Intelligent management system and method after liver transplantation based on multi-modal monitoring and AI

Abstract

The invention discloses a liver transplantation post-operation intelligent management system and method based on multi-modal monitoring and AI. The method comprises the following steps that S1, physiological parameters, biomarker data, iconography data and self-reporting data of a patient are collected in real time; s2, carrying out preprocessing, feature extraction and fusion on the collected data, and predicting the postoperative complication risk of the patient by using a risk prediction model; s3, generating a personalized management strategy by using a reinforcement learning algorithm according to a risk prediction result, and providing early warning information and intervention suggestions for medical personnel and patients; s4, according to the postoperative management plan and the doctor's advice of the patient, a fixed-point medication reminding function is set, a personalized medication scheme is generated, and medication compliance of the patient is monitored; and S5, the patient makes an appointment for remote medical services, selects appropriate doctors and time to carry out video inquiry or online consultation, and records inquiry content.

Description

Technical Field

[0001] The present invention belongs to the field of medical health management technology, and specifically relates to an intelligent management system and method for liver transplantation based on multimodal monitoring and AI. Background Art

[0002] Liver transplantation is an effective means of treating end-stage liver disease, but postoperative management is crucial. However, after liver transplantation, patients still face a variety of risks, such as rejection, infection, liver function recovery, etc., which require long-term and accurate monitoring and management. Current postoperative management of liver transplantation mainly relies on regular clinical examinations and subjective feedback from patients. It lacks real-time, comprehensive and accurate monitoring methods, making it difficult to detect potential problems in a timely manner and take effective intervention measures. In addition, postoperative medication management is complicated, and patients are prone to postoperative complications due to forgetting to take medication or taking medication incorrectly. Summary of the invention

[0003] To solve the above problems, the present invention proposes an intelligent management system and method for post-liver transplantation based on multimodal monitoring and AI.

[0004] To achieve the above object, the present invention adopts the following technical solutions: An intelligent management method after liver transplantation based on multimodal monitoring and AI, characterized by comprising the following steps: S1. Real-time collection of patients’ physiological parameters, biomarker data, imaging data and self-reported data through multimodal monitoring modules; The specific steps of step S1 are: S11. Physiological parameter monitoring: patients wear smart bracelets or smart clothing to monitor and collect physiological parameters in real time; S12. Biomarker monitoring: Regularly collect blood and urine samples from patients to test biomarkers, and combine with portable testing equipment to achieve real-time testing of some indicators; S13. Imaging monitoring: Use imaging methods to regularly check liver morphology and blood flow, extract key features through image processing and store them in the system database; S14. Patient self-reported data: patients recorded postoperative symptoms, medication status, and psychological status through mobile phone applications or web pages; S2, the data processing and analysis module preprocesses, extracts and fuses the collected data, and uses the risk prediction model to predict the risk of postoperative complications of patients; The specific steps of step S2 are: S21. Data preprocessing: cleaning, normalizing and filling missing values ​​of the collected multi-source data; S22. Feature extraction and fusion: Based on machine learning algorithms, key features are extracted from physiological parameters, biomarker data, imaging data, and patient self-reported data, and multimodal data fusion is performed to construct a comprehensive patient postoperative status feature vector. The fusion calculation formula is: , where is the comprehensive feature vector of the patient's postoperative status; is the physiological parameter characteristic; for biomarker characteristics; For imaging features; The characteristics of the data were self-reported by patients; S23. Constructing a risk prediction model: Constructing a risk prediction model for postoperative complications by combining a convolutional neural network and a recurrent neural network; wherein the convolutional neural network is used to extract spatial features, and the recurrent neural network is used to process time series data; S24. Model training: Use historical data to train the risk prediction model and minimize the loss function L: , where θ is the model parameter; N is the number of samples; i is the sample number; R i is the true risk value of the i-th sample; is the predicted risk value of the i-th sample; S25, risk prediction: Use the trained risk prediction model to predict the risk of the newly input comprehensive patient postoperative status feature vector F, and the calculation formula is: , where The risk prediction results; Prediction processing for risk prediction models; S3. Based on the risk prediction results and combined with the individual characteristics of the patient, the intelligent feedback and intervention module uses reinforcement learning algorithms to generate personalized management strategies and provide early warning information and intervention suggestions to medical staff and patients; S4. Through the fixed-point medication reminder and medication method module, according to the patient's postoperative management plan and doctor's advice, set the fixed-point medication reminder function, generate personalized medication plans, and monitor the patient's medication compliance; The specific steps of step S4 are: S41. According to the patient's postoperative management plan and doctor's orders, use mobile phone applications, smart bracelets or smart speakers to remind patients to take medication regularly and record the time of taking medication; S42. Automatically generate personalized medication plans based on the patient's real-time monitoring data and risk prediction results. The medication plans include drug types, dosages, and medication schedules. S43. Monitor patients’ medication compliance through self-reported data and medication reminder records on smart devices; if it is found that the patient does not take the medication on time, a reminder will be automatically sent to the medical staff and a second reminder will be sent to the patient; S44. Connect with hospital pharmacies or third-party drug delivery platforms to provide drug delivery services to patients so that they can receive the required drugs on time; S5. Patients make appointments for telemedicine services through the telemedicine appointment function module, select appropriate doctors and time for video consultation or online consultation, and record the content of the consultation.

[0005] Preferably, the physiological parameters in step S11 include heart rate, blood pressure, body temperature and exercise data; the biomarkers in step S12 include liver function indicators and immune indicators; the liver function indicators include transaminase, bilirubin and protein; the immune indicators include lymphocyte levels and blood concentrations of immunosuppressive drugs; the imaging methods in step S13 include ultrasound, electronic computed tomography and magnetic resonance imaging; the postoperative symptoms in step S14 include pain, fatigue and appetite.

[0006] Preferably, the individual characteristics of the patient in step S3 include age, gender and underlying diseases; the personalized management strategy includes medication adjustment recommendations, rehabilitation training plans, nutritional guidance and psychological intervention.

[0007] Preferably, the specific steps of step S5 are: S51. Appointment application: The patient selects the doctor to be consulted and the appointment time through the mobile application or the web page, and submits the appointment application. The appointment application is expressed as: A=(D,T), where A is the appointment application; D is the doctor to be consulted; T is the appointment time; S52, appointment reminder: within Δt before the appointment time, a reminder notification is sent to the patient via SMS or application, and the reminder notification is expressed as: R="Your scheduled telemedicine consultation will start after Δt, please prepare", where R is the reminder notification; S53, Video consultation: At the appointment time, the patient communicates with the doctor face-to-face remotely through the system's video consultation function. The video consultation is expressed as: V = VideoConsultation(D, T), where V is the video consultation; VideoConsultation(·) is the video record; S54. Consultation record: The system automatically records the content of each telemedicine consultation. The consultation record includes the doctor's diagnosis, suggestions and the patient's questions. The consultation record is used by medical staff to check the patient's consultation history at any time to track the patient's postoperative recovery. The consultation record is expressed as: C=Record(V), where C is the consultation record; Record(·) is the text and image record.

[0008] Preferably, the reservation method in step S51 includes reservation by department, reservation by doctor and reservation by time period.

[0009] An intelligent management system for liver transplantation based on multimodal monitoring and AI, including: Multimodal monitoring module, used to collect patients’ physiological parameters, biomarker data, imaging data and self-reported data in real time; The data processing and analysis module is used to pre-process, extract features and fuse the collected data, and use the risk prediction model to predict the risk of postoperative complications of patients; Intelligent feedback and intervention module, which is used to generate personalized management strategies based on risk prediction results and individual characteristics of patients using reinforcement learning algorithms, and provide early warning information and intervention suggestions to medical staff and patients; The module of fixed-point medication reminder and medication method is used to set the fixed-point medication reminder function according to the patient's postoperative management plan and doctor's advice, generate personalized medication plan, and monitor the patient's medication compliance; The telemedicine appointment function module is used for patients to make appointments for telemedicine services, select appropriate doctors and times for video consultations or online consultations, and record the content of the consultations.

[0010] After adopting the above technical solution, the present invention has the following beneficial effects: The intelligent management method for liver transplantation based on multimodal monitoring and AI of the present invention integrates multiple monitoring means (including wearable devices, biomarker monitoring, imaging monitoring, etc.) with artificial intelligence technology to achieve real-time monitoring of the patient's postoperative recovery, risk warning and personalized management strategy formulation, and ensures that the patient takes the medicine on time, in the right amount and correctly through the fixed-point medication reminder function. In addition, the intelligent management system for liver transplantation based on multimodal monitoring and AI of the present invention is mainly aimed at the management of patients after liver transplantation surgery. Through multimodal monitoring, artificial intelligence technology and fixed-point medication reminder functions, it realizes comprehensive monitoring, precise management and personalized medication guidance of the patient's postoperative recovery, and supports telemedicine appointment functions, which facilitates remote consultation between patients and medical staff. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Figure 1 This is a flow chart of the intelligent management method after liver transplantation based on multimodal monitoring and AI of the present invention; Figure 2 The figure is a flowchart of the intelligent management system after liver transplantation based on multimodal monitoring and AI of the present invention. DETAILED DESCRIPTION

[0012] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0013] like Figure 1 and Figure 2 As shown, a method for intelligent management after liver transplantation based on multimodal monitoring and AI includes the following steps: S1, collecting the patient's physiological parameters, biomarker data, imaging data and self-reported data in real time through the multimodal monitoring module 1; The specific steps of step S1 are: S11. Physiological parameter monitoring: patients wear smart bracelets or smart clothing to monitor and collect physiological parameters in real time; The physiological parameters in step S11 include heart rate, blood pressure, body temperature and exercise data; S12. Biomarker monitoring: Regularly collect blood and urine samples from patients to test biomarkers, and combine with portable testing equipment to achieve real-time testing of some indicators; The biomarkers in step S12 include liver function indicators and immune indicators; the liver function indicators include transaminase, bilirubin and protein; the immune indicators include lymphocyte levels and blood concentrations of immunosuppressive drugs; S13. Imaging monitoring: Use imaging methods to regularly check liver morphology and blood flow, extract key features through image processing and store them in the system database; The imaging method in step S13 includes ultrasound, computerized tomography and magnetic resonance imaging; S14. Patient self-reported data: patients recorded postoperative symptoms, medication status, and psychological status through mobile phone applications or web pages; The postoperative symptoms in step S14 include pain, fatigue and appetite; S2, data processing and analysis module 2 preprocesses, extracts and fuses the collected data, and uses the risk prediction model to predict the risk of postoperative complications of patients; The specific steps of step S2 are: S21. Data preprocessing: cleaning, normalizing and filling missing values ​​of the collected multi-source data; S22. Feature extraction and fusion: Based on machine learning algorithms, key features are extracted from physiological parameters, biomarker data, imaging data, and patient self-reported data, and multimodal data fusion is performed to construct a comprehensive patient postoperative status feature vector. The fusion calculation formula is: , where is the comprehensive feature vector of the patient's postoperative status; is the physiological parameter characteristic; for biomarker characteristics; For imaging features; The characteristics of the data were self-reported by patients; S23. Constructing a risk prediction model: Constructing a risk prediction model for postoperative complications by combining a convolutional neural network and a recurrent neural network; wherein the convolutional neural network is used to extract spatial features, and the recurrent neural network is used to process time series data; S24. Model training: Use historical data to train the risk prediction model and minimize the loss function L: , where θ is the model parameter; N is the number of samples; i is the sample number; R i is the true risk value of the i-th sample; is the predicted risk value of the i-th sample; S25, risk prediction: Use the trained risk prediction model to predict the risk of the newly input comprehensive patient postoperative status feature vector F, and the calculation formula is: , where The risk prediction results; Prediction processing for risk prediction models; S3. Based on the risk prediction results and combined with the individual characteristics of the patient, the intelligent feedback and intervention module 3 generates personalized management strategies using reinforcement learning algorithms and provides early warning information and intervention suggestions to medical staff and patients; The individual characteristics of the patient in step S3 include age, gender and underlying diseases; the personalized management strategy includes medication adjustment suggestions, rehabilitation training plans, nutritional guidance and psychological intervention; S4, through the fixed-point reminder medication and medication method module 4, according to the patient's postoperative management plan and doctor's advice, set the fixed-point medication reminder function, generate a personalized medication plan, and monitor the patient's medication compliance; The specific steps of step S4 are: S41. According to the patient's postoperative management plan and doctor's orders, use mobile phone applications, smart bracelets or smart speakers to remind patients to take medication regularly and record the time of taking medication; S42. Automatically generate personalized medication plans based on the patient's real-time monitoring data and risk prediction results. The medication plans include drug types, dosages, and medication schedules. S43. Monitor patients’ medication compliance through self-reported data and medication reminder records on smart devices; if it is found that the patient does not take the medication on time, a reminder will be automatically sent to the medical staff and a second reminder will be sent to the patient; S44. Connect with hospital pharmacies or third-party drug delivery platforms to provide drug delivery services to patients so that they can receive the required drugs on time; S5. The patient makes an appointment for telemedicine service through telemedicine appointment function module 5, selects a suitable doctor and time for video consultation or online consultation, and records the consultation content; The specific steps of step S5 are: S51. Appointment application: The patient selects the doctor to be consulted and the appointment time through the mobile application or the web page, and submits the appointment application. The appointment application is expressed as: A=(D,T), where A is the appointment application; D is the doctor to be consulted; T is the appointment time; In step S51, the reservation mode includes reservation by department, reservation by doctor and reservation by time period; S52, appointment reminder: within Δt before the appointment time, a reminder notification is sent to the patient via SMS or application, and the reminder notification is expressed as: R="Your scheduled telemedicine consultation will start after Δt, please prepare", where R is the reminder notification; S53, Video consultation: At the appointment time, the patient communicates with the doctor face-to-face remotely through the system's video consultation function. The video consultation is expressed as: V = VideoConsultation(D, T), where V is the video consultation; VideoConsultation(·) is the video record; S54. Consultation record: The system automatically records the content of each telemedicine consultation. The consultation record includes the doctor's diagnosis, suggestions and the patient's questions. The consultation record is used by medical staff to check the patient's consultation history at any time to track the patient's postoperative recovery. The consultation record is expressed as: C=Record(V), where C is the consultation record; Record(·) is the text and image record.

[0014] An intelligent management system for liver transplantation based on multimodal monitoring and AI, including: Multimodal monitoring module 1, used to collect patients' physiological parameters, biomarker data, imaging data and self-reported data in real time; Data processing and analysis module 2 is used to preprocess, extract features and fuse the collected data, and use the risk prediction model to predict the risk of postoperative complications of patients; Intelligent feedback and intervention module 3 is used to generate personalized management strategies based on risk prediction results and individual characteristics of patients using reinforcement learning algorithms, and to provide early warning information and intervention suggestions to medical staff and patients; The fixed-point medication reminder and medication method module 4 is used to set the fixed-point medication reminder function according to the patient's postoperative management plan and doctor's advice, generate a personalized medication plan, and monitor the patient's medication compliance; The telemedicine reservation function module 5 is used for patients to make reservations for telemedicine services, select appropriate doctors and time for video consultation or online consultation, and record the consultation content.

[0015] Example

[0016] The following is a specific embodiment of the present invention: Patient Xiao Zhang received a liver transplant in a hospital and wore a smart bracelet after the operation. The hospital also equipped him with a portable biomarker detection device. In the first week after the operation, the smart bracelet monitored in real time that Xiao Zhang's heart rate was abnormally elevated, and the portable detection device detected a slight increase in his transaminase level. The system transmits these data to the data processing and analysis module 2. After feature extraction and fusion, the risk prediction model determines that Xiao Zhang may have a risk of mild rejection. The system immediately sends an early warning message to the medical staff and Xiao Zhang. The medical staff adjusted Xiao Zhang's immunosuppressant dosage according to the system's recommendations and provided Xiao Zhang with detailed rehabilitation guidance through telemedicine.

[0017] At the same time, the system set up a fixed-point medication reminder function according to Xiao Zhang's postoperative management plan, reminding Xiao Zhang to take immunosuppressants at 8 am and 8 pm every day. The system also automatically generated a personalized medication plan, adjusted the drug dosage according to Xiao Zhang's liver function indicators, and recorded Xiao Zhang's medication time through a mobile phone application. If Xiao Zhang did not take the medicine on time, the system would automatically send a reminder to the medical staff and a second reminder to Xiao Zhang. Through the system's drug delivery service, Xiao Zhang received the required drugs on time, ensuring the continuity of medication.

[0018] In addition, Xiao Zhang felt unwell in the third week after the operation, so he made an appointment for telemedicine service through the system and chose to have a video consultation with a liver transplant expert at 9 am the next day. The system sent a reminder to Xiao Zhang one hour before the appointment time, and Xiao Zhang attended the video consultation on time. The doctor further adjusted the rehabilitation plan based on the system's real-time data and Xiao Zhang's description, and provided psychological counseling. Xiao Zhang carried out postoperative rehabilitation according to the personalized management strategy provided by the system, and recovered smoothly after the operation.

[0019] The above is only a preferred specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed by the present invention should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. An intelligent management method for post-liver transplantation based on multimodal monitoring and AI, characterized in that: The following steps are involved: S1. Real-time collection of patients’ physiological parameters, biomarker data, imaging data and self-reported data through multimodal monitoring modules; The specific steps of step S1 are: S11. Physiological parameter monitoring: patients wear smart bracelets or smart clothing to monitor and collect physiological parameters in real time; S12. Biomarker monitoring: Regularly collect blood and urine samples from patients to test biomarkers, and combine with portable testing equipment to achieve real-time testing of some indicators; S13. Imaging monitoring: Use imaging methods to regularly check liver morphology and blood flow, extract key features through image processing and store them in the system database; S14. Patient self-reported data: patients recorded postoperative symptoms, medication status, and psychological status through mobile phone applications or web pages; S2, the data processing and analysis module preprocesses, extracts and fuses the collected data, and uses the risk prediction model to predict the risk of postoperative complications of patients; The specific steps of step S2 are: S21. Data preprocessing: cleaning, normalizing and filling missing values ​​of the collected multi-source data; S22. Feature extraction and fusion: Based on machine learning algorithms, key features are extracted from physiological parameters, biomarker data, imaging data, and patient self-reported data, and multimodal data fusion is performed to construct a comprehensive patient postoperative status feature vector. The fusion calculation formula is: , where is the comprehensive feature vector of the patient's postoperative status; is the physiological parameter characteristic; for biomarker characteristics; For imaging features; The characteristics of the data were self-reported by patients; S23. Constructing a risk prediction model: Constructing a risk prediction model for postoperative complications by combining a convolutional neural network and a recurrent neural network; wherein the convolutional neural network is used to extract spatial features, and the recurrent neural network is used to process time series data; S24. Model training: Use historical data to train the risk prediction model and minimize the loss function L: , where θ is the model parameter; N is the number of samples; i is the sample number; R i is the true risk value of the i-th sample; is the predicted risk value of the i-th sample; S25, risk prediction: Use the trained risk prediction model to predict the risk of the newly input comprehensive patient postoperative status feature vector F, and the calculation formula is: , where The risk prediction results; Prediction processing for risk prediction models; S3. Based on the risk prediction results and combined with the individual characteristics of the patient, the intelligent feedback and intervention module uses reinforcement learning algorithms to generate personalized management strategies and provide early warning information and intervention suggestions to medical staff and patients; S4. Through the fixed-point medication reminder and medication method module, according to the patient's postoperative management plan and doctor's advice, set the fixed-point medication reminder function, generate personalized medication plans, and monitor the patient's medication compliance; The specific steps of step S4 are: S41. According to the patient's postoperative management plan and doctor's orders, use mobile phone applications, smart bracelets or smart speakers to remind patients to take medication regularly and record the time of taking medication; S42. Automatically generate personalized medication plans based on the patient's real-time monitoring data and risk prediction results. The medication plans include drug types, dosages, and medication schedules. S43. Monitor patients’ medication compliance through self-reported data and medication reminder records on smart devices; if it is found that the patient does not take the medication on time, a reminder will be automatically sent to the medical staff and a second reminder will be sent to the patient; S44. Connect with hospital pharmacies or third-party drug delivery platforms to provide drug delivery services to patients so that they can receive the required drugs on time; S5. Patients make appointments for telemedicine services through the telemedicine appointment function module, select appropriate doctors and time for video consultation or online consultation, and record the content of the consultation.

2. The intelligent management method after liver transplantation based on multimodal monitoring and AI as claimed in claim 1, characterized in that: The physiological parameters described in step S11 include heart rate, blood pressure, body temperature and exercise data; the biomarkers described in step S12 include liver function indicators and immune indicators; the liver function indicators include transaminase, bilirubin and protein; the immune indicators include lymphocyte levels and blood concentrations of immunosuppressive drugs; the imaging methods described in step S13 include ultrasound, electronic computed tomography and magnetic resonance imaging; the postoperative symptoms described in step S14 include pain, fatigue and appetite.

3. The intelligent management method after liver transplantation based on multimodal monitoring and AI as claimed in claim 1, characterized in that: The individual characteristics of the patient in step S3 include age, gender and underlying diseases; the personalized management strategy includes medication adjustment recommendations, rehabilitation training plans, nutritional guidance and psychological intervention.

4. The intelligent management method after liver transplantation based on multimodal monitoring and AI as claimed in claim 1, characterized in that: The specific steps of step S5 are: S51. Appointment application: The patient selects the doctor to be consulted and the appointment time through the mobile application or the web page, and submits the appointment application. The appointment application is expressed as: A=(D,T), where A is the appointment application; D is the doctor to be consulted; T is the appointment time; S52, appointment reminder: within Δt before the appointment time, a reminder notification is sent to the patient via SMS or application, and the reminder notification is expressed as: R="Your scheduled telemedicine consultation will start after Δt, please prepare", where R is the reminder notification; S53, Video consultation: At the appointment time, the patient communicates with the doctor face-to-face remotely through the system's video consultation function. The video consultation is expressed as: V = VideoConsultation(D, T), where V is the video consultation; VideoConsultation(·) is the video record; S54. Consultation record: The system automatically records the content of each telemedicine consultation. The consultation record includes the doctor's diagnosis, suggestions and the patient's questions. The consultation record is used by medical staff to check the patient's consultation history at any time to track the patient's postoperative recovery. The consultation record is expressed as: C=Record(V), where C is the consultation record; Record(·) is the text and image record.

5. The intelligent management method after liver transplantation based on multimodal monitoring and AI as claimed in claim 1, characterized in that: The reservation methods in step S51 include reservation by department, reservation by doctor and reservation by time period.

6. An intelligent management system for post-liver transplantation based on multimodal monitoring and AI, characterized in that: include: Multimodal monitoring module, used to collect patients’ physiological parameters, biomarker data, imaging data and self-reported data in real time; The data processing and analysis module is used to pre-process, extract features and fuse the collected data, and use the risk prediction model to predict the risk of postoperative complications of patients; Intelligent feedback and intervention module, which is used to generate personalized management strategies based on risk prediction results and individual characteristics of patients using reinforcement learning algorithms, and provide early warning information and intervention suggestions to medical staff and patients; The module of fixed-point medication reminder and medication method is used to set the fixed-point medication reminder function according to the patient's postoperative management plan and doctor's advice, generate personalized medication plan, and monitor the patient's medication compliance; The telemedicine appointment function module is used for patients to make appointments for telemedicine services, select appropriate doctors and times for video consultations or online consultations, and record the content of the consultations.

Images