-time point model (integration of three time points)

[0250] (1) Screening of key variables

[0251] The 11 differential indicators identified by integrating multi-time point data through univariate regression are shown in Table 6.

[0252] Table 6

[0253]

[0254] Figure 9A A graph of coefficient paths in the Lasso regression results for variables detected at multiple time points; Figure 9B This is the cross-validation error curve in the Lasso regression results for variables detected at multiple time points.

[0255] Lasso regression was performed on 11 differential variables, and 9 key variables were identified: CD8 count 7 days before the onset of diarrhea symptoms. + T%, absolute NK cell count 7 days before the onset of diarrhea symptoms, and CD4 count 7 days before the onset of diarrhea symptoms. + CD28 + / CD4 + T%, CD4 count 7 days before diarrhea symptoms appear+ Th2 / CD4 + T%, TNFR1 at week 4 after allo-HSCT, absolute NK cell count at week 4 after allo-HSCT, GM-CSF at week 5 after allo-HSCT, absolute NK cell count at week 5 after allo-HSCT, CD3 at week 5 after allo-HSCT + CD38 + T%.

[0256] Multi-time point model: Integrating key detection variables (a total of 9, including CD8) at 7 days before the onset of diarrhea symptoms, 4 weeks after allo-HSCT, and 5 weeks after allo-HSCT. + T%, absolute NK cell count, CD4 + CD28 + / CD4 + T%, CD4+Th2 / CD4+T%, TNFR1 at week 4 after allo-HSCT, absolute NK cell count at week 4 after allo-HSCT, GM-CSF at week 5 after allo-HSCT, absolute NK cell count at week 5 after allo-HSCT, CD3 at week 5 after allo-HSCT + CD38 + T%), calculate the test score, test score = 2.6714 + (CD8+) days prior to the onset of diarrhea symptoms. + T%) × (-0.0116) + Absolute NK cell count 7 days before the onset of diarrhea symptoms × 0.0049 + (CD4+ 7 days before the onset of diarrhea symptoms) + CD28 + / CD4 + %)×(-0.0217)+(CD4 count 7 days before the onset of diarrhea symptoms) + Th2 / CD4 + T%)×(-0.0256)+(TNFR1 content at week 4 after allo-HSCT)×(-0.0005)+(absolute NK cell count at week 4 after allo-HSCT)×(0.0006)+(GM-CSF content at week 5 after allo-HSCT)×(-1.6570)+(absolute NK cell count at week 5 after allo-HSCT)×(0.0038)+(CD3+ at week 5 after allo-HSCT)×(-0.0006)+(GM-CSF content at week 5 after allo-HSCT)×(-1.6570)+(absolute NK cell count at week 5 after allo-HSCT)×(0.0038)+(CD3+ at week 5 after allo-HSCT)×(-0.0006)×(-0.0006)×(-0.0006)×(GM-CSF content at week 5 after allo-HSCT ...GM-CSF content at week 5 after allo- + CD38 + T%)×(0.0243), then the "detection score" is used to build the model, and the model formula is:

[0257] logit(P2 (infection)) = 0.238 + 1.632 × (test score)

[0258] The model's cut-off threshold is 0.379. When the detection probability P2 > 0.379, it is considered a high risk of infection; otherwise, it is considered a high risk of GVHD.

[0259] (2) Model performance

[0260] Figure 10A A nomogram of the multi-timepoint detection model performance (constructed based on detection scores);

[0261] Figure 10B The calibration curve for the multi-time point detection model performance (validated by 1000 Bootstrap resampling, C-statistic = 0.884 after correction).

[0262] Figure 10C The decision curve in the performance of the multi-time-point detection model;

[0263] Figure 10D The ROC curve for the multi-time-point detection model performance is shown. The detection score AUC = 0.908 (CUT-OFF = 0.379, sensitivity 87.5%, specificity 79.4%).

[0264] Speculative power: The ROC curve AUC = 0.908 (95% CI 0.836, 0.981), which is significantly higher than that of the single-time point model;

[0265] Threshold and diagnostic efficacy: Optimal cut-off = 0.379, sensitivity 87.5%, specificity 79.4%;

[0266] Calibration accuracy: Validated by 1000 Bootstrap resampling cycles, the corrected C-statistic is 0.884 (95% CI 0.880-0.889), demonstrating excellent consistency of the calibration curve.

[0267] Clinical utility: Decision curves show that the benefit range (threshold probability 0.05-0.9) and degree of benefit of model-guided clinical intervention are both superior to single-point models.

[0268] Example 3 This example provides a clinical application example of the identification method of the present invention in Example 2.

[0269] Case 1: Example of Differentiation of Gastrointestinal Graft-versus-Host Disease Diarrhea (GI-GVHD)

[0270] Patient information: Male, 10 years old, underwent allo-HSCT for acute lymphoblastic leukemia. On the 28th day after transplantation (4 weeks after allo-HSCT), he developed diarrhea (4 times a day, loose stools, no bloody stools). The specific results of the biomarkers test are shown in Table 7.

[0271] Table 7. Biometric detection results of subjects in Example 3-1

[0272]

[0273] Substitute the above biomarker detection data into the multi-time-point detection model described in this invention, and perform calculations and verifications according to the following steps:

[0274] ①Predicted score = 2.6714

[0275] + 62 × (-0.0116) = -0.719

[0276] + 35 × 0.0049 = +0.172

[0277] + 15 × (-0.0217) = -0.326

[0278] + 80 × (-0.0256) = -2.048

[0279] + 1100 × (-0.0005) = -0.550

[0280] + 68 × 0.0006 = +0.0408

[0281] + 0.8 × (-1.6570) = -1.326

[0282] + 45 × 0.0038 = +0.171

[0283] + 50 × 0.0243 = +1.215

[0284] =-0.699

[0285] Detection probability calculation:

[0286] logit(P2) =0.238 - 1.632 × 0.699 = -0.903

[0287] P2 = e^-0.903 / (1 + e^-0.903) ≈ 0.288

[0288] Model determination:

[0289] P2 = 0.288 < 0.379, according to the method of the present invention, the cause of diarrhea in the subject was determined to be a high risk of gastrointestinal graft-versus-host disease.

[0290] Clinical validation: Subsequent endoscopic biopsy showed crypt destruction (Lerner grade II), confirming a diagnosis of GI-GVHD, consistent with the determination result of the method of this invention. This demonstrates that the method described in this invention can accurately identify gastrointestinal graft-versus-host disease-related diarrhea in the early stages of diarrhea. After administration of enhanced immunosuppressive therapy (methylprednisolone combined with tacrolimus), the diarrhea resolved within 3 days.

[0291] Case 2: Example of Differential Diarrhea Caused by Enterovirus Infection

[0292] Patient information: Female, 15 years old, underwent allo-HSCT for myelodysplastic syndrome. On the 35th day after transplantation (5 weeks after allo-HSCT), she developed diarrhea (5 times a day, loose stools, with a small amount of blood in the stool). The test results are shown in Table 8.

[0293] Table 8. Biometric detection results of subjects in Example 3-2

[0294]

[0295] Substitute the above biomarker detection data into the multi-time-point detection model described in this invention, and perform calculations and verifications according to the following steps:

[0296] Detection score = 2.6714

[0297] + 30 × (-0.0116) = -0.348

[0298] + 90 × 0.0049 = +0.441

[0299] + 30 × (-0.0217) = -0.651

[0300] + 25 × (-0.0256) = -0.640

[0301] + 800 × (-0.0005) = -0.400

[0302] + 120 × 0.0006 = +0.072

[0303] + 0.4 × (-1.6570) = -0.663

[0304] + 110 × 0.0038 = +0.418

[0305] + 20 × 0.0243 = +0.486

[0306] =1.386

[0307] Detection probability calculation:

[0308] logit(P2) = 0.238 + 1.632 × 1.386 = 2.501

[0309] P2= e^2.501 / (1 + e^2.501) ≈ 0.924

[0310] Model determination: P2 = 0.924 > 0.379, according to the method of the present invention, the cause of diarrhea in this subject is determined to be a high risk of enterovirus infection.

[0311] Clinical validation: Peripheral blood CMV DNA A positive test result (1.2 × 10^5 copies / mL) confirmed CMV enteritis, consistent with the findings of the method described in this invention. This demonstrates that the method described in this invention can effectively identify diarrhea following allogeneic hematopoietic stem cell transplantation caused by enterovirus infection. After reducing the dosage of immunosuppressants and administering ganciclovir antiviral treatment, the diarrhea subsided within one week, and CMV was detected. DNA Negative.

[0312] As described above, this invention, validated based on 58 clinical samples, demonstrates that a detection model based on dynamic monitoring of immune cell subsets and cytokines can effectively differentiate between allo-HSCT-related G-GVHD and viral infection-related diarrhea. Specifically, the multi-timepoint model (AUC=0.908) exhibits significantly superior differential diagnostic efficacy compared to the single-timepoint model (AUC=0.799) and traditional pathological biopsy, and offers advantages such as being minimally invasive, low-risk, early-stage, and standardized. It can serve as an important tool for clinical differential diagnosis, providing a basis for precise treatment of patients with diarrhea after allo-HSCT.

[0313] Example 4 This example provides batch verification data of the identification method of the present invention in Example 1.

[0314] Data validation: We further validated the detection method described in this invention on a validation set independent of the model construction samples.

[0315] This embodiment selected 300 patients who underwent allogeneic hematopoietic stem cell transplantation (allo-HSCT) between 2024 and 2025 as an independent validation set, including 180 patients with graft-versus-host disease (GVHD) and 120 patients with virus-associated diarrhea. The baseline characteristics of the two groups are detailed in Table 9.

[0316] Table 9

[0317]

[0318]

[0319] All validation samples were tested according to the method described in this invention, based on the results of multiple immune cell subset ratios and cytokine detection, as shown in Table 10. The absolute NK count unit is cells / μL, and the units for TNFR1 and GM-CSF are pg / mL.

[0320] Table 10 Validation Dataset

[0321]

[0322]

[0323]

[0324]

[0325]

[0326]

[0327]

[0328]

[0329]

[0330]

[0331]

[0332] Calculate the test score based on the data in Table 10: Test score = 2.6714 + (CD8+) / (CD8+ from 7 days prior to the onset of diarrhea symptoms) + T%) × (-0.0116) + Absolute NK cell count 7 days before the onset of diarrhea symptoms × 0.0049 + (CD4+ 7 days before the onset of diarrhea symptoms) + CD28 + / CD4 + %)×(-0.0217)+(CD4 count 7 days before the onset of diarrhea symptoms) + Th2 / CD4 + T%)×(-0.0256)+(TNFR1 content at week 4 after allo-HSCT)×(-0.0005)+(absolute NK cell count at week 4 after allo-HSCT)×(0.0006)+(GM-CSF content at week 5 after allo-HSCT)×(-1.6570)+(absolute NK cell count at week 5 after allo-HSCT)×(0.0038)+(CD3+ at week 5 after allo-HSCT)×(-0.0006)+(GM-CSF content at week 5 after allo-HSCT)×(-1.6570)+(absolute NK cell count at week 5 after allo-HSCT)×(0.0038)+(CD3+ at week 5 after allo-HSCT)×(-0.0006)×(-0.0006)×(-0.0006)×(GM-CSF content at week 5 after allo-HSCT ...GM-CSF content at week 5 after allo- + CD38 +T%)×(0.0243), then the "detection score" is further substituted into the fixed logistic regression model formula:

[0333] logit(P2 (infection)) = 0.238 + 1.632 × (test score)

[0334] Based on this, the probability of infection is obtained, and 0.379 is used as a unified discrimination threshold. When the detection probability is greater than 0.379, it is judged as a high risk of viral infection; otherwise, it is judged as a high risk of GVHD.

[0335] As shown in Table 10, the total number of samples was 300, with 180 cases in the GVHD group and 120 cases in the infection group.

[0336] Core performance metrics (threshold = 0.379):

[0337] Overall accuracy: 86.0%, GVHD identification accuracy: 78.9%, infection identification accuracy: 96.7%.

[0338] Figure 11 ROC curves for the multi-timepoint detection model validated in 300 cases. Figure 11 As can be seen, in the 300-case validation set, the area under the ROC curve (AUC) of the detection model reached approximately 0.90, demonstrating good overall discrimination efficacy; at a threshold of 0.379, the sensitivity was 96.7% and the specificity was 78.9%; the multi-time point detection model of the present invention showed excellent performance in the validation of 300 samples, with high overall accuracy and low false positive rate, and has good clinical application value.

[0339] In summary, the detection model constructed based on multiple indicators of immune cell subsets and cytokines in this invention has been independently validated in 58 initial clinical samples and 300 allo-HSCT patients, demonstrating its ability to stably and effectively differentiate between G-GVHD and viral infection-related diarrhea after allo-HSCT. This method offers significant advantages such as being minimally invasive, low-risk, reproducible, highly standardized, and capable of early warning. It is suitable for routine clinical testing conditions and can provide a reliable basis for accurate subtyping diagnosis and individualized treatment decisions in post-transplant diarrhea patients, showing promising prospects for clinical promotion and industrial application.

Claims

1. A kit for identifying the pathogenic cause of diarrhea after allogeneic hematopoietic stem cell transplantation, characterized by, It includes reagents for detecting an individual's biomarkers; The biomarkers include: absolute NK cell count, CD4 + CD28 + T cells account for CD4+ + Percentage of T lymphocytes; The biomarkers were detected within seven days before the onset of diarrhea symptoms. The biomarkers further include the following group biomarkers: Group 1: CD3 levels detected within 7 days prior to the onset of diarrhea symptoms + CD8 + Percentage of T cells among lymphocytes, CD4 + Type 2 helper T cells account for a significant portion of CD4+. + Percentage of T lymphocytes; Second group: TNFR1 levels and absolute NK cell counts were measured within 4 weeks after allo-HSCT; Group 3: GM-CSF levels, absolute NK cell count, and CD3 count measured within 5 weeks after allo-HSCT. + CD38 + T occupies CD3 + Percentage of T lymphocytes; The reagents include antibodies and reagents for detecting cytokine levels; The antibodies comprise three groups of antibodies: The first group of antibodies includes fluorescently labeled CD3, CD56, CD16, CD45, CD4, CD19, and CD8 antibodies, labeled in the order of FITC, PE, PE, PerCP-Cy5.5, PE-Cy7, APC, and APC-Cy7. These antibodies are added to a flow cytometry absolute counting tube containing a single-cell suspension of the sample. This first group of antibodies is used to detect CD8+T%, NK cell absolute count, and other abnormal cell counts. The second group of antibodies includes fluorescently labeled CD183, CD3, CD4, CD196, and CD8 antibodies, labeled in the following order: PE, PerCP-Cy5.5, PE-Cy7, APC, and APC-Cy7. These antibodies are added to a standard flow cytometry tube containing a single-cell suspension of the test sample. This second group of antibodies is used to detect CD4+. + Th2 / CD4 + T%, CD4 + CD28 + / CD4 + T% The third group of antibodies includes fluorescently labeled CD8, CD38, CD3, CD4, and CD28 antibodies. The fluorescent labeling order of each antibody is FITC, PE, PerCP-Cy5.5, PE-Cy7, and APC. These antibodies are added to a standard flow cytometry tube in which the sample to be tested is in a single-cell suspension state. The reagents used to detect cytokine levels include: A reagent for detecting TNFR1 and GM-CSF levels at the protein level.

2. The kit of claim 1, wherein The test samples for the individuals being tested include peripheral blood samples.

3. The kit of claim 2, wherein The reagents include a cell preservation solution for preserving individual peripheral blood samples to obtain a cell suspension suitable for the detection of the biomarkers; Each 100ml of the cell protection solution contains 1-10mL of compound electrolyte glucose injection, 1-10mL of human serum albumin injection, 1-10mL of compound amino acid injection, 10-20μM salvianolic acid B, 20-60μM rhodioloside, 1-10μM ginkgo biloba extract, and is supplemented with compound electrolyte injection.

4. A system for identifying the pathogenic cause of diarrhea after allogeneic hematopoietic stem cell transplantation, characterized in that, The identification system includes: The data acquisition module is used to acquire detection data of biological indicators in peripheral blood samples from individuals at multiple time points after allogeneic hematopoietic stem cell transplantation. These biological indicators include: absolute NK cell count, CD4 + CD28 + T cells account for CD4+ + Percentage of T lymphocytes; The biomarkers were detected within seven days before the onset of diarrhea symptoms. The biomarkers further include the following group biomarkers: Group 1: CD3 levels detected within 7 days prior to the onset of diarrhea symptoms + CD8 + percentage of T cells among lymphocytes, CD4 + Type 2 helper T cells account for a significant portion of CD4+. + Percentage of T lymphocytes; Second group: TNFR1 levels and absolute NK cell counts were measured within 4 weeks after allo-HSCT; Third quartile: GM-CSF levels, NK absolute counts, CD3 + CD38 + T cells out of CD3 + T lymphocyte percentage; The processing and analysis module stores a detection model, which is used to input the detection data of the biomarkers into the detection model and output the detection probability of the individual developing gastrointestinal graft-versus-host disease or enterovirus infection.

5. The authentication system of claim 4, wherein, The causes of diarrhea include gastrointestinal graft-versus-host disease and / or enterovirus infection.

6. The authentication system of claim 4, wherein, Identify the causes of diarrhea after allogeneic hematopoietic stem cell transplantation based on individual biomarkers; The detection data of the aforementioned biomarkers are calculated according to the following steps: Step 1: Calculate the test score = 2.6714 + A × (-0.0116) + B × 0.0049 + C × (-0.0217) + D × (-0.0256) + E × (-0.0005) + F × (0.0006) + G × (-1.6570) + H × (0.0038) + I × 0.0243; Where A represents the absolute NK cell count measured within seven days prior to the onset of diarrhea symptoms, and B represents the CD4 count measured within seven days prior to the onset of diarrhea symptoms. + CD28 + T cells account for CD4+ + T lymphocyte percentage; C represents CD4 count measured within 7 days prior to the onset of diarrhea symptoms. + Type 2 helper T cells account for a significant portion of CD4+. + T lymphocyte percentage, and D represents CD3 count measured within 7 days prior to the onset of diarrhea symptoms. + CD8 + The percentage of T cells among lymphocytes; E represents the TNFR1 level measured within 4 weeks after allo-HSCT; F represents the absolute NK cell count measured within 4 weeks after allo-HSCT; G represents the GM-CSF level measured within 5 weeks after allo-HSCT; H represents the absolute NK cell count measured within 5 weeks after allo-HSCT; and I represents the CD3+ level measured within 5 weeks after allo-HSCT. + CD38 + T occupies CD3 + Percentage of T lymphocytes; Step 2: Calculate the linear detection value Step 3: Convert linear detection values ​​into detection probabilities Where e is a constant, e = 2.71828; when the detection probability P2 > 0.379, the cause of diarrhea is determined to be enterovirus infection; when the detection probability P2 ≤ 0.379, the cause of diarrhea is determined to be gastrointestinal graft-versus-host disease.

Images