System and method for predicting graft rejection

Abstract

A computer-implemented method and system for predicting graft rejection are disclosed herein. The method and system can calculate, based on a score, whether graft rejection will occur in a transplant recipient or the probability of prediction of the degree to which graft rejection will occur. The score can be calculated based on transplant recipient data and one or more parameter weights. The transplant recipient data can include cell-free DNA from the transplant donor (dd-cfDNA) and one or more other types of parameters, such as one or more clinical parameters, one or more functional parameters, one or more immunological parameters, one or more transplant recipient characteristics, or one or more transplant characteristics. In some embodiments, the parameters used to calculate the probability of prediction do not include histological parameters.

Description

Technical Field

[0001] (Cross - Reference to Related Applications) This application claims the benefit of U.S. Provisional Patent Application No. 63 / 348,971, filed on June 3, 2022, and U.S. Provisional Patent Application No. 63 / 358,484, filed on July 5, 2022, both of which are hereby incorporated by reference in their entirety.

[0002] (Field of the Invention) The present disclosure generally relates to systems and methods for determining the status of an allograft, including predicting graft rejection.

Background Art

[0003] The transplantation of cells, tissues, parts or the whole of organs is a life - saving medical procedure when an individual experiences acute organ failure or suffers from some malignant tumor. Many organs, including but not limited to the heart, kidney, liver, lung, and pancreas, can be successfully transplanted, and one of the most common types of organ transplantation performed today is kidney transplantation.

[0004] When non - self (allogeneic) cells, tissues, or organs (allografts) are transplanted into a recipient, the recipient's immune system recognizes that the allograft is foreign to the body and activates various mechanisms to reject the allograft. Therefore, it is necessary to medically suppress such an immune response to minimize the risk of graft rejection. After transplantation, the status of the transplant can be monitored by various clinical diagnostic tests that may include invasive histopathological evaluation of transplant biopsy tissue. However, histopathological evaluation (e.g., biopsy) only provides information about the current rejection status of the allograft and lacks the ability to provide prognosis and / or predictive risk assessment regarding future allograft dysfunction and / or allograft rejection.

[0005] A physician or medical expert may have their own criteria for weighting the contribution of measured information to the overall assessment of the status of an allograft. The measured information can be collected from various clinical laboratory diagnostic tests, demographic information, and other parameters. An objective and consistent quantification of the risk of graft rejection would be useful to guide treatment options such as immunosuppressive therapy and daily clinical care, possibly in comparison to a reference set of allograft statuses.

[0006] There is a need for a system and method for predicting graft rejection based on an objective, comprehensive, and consistent relative assessment that ultimately improves diagnostic accuracy and reliability, as provided by the present disclosure. SUMMARY OF THE INVENTION

[0007] A computer-implemented method for determining the risk of graft rejection using a machine learning system is disclosed. The method includes receiving, via a computer or an input function, recipient data of a transplant recipient including a set of parameters, wherein the set of parameters includes donor-derived cell-free DNA (dd-cfDNA); receiving one or more parameter weights; calculating a score based on the recipient data and the one or more parameter weights; and calculating, based on the score, a prediction probability of whether or not graft rejection will occur or the degree to which graft rejection will occur in the transplant recipient. Additionally or alternatively, in some embodiments, the set of parameters of the recipient data further includes one or more clinical parameters including the time from transplantation to evaluation, one or more functional parameters including estimated glomerular filtration rate (eGFR), creatinine, or proteinuria, one or more immunological parameters including the mean fluorescence intensity of donor-specific antibodies or the number of anti-human leukocyte antigen (HLA) mismatches, one or more recipient characteristics including the age of the transplant recipient or donor organ infection information, or one or more transplant characteristics including previous transplant information or previous rejection information. Additionally or alternatively, in some embodiments, the set of parameters does not include histological parameters. Additionally or alternatively, in some embodiments, the computer-implemented method further includes generating a prediction of the recipient data within a reference set, the reference set including one or more other transplant recipients having one or more common characteristics. Additionally or alternatively, in some embodiments, calculating the score includes multiplying each parameter weight by the corresponding parameter of the recipient data for the parameter weight and calculating the score from the sum of the multiplications. Additionally or alternatively, in some embodiments, calculating the prediction probability includes determining the intercept of a multivariable logistic regression model and calculating the prediction probability from the score and the intercept.Additionally or alternatively, in some embodiments, one or more parameter weights are obtained from a machine learning model trained to: obtain a cohort dataset including a first set of model parameters of transplant recipients within a cohort and cohort graft rejection information, wherein the first set of model parameters includes dd-cfDNA, analyze the first set of model parameters for a relationship between the cohort dataset and the corresponding cohort graft rejection information, wherein the first set of model parameters is analyzed individually, select a second set of model parameters from the first set of model parameters, wherein the second set of model parameters meets one or more first criteria, select a third set of model parameters from the second set of model parameters, wherein the third set of model parameters includes independent variables related to graft rejection and meets one or more second criteria, and generate one or more parameter weights corresponding to the third set of model parameters of the cohort dataset. Additionally or alternatively, in some embodiments, the first set of model parameters further includes one or more of: one or more clinical parameters including kidney graft dysfunction, time since last graft rejection, or time from transplant to assessment; one or more functional parameters including estimated glomerular filtration rate (eGFR) or proteinuria; one or more immunological parameters including mean fluorescence intensity of donor-specific antibodies or number of anti-human leukocyte antigen (HLA) mismatches; one or more recipient and donor characteristics including recipient age, recipient gender, or donor organ infection information; or one or more transplant characteristics including donor age, donor gender, donor type, previous transplant information, cold ischemia time, or dual transplant kidney information. Additionally or alternatively, in some embodiments, the second set of model parameters includes dd-cfDNA, allograft dysfunction, recent graft rejection information, time from transplant to assessment, estimated glomerular filtration rate (eGFR), proteinuria, mean fluorescence intensity of donor-specific antibodies, recipient age, recipient gender, donor age, donor gender, donor type, previous transplant information, cold ischemia time, dual transplant information, and number of anti-human leukocyte antigen (HLA) mismatches.Additionally or alternatively, in some embodiments, a third set of model parameters includes dd-cfDNA, estimated glomerular filtration rate (eGFR), graft dysfunction, recent graft rejection information, and mean fluorescence intensity of donor-specific antibodies. Additionally or alternatively, in some embodiments, a machine learning model trained to analyze a first set of model parameters for relevance includes a machine learning model trained to analyze whether or not the parameters of the first set of model parameters identify the presence or absence of graft rejection in cohort graft rejection information or the degree to which the presence or absence of graft rejection is identified. Additionally or alternatively, in some embodiments, one or more first criteria or one or more second criteria include model parameters having a confidence interval above a confidence interval threshold or a p-value below a p-value threshold. Additionally or alternatively, in some embodiments, the confidence interval threshold is 95% and the p-value threshold is 0.2. Additionally or alternatively, in some embodiments, one or more first criteria include that the number of model parameters within a second set of model parameters is less than a threshold number. Additionally or alternatively, in some embodiments, a machine learning model trained to analyze a first set of model parameters for relevance includes a machine learning model trained to reduce the dimensionality of a cohort data set based on the presence or absence of graft rejection in cohort graft rejection information. Additionally or alternatively, in some embodiments, a machine learning model trained to analyze a first set of model parameters for relevance includes a machine learning model trained to determine the relevance between dd-cfDNA and one or more of the causes of end-stage renal disease or types of graft rejection. Additionally or alternatively, in some embodiments, a machine learning model trained to analyze a first set of model parameters includes a machine learning model trained to reduce the dimensionality of a cohort data set based on the type of graft rejection.Additionally or alternatively, in some embodiments, a machine learning model trained to select a third set of model parameters performs a backward selection by individually analyzing whether or not the parameters of the second set of model parameters identify the presence or absence of graft rejection in the cohort graft rejection information or the degree to which the presence or absence of graft rejection is identified, and compares the individual analyses to select a third set of model parameters. Additionally or alternatively, in some embodiments, one or more second criteria include that the number of model parameters within the third set of model parameters is less than a threshold number.

[0008] A system for classifying the state of a graft is disclosed. The system receives recipient data of a transplant recipient including a set of parameters including donor-derived cell-free DNA (dd-cfDNA), receives one or more parameter weights, calculates a score based on the recipient data and the one or more parameter weights, and based on the score, calculates a predicted probability of whether or not graft rejection will occur or the degree to which graft rejection will occur in the transplant recipient, and may include a scoring unit. The system may include one or more data processors and a non-transitory computer-readable storage medium including instructions that, when executed on the one or more data processors, cause the one or more data processors to execute some or all of one or more of the methods disclosed herein. In some embodiments, the system includes a non-transitory computer-readable storage medium including instructions that, when executed on the one or more data processors, cause the one or more data processors to execute some or all of one or more of the methods and / or some or all of one or more of the processes disclosed herein.

[0009] A computer program product is disclosed. The computer program product may be embodied in a tangible machine-readable storage medium and includes instructions configured to cause one or more data processors to execute some or all of one or more of the methods disclosed herein. BRIEF DESCRIPTION OF THE DRAWINGS

[0010]

Figure 1

Figure 2

Figure 3A

Figure 3B

Figure 3C

Figure 3D

Figure 4A

Figure 4B

Figure 5

Figure 6

Figure 7

DETAILED DESCRIPTION OF THE INVENTION

[0011] The present disclosure is at least partially based on the development of a computer-implemented method and system for predicting graft rejection.

[0012] By predicting the risk of allograft rejection in transplant recipients based on an objective, comprehensive, and consistent relative assessment, classification and treatment of transplant recipients according to the overall risk of allograft rejection become possible, thereby improving daily clinical care, avoiding unnecessary invasive treatments, and prolonging allograft engraftment.

[0013] Whether or not graft rejection occurs in a transplant recipient or the predicted probability of the degree to which graft rejection occurs can be calculated based on a score. The score can be calculated based on transplant recipient data and one or more parameter weights. The transplant recipient data can include donor-derived cell-free DNA (dd-cfDNA) and one or more other types of parameters, such as one or more clinical parameters, one or more functional parameters, one or more immunological parameters, one or more transplant recipient characteristics, or one or more transplant characteristics. In some embodiments, the parameters used to calculate the predicted probability do not include histological parameters.

[0014] The prediction probability can be used by a physician or medical expert. The physician or medical expert can input transplant recipient data into the interface. The scoring unit can receive the transplant recipient data and calculate a score based on the transplant recipient data and one or more parameter weights. The scoring unit can receive the parameter weights via a trained machine learning model. Also, the scoring unit can calculate a prediction probability. The prediction probability can be a binary representation (e.g., yes or no) or a quantitative value (e.g., 80%) indicating the degree to which graft rejection will occur in the transplant recipient. The prediction probability can be presented on a graphical user interface that is displayed to the physician or medical expert. The prediction probability can be an objective measure that complements or alternatively is used instead of the evaluation by the physician or medical expert.

[0015] The prediction probability can be provided by a medical analysis tool that is readily accessible to the physician or medical expert. The medical analysis tool can display the prediction probability, the prediction, or both on the user interface. The prediction probability can be an accurate and quantifiable measure of the state of the allograft.

[0016] By quantifying the prediction, the prediction probability and the prediction can become objective and more consistent. The prediction probability can be used to accurately compare the state of the allograft at one point in time with another point in time. Additionally or alternatively, the quantification can be used as guidance for determining treatment options and related timings. To provide consistency, reliability, and granularity, embodiments of the present disclosure can include computer-implemented tools and methods for evaluating measurements from transplant recipients. A systematic evaluation can help better characterize the response of the transplant recipient to treatment and can help provide information for the subsequent management of the transplant recipient. The results of the machine learning model and the medical analysis tool can be more reproducible such that the variability between transplant recipients or different measurement times for a given transplant recipient is reduced.

[0017] One or more parameter weights can be generated by a machine learning model trained to obtain a dataset including a set of parameters and graft rejection information from transplant recipients within a cohort. The machine learning model can analyze the parameters and perform one or more selection steps to select a subset of the parameters. The machine learning model can then generate parameter weights corresponding to the selected subset of parameters.

[0018] The machine learning model can be trained to select parameters related to the status of an allograft (e.g., the likelihood of predicted failure). The machine learning model can receive a cohort dataset from a cohort such as a derived cohort. The cohort dataset can include parameters that may or may not be related to allograft failure. In some embodiments, the parameters can have different degrees of relevance. The machine learning model can select the parameters with the highest degree of relevance to graft rejection. In some embodiments, the machine learning model can generate parameter weights according to the degree of relevance. The parameters most relevant to allograft failure can be weighted more.

[0019] The following description is presented to enable a person skilled in the art to make and use various embodiments. The description of specific devices, technologies, and applications is provided by way of example only. These examples are provided merely to add context and assist in understanding the described examples. Thus, it will be apparent to those skilled in the art that the described examples can be practiced without some or all of the specific details. Other applications are possible and the following examples should not be construed as limiting. Various modifications to the examples described herein will be readily apparent to those skilled in the art, and the general principles defined herein can be applied to other examples and applications without departing from the spirit and scope of the various embodiments. The various embodiments are not limited to the examples described and shown herein, but should be given a scope consistent with the claims.

[0020] Various techniques and process flow steps are described in detail with reference to the examples shown in the accompanying drawings. In the following description, numerous specific details are set forth in order to provide a thorough understanding of one or more aspects and / or features described or referenced herein. However, it will be apparent to those skilled in the art that one or more aspects and / or features described or referenced herein may be practiced without some or all of these specific details. In other instances, well-known process steps and / or structures have not been described in detail in order not to obscure some of the aspects and / or features described or referenced herein.

[0021] In the following description of the examples, reference is made to the accompanying drawings that form a part hereof, and in which are shown by way of illustration specific examples that may be practiced. It is to be understood that other examples may be used and structural changes may be made without departing from the scope of the disclosed examples.

[0022] The terms used in the description of the various embodiments described herein are for the purpose of describing particular embodiments only and are not intended to be limiting. As used in the description of the various described embodiments and the appended claims, the singular forms "a", "an", and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. The term "and / or" as used herein refers to any and all possible combinations of one or more of the associated listed items and includes them. It will also be understood that the terms "includes", "including", "comprises", and / or "comprising", when used herein, specify the presence of the stated features, integers, steps, operations, elements, and / or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof.

[0023] Exemplary System for Predicting Graft Rejection FIG. 1 shows an exemplary system 100 for predicting graft rejection according to an embodiment of the present disclosure. The system 100 may include an interface 160 and a scoring unit 170. Examples of the present disclosure may include some or all of the components shown in the figures, or other components not shown in the figures. The system 100 may be, for example, a medical analysis tool. A physician or medical professional can use the medical analysis tool to help monitor and / or classify the status of an allograft in a transplant recipient, and to monitor and / or propose adjustments to immunosuppressive therapy administered to or to be administered to the transplant recipient. Monitoring the status of the allograft includes analyzing various aspects that provide useful information regarding the physiological or health status of the allograft. The method of the present disclosure can be used to predict the probability of graft rejection based on measurement information indicating the status of the allograft. The predicted probability can be a quantitative percentage reflecting the probability of graft rejection occurring in the transplant recipient.

[0024] The interface 160 can receive user input of transplant recipient data (e.g., input from a physician or medical professional). Exemplary transplant recipient data can include information from laboratory tests regarding donor-derived cell-free DNA (dd-cfDNA), creatinine levels in plasma, serum, and / or urine at one or more time points, proteinuria, information regarding estimated glomerular filtration rate, time to post-transplant evaluation, transplant recipient characteristics such as age and gender, information regarding previous transplants, information regarding previous graft rejection events, and the like. The interface 160 will be described in more detail below.

[0025] The scoring unit 170 can be a tool for evaluating transplant recipient data (e.g., information from laboratory tests obtained from the blood and / or urine samples of the transplant recipient, or information from a biopsy). The scoring unit 170 can receive one or more parameter weights 190 (e.g., obtained from the machine learning model 150 shown in FIG. 4A), and can also receive transplant recipient data from the interface 160. The scoring unit 170 can calculate the status of the allograft of the transplant recipient. The status of the allograft can include a predicted probability or prediction. The scoring unit 170 can output the predicted probability and / or prediction 180.

[0026] FIG. 2 shows the predicted probability of graft rejection in a transplant recipient according to an embodiment of the present disclosure A flowchart of an exemplary method for calculating and / or generating a prediction of transplant recipient data in a reference set is shown. Method 200 may include step 202 where system 100 may receive transplant recipient data. The transplant recipient data may be received, for example, via interface 160. The transplant recipient data may be data related to a transplant recipient. The transplant recipient data may include measurement information from the transplant recipient. Exemplary measurement information may include creatinine levels in plasma, serum, and / or urine, proteinuria, urinary albumin, urinary microalbumin, urinary protein, estimated glomerular filtration rate (eGFR), urinary albumin-creatinine ratio, blood urea nitrogen, serum sodium, serum potassium, serum chloride, serum bicarbonate, serum calcium, serum albumin, complete blood count panel, liver function panel, lipid profile panel, coagulation panel, magnesium, phosphorus, brain natriuretic peptide, hemoglobin, uric acid, endostatin, number of HLA mismatches, mean fluorescence intensity (MFI) of donor-specific antibodies (DSA), and transplant recipient characteristics such as the age of the transplant recipient, time from transplant to evaluation, whether the transplant recipient has previously received a transplant, information regarding systemic infections such as viral infections, e.g., information regarding BK virus infection, but is not limited thereto. Optionally, the measurement information may also include information regarding whether the transplant recipient has previously experienced a graft rejection event. In some embodiments, the evaluation time may be the period (e.g., number of years) from when the measurement information is extracted until the transplant is performed.

[0027] In step 204, the system may receive one or more parameter weights. The one or more parameter weights may be received, for example, from a machine learning model. As will be described in more detail below, the machine learning model may be trained to generate parameter weights based on various input parameters such as a cohort dataset of transplant recipients and corresponding cohort graft rejection information. In various embodiments, the machine learning model or algorithm for analyzing various input parameters may include, but is not limited to, regularized classification models such as logistic regression, Ridge, Lasso, and Elastic Net, shortest shrinkage centroid, gradient boosting machine, random forest, support vector machine, k-nearest neighbor, neural network, and the like.

[0028] In step 206, the system may calculate a score using the scoring unit 170 based on the transplant recipient data and the parameter weights. The transplant recipient data may include a set of parameters. The set of parameters may include at least donor-derived cell-free DNA (dd-cfDNA). In some embodiments, the set of parameters may include parameters related to information on whether the transplant recipient has previously experienced a graft rejection event. The parameter weights may correspond to the set of parameters. In some embodiments, each parameter may have a corresponding parameter weight. In some embodiments, each parameter weight may indicate the degree of relevance of the parameter to graft rejection. For example, a higher parameter weight may indicate a higher relevance of the parameter to graft rejection. The score may be calculated based on the transplant recipient data and the parameter weights. In some embodiments, the score may be calculated by summing a plurality of multiplications. Each multiplication may be for each parameter weight and may include multiplying the corresponding parameter of the transplant recipient data by the parameter weight. The generation of the weights will be described in more detail below.

[0029] In step 208, the system can determine the state of the allograft, such as the health state. The health state can be a state indicating allograft tolerance or allograft rejection, or the possibility thereof. In some embodiments, determining the state of the allograft can include calculating a predicted graft rejection probability and / or generating a prediction. The prediction probability can be a numerical value (e.g., a percentage) indicating whether or not graft rejection will occur or the degree to which graft rejection will occur in the transplant recipient based on a score. In some embodiments, calculating the prediction probability includes calculating a probability from a score. In some embodiments, calculating the prediction probability includes determining an intercept of a multivariable logistic regression model and calculating the prediction probability from the score and the intercept. The prediction probability can be represented by a numerical value, a level or category indicating a numerical value, a binary indicator, etc. In some embodiments, the prediction probability and / or the prediction can indicate the current risk of graft rejection in the transplant recipient.

[0030] The prediction can be a visual aid used to explain the prediction probability of transplant recipients in the reference set. The reference set can include information from one or more other transplant recipients having one or more common characteristics, such as the time until post-transplant evaluation.

[0031] Embodiments of the present disclosure can include repeating one or more steps of method 200 and / or method 400 (described below). Although the description and figures show specific steps of the method in a particular order, the steps of the method may be performed in other orders not described or illustrated. Additionally or alternatively, embodiments of the present disclosure can include performing all, some, or none of the steps of method 200 and / or method 400, as appropriate. Further, although a particular component, device, or system is described as performing the steps of method 200 and / or method 400, any suitable combination of components, devices, or systems (including those not explicitly disclosed) can be used to perform the steps.

[0032] In some embodiments, the system can output a prediction probability and / or a prediction, such as by displaying the prediction probability and / or the prediction on a user interface. The user interface can be easily accessible and can be included in a medical analysis tool that immediately provides the status of the allograft to a physician or medical professional. FIG. 3A shows an exemplary user interface that displays a text box and an input box, according to an embodiment of the present disclosure. The user interface 300 can be a user interface (UI) displayed on a display of a device (e.g., a mobile phone, a tablet, a laptop computer, etc.). In some embodiments, the user interface 300 can be accessed by navigating the user to a website or an application page.

[0033] The user interface 300 may include one or more text boxes 304, one or more input boxes 308, or both. The text of a given text box 304 may be associated with the type of information input by the user into the corresponding input box 308. For example, text box 304A may display "Time from transplantation to evaluation (years)". The system may receive input from a user (e.g., a physician, nurse, medical assistant, medical professional, etc.) into the corresponding input box 308A. As another example, text box 304B may display "Patient's age (years)", and the corresponding input box 308B may be used by the user to input corresponding transplant recipient data. The transplant recipient data may include a (third) set of parameters used to calculate the probability of whether or not graft rejection occurs or the degree to which graft rejection occurs in the transplant recipient. As shown in the figure, exemplary transplant recipient data includes the following parameters, namely, the time from transplantation to evaluation ("Time from transplantation to evaluation (years)"), the age of the transplant recipient ("Patient's age (years)"), whether the transplant recipient has previously received a transplant, e.g., a previous kidney transplant ("Previous kidney transplant"), information regarding a systemic infection, e.g., an infection by BK virus ("BK virus (log)"), whether there has been an onset of a recent rejection reaction in the transplant recipient ("Onset of recent rejection reaction"), eGFR ("eGFR mL / min / 1.73m 2 "), information regarding one or more of serum, plasma, and / or urine creatinine values at one or more time points ("Previous creatinine (mg / L)"), current creatinine value ("Current creatinine (mg / L)"), proteinuria value ("Proteinuria (g / g)"), number of HLA mismatches ("HLA mismatch"), MFI DSA ("MFI DSA"), or percentage of dd-cfDNA ("ddcfDNA (%)").

[0034] The user interface 300 may include one or more text boxes 306, such as an evaluation time, transplant recipient characteristics, transplant recipient data, transplant characteristics, recent rejection, functional parameters, and / or immunological parameters, as text boxes 304.

[0035] Additionally or alternatively, the user interface may include one or more graphical user interface buttons 310. The user interface buttons 310 and the input box 308 may be interactive, enabling the user to input data and parameters such as an evaluation time, transplant recipient characteristics, transplant recipient data, transplant characteristics, recent rejection, functional parameters, and / or immunological parameters. After the user inputs information into the input box 308, the user can operate the system to calculate a prediction probability and / or generate a prediction 180, for example, by clicking on a "send" graphical user interface button 310. For example, the direction and magnitude of the data and / or parameters projected in the form of a two-dimensional plot as shown in FIGS. 3B-3D may provide further insight for elucidating or interpreting the mechanism of the transplant recipient's rejection reaction and guiding the corresponding patient management and treatment strategies.

[0036] FIG. 3B shows an exemplary user interface displaying exemplary prediction probabilities and predictions according to an embodiment of the present disclosure. The user interface 350 may include text boxes and input boxes similar to the user interface 300 (of FIG. 3A). In the example shown in FIG. 3B, the transplant recipient data has the parameters shown in Table 1.

[0037] [Table 1]

[0038] The user interface 350 may further comprise one or more text boxes and / or one or more graphical representations. The text box 314 may provide an output of the predicted probability, and the graphical representation 316 may provide a visual aid for the prediction. For example, as shown in the figure, from the exemplary parameters and transplant recipient data, the probability of rejection in a transplant patient is 0.8%.

[0039] Figures 3C and 3D show exemplary user interfaces for displaying the predicted probabilities and predictions of additional exemplary transplant recipients according to embodiments of the present disclosure. In Figure 3C, the exemplary transplant recipient is a 50-year-old woman who received a transplant two years ago. The transplant recipient has not experienced a recent rejection or, for example, a donor organ infection by BK virus, but has previously experienced a kidney transplant. The woman's previous creatinine was 0.7 mg / L, but her current creatinine is 1.2 mg / L, indicating kidney dysfunction. The woman's eGFR is 48 mL / min / 1.73m 2 and the proteinuria is 1 g / g. The number of HLA mismatches is 3, the anti-HLA DSA MFI is 7000, and the dd-cfDNA is 2%. The system may calculate a predicted probability of 80.9% and classify that the risk of rejection in this woman is high.

[0040] In Figure 3D, the exemplary transplant recipient is a 52-year-old woman who received a second kidney transplant but has not recently experienced a rejection. The time from transplant to evaluation was 3 years. The transplant recipient's previous and current creatinine values are 0.5 mg / L and 0.8 mg / L, respectively, and the eGFR is 48 mL / min / 1.73m 2 and the proteinuria is 0.5 g / g. The number of HLA mismatches is 3. The patient has no MFI DSA and the dd-cfDNA is 1%. The predicted probability calculated for this transplant recipient is 21%, as shown in the figure.

[0041] As described above, system 100 (e.g., a medical analysis tool) may generate one or more parameter weights. In some embodiments, the one or more parameter weights may be generated by system 100. The parameter weights are used to calculate the predicted probability or prediction of graft rejection for a given transplant recipient.

[0042] FIG. 4A shows an exemplary system for generating parameter weights according to an embodiment of the present disclosure. System 450 may include biomarker unit 120, baseline feature unit 130, database 140, and machine learning model 150.

[0043] Biomarker unit 120 may be configured to measure one or more features of a sample from a transplant recipient. Exemplary features can include, but are not limited to, donor-derived cell-free DNA (dd-cfDNA), anti-human leukocyte antigen (HLA) donor-specific antibody (DSA), and creatinine levels. Dd-cfDNA may be expressed as a percentage of total cell-free DNA or as an absolute concentration. Anti-HLA DSA may refer to the presence of donor-specific anti-HLA antibodies in the transplant recipient. Creatinine levels refer to the amount of chemical waste filtered by the kidneys of the transplant recipient. In some embodiments, the blood sample and corresponding features can be obtained simultaneously with the biopsy.

[0044] In some embodiments, the characteristics of a sample from a transplant recipient are determined according to an experimental (laboratory) workflow involving extracting cell-free DNA, i.e., cell-free DNA containing cell-free DNA from the transplant recipient and the graft, from a blood, plasma, serum, and / or urine sample obtained from transplant recipients in a derivation cohort and a validation cohort (cohorts are discussed in more detail below). In some embodiments, the level or amount of donor-derived cell-free DNA (dd-cfDNA) can be determined with targeted amplification and targeted high-throughput sequencing of selected polymorphic markers, e.g., selected single nucleotide polymorphisms (SNPs), as described respectively in U.S. Patent Application No. 14 / 658,061, filed on March 13, 2015, and U.S. Patent Application No. 17 / 351,040, filed on June 17, 2021, both of which are hereby incorporated by reference in their entirety. A polymorphic marker represents a locus at which two or more alternative nucleic acid sequences or alleles exist due to changes in one or more bases, one or more insertions, one or more repeats, one or more deletions, and their diversities. Thus, in some embodiments, the level or amount of donor-derived cell-free DNA (dd-cfDNA) can be determined with targeted analysis of alternative polymorphic markers such as short tandem repeats (STRs), restriction fragment length polymorphisms (RFLPs), variable number tandem repeats (VNTRs), hypervariable regions, minisatellites, dinucleotide repeats, trinucleotide repeats, tetranucleotide repeats, simple sequence repeats, and insertion elements. In some embodiments, dd-cfDNA can be quantified as a percentage of total cfDNA or as an absolute concentration. In some embodiments, the level or amount and characteristics of dd-cfDNA in a sample from a transplant recipient can be determined after receiving experimental data, e.g., sequencing reads, or other data-related information, e.g., quality control-related data, results, genotype information, SNP mutation rates, etc., from a database or other non-experimental source.

[0045] Additionally or alternatively, biomarker unit 120 may determine the mean fluorescence intensity (MFI) of DSA. The MFI may indicate the donor-specific antibody strength. In some embodiments, the beads can be classified according to a predetermined normalized MFI, such as less than 500, 500 - 3000, 3000 - 6000, and greater than 6000. A single antigen flow bead assay can be used to determine whether there are DSA present against one or more antigens (e.g., HLA-A, HLA-B, HLA-Cw, HLA-DR, HLA-DQ, HLA-DP, etc.). Embodiments of the present disclosure may include using HLA typing of the transplant recipient and donor to identify HLA antigens. Parameters for predicting the probability of graft rejection may include one or more immunological variables such as circulating anti-HLA DSA.

[0046] In some embodiments, biomarker unit 120 may measure or calculate one or more functional parameters. One non-limiting exemplary functional parameter is estimated glomerular filtration rate (eGFR). Glomerular filtration rate (GFR) refers to the amount of blood filtered by the kidneys and can indicate its renal function. Another non-limiting exemplary functional parameter is proteinuria, which is the level of excess protein in the urine of the transplant recipient. Yet another exemplary functional parameter is the time from transplantation to (risk) assessment. In some embodiments, the time of risk assessment may be the time when a biopsy or blood sample is extracted or measured from the transplant recipient.

[0047] The baseline feature unit 130 can be configured to receive, store, and / or determine one or more baseline features specific to a given transplant recipient. Exemplary baseline features can include, but are not limited to, donor comorbidities, recipient comorbidities, recipient characteristics such as age, gender, height, weight, glomerulopathy, polycystic kidney disease, diabetes, causes of end-stage renal disease including vascular disease and not limited thereto, and other previous transplant information (including information regarding previous transplants), dual transplant information (including dual kidney transplant information), donor characteristics (e.g., age, gender, height, weight, whether deceased, expanded criteria donor, etc.), and transplant characteristics (e.g., cold ischemia time, HLA-A / B / DR mismatch, ABO-incompatible transplant, graft weight, etc.). The parameters of the predicted probability of graft rejection can include one or more baseline features.

[0048] The database 140 can be configured to store one or more data sets based on data from the biomarker unit 120 and / or the baseline feature unit 130. For example, the database 140 can store cohort data sets such as a derived cohort data set and a validation cohort data set. The cohort data set can include measurement information from samples obtained from transplant recipients who were recipients of allografts (e.g., organ grafts, tissue grafts, cell grafts, etc.) from donors. The derived cohort data set can include data from transplant recipients having a first feature (e.g., who received a transplant during a first period), and the validation cohort data set can include data from transplant recipients having a second feature (e.g., who received a transplant during a second period). In some embodiments, one or more parameters (e.g., baseline features, average dd-cfDNA level or concentration, diagnosis, graft rejection information, etc.) of the transplant recipients within the derived cohort can be similar (e.g., average value ± a specific percentage deviation) to the baseline features of the transplant recipients within the validation cohort.

[0049] Examples of transplant organs include, for example, the heart, kidney, lung, liver, pancreas, cornea, organ system, angiogenesis composite allograft, intestinal graft, or other solid organs, and combinations thereof. Examples of grafts that a recipient receives from a donor also include, for example, other allografts such as bone marrow grafts, pancreatic islet cells, stem cells, skin tissue, skin cells or xenografts.

[0050] In some embodiments, the allograft is a cellular allograft, for example, a graft comprising allogeneic cells derived from a donor. These include cells directly harvested from the donor for administration to the recipient, cells harvested from the donor and genetically engineered before administration to the recipient, cells harvested from the donor and cultured before administration to the recipient, cells harvested from the donor and subjected to a manufacturing process before administration to the recipient, and any combination thereof, but are not limited thereto. The cells can also be stored ( "off-the-shelf" cells) before administration to the recipient.

[0051] The recipient of the transplantation can be administered one or more various allogeneic cells. Examples of allogeneic cells include, but are not limited to, blood cells, stem cells, cardiomyocytes, neurons, lymphocytes, NK cells, NKT cells, Treg cells, macrophages, dendritic cells, and pancreatic islet cells. In some embodiments, the allogeneic cells are allogeneic blood cells. Examples of allogeneic blood cells include hematopoietic stem cells (HSCs), T cells, B cells, and CAR T cells, NK cells, NKT cells, or TILs. In some embodiments, the allogeneic cells are allogeneic T cells. In some embodiments, the allogeneic cells are administered as bone marrow, cord blood, or purified allogeneic cells. In some embodiments, the allogeneic cells are bone marrow cells. In some embodiments, the allogeneic cells are cord blood cells. In some embodiments, the allogeneic cells are allogeneic CAR T cells, allogeneic universal CAR T cells (i.e., when the CAR binds to an antibody that binds to a specific antigen), allogeneic split CAR T cells (i.e., when a dimerizer activates CAR T cell function), allogeneic activatable CAR T cells, allogeneic suppressible CAR T cells, allogeneic multivalent CAR T cells (i.e., when the CAR must bind to multiple specific antigens and / or agents to induce T cell activation), allogeneic tumor-infiltrating lymphocytes, allogeneic regulatory T cells, allogeneic genetically engineered T cells, or allogeneic T cells having a genetically engineered or synthetic T cell receptor (TCR), virus-specific T cells (e.g., EBV, HPV, BKV, CMV, etc.), antigen-specific T cells, neoantigen-specific T cells, or any cells isolated from a donor. In some embodiments, the allogeneic cells are derived from a donor. In some embodiments, the allogeneic cells are a cell allograft.

[0052] Database 140 may also store cohort graft rejection information. The cohort graft rejection information may include the outcome of the rejection reaction of the transplant recipient in the corresponding cohort. In some embodiments, the outcome of the rejection reaction may be based on whether the allograft history of the transplant recipient exhibits a pattern of lesions that meet the Banff criteria for active AMR (A-AMR), chronic active AMR (CA-AMR), acute TCMR, or chronic active TCMR (CA-TCMR). The cohort dataset obtained from database 140 may include data corresponding to individual transplant recipients of the cohort and / or representative data (e.g., mean values, averages, medians, percentages, etc.) of a group of transplant recipients of the cohort (including all or less than all of the transplant recipients within the cohort).

[0053] Machine learning model 150 may be configured to receive a cohort dataset from database 140, analyze one or more sets of parameters associated with the dataset, select one or more sets of parameters, and generate one or more parameter weights. Machine learning model 150 may be trained to identify a relationship between the cohort dataset and the state of the allograft and then generate parameter weights based on the identified relationship. Machine learning model 150 may be able to narrow down the set of parameters used in generating the weights by performing one or more selection steps.

[0054] Figure 4B shows a flowchart of an exemplary method executed by a machine learning model according to an embodiment of the present disclosure. Method 400 may include, at step 402, obtaining a cohort dataset, for example, from database 150. The cohort dataset may include a first set of parameters of transplant recipients in the cohort and cohort rejection information. The cohort dataset may be data obtained by one or more units such as biomarker unit 120 and baseline feature unit 130. The first set of parameters may include, without limitation, dd-cfDNA and one or more of clinical parameters, functional parameters, immunological parameters, recipient characteristics, or transplant characteristics.

[0055] Clinical parameters may include allograft dysfunction, time since the last rejection if a previous rejection occurred, or time from transplantation to evaluation. Allograft dysfunction, for example, kidney allograft dysfunction, may refer to an increase in serum creatinine of more than 0.3 mg / liter or more than 50% from baseline serum creatinine. Baseline serum creatinine may be, for example, the lowest serum creatinine value for a given transplant recipient during the month prior to evaluation, or the last known serum creatinine.

[0056] Functional parameters may include eGFR, creatinine, or proteinuria. Immunological parameters may include the mean fluorescence intensity of anti-human leukocyte antigen (HLA) donor-specific antibodies or the number of HLA mismatches. Recipient characteristics may include the age or gender of the recipient. Transplant characteristics may include the mass of the graft, the age of the donor, the gender of the donor, the type of donor (living-related, deceased-related, living-unrelated, deceased-unrelated), the weight of the donor, the height of the donor, previous transplant information, for example, previous kidney transplant information, cold ischemia time, or dual transplant information, for example, dual transplant kidney information.

[0057] In step 404, the machine learning model 150 may analyze a first set of parameters for the relevance between the cohort dataset and the corresponding cohort graft rejection information. For each parameter in the first set of parameters, the machine learning model may determine whether there is a relevance with the corresponding cohort graft rejection information. This step may include analyzing whether the parameters of the first set of parameters identify the presence or absence of graft rejection in the rejection information. In some embodiments, the first set of parameters may be analyzed individually from each other. That is, each parameter may be analyzed without considering other parameters within the first set of parameters.

[0058] As a non-limiting example, the cohort may be a derived cohort including 637 transplant recipients. The mean values from the cohort dataset are shown in Table 2. The machine learning model 150 may analyze the diagnosis of the transplant recipient to determine that a particular parameter is not associated with graft rejection. For example, out of 637 transplant recipients, 85 were diagnosed with active AMR, 23 were diagnosed with chronic AMR, 7 were diagnosed with inactive AMR, 16 were diagnosed with acute TCMR, 3 were diagnosed with chronic active TCMR, 7 were diagnosed with mixed rejection reaction, 12 were diagnosed with borderline lesions, 14 were diagnosed with viral nephritis, 12 were diagnosed with glomerulonephritis without rejection reaction, 12 were diagnosed with FSGS, 219 were diagnosed with IFTA, and 227 were diagnosed with no specific lesions. The machine learning model may determine that there is little or no distinction in the presence or absence of graft rejection. As another example, the machine learning model 150 may analyze the anti-HLA DSA MFI and determine that this parameter is associated with graft rejection. For example, out of 637 transplant recipients, the anti-HLA DSA MFI of 383 transplant recipients was less than 500, less than 194 transplant recipients were between 500 and 3000, 27 transplant recipients were between 3000 and 6000, and 33 transplant recipients were over 6000. The machine learning model may determine that there is a relevance between the anti-HLA DSA MFI parameter and the presence of graft rejection in the cohort dataset.

[0059] dd-cfDNA levels or concentrations that exceed a threshold or cutoff value may be associated with graft rejection. In some embodiments, analyzing a first set of model parameters for association includes determining an association between the level or concentration of dd-cfDNA and one or more of a cause of end-stage renal disease or a type of graft rejection.

[0060] In some embodiments, the machine learning model 150 may determine that there is a relatively low association between dd-cfDNA and one or more histological parameters. Embodiments of the present disclosure include a set of parameters used for calculating scores and prediction probabilities and / or generating predictions as not including histological parameters.

[0061] In step 406, the machine learning model 150 may select a second set of parameters from the first set of parameters. The second set of parameters may be selected based on meeting one or more (first) criteria. The machine learning model 150 may select one parameter at a time for evaluation. For example, the machine learning model 150 may evaluate a first parameter to determine whether it meets a first criterion. The machine learning 150 may then evaluate a second parameter to determine whether it meets the first criterion. In some embodiments, the analysis (step 404) and selection (step 406) of the first set of parameters may be performed together such that the results from the analysis lead to the selection.

[0062] The selection of parameters (including the first parameter set, the second parameter set, and / or the third parameter set) can be related to the selected parameters and graft rejection. The level of association between the parameters and graft rejection can be represented by one or more statistical probabilities such as odds ratio (OR), confidence interval (CI), or p-value. The odds ratio can quantify the strength of the association between the parameters and graft rejection. The confidence interval can represent the probability that the parameter is within a specific interval around the mean plus or minus the standard deviation. For example, a 95% confidence interval can mean that 95% of the data for a given parameter has values around the mean plus or minus the standard deviation.

[0063] The p-value can be a measure of the probability that the difference between two data sets is real. The p-value can represent the difference between the presence (first data set) and absence (second data set) of graft rejection. A smaller p-value can correspond to a larger difference. For example, the smaller the p-value of a corresponding parameter, the greater the discrimination amount of the presence or absence of graft rejection in the rejection information of that parameter.

[0064] One or more first criteria can include parameters having a confidence interval above a confidence interval threshold, a p-value below a p-value threshold, an odds ratio greater than an odds ratio threshold, or a combination thereof. The threshold or thresholds generally refer to any predetermined level or range of levels indicating the relevance of the parameter to the presence or absence of the risk of graft rejection in the transplant recipient. The threshold can take various forms. It can be a single cut-off value such as a median or an average value.

[0065] The confidence interval threshold and the p-value threshold can be predetermined values such as 95% and 0.2, respectively. In other words, parameters having a confidence level greater than 95% and a p-value less than 0.2 can be selected to form a second set of parameters.

[0066] Embodiments of the present disclosure may include other criteria used to select a second set of parameters, such as that the number of parameters in the second set of parameters is less than a threshold number of parameters. Parameters that are not selected may not be considered in subsequent steps and thus may not be included in the calculation of the prediction probability. In some embodiments, the number of parameters in the second set of parameters may be less than the number of parameters in the first set of parameters.

[0067] FIG. 5 shows an exemplary data table of a second set of parameters of a cohort data set selected from a first set of parameters according to an embodiment of the present disclosure. Exemplary parameters of the second set of parameters may include, but are not limited to, dd-cfDNA level or concentration, allograft dysfunction, recent graft rejection information, time from transplantation to evaluation, eGFR, proteinuria, mean fluorescence intensity of donor-specific antibodies against human leukocyte antigen, recipient age, recipient gender, donor age, donor gender, donor type, previous kidney transplantation information, cold ischemia time, dual kidney transplantation information, and number of HLA mismatches. Further parameters may include information from biopsies.

[0068] Referring back to FIG. 4B, at step 408, the machine learning model may select a third set of parameters from a second set of parameters. The third set of parameters may include independent variables related to graft rejection. In some embodiments, the third set of parameters may be selected based on meeting one or more second criteria. In some embodiments, the selection of the third set of parameters may include performing a backward selection that individually analyzes the parameters of the second set of parameters to determine whether the parameters identify the presence or absence of graft rejection in the rejection information. The individual analysis may include temporarily removing the parameter of interest to determine whether the parameter of interest affects the level of relevance. In some embodiments, each parameter may be analyzed and then the analyses compared to select the third set of parameters. For example, a first analysis may include temporarily removing the eGFR parameter and determining that the level of relevance with the remaining parameter set is lower. The machine learning model 150 may select the eGFR parameter as part of the third set of parameters. A second analysis may include temporarily removing the time from transplantation to biopsy parameter and determining that the level of relevance did not change significantly, such that the machine learning model 150 may not select the time from transplantation to biopsy parameter as part of the third set of parameters.

[0069] The one or more second criteria may include parameters having a confidence interval above a confidence interval threshold or a p-value below a p-value threshold. In some embodiments, the confidence interval threshold and the p-value threshold may be predetermined values such as 95% and 0.2, respectively. In other words, parameters having a confidence level greater than 95% and a p-value less than 0.2 may be selected to form the second set of parameters. In some embodiments, the one or more second criteria may be the same as the one or more first criteria. In other embodiments, the one or more second criteria may have a p-value threshold that is less than the p-value threshold of the one or more first criteria.

[0070] Additionally or alternatively, other criteria may be used to select a third set of parameters. For example, the machine learning model 150 may select the third set of parameters based on the number of parameters in the third set of parameters being less than a predetermined number of parameters. The selected third set of parameters may be, for example, the parameters having the lowest p-value among the second set of parameters (e.g., the N parameters having the lowest p-value). The unselected parameters may not be considered in subsequent steps and thus may not be included in the calculation of the prediction probability. In some embodiments, the number of parameters in the third set of parameters may be less than the number of parameters in the second set of parameters.

[0071] The machine learning model 150 can determine that the third set of parameters is highly associated with graft rejection. For example, the machine learning model 150 can select the parameters with the highest association with graft rejection. These parameters can include, for example, specific baseline characteristics (e.g., the age of the transplant recipient, previous transplant information, time from transplantation to evaluation), immunological variables (e.g., the number of HLA mismatches, dd-cfDNA, anti-HLA DSA information), and information regarding systemic infections by viruses, bacteria, fungi, and parasites that the transplant recipient may have contracted during or after the transplantation process, such as BK virus information. Infectious agents that can cause systemic infections and whose presence or level can be tested can include, but are not limited to, viruses such as cytomegalovirus, Epstein-Barr virus, anelloviridae, and BK virus; bacteria such as Pseudomonas aeruginosa, Enterobacteriaceae, Nocardia, Streptococcus pneumonia, Staphylococcus aureus, Legionella; fungi such as Candida, Aspergillus, Cryptococcus, Pneumocystis carinii; or parasites such as Toxoplasma gondii.

[0072] Figure 6 shows an exemplary data table of the third set of parameters of a cohort data set selected from the second set of parameters according to an embodiment of the present disclosure. The exemplary third set of parameters can include parameters such as (but not limited to) dd-cfDNA level or concentration, allograft dysfunction, recent graft rejection information, and anti-HLA DSA MFI.

[0073] In step 410 (of FIG. 4B), the machine learning model may generate one or more parameter weights corresponding to a third set of parameters of the cohort dataset. For example, the machine learning model may generate seven parameter weights for seven parameters of the third set of parameters. The parameter weights may be selected and may correspond to the relevance of the individual parameters to the state of the allograft (e.g., the predicted likelihood of allograft failure). For example, the first parameter may have a corresponding first parameter weight, and the second parameter may have a corresponding second parameter weight. The first parameter may have a higher degree of relevance to the prediction probability such that the first parameter weight has a higher value than the second parameter weight. Other relationships and methods for determining the parameter weights may also be used. The parameter weights may be calculated using a multivariable logistic regression model.

[0074] Embodiments of the present disclosure may include training a machine learning model. The machine learning model may be trained by using a derived dataset for a derived cohort when analyzing a first set of parameters. The machine learning model may receive the derived dataset for the derived cohort and corresponding derived cohort graft rejection information. The machine learning model may also receive the first set of parameters and analyze the first set of parameters for the relevance between the derived dataset and the corresponding (derived cohort) graft rejection information. The first set of parameters may be analyzed individually.

[0075] During the training phase, the machine learning model may select a second set of parameters from a first set of parameters, where the second set of parameters meets one or more first criteria. The machine learning model may then select a third set of parameters from the second set of parameters. The selected third set of parameters may include independent variables related to graft rejection and meet one or more second criteria. In some embodiments, the selection of parameters for the third set may be based on relevance. In some embodiments, the parameters selected for the third set may have a higher relevance than the unselected parameters. The machine learning model may use the third set of parameters and, as described above, generate one or more parameter weights corresponding to the third set of parameters (of the derived dataset). The machine learning model can calculate a derived score based on the third set of parameters of the derived dataset and the calculated parameter weights.

[0076] During the validation phase, the machine learning model may be tested by applying one or more parameter weights generated from the derived dataset (during the training phase) to the third set of parameters of the validation dataset. The validation phase is used to validate the trained model and ensure that the output of the machine learning model corresponds to the data used to train the model. The machine learning model can calculate a validation score and / or the state (e.g., predicted probability) of the corresponding allograft based on the third set of parameters of the validation dataset and the parameter weights. The performance of the machine learning model can be determined by comparing the derived score and the validation score. In some embodiments, the performance may be based on the discrimination of the scores and the area under the receiver operating characteristic (ROC) curve.

[0077] Embodiments of the present disclosure may include implementing one or more rules regarding the selection of parameters. The rules may be implemented by criteria for one or more selection steps. For example, the criteria may include a threshold number of parameters, a threshold relevance, a threshold number of cohorts, etc.

[0078] If the machine learning model is not sufficiently trained, training data (e.g., derived dataset) can be modified to provide feedback to the model. The output of the machine learning model between training iterations can be evaluated by a physician or medical expert to determine which data within the training data should be modified. The physician or medical expert can modify specific data in areas where there is potential for improvement, such as which parameters should have a higher relevance.

[0079] Examples of administration of immunosuppressive therapy Immunosuppressive therapy generally refers to the administration of an immunosuppressive agent or other therapeutic agent that suppresses the immune response in a transplant recipient. Examples of immunosuppressive agents include, for example, anticoagulants, antimalarials, heart medications, non-steroidal anti-inflammatory drugs (NSAIDs) and steroids, such as Ace inhibitors, aspirin, azathioprine, B7RP-1-fc, β-blockers, brequinar sodium, Campath-1H, celecoxib, chloroquine, corticosteroids, coumadin, cyclophosphamide, cyclosporin A, DHEA, deoxypeganine, dexamethasone, diclofenac, drobid, etodolac, everolimus, FK778, felodene, fenoprofen, flurbiprofen, heparin, hydralazine, hydroxychloroquine, CTLA-4 or LFA3 immunoglobulin, ibuprofen, indomethacin, ISAtx-247, ketoprofen, ketorolac, leflunomide, meclofenamate, mefenamic acid, mepacrine, 6-mercaptopurine, meloxicam, methotrexate, mizoribine, mycophenolate mofetil, naproxen, oxaprozin, plakenil, NOX-100, prednisone, methylprednisolone, rapamycin (sirolimus), sulindac, tacrolimus (FK506), thymoglobulin, tolmetin, trespolimus, UO126, etc., and antibodies, such as alpha lymphocyte antibody, adalimumab, anti-CD3 antibody, anti-CD25 antibody, anti-CD52 antibody, anti-IL2R antibody, and anti-TAC antibody, basiliximab, daclizumab, etanercept, hu5C8, infliximab, OKT4 and natalizumab, etc.

[0080] In some embodiments, the prediction probability or the variance of the prediction probability not changing over time may indicate that there is no need to adjust the immunosuppressive therapy administered to the transplant recipient, or that the administered immunosuppressive therapy can be maintained. The decision to maintain the immunosuppressive therapy administered to the transplant recipient may be based on additional clinical factors such as the health status of the transplant recipient. In some embodiments, the immunosuppressive therapy administered to the transplant recipient is maintained.

[0081] In some embodiments, adjusting the immunosuppressive therapy includes changing the type or form of the immunosuppressive agent or other immunosuppressive therapy administered to the transplant recipient. In some embodiments, if the transplant recipient is not receiving immunosuppressive therapy, the methods of the present disclosure may indicate a need to initiate administration of immunosuppressive therapy to the transplant recipient.

[0082] Other transplant-related therapies include treatments or therapies other than transplantation or immunosuppressive therapy administered to the transplant recipient to promote engraftment of the graft or to treat transplant-related symptoms (e.g., cytokine release syndrome, neurotoxicity). Examples of other transplant-related therapies include, but are not limited to, administration of antibodies, antigen-targeted ligands, non-immunosuppressive drugs, and other agents that stabilize or destabilize graft components that are important for transplant activity or that directly activate or inhibit one or more transplant activities. These activities may include the ability to induce an immune response, the ability to recognize specific antigens, the ability to replicate, and / or the ability to induce repair of damaged tissue. Adjusting the immunosuppressive therapy can be combined with adjusting, initiating, or discontinuing other transplant-related therapies.

[0083] The method of the present disclosure can predict the probability of graft rejection in a transplant recipient or provide a prediction of risk in a reference set of transplant recipients. The predicted probability and / or prediction can be used to inform the need to adjust the monitoring of the transplant recipient. Generally, the change over time of the predicted risk is useful in determining the need to adjust the monitoring of the transplant recipient. In some embodiments, determining the state of the graft as described above is useful in determining the need to adjust the monitoring of the transplant recipient.

[0084] Depending on the state of the transplant, the monitoring of the transplant recipient can be appropriately adjusted. For example, the monitoring can be adjusted by increasing or decreasing the frequency of monitoring as appropriate. The monitoring can be adjusted by changing the means of monitoring, for example, by changing the metric used to monitor the transplant recipient.

[0085] Exemplary system for calculating the predicted probability of graft rejection or generating a prediction The systems and methods described above can be implemented by a device. FIG. 7 shows an exemplary device that implements the previously disclosed systems and methods according to an embodiment of the present disclosure. The device 702 can be a portable electronic device such as a mobile phone, a tablet computer, a laptop computer, or a wearable device. The device 702 can include a processor 704 (e.g., a central processing unit (CPU), a graphics processing unit (GPU), or both), a main memory 706 (e.g., a dynamic random access memory (DRAM) such as a read-only memory (ROM), a flash memory, a synchronous DRAM (SDRAM), or a Rambus DRAM (RDRAM)), and a static memory 708 (e.g., a flash memory, a static random access memory (SRAM), etc.), which can communicate with each other via a bus 710.

[0086] Device 702 may also include a display 712, an input / output device 714 (e.g., a touch screen), a transceiver 716, and a storage 718. Storage 718 includes a machine-readable medium 720 storing one or more instruction sets 724 (e.g., software) embodying any one or more of the methods or functions described herein. The software may also be wholly or at least partially present in main memory 706 and / or in processor 704 during execution by computer 702, and main memory 706 and processor 704 also constitute machine-readable media. The software may further be transmitted or received over a network via network interface device 722.

[0087] Although machine-readable medium 720 is shown as a single medium in one embodiment, the term "machine-readable medium" should be interpreted to include a single medium or multiple media (e.g., a centralized or distributed database and / or associated caches and servers) storing one or more instruction sets. The term "machine-readable medium" should also be interpreted to include any medium capable of storing, encoding, or carrying a set of instructions for machine execution and causing any one or more of the methods of the present invention to be executed by a machine. Thus, the term "machine-readable medium" should be interpreted to include, but not be limited to, solid-state memory, optical and magnetic media, and carrier wave signals.

[0088] The systems, methods, and data described herein may be stored in storage 718, main memory 706, static memory 708, or any combination thereof. Display 712 may be used to present a user interface to a physician or medical professional, and input / output device 714 may be used to receive input (e.g., clicking on a graphic representing a microblog) from a physician or medical professional. Transceiver 716 may be configured to communicate with a network, for example.

[0089] Examples of the present disclosure are fully described with reference to the accompanying drawings, it should be noted that various changes and modifications will be apparent to those skilled in the art. Such changes and modifications should be understood to be included within the scope of the examples of the present disclosure as defined by the appended claims.

Claims

1. A computer-implemented method for determining the risk of graft rejection using a machine learning system, comprising: receiving, via a computer or an input function, recipient data of a transplant recipient including a set of parameters, wherein the set of parameters includes donor-derived cell-free DNA (dd-cfDNA); receiving one or more parameter weights; calculating a score based on the transplant recipient data and the one or more parameter weights; and calculating a prediction probability of whether or not graft rejection will occur or the degree to which graft rejection will occur in the transplant recipient based on the score.

2. The set of parameters of the transplant recipient data further includes: one or more clinical parameters including the time from transplantation to evaluation; one or more functional parameters including estimated glomerular filtration rate (eGFR), creatinine, proteinuria, or a combination thereof; one or more immunological parameters including the mean fluorescence intensity of donor-specific antibodies, the number of anti-human leukocyte antigen (HLA) mismatches, or both; one or more transplant recipient characteristics including the age of the transplant recipient, donor organ infection information, or both; previous transplant information, previous rejection information, or one or more transplant characteristics including a combination thereof, the computer-implemented method according to claim 1.

3. The computer-implemented method according to claim 1, wherein the set of parameters does not include histological parameters.

4. The computer-implemented method according to claim 1, further comprising generating a prediction of the transplant recipient data within a reference set, the reference set including one or more other transplant recipients having one or more common characteristics.

5. The computer-implemented method according to claim 4, wherein the generated prediction is used to interpret the mechanism of graft rejection and / or to guide treatment.

6. Calculating the score comprises for each parameter weight, multiplying the parameter weight by the corresponding parameter of the transplant recipient data, and calculating the score from the sum of the multiplications, the computer-implemented method according to claim 1.

7. Calculating the prediction probability comprises determining the intercept of a multivariate logistic regression model, and calculating the prediction probability from the score and the intercept, the computer-implemented method according to claim 1.

8. The one or more parameter weights are obtaining a cohort data set comprising a first set of model parameters of transplant recipients in a cohort and cohort graft rejection information, analyzing the first set of model parameters for a relationship between the cohort data set and the corresponding cohort graft rejection information, selecting one or more subsequent sets of model parameters from the first set of model parameters or a preceding set of model parameters, selecting a last set of model parameters from the one or more subsequent sets of model parameters, the last set of model parameters comprising independent variables related to graft rejection and satisfying one or more second criteria, received from a machine learning model trained to generate the one or more parameter weights corresponding to the last set of model parameters of the cohort data set, the computer-implemented method according to claim 1.

9. The first set of the model parameters, or one or more subsequent sets of the model parameters, or both, include dd-cfDNA, and the one or more subsequent sets of the model parameters satisfy one or more first criteria, the computer-implemented method according to claim 8.

10. The first set of the model parameters one or more clinical parameters including graft dysfunction, time from the last graft rejection, time from transplantation to evaluation, or a combination thereof; one or more functional parameters including estimated glomerular filtration rate (eGFR), creatinine, proteinuria, or a combination thereof; one or more immunological parameters including the mean fluorescence intensity of donor-specific antibodies, the number of anti-human leukocyte antigen (HLA) mismatches, or a combination thereof; one or more recipient and donor characteristics including the recipient's age, the recipient's gender, donor organ infection information, or a combination thereof, the donor's age, the donor's gender, the donor's type, previous transplantation information, cold ischemia time, dual transplant kidney information, or one or more transplantation characteristics including a combination thereof, further comprising one or more of those according to claim 8, the computer-implemented method.

11. The one or more subsequent sets of the model parameters include one or more of dd-cfDNA, graft dysfunction, recent graft rejection information, time from transplantation to evaluation, estimated glomerular filtration rate (eGFR), proteinuria, mean fluorescence intensity of donor-specific antibodies, recipient's age, recipient's gender, donor's age, donor's gender, donor's type, previous transplantation information, cold ischemia time, dual transplant information, anti-human leukocyte antigen (HLA) mismatch, or a combination thereof, the computer-implemented method according to claim 8.

12. The computer-implemented method of claim 8, wherein the last set of the model parameters includes one or more of dd-cfDNA, estimated glomerular filtration rate (eGFR), graft dysfunction, recent graft rejection information, mean fluorescence intensity of donor-specific antibodies, proteinuria, or a combination thereof.

13. The computer-implemented method of claim 8, wherein analyzing the first set of the model parameters for relevance includes analyzing whether the parameters of the first set of the model parameters identify the presence or absence of graft rejection in the cohort graft rejection information or the degree of identifying the presence or absence of graft rejection.

14. The computer-implemented method of claim 8, wherein the first set of the model parameters is analyzed individually.

15. The computer-implemented method of claim 8 or 9, wherein the one or more first criteria or the one or more second criteria include model parameters having a confidence interval above a confidence interval threshold or a p-value below a p-value threshold.

16. The computer-implemented method of claim 15, wherein the confidence interval threshold is 95% and the p-value threshold is 0.

2.

17. The computer-implemented method of claim 9, wherein the one or more first criteria include that the number of model parameters in the set of the one or more subsequent model parameters is less than a threshold number.

18. The computer-implemented method of claim 8, wherein analyzing the first set of the model parameters for relevance includes reducing the dimension of the cohort data set based on the presence or absence of graft rejection in the cohort graft rejection information.

19. Analyzing the first set of the model parameters regarding the relatedness includes determining the relatedness between the dd-cfDNA and one or more of the cause of end-stage renal disease, the type of graft rejection, or a combination thereof. The computer-implemented method according to claim 8.

20. Analyzing the first set of the model parameters includes reducing the dimension of the cohort data set based on the type of graft rejection. The computer-implemented method according to claim 8.

21. Selecting the last set of the model parameters is performing backward selection by analyzing whether the parameters of the one or more subsequent sets of the model parameters identify the presence or absence of graft rejection in the cohort graft rejection information, or the degree of identifying the presence or absence of graft rejection, and comparing the individual analyses and selecting the last set of the model parameters. The computer-implemented method according to claim 8.

22. The one or more second criteria include that the number of model parameters in the last set of the model parameters is less than a threshold number. The computer-implemented method according to claim 8.

23. A system for classifying the state of a graft, comprising a scoring unit, receiving recipient data of a transplant recipient including a set of parameters, the set of parameters including donor-derived cell-free DNA (dd-cfDNA), receiving one or more parameter weights, calculating a score based on the transplant recipient data and the one or more parameter weights, and calculating a prediction probability of whether graft rejection occurs or the degree to which graft rejection occurs in the transplant recipient based on the score. A system comprising a scoring unit.

24. the set of parameters of the transplantation recipient data being one or more clinical parameters including the time from after transplantation to evaluation; one or more functional parameters including estimated glomerular filtration rate (eGFR), creatinine, proteinuria, or a combination thereof; one or more immunological parameters including the mean fluorescence intensity of donor - specific antibodies, the number of anti - human leukocyte antigen (HLA) mismatches, or both; one or more transplantation recipient characteristics including the age of the transplantation recipient, donor organ infection information, or both; The system according to claim 23, further comprising one or more of one or more transplantation characteristics including previous transplantation information, previous rejection information, or both.

25. The system according to claim 23, wherein the set of parameters does not include histological parameters.

26. The system according to claim 23, further comprising a unit for generating a prediction of the transplantation recipient data in a reference set, the reference set including one or more other transplantation recipients having one or more common characteristics.

27. The system according to claim 26, wherein the generated prediction is used to interpret the mechanism of graft rejection and / or to guide treatment.

28. calculating the score being for each parameter weight, multiplying the parameter weight by the corresponding parameter of the transplantation recipient data, and calculating the score from the sum of the multiplications. The system according to claim 23.

29. calculating the prediction probability being determining the intercept of a multivariable logistic regression model, and calculating the prediction probability from the score and the intercept. The system according to claim 23.

30. the one or more parameter weights are obtaining a cohort dataset that includes a first set of model parameters of transplant recipients in a cohort and cohort graft rejection information, analyzing the first set of model parameters for a relevance between the cohort dataset and the corresponding cohort graft rejection information, selecting one or more subsequent sets of model parameters from the first set of model parameters or a preceding set of model parameters, selecting a last set of model parameters from the one or more subsequent sets of model parameters, the last set of model parameters including independent variables related to graft rejection and satisfying one or more second criteria, The system according to claim 23, received from a machine learning model trained to generate the one or more parameter weights corresponding to the last set of model parameters of the cohort dataset.

31. The system according to claim 30, wherein the first set of model parameters or one or more subsequent sets of model parameters, or both, include dd-cfDNA, and the one or more subsequent sets of model parameters satisfy one or more first criteria.

32. The first set of model parameters is one or more clinical parameters including graft dysfunction, time from last graft rejection, time from transplantation to evaluation, or a combination thereof; one or more functional parameters including estimated glomerular filtration rate (eGFR), creatinine, proteinuria, or a combination thereof; one or more immunological parameters including mean fluorescence intensity of donor-specific antibodies, number of anti-human leukocyte antigen (HLA) mismatches, or a combination thereof; one or more recipient and donor characteristics including recipient age, recipient gender, donor organ infection information, or a combination thereof; One or more transplantation characteristics including the age of the donor, the gender of the donor, the type of the donor, previous transplantation information, cold ischemia time, dual transplanted kidney information, or a combination thereof, and further including one or more of them, the system according to claim 30. The system according to claim 30, further comprising one or more of one or more transplantation characteristics including the age of the donor, the gender of the donor, the type of the donor, previous transplantation information, cold ischemia time, dual transplanted kidney information, or a combination thereof.

33. One or more subsequent sets of the model parameters include one or more of dd-cfDNA, graft dysfunction, recent graft rejection information, time from transplantation to evaluation, estimated glomerular filtration rate (eGFR), proteinuria, mean fluorescence intensity of donor-specific antibodies, age of the recipient, gender of the recipient, age of the donor, gender of the donor, type of the donor, previous transplantation information, cold ischemia time, dual transplantation information, anti-human leukocyte antigen (HLA) mismatch, or a combination thereof, the system according to claim 30.

34. The last set of the model parameters includes one or more of dd-cfDNA, estimated glomerular filtration rate (eGFR), graft dysfunction, recent graft rejection information, mean fluorescence intensity of donor-specific antibodies, proteinuria, or a combination thereof, the system according to claim 30.

35. Analyzing the first set of the model parameters for relevance includes analyzing whether the parameters of the first set of the model parameters identify the presence or absence of graft rejection in the cohort graft rejection information or the degree of identifying the presence or absence of graft rejection, the system according to claim 30.

36. The one or more first criteria or the one or more second criteria include model parameters having a confidence interval above a confidence interval threshold or a p-value less than a p-value threshold, the system according to claim 30 or 31.

37. The confidence interval threshold is 95% and the p-value threshold is 0.2, the system according to claim 36.

38. The system of claim 31, wherein the one or more first criteria include that the number of model parameters within the set of one or more subsequent model parameters is less than a threshold number. **Claim 39** The system of claim 30, wherein the model parameters of the first set are analyzed individually. **Claim 40** The system of claim 30, wherein analyzing the first set of model parameters for relevance includes reducing the dimensionality of the cohort data set based on the presence or absence of graft rejection in the cohort graft rejection information. **Claim 41** The system of claim 30, wherein analyzing the first set of model parameters for relevance includes determining the relevance of the dd-cfDNA with one or more of a cause of end-stage renal disease, a type of graft rejection, or a combination thereof. **Claim 42** The system of claim 30, wherein analyzing the first set of model parameters includes reducing the dimensionality of the cohort data set based on the type of graft rejection. **Claim 43** Selecting the last set of the model parameters includes performing a backward selection by analyzing whether the parameters of the set of one or more subsequent model parameters identify the presence or absence of graft rejection in the cohort graft rejection information, or the degree to which the presence or absence of graft rejection is identified, and selecting the last set of the model parameters by comparing individual analyses. **Claim 44** The system of claim 30, wherein the one or more second criteria include that the number of model parameters within the last set of the model parameters is less than a threshold number. **Claim 45** A non-transitory computer-readable storage medium for determining the risk of graft rejection using a machine learning system, the medium storing one or more programs which, when executed by one or more processors of an electronic device having a display, cause the electronic device to, receive recipient data of a transplant recipient including a set of parameters via a computer or an input function, the set of parameters including donor-derived cell-free DNA (dd-cfDNA); receive one or more parameter weights; calculate a score based on the transplant recipient data and the one or more parameter weights; and calculate a prediction probability of whether or not graft rejection will occur or the degree to which graft rejection will occur in the transplant recipient based on the score. A non-transitory computer-readable storage medium comprising instructions to cause execution.

46. The set of parameters of the transplant recipient data is one or more clinical parameters including the time from after transplantation to evaluation; one or more functional parameters including estimated glomerular filtration rate (eGFR), creatinine, proteinuria, or a combination thereof; one or more immunological parameters including the mean fluorescence intensity of donor-specific antibodies, the number of anti-human leukocyte antigen (HLA) mismatches, or both; one or more transplant recipient characteristics including the age of the transplant recipient, donor organ infection information, or both; The computer-readable storage medium according to claim 45, further comprising one or more of one or more transplant characteristics including previous transplant information, previous rejection information, or both.

47. The computer-readable storage medium according to claim 45, wherein the set of parameters does not include histological parameters.

48. The computer-readable storage medium according to claim 45, further comprising generating a prediction of the transplant recipient data within a reference set, the reference set including one or more other transplant recipients having one or more common characteristics.

49. The computer-readable storage medium according to claim 48, wherein the generated prediction is used to interpret the mechanism of graft rejection and / or to guide treatment.

50. Calculating the score comprises for each parameter weight, multiplying the corresponding parameter of the transplant recipient data by the parameter weight, and calculating the score from the sum of the multiplications, the computer-readable storage medium according to claim 45.

51. Calculating the prediction probability comprises determining an intercept of a multivariate logistic regression model, and calculating the prediction probability from the score and the intercept, the computer-readable storage medium according to claim 45.

52. The one or more parameter weights obtain a cohort data set including a first set of model parameters of transplant recipients in a cohort and cohort graft rejection information, analyze the first set of model parameters for a relationship between the cohort data set and the corresponding cohort graft rejection information, select one or more subsequent sets of model parameters from the first set of model parameters or a preceding set of model parameters, select a last set of model parameters from the one or more subsequent sets of model parameters, the last set of model parameters including independent variables related to graft rejection and satisfying one or more second criteria, The computer-readable storage medium according to claim 45, received from a machine learning model trained to generate the one or more parameter weights corresponding to the last set of the model parameters of the cohort dataset.

53. The computer-readable storage medium according to claim 52, wherein the first set of the model parameters or one or more subsequent sets of the model parameters, or both, include dd-cfDNA, and the one or more subsequent sets of the model parameters satisfy one or more first criteria.

54. The first set of the model parameters is one or more clinical parameters including graft dysfunction, time from last graft rejection, time from transplantation to evaluation, or a combination thereof; one or more functional parameters including estimated glomerular filtration rate (eGFR), creatinine, proteinuria, or a combination thereof; one or more immunological parameters including mean fluorescence intensity of donor-specific antibodies, number of anti-human leukocyte antigen (HLA) mismatches, or a combination thereof; one or more recipient and donor characteristics including recipient age, recipient gender, donor organ infection information, or a combination thereof; donor age, donor gender, donor type, previous transplantation information, cold ischemia time, dual transplant kidney information, or one or more transplantation characteristics including a combination thereof, and further includes one or more of the foregoing, the computer-readable storage medium according to claim 52.

55. One or more subsequent sets of the model parameters include one or more of dd-cfDNA, graft dysfunction, recent graft rejection information, time from transplantation to evaluation, estimated glomerular filtration rate (eGFR), proteinuria, mean fluorescence intensity of donor-specific antibodies, recipient age, recipient gender, donor age, donor gender, donor type, previous transplantation information, cold ischemia time, dual transplantation information, anti-human leukocyte antigen (HLA) mismatch, or combinations thereof, the computer-readable medium according to claim 52.

56. The last set of the model parameters includes one or more of dd-cfDNA, estimated glomerular filtration rate (eGFR), graft dysfunction, recent graft rejection information, mean fluorescence intensity of donor-specific antibodies, proteinuria, or combinations thereof, the computer-readable medium according to claim 52.

57. Analyzing the first set of the model parameters for relevance includes analyzing whether the parameters of the first set of the model parameters identify the presence or absence of graft rejection in the cohort graft rejection information or the degree of identifying the presence or absence of graft rejection, the computer-readable medium according to claim 52.

58. The one or more first criteria or the one or more second criteria include model parameters having a confidence interval above a confidence interval threshold or a p-value below a p-value threshold, the computer-readable medium according to claim 52 or 53.

59. The confidence interval threshold is 95% and the p-value threshold is 0.2, the computer-readable medium according to claim 58.

60. The one or more first criteria include that the number of model parameters in the set of the one or more subsequent model parameters is less than a threshold number, the computer-readable medium according to claim 53.

61. Analyzing the first set of model parameters for relevance includes reducing the dimension of the cohort dataset based on the presence or absence of graft rejection in the cohort graft rejection information, the computer-readable medium according to claim 52.

62. Analyzing the first set of model parameters for relevance includes determining the relevance of the dd-cfDNA with one or more of the cause of end-stage renal disease, the type of graft rejection, or a combination thereof, the computer-readable medium according to claim 52.

63. Analyzing the first set of model parameters includes reducing the dimension of the cohort dataset based on the type of graft rejection, the computer-readable medium according to claim 52.

64. Selecting the last set of the model parameters includes Performing a backward selection by analyzing whether the parameters of the one or more subsequent sets of model parameters identify the presence or absence of graft rejection in the cohort graft rejection information, or the degree of identifying the presence or absence of graft rejection, and Comparing the individual analyses and selecting the last set of the model parameters, the computer-readable medium according to claim 52.

65. The one or more second criteria include that the number of model parameters in the last set of the model parameters is less than a threshold number, the computer-readable medium according to claim 52.

Images