System for predicting transplant survival

Abstract

The invention relates to medicine, and more particularly to organ transplantation, and is intended to improve personalized prediction of transplant survival in the case of solid organ transplantation, using artificial intelligence. A system for the personalized prediction of transplant survival in the case of solid organ transplantation, which utilizes artificial intelligence, comprises the following, functionally interconnected via the Internet: medical institutions, recipient and transplant databases, a data-server base, and an administrative computer-programmable device with a data display monitor. The system further comprises an artificial intelligence-based predictive computer platform for solid-organ recipients which is designed to be capable of determining the best donor-recipient pairing and transplant survival on the basis of different clinical, instrumental, laboratory, histological and genetic features, and is integrated with the data-server base. The invention makes it possible to increase the functional capabilities of personalized prediction of transplant survival in the case of solid organ transplantation.

Description

[0001] G16B40 / 20

[0002] G16H20 / 40

[0003] A personalized system for predicting graft survival in solid organ transplantation using artificial intelligence

[0004] The invention relates to medicine, in particular to transplantology, and is intended to improve personalized prediction of graft survival in solid organ transplantation using artificial intelligence.

[0005] Despite significant advances, challenges such as donor organ shortages, chronic immune damage, and post-transplant complications continue to limit the potential of transplantation. This underscores the need for further improvements in this field. In response to these challenges, artificial intelligence (AI) has been implemented to optimize the organ allocation and transplantation process. The development of AI has been made possible by the availability of medical data and improved analytical methods, leading to its exponential growth in various fields of medicine. Recent advances in AI have revived interest in its use to solve complex transplantation problems.

[0006] The prior art, in particular document US11754824B2, 12.09.2023, describes an intelligent device for vital microscopy IVM. The IVM device includes: a receiver configured to receive at least one IVM image of human microcirculation, MS, of an organ surface; a training processor connected to the receiver and configured to: process at least one IVM image and extract at least one MS variable therefrom and identify from the extracted at least one MS variable at least one IVM image of at least one of: the underlying cause of the observed abnormality, an intervention, a disease state, a disease diagnosis, a medical condition of a human; the presence of a pathogen; and an output connected to the training processor and configured to output an identification.

[0007] However, this device is not intended for personalized prediction of graft survival.

[0008] In the prior art, document US10783632B2, dated September 22, 2020, discloses a machine learning system and method for predicting wound healing, such as diabetic foot ulcers or other wounds, as well as for assessment implementations, such as image segmentation into wounded and non-wounded regions. Systems for assessing or predicting wound healing may include a light detector element configured to collect light of at least a first wavelength reflected from a tissue region, including a wound, and one or more processors configured to generate an image based on a signal from the light detector.an element having pixels depicting a tissue region, determining reflectance intensity values ​​for at least a subset of the pixels, determining one or more quantitative characteristics of the subset of the plurality of pixels based on the reflectance intensity values, and generating a predicted or estimated healing parameter associated with the wound over a predetermined time interval.

[0009] While this system can predict wound healing, it also cannot personally predict graft survival in solid organ transplants using artificial intelligence.

[0010] The prior art also includes document AU2020277267B2, 05.10.2023, which discloses a system and method for detecting, predicting, or monitoring the condition or outcome of a transplant in a transplant recipient. In some aspects, the method for detecting or predicting the condition of a transplant recipient includes a) obtaining a sample, wherein the sample contains another gene expression product of the transplant recipient; b) conducting an analysis to determine the expression level of one or more gene expression products in the transplant recipient; and c) detecting or predicting the condition of the transplant recipient by applying an algorithm to the expression level determined in step b), wherein the algorithm is a classifier capable of distinguishing between at least two conditions that are not normal conditions, and wherein one of the at least two conditions is graft rejection or graft dysfunction.

[0011] Although this system can predict graft survival in personalized organ transplantation, it has drawbacks such as limited functionality, complexity of implementation, and lack of integration with artificial intelligence predictive models.

[0012] The declared system is aimed at solving some of the problems and shortcomings of known analogues.

[0013] The stated technical solution for predicting organ transplantation based on AI is aimed at the following: 1) The implementation of an AI-based prediction platform into clinical practice for recipients of kidneys and other solid organs (from both living and deceased donors) is aimed at accurately predicting early deterioration of allograft function and survival outcomes before and after transplantation.

[0014] 2) The second aspect involves integrating the developed transplant registry platform with various transplant hospitals. This integration aims to unify and optimize transplant data analysis using artificial intelligence tools, thereby assisting physicians in decision-making and improving transplant survival rates.

[0015] (i) The claimed system involves the use of a prediction algorithm developed using a large dataset of living donor kidney transplants, which has demonstrated improved prediction accuracy compared to previous studies focusing on graft prediction, which relied on a limited number of deceased donor kidney transplants observed over a long period (1 to 5 years).

[0016] (ii) Data dimensionality reduction using a prediction algorithm framework to manage a high-dimensional and complex heart / kidney transplant dataset and the inclusion of histological images obtained from both donor and recipient for transplant data improved the early assessment of functional impairment or rejection of heart / kidney allografts.

[0017] (iii) An innovative method for non-overlapping patient segmentation provides a detailed view of feature importance over time while circumventing biases from previous cohorts;

[0018] (iv) Integrative AI-based analysis of various clinical, instrumental, laboratory, histological, and genetic characteristics. This discovery enables the selection of the most important characteristics for predicting heart / kidney transplant outcomes within a specific time window.

[0019] (v) A personalized AI-based prediction system for graft survival in solid organ transplantation utilizes large cohorts of patients with different ethnic datasets (White, Asian, and Black) to investigate potential mechanisms of decline in cardiac and renal function in different ethnic groups.

[0020] (vi) The practical outcomes of the submitted technical solution are: A) To provide a cloud-based data-driven decision support (CBD) application to assist cardiologists and nephrologists in the heart / kidney allocation process by determining the best donor-recipient pair and graft survival.

[0021] B) This methodology allows for the connection of several transplant centers in Kazakhstan or other countries to a network for the collection / exchange of transplant data and dynamic analysis.

[0022] The stated technical result also strengthens the system's functionality for personalized graft survival prediction in solid organ transplants using artificial intelligence, increasing data collection and simplifying system management, while providing rapid user access to graft survival data.

[0023] In order to solve the problems of the prior art as described above, an embodiment of the present invention is a system for personalized prediction of graft survival in solid organ transplantation using artificial intelligence, comprising medical institutions functionally interconnected via the Internet of Things technology with a recipient and transplant database, a database with a data server, an administrative computer-programmable device with a data display monitor, characterized in that it contains a computer platform for prediction using artificial intelligence in clinical practice for recipients of kidneys and other solid organs, configured to determine the best pair of donor and recipient and graft survival based on various clinical, instrumental, laboratory, histological and genetic features, integrated with a database with a data server,a transplant registry and a classifier of patients in need of transplantation located in a database with a data server, transplant hospitals that are connected to the database with a data server through a computer-programmable device for medical workers configured to manage the data of a transplant, a patient, a donor and a recipient, a user access device for recipients configured to receive and transmit patient data from the database with a data server, clinical diagnostic centers or laboratories connected to the database with a data server, and configured to transmit recipient data according to the individual identification data of the recipient, wherein the administrative computer-programmable device is configured to manage users, including registration, authorization and granting access rights to the system, the computer forecasting platform is also configured to maintain a registry for a transplant,with the display of the recipient and donor waiting list, based on machine and deep learning, an accurate prognosis of graft survival in solid organ transplantation from donor to recipient was made; based on a classifier of patients in need of transplantation, the most needy patient was identified (donor allocation system), who urgently needs a transplant; based on genetic analysis, the probability of rejection and graft compliance before organ transplantation from donor to recipient was determined; based on the recipient analysis data obtained from clinical diagnostic centers or laboratories, immunosuppression was determined and the dose of the recipient's personalized immunosuppressant drug was adjusted.

[0024] In one embodiment of a system for personalized prediction of graft survival in solid organ transplantation using artificial intelligence, a computer prediction platform based on deep machine learning is designed to make an assumption and forecast of graft survival in solid organ transplantation from a donor to a recipient for one, three, and five years.

[0025] In one embodiment of the system for personalized prediction of graft survival in solid organ transplantation using artificial intelligence, the computer prediction platform, based on graft, patient, donor, and recipient data, is designed with the ability to generate pathomorphology and Banff / Tissue Staining classification of tissue obtained by biopsy.

[0026] In one embodiment of a system for personalized prediction of graft survival in solid organ transplantation using artificial intelligence, the computer prediction platform comprises an allocation calculator accessible to a computer-programmable device for healthcare professionals that is configured to calculate a composite lung allocation index, an end-stage liver disease model, estimated panel reactive antibodies, estimated post-transplant survival, end-stage liver disease in children, and a kidney donor risk index.

[0027] In one implementation of a system for personalized prediction of graft survival in solid organ transplants using artificial intelligence, the database and server are linked to state medical care and control centers.

[0028] In one implementation of a system for personalized prediction of graft survival in solid organ transplants using artificial intelligence, the database and server are linked to state medical care and control centers.

[0029] The stated solution is explained by the following figures of the drawing.

[0030] Fig. 1 Functional diagram of the implementation of a system for personalized prediction of graft survival in solid organ transplantation using artificial intelligence.

[0031] Fig. 2. Schematic diagram of a system for personalized prediction of graft survival in solid organ transplantation using artificial intelligence.

[0032] Fig. 3. Schematic diagram of the function of the computer prediction platform in the system for personalized prediction of graft survival in solid organ transplantation using artificial intelligence.

[0033] Conventional designations adopted on the drawing figures:

[0034] 1 - a system for personalized prediction of graft survival in solid organ transplantation using artificial intelligence,

[0035] 2 - medical institutions,

[0036] 3 - recipient and transplant database,

[0037] 4 - database with data server,

[0038] 5 - administrative computer-programmable device,

[0039] 6 - computer forecasting platform,

[0040] 7 - transplant registry,

[0041] 8 - classifier of patients requiring transplantation,

[0042] 9 - transplant hospitals, 10 - computer-programmable device for medical workers,

[0043] 11 - user access device for recipients,

[0044] 12 - recipient,

[0045] 13 - clinical diagnostic centers or laboratories,

[0046] 14 - distribution calculator,

[0047] 15 - state centers for medical care and control.

[0048] According to Fig. 1-3, the claimed system (1) for personalized prediction of transplant survival in solid organ transplantation using artificial intelligence, includes medical institutions (2) functionally interconnected via the Internet of Things technology with a database of the recipient and the transplanted person (3), a database with a data server (4), an administrative computer-programmable device (5) with a data display monitor, characterized in that it contains a computer prediction platform (6) using artificial intelligence in clinical practice for recipients of kidneys and other solid organs, configured to determine the best pair of donor and recipient and transplant survival based on various clinical, instrumental, laboratory, histological and genetic features, integrated with a database with a data server, a transplant registry (7) and a patient classifier (8),in need of transplantation located in a database with a data server (4), transplant hospitals (9), which through a computer-programmable device for medical workers (10) configured to manage the data of the transplant, patient, donor and recipient, are connected to the database with a data server (4), a user device (11) for accessing recipients (12), configured to receive and transmit patient data from the database with a data server (4), clinical diagnostic centers or laboratories (13) connected to the database with a data server (4), and, and printed out with the ability to transmit recipient data (12) according to the individual identification data of the recipient, wherein the administrator computer-programmable device (5) is configured to manage users, including registration, authorization and granting access rights to the system,The computer prediction platform (6) is also configured to maintain a registry for the transplant, displaying the waiting list of the recipient and the donor, to make an assumption and forecast of the graft survival rate in the transplantation of solid organs from a donor to a recipient based on machine deep learning, to determine the most needy patient who needs a transplant urgently based on a classifier of patients in need of a transplant, to determine the probability of rejection and the compatibility of the transplant before the transplantation of organs from a donor to a recipient based on genetic analysis, to determine immunosuppression and make a dose adjustment of the recipient's personalized immunosuppressant drug based on the recipient's analysis data obtained from clinical diagnostic centers or laboratories.

[0049] In one embodiment of the system (1) for personalized prediction of graft survival in solid organ transplantation using artificial intelligence, the computer prediction platform (5), based on machine deep learning, is designed with the ability to make an assumption and forecast of graft survival in solid organ transplantation from a donor to a recipient for one, three and five years.

[0050] In one embodiment of the system (1) for personalized prediction of graft survival in solid organ transplantation using artificial intelligence, the computer prediction platform (5), based on graft, patient, donor and recipient data, is designed with the ability to generate pathomorphology and Banff / Tissue Staining and Biopsy classification.

[0051] In one embodiment of the system (1) for personalized prediction of graft survival in solid organ transplantation using artificial intelligence, the computer prediction platform (5) comprises a distribution calculator (14) accessible to a computer-programmable device for healthcare professionals (10) that is configured to calculate a composite lung distribution index, a model of end-stage liver disease, calculated panel reactive antibodies, estimated survival after transplantation, end-stage liver disease in children, and a kidney donor risk index.

[0052] In one embodiment of the system for personalized prediction of graft survival in solid organ transplantation using artificial intelligence, the database with the data server (4) is linked to integrated state centers for medical care and control (15).

[0053] The method of operation of the personalized system for predicting graft survival in solid organ transplantation using artificial intelligence is characterized by the fact that the method includes the following steps

[0054] Collects recipient, donor and transplant data, including clinical, instrumental, laboratory, histological and genetic characteristics, using a server-based database linked to a transplant registry and a classifier of patients in need of transplantation.

[0055] Analyzes recipient data received from clinical diagnostic centers and laboratories and determines individual characteristics that influence the prognosis of transplant survival.

[0056] Conducts a preliminary assessment of the probability of transplant rejection through genetic analysis of the compatibility of the donor and recipient.

[0057] Selects the most suitable donor-recipient pair based on registry data and a classifier of patients requiring transplantation to minimize the risk of rejection and increase the likelihood of a successful transplant.

[0058] Conducts personalized graft survival prognosis using a computer-based prediction platform that utilizes machine and deep learning algorithms to generate one-, three-, and five-year graft survival prognosis.

[0059] Adjusts the recipient's immunosuppressant drug dosage based on the results of analysis obtained from the recipient's clinical status and provides automatic transfer of immunosuppression data to the server database.

[0060] Displays current and predicted transplant data, including biopsy results, histology, pathology, and Banff / tissue stain classification, on a computer-programmable device for healthcare professionals.

[0061] Supports clinical control and registration of recipient data in the registry, using the administrator device for access management, including authorization, registration and assignment of access rights to system users.

[0062] Result: The method provides an automated and personalized prognosis of graft survival in solid organ transplantation based on a comprehensive data analysis, which allows for increased prognosis accuracy, optimized immunosuppressant dosage, and improved overall quality of medical care for recipients.

[0063] The proposed system also includes two predictive models for solid organ transplantation from living and deceased donors. Tested on over 29,000 patients, the system offers personalized predictions of graft outcome five years after transplantation, as well as donor-recipient compatibility. Three different interfaces have been created: for patients, for healthcare professionals, and for the administrator. Transplantologists have full access to patient information, and patients can have tests performed at their place of residence. Using AI algorithms, the analysis is automatically completed using the patient's Individual Identification Number (IIN). This IIN will be integrated with government agencies such as DamuMed (a platform for accessing medical organizations to schedule appointments, arrange home visits, and view medical records) and the Republican Center for Electronic Healthcare (RCEH) and will be accessible to the doctor.

Claims

Formula 1. A system for personalized prediction of graft survival in solid organ transplantation using artificial intelligence, comprising medical institutions functionally interconnected via the Internet of Things technology with a recipient and transplantant database, a database with a data server, an administrative computer-programmable device with a data display monitor, characterized in that it contains a computer platform for prediction using artificial intelligence in clinical practice for recipients of kidneys and other solid organs, designed with the ability to determine the best pair of donor and recipient and graft survival based on various clinical, instrumental, laboratory, histological and genetic characteristics, integrated with a database with a data server, a transplant registry and a classifier of patients in need of transplantation located in the database with a data server, transplant hospitals,which, through a computer-programmable device for medical workers, configured to manage the data of the transplant, patient, donor and recipient, are connected to a database with a data server, a user access device for recipients, configured to receive and transmit patient data from the database with a data server, clinical diagnostic centers or laboratories connected to the database with a data server, and, moreover, printed with the ability to transmit recipient data according to the individual identification data of the recipient, wherein the administrative computer-programmable device is configured to manage users, including registration, authorization and granting access rights to the system, the computer forecasting platform is also configured to maintain a registry for the transplant, with the display of the waiting list of the recipient and the donor,to predict graft survival in solid organ transplants from donor to recipient using machine and deep learning, Based on the classifier of patients in need of transplantation, determine the most needy patient who needs transplantation urgently; Based on genetic analysis, determine the probability of rejection and the compatibility of the graft before transplantation of organs from a donor to a recipient; Based on the recipient's analysis data obtained from clinical diagnostic centers or laboratories, determine immunosuppression and adjust the dose of the recipient's personalized immunosuppressant drug.

2. The system according to paragraph 1, characterized in that the computer forecasting platform, based on machine and deep learning, is designed with the ability to make an assumption and forecast the survival of the transplant during the transplantation of solid organs from a donor to a recipient for one, three and five years.

3. The system according to claim 1, characterized in that the computer platform for forecasting based on data from the transplant, patient, donor and recipient is designed with the ability to generate pathomorphology and Banff / Tissue Staining classification using computer vision algorithms.

4. The system of claim 1, wherein the computerized prediction platform comprises a solid organ allocation calculator accessible to a computer-programmable device for healthcare professionals that is configured to calculate a composite lung allocation index, an end-stage liver disease model, estimated panel reactive antibodies, estimated survival after transplantation, end-stage liver disease in children, and a kidney donor risk index.

5. The system according to paragraph 1, characterized in that the database with the data server is connected and integrated with state centers for medical care and control.

Images