CN-121971079-A - Kidney rejection monitoring system based on wireless implantable biosensor

CN121971079ACN 121971079 ACN121971079 ACN 121971079ACN-121971079-A

Abstract

The invention relates to the technical field of biological sensors and wireless health monitoring, and particularly discloses a renal rejection monitoring system based on a wireless implantable biological sensor. The system comprises a wireless implantable biological sensing array, an in-vivo signal preprocessing unit, an in-vitro data receiving and synchronizing device, a multi-mode heterogeneous data fusion engine and a time space map neural network prediction model, wherein a dynamic immune-metabolism knowledge map is constructed by collecting immune and metabolism multi-dimensional biomarkers around a transplanted kidney in real time, and accurate prediction and grading early warning of rejection risks are realized by utilizing space cascading and time evolution rules of time space map neural network modeling signals. By adopting the technical scheme, the sub-clinical rejection signals can be identified before the abnormal of the conventional indexes such as creatinine and the like, the window is intervened in advance, meanwhile, the long-term noninvasive monitoring is realized by means of the wireless low-power-consumption design, and the accuracy and the intelligent level of postoperative management are improved.

Inventors

WU DONGJUAN
XU TONG
MA SHUAIJUN
SUN MEIGUI
ZHANG MENGJIAO
HOU MIN
YANG DAN

Assignees

中国人民解放军空军军医大学

Dates

Publication Date: 20260505
Application Date: 20260403

Claims (10)

1. A wireless implantable biosensor-based renal rejection monitoring system, comprising: A wireless implantable biosensing array configured to acquire, in real-time, multi-dimensional biomarker signals associated with immunology and metabonomics in a tissue microenvironment surrounding a transplanted kidney, the multi-dimensional biomarker signals including cytokine concentration dynamic change data and metabolite concentration fluctuation data; The in-vivo signal preprocessing unit is electrically connected with the wireless implantable biological sensing array and is configured to filter, amplify and analog-to-digital convert an original biological signal and transmit the processed digital signal to the outside of the body through a wireless communication protocol; The in-vitro data receiving and synchronizing device is configured to receive the wireless signals from the in-vivo signal preprocessing unit, and align time axes of different types of biomarker data according to uniform time stamps to form a structured time sequence data stream; The multi-mode heterogeneous data fusion engine is configured to map the aligned immunological data and metabonomic data into a unified knowledge graph frame and construct a dynamic heterogeneous graph structure reflecting the interaction relationship among immune cells, tubular epithelial cells and metabolic pathways; And the space-time diagram neural network prediction model is configured to be used for simultaneously modeling cascade propagation characteristics of biological signals in a space dimension and a latent evolution rule in a time dimension based on the dynamic heterogram structure and outputting a risk prediction result of renal rejection.
2. The wireless implantable bio-sensor based renal rejection monitoring system of claim 1, wherein: The wireless implantable biological sensing array adopts a flexible biocompatible substrate as a physical carrier, wherein the substrate is made of polyimide or polydimethylsiloxane and is used for being attached to the surface of a capsule of an transplanted kidney; The flexible biocompatible substrate is integrated with a plurality of miniaturized biosensing elements, including a specific immunosensor for interleukin six, a specific immunosensor for tumor necrosis factor alpha, an enzymatic electrochemical sensor for glucose and a mediator-modified electrochemical sensor for lactic acid; The specific immunosensor comprises a gold electrode and a self-assembled monolayer modified on the surface of the gold electrode, wherein the self-assembled monolayer consists of short-chain alkane with a thiol group, and a capture antibody is covalently connected to the terminal carboxyl of the self-assembled monolayer through carbodiimide chemical reaction; the enzymatic electrochemical sensor detects the oxidation current of hydrogen peroxide by utilizing the enzymatic reaction of the electrode surface; the outer layers of the miniaturized biological sensitive elements are all encapsulated with polymer nanometer aperture membranes for filtering macromolecular protein interference.
3. The wireless implantable bio-sensor based renal rejection monitoring system of claim 1, wherein: the in-vivo signal preprocessing unit is internally provided with a precise low-noise instrument amplifying circuit, a second-order Butterworth low-pass filter, a successive approximation type analog-to-digital converter and a microcontroller for executing an adaptive sampling strategy; The precise low-noise instrument amplifying circuit is configured to perform primary amplification on microampere-level current signals or millivolt-level voltage signals output by the wireless implantable biological sensing array; The second-order Butterworth low-pass filter is connected to the output end of the precise low-noise instrument amplification circuit and is used for removing high-frequency physiological noise and electromagnetic interference; The microcontroller is configured with a sleep mode and a sampling mode in which each sensor channel is activated sequentially, and the microcontroller is configured to execute a signal rate-of-change based task scheduling logic: When the concentration of the biomarker is monitored to be in a stable state, the sampling period is prolonged to a preset first duration; when the instantaneous change rate of the biomarker is monitored to exceed the preset safety slope, immediately increasing the sampling frequency to a preset second time period, wherein the first time period is longer than the second time period.
4. The wireless implantable bio-sensor based renal rejection monitoring system of claim 1, wherein: the in-vitro data receiving and synchronizing device adopts a double-channel time synchronizing mechanism and comprises a built-in high-precision real-time clock, an event trigger mark module and a buffer array; the high-precision real-time clock is configured to mark an absolute timestamp for the original biomarker data received for each frame; the event triggering mark module is configured to automatically record the starting and ending time of the physiological fluctuation when the instantaneous change rate of the biomarker concentration exceeds a preset threshold value; the buffer array is configured to store heterogeneous data of different sampling frequencies and perform resampling logic; The resampling logic is configured to process the electrochemical signal acquired at high frequency and the immune signal acquired at low frequency by adopting a linear interpolation algorithm or a spline interpolation algorithm, so that the interleukin six data, the tumor necrosis factor alpha data, the glucose data and the lactic acid data are aligned under the same time axis scale to eliminate sampling time lag.
5. The wireless implantable bio-sensor based renal rejection monitoring system of claim 1, wherein: When the dynamic heterogram structure is constructed, the multi-modal heterogeneous data fusion engine defines three core node types, namely an immune node representing cytokines and immune cell activation states thereof, a metabolic node representing metabolite concentration level and a tissue function node representing metabolic activity of kidney parenchymal cells; the multi-mode heterogeneous data fusion engine is configured with an attention mechanism module, and the attention mechanism module is configured to calculate the mutual information quantity between any two associated nodes and distribute attention scores as side weights according to the mutual information quantity; when the co-evolution characteristics between immune and metabolic nodes are enhanced, the attention score increases accordingly; The multimodal heterogeneous data fusion engine is further configured with a global bias introduction module for introducing external clinical benchmark data including patient age, postoperative days, and a basic blood concentration of an immunosuppressant as global bias items into the dynamic iso-patterning structure.
6. The wireless implantable bio-sensor based renal rejection monitoring system of claim 1, wherein: the space-time diagram neural network prediction model is formed by alternately stacking a spatial feature extraction layer and a temporal feature memory layer; The space feature extraction layer adopts a graph rolling network architecture and is configured to simulate the diffusion process of immune activation signals in a kidney tissue microenvironment by aggregating the feature information of adjacent nodes in the dynamic abnormal graph, the hidden state update of each node depends on the feature vector of the current moment of the node and the weighted summation of the feature vectors of all neighbor nodes, and the calculation of the weight refers to the biological interaction strength among the nodes; the time characteristic memory layer adopts a long-period memory network structure and is configured to capture a subclinical fluctuation mode of rejection latency; The long-term memory network comprises an input gate, a forgetting gate and an output gate, and immune early warning signals are reserved and physiological fluctuation noise is discarded by maintaining long-term cell state vectors.
7. The wireless implantable bio-sensor based renal rejection monitoring system of claim 1, wherein: The space-time diagram neural network prediction model is configured with a pre-training module and a risk classification module based on contrast learning; The pre-training module is configured to learn general characteristic characterization of biomarker fluctuation on a historical data set and perform fine adjustment based on real-time data of a current patient, and a model loss function of the model loss function comprises a classification prediction loss and a space-time consistency constraint term, wherein the space-time consistency constraint term requires that a difference value between risk prediction values output by the model in adjacent time periods is smaller than a preset smoothing threshold value; the risk classification module is connected with an output layer of the space-time diagram neural network prediction model, is configured to normalize the rejection probability score, classifies risks into three levels of low risk, medium risk and high risk according to normalized values, and generates corresponding clinical intervention suggestions for different risk levels.
8. The wireless implantable bio-sensor based renal rejection monitoring system of claim 7, wherein: the system also comprises a user interaction terminal, wherein the user interaction terminal is provided with an abnormal event backtracking module and a remote medical cooperation module; The abnormal event backtracking module is configured to automatically call all associated biomarker data and evolution snapshots of the dynamic heterogeneous graph structure within 7 days of history when the space-time diagram neural network prediction model judges that high risk rejection tendency exists, and visually display an abnormal evolution path of edge weights in the dynamic heterogeneous graph structure; The remote medical collaboration module is configured to transmit the risk rejection report and the associated atlas snapshot to the cloud end in an encrypted manner, and call the large-scale language model to automatically generate a pathology analysis report in combination with the medical literature database.
9. The wireless implantable bio-sensor based renal rejection monitoring system of claim 8, wherein: the system also comprises a self-checking circuit and a self-cleaning module; Each detection site of the wireless implantable biological sensor array is provided with a main and standby double-way sensor; The self-checking circuit is integrated in the in-vivo signal preprocessing unit and is configured to evaluate the contact impedance of the sensing electrode and the tissue interface in real time, and when the contact impedance value corresponding to the main sensor deviates from a preset range, the control system is automatically switched to a standby sensor channel; The self-cleaning module comprises a miniature ultrasonic transducer integrated on the surface of an electrode, and is configured to generate mechanical vibration with micron-scale amplitude so as to remove fibrin and cell debris attached on the surface of a sensing membrane, wherein the triggering period of the mechanical vibration is dynamically determined by the in-vivo signal preprocessing unit according to the reduction degree of signal to noise ratio.
10. The wireless implantable bio-sensor based renal rejection monitoring system of claim 9, wherein: the system also comprises a digital twin engine and a clinical decision deduction terminal; The digital twin engine is configured to construct a four-dimensional space-time model of transplanted kidney in a virtual space according to the acquired multi-modal heterogeneous data, the geometrical topological parameters of the kidney anatomy structure and the partial differential equation set of the physiological and biochemical reaction; The digital twin engine takes the multi-dimensional biomarker signals acquired in real time as boundary conditions of the partial differential equation set, reduces concentration gradient distribution of immune molecules in a kidney tissue gap through numerical simulation, generates simulation data and inputs the simulation data into the space-time diagram neural network prediction model; The clinical decision deduction terminal is configured to receive a simulated medication regimen input by a doctor and call the digital twin engine and the space-time diagram neural network prediction model to deduce an expected response curve of a kidney immune state under the simulated medication regimen.

Description

Kidney rejection monitoring system based on wireless implantable biosensor Technical Field The invention belongs to the technical field of biological sensors and wireless health monitoring, and particularly relates to a renal rejection monitoring system based on a wireless implantable biological sensor. Background With the cross-fusion of biomedical sensors and clinical monitoring techniques, wireless implantable monitoring systems have become a sophisticated means of post-organ transplant management. The real-time grasping of the physiological function state of the transplanted organ has important significance for reducing postoperative complications and prolonging the survival time of the transplanted organ. In the kidney transplantation rehabilitation process, a high-precision sensing element is utilized to capture biochemical signals in a human body, so that a direct objective basis can be provided for early detection of kidney function damage and rejection reaction, and the mode of clinical diagnosis and treatment from experience driving to data driving is promoted. The monitoring system for renal rejection is focused on the construction of a comprehensive evaluation system covering metabolic pathways and immune responses through real-time perception of multi-modal biological information. The key objective in the direction is to cooperatively acquire key physiological parameters in a receptor by using a miniaturized sensing matrix, and to aggregate the key physiological parameters to an analysis end through an efficient data transmission protocol. The system often requires pretreatment and feature alignment of these complex biological signals in an effort to accurately characterize the viability and immune rejection strength of the transplanted kidney in complex in vivo environments. In the prior art, the monitoring means mainly depend on macroscopic clinical indexes such as creatinine elevation or urine volume change, and the diagnosis feedback is often behind the actual occurrence time of pathological injury, so that the window period of therapeutic intervention is limited. Traditional analysis methods generally perform threshold matching only on single-dimensional data, lack deep analysis of heterogeneous association between metabolites and cytokines, and are prone to misjudgment when infection or other complications occur to the receptor. The existing algorithm model shows obvious insufficient space-time characteristic capturing capability when processing long-time sequence and dynamic evolution immune signals, and cannot simulate the nonlinear diffusion process of immune storm in a biological network. These defects cause that tiny pathological fluctuations in the subclinical level are difficult to be recognized early, thereby affecting the accuracy and scientificity of early warning decisions, and a renal rejection monitoring system based on a wireless implantable biosensor is expected. Disclosure of Invention The invention aims to provide a renal rejection monitoring system based on a wireless implantable biosensor, which can solve the problems in the background art. In order to achieve the above purpose, the technical scheme adopted by the invention is as follows: A renal rejection monitoring system based on a wireless implantable biosensor comprises a wireless implantable biological sensing array, an in-vivo signal preprocessing unit, an in-vitro data receiving and synchronizing device, a multi-mode heterogeneous data fusion engine and a time space diagram neural network prediction model; the wireless implantable biological sensing array is configured to collect multi-dimensional biomarker signals associated with immunology and metabonomics in real time in a tissue microenvironment surrounding a transplanted kidney, and the multi-dimensional biomarker signals comprise cytokine concentration dynamic change data and metabolite concentration fluctuation data; the in-vivo signal preprocessing unit is electrically connected with the wireless implantable biological sensing array and is configured to filter, amplify and analog-to-digital convert an original biological signal and transmit the processed digital signal to the outside of the body through a wireless communication protocol; The in-vitro data receiving and synchronizing device is configured to receive the wireless signals from the in-vivo signal preprocessing unit, and perform time axis alignment on different types of biomarker data according to the uniform time stamp to form a structured time sequence data stream; The multi-modal heterogeneous data fusion engine is configured to map the aligned immunological data and metabonomic data into a unified knowledge graph frame to construct a dynamic heterogeneous graph structure reflecting the interaction relationship among immune cells, tubular epithelial cells and metabolic pathways; the space-time diagram neural network prediction model is configured to be used for simultaneously modeling casca