Thoracic surgery postoperative patient health data remote processing system based on Internet of Things

Abstract

The invention relates to the technical field of medical internet of things and remote monitoring, in particular to a thoracic surgery postoperative patient health data remote processing system based on the internet of things. The system comprises a baseline modeling module, a data acquisition module, a physiological state deduction module and a feedback control module. The system acquires baseline parameters and obtains body surface and intracavity real pressure data in real time for comparison and deduction; the core of the method is that a physiological mechanism kinetic equation is embedded in a loss function of a state estimation model as a hard constraint, a predicted value is corrected by using a fluid mechanics law, and a pressure residual feature is extracted to generate an intervention instruction; according to the method, the defects that a pure data driving model is easy to overfit and violates common sense are overcome, and accurate extraction of organ compliance parameters from strong noise and quantitative evaluation of functions are realized.

Description

Technical Field

[0001] This invention relates to the field of medical Internet of Things and remote monitoring technology, specifically to an Internet of Things-based remote processing system for postoperative health data of thoracic surgery patients. Background Technology

[0002] In the current postoperative monitoring environment of thoracic surgery, IoT data acquisition terminals continuously acquire patients' surface physiological time-series data and intracavitary real pressure data. Such data are often accompanied by significant individual differences and sudden physiological noise. To process the aforementioned health data, existing solutions generally employ purely data-driven deep learning models and isolated anomaly detection mechanisms, analyzing vital signs and chest drainage data separately. While this approach can achieve basic monitoring, the lack of physical mechanism constraints makes the model prone to overfitting when processing high-concurrency noise data, such as turning over and coughing, leading to predictions that violate fluid dynamics principles. Furthermore, it is susceptible to triggering numerous invalid false alarms due to random fluctuations, increasing the system's ineffective monitoring load. In addition, its alarms only remain at the level of superficial data and cannot provide decision support for addressing the underlying pathology. Therefore, how to effectively filter out physiological and sensor artifacts and quantify latent pathological features based on physical constraints to improve the accuracy of postoperative condition prediction and medical intervention has become an urgent technical problem to be solved. Summary of the Invention

[0003] The purpose of this invention is to provide an Internet of Things (IoT)-based remote processing system for postoperative health data of thoracic surgery patients, addressing the following technical problems: Existing postoperative data processing technologies for thoracic surgery suffer from a lack of physical mechanism constraints and difficulty in providing root cause pathological decision support when facing high levels of physiological noise. There is an urgent need for an IoT-based remote processing system for postoperative health data of thoracic surgery patients that can effectively filter out physiological and sensor artifacts based on physical constraints, quantify latent pathological characteristics, and generate intervention instructions containing pathological causal explanations. The objective of this invention can be achieved through the following technical solutions: Baseline modeling module, IoT data acquisition terminal, physiological state projection server, and feedback control terminal; The baseline modeling module is used to acquire organ function test data and physiological structure change range data of the target object; establish a parameterized baseline model based on the organ function test data and the physiological structure change range data; and extract structured boundary parameters from the parameterized baseline model. The IoT data acquisition terminal is used to collect the surface physiological time-series data and intracavitary real pressure data of the target object in real time, and send the surface physiological time-series data and intracavitary real pressure data to the physiological state inference server. The physiological state extrapolation server is configured with a state estimation model, the loss function of which embeds physiological mechanism dynamic equations as hard constraints. The physiological state extrapolation server uses the state estimation model, taking the body surface physiological time-series data as the driving force, and combines the parameterized baseline model and the structured boundary parameters to extrapolate and calculate the theoretical pressure change data within the cavity; and performs difference fitting calculation between the actual pressure data within the cavity and the theoretical pressure change data within the cavity to extract pressure residual features. The feedback control terminal is used to generate intervention commands based on the pressure residual characteristics, and to feed back the new time series data generated after executing the intervention commands to the baseline modeling module to update the parameterized baseline model.

[0004] Furthermore, the feedback control terminal is specifically used for: The pressure residual characteristics are conditionally determined; In response to the pressure residual feature satisfying a preset periodic offset threshold condition, the pressure residual feature is decoupled and analyzed, and implicit key parameters containing physiological and pathological correlations are output. In response to the pressure residual characteristic not meeting the preset periodic offset threshold condition, a normal state maintenance signal is output.

[0005] Furthermore, the implicit key parameters include organ re-expansion degree scoring data and cavity air leakage activity index; The feedback control terminal is also used to generate the intervention instruction containing pathological causal explanation based on the extracted organ re-expansion degree score data and the cavity air leakage activity index.

[0006] Furthermore, the target group is post-thoracic surgery patients; the baseline modeling module is specifically used for: Preoperative pulmonary function test data of the patients who underwent thoracic surgery were obtained as organ function test data. The surgical resection extent data is obtained as the physiological structural change extent data; The initial pleural surface area and the expected remaining vital capacity are used as the structured boundary parameters.

[0007] Furthermore, the IoT data acquisition terminal includes multiple physiological parameter acquisition units for acquiring the body surface physiological time-series data and the intracavitary real pressure data; the body surface physiological time-series data includes at least chest and abdominal displacement tension data, time-series blood oxygen saturation data, and digital monitoring drainage volume data.

[0008] Furthermore, the physiological mechanism dynamic equation is a nonlinear pleural cavity aerodynamic equation, used to describe the physiological constraint relationship between the theoretical pressure within the cavity and the real-time volume and the rate of volume change; the body surface physiological time-series data includes real-time volume data and volume change rate data; the state estimation model is a physical information neural network model or a Bayesian filtering model.

[0009] Furthermore, when the physiological state extrapolation server performs extrapolation calculations using the state estimation model, it is also used for: The predicted values ​​of the data representation are corrected by fluid dynamics laws using the nonlinear pleural cavity aerodynamic equation. Real-time organ compliance parameters are extracted from the data stream containing non-target environmental noise, and continuous quantitative assessment results of organ function are generated based on the real-time organ compliance parameters.

[0010] Furthermore, the intervention instructions include: Breathing regulation commands used to guide the target subject to actively clear sputum, or device control commands used to limit the device from increasing the oxygen concentration.

[0011] Compared with the prior art, the present invention has the following beneficial effects: 1. This system uses the state estimation model in the physiological state extrapolation server and embeds the physiological mechanism dynamics equation as a hard constraint in its loss function. During the extrapolation calculation, the nonlinear pleural cavity aerodynamic equation is used to correct the predicted values ​​of the data appearance by the constraints of fluid dynamics laws. This technical feature effectively overcomes the defects of traditional pure data-driven models that are prone to overfitting and producing predictions that violate the common sense of fluid dynamics when dealing with high concurrency noise. It can accurately extract real-time organ compliance parameters from the data stream containing non-target environmental noise and generate continuous quantitative assessment results of organ function accordingly. 2. This system utilizes a feedback control terminal to perform conditional judgments on the pressure residual features extracted from the fitting calculation. When the pressure residual features do not meet the preset periodic offset threshold condition, the system outputs a normal state maintenance signal. Only when the preset periodic offset threshold condition is met is the pressure residual feature decoupled for analysis. This technical feature can effectively identify and filter out single or accidental pressure residual fluctuations caused by sudden physiological changes, avoiding a large number of invalid false alarms triggered by isolated time-series data anomalies, thereby effectively reducing the frequency of invalid alarms in the system. 3. After decoupling the pressure residual characteristics, this system can output implicit key parameters containing physiological and pathological correlations, specifically including organ re-expansion degree scoring data and cavity air leakage activity index. Based on these parameters, it generates intervention instructions containing pathological causal explanations, such as respiratory regulation instructions to guide the target subject to actively expectorate, or equipment control instructions to limit the increase of oxygen concentration. This technical feature breaks through the limitations of existing monitoring schemes that only stay at the level of superficial data alarms, establishes a clear clinical pathological causal link, and provides automated decision support and closed-loop response mechanism that directly reaches the root pathology for actual medical intervention. Attached Figure Description

[0012] The present invention will be further explained below with reference to the accompanying drawings and embodiments: Figure 1 This is a system block diagram of the present invention. Detailed Implementation

[0013] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.

[0014] Example 1: Please see Figure 1 A remote processing system for postoperative thoracic surgery patient health data based on the Internet of Things includes: Baseline modeling module, IoT data acquisition terminal, physiological state projection server, and feedback control terminal; Among them, the baseline modeling module is used to acquire organ function test data and physiological structure change range data of the target object; establish a parameterized baseline model based on organ function test data and physiological structure change range data, and extract structured boundary parameters from the parameterized baseline model; The Internet of Things (IoT) data acquisition terminal is used to collect real-time physiological time-series data of the target object's body surface and real intracavitary pressure data, and send the physiological time-series data of the body surface and real intracavitary pressure data to the physiological state inference server. The physiological state extrapolation server is equipped with a state estimation model. The loss function of the state estimation model embeds physiological mechanism dynamic equations as hard constraints. The physiological state extrapolation server uses the state estimation model, with body surface physiological time-series data as the driving force, and combines a parameterized baseline model and structured boundary parameters to extrapolate and calculate the theoretical pressure change data in the cavity. The server then performs difference fitting calculation between the actual pressure data in the cavity and the theoretical pressure change data in the cavity to extract the pressure residual features, hereinafter referred to as pressure residuals. The feedback control terminal is used to generate intervention commands based on the pressure residual characteristics and to feed back the new time series data generated after the intervention commands are executed to the baseline modeling module to update the parameterized baseline model.

[0015] This embodiment provides a modeling and state inference mechanism for a remote processing system of postoperative patient health data in thoracic surgery based on the Internet of Things; specifically, the baseline modeling module obtains the objective test results of the target object and maps them into a virtual structure with boundary constraints; The IoT data acquisition terminal continuously collects external physiological characteristics and internal physical pressure, and feeds them to the physiological state estimation server. The state estimation model in the server is strictly constrained by the embedded physiological mechanism dynamic equation. The characteristic data is substituted into the equation as variables to deduce the theoretical pressure change data in the cavity under an ideal state. The system compares and fits the measured intracavitary pressure data with the theoretical pressure change data to extract the pure pressure residual features. The feedback control terminal analyzes these features and outputs corresponding instructions to update the model in a closed loop. If the IoT data acquisition terminal loses connection or data packets during this process, resulting in empty physiological time series data on the body surface, the system will automatically freeze the current simulation state, downgrade to call the historical baseline of the previous cycle to maintain monitoring, and send a hardware troubleshooting alarm to the front end until the signal is restored. For example, a scenario is set up with a 65-year-old male patient undergoing left lower lobectomy, focusing on continuous chest drainage and respiratory function monitoring within 48 hours post-surgery. The system retrieves the patient's preoperative pulmonary function test data and the specific extent of the left lower lobectomy, establishes a dedicated baseline model, and extracts data such as the initial pleural surface area. Expected remaining lung capacity Structured boundary parameters; postoperatively, wearable devices and digital drainage bottles monitor the patient's dynamic blood oxygenation, heart rate, and intrathoracic pressure in real time. The data is transmitted to the server; under the constraints of the fluid dynamics equations, the model calculates the theoretically expected pressure changes. The system calculates the difference between the two. This refers to the pressure residual characteristics, and based on these characteristics, intervention instructions for expectoration or postural adjustment are issued. The physiological feedback after the patient executes these instructions reshapes the baseline model again. The purpose of this mechanism is to correct the limitations of simplifying the human physiological system into a single time-series data source. By constraining IoT data with physical equations, it can quantify the invisible intrapleural biomechanical state in real time under strong noise and individual physiological differences.

[0016] Example 2: The feedback control terminal is specifically used for: Conditional determination of pressure residual characteristics; In response to the pressure residual features satisfying the preset periodic offset threshold condition, the pressure residual features are decoupled and analyzed, and implicit key parameters containing physiological and pathological correlations are output. In response to the pressure residual characteristics not meeting the preset periodic offset threshold condition, a normal state maintenance signal is output.

[0017] Hidden key parameters include organ re-expansion score data and cavity air leakage activity index; The feedback control terminal is also used to generate intervention instructions that include pathological causal explanations based on the extracted organ re-expansion degree score data and cavity air leakage activity index.

[0018] Regarding the solution in the previous embodiment, when faced with sudden physiological noises such as postoperative patients frequently turning over or slight coughing, single or accidental pressure residual fluctuations can easily trigger intervention, leading to a large number of invalid false alarms and increasing the invalid monitoring load of the system. Therefore, this embodiment introduces a residual decoupling step based on periodic offset determination; specifically, the feedback control terminal continuously monitors the pressure residual characteristics. The system compares the feature with a preset time window and offset amplitude. If the feature shows a regular offset across cycles and meets the preset periodic offset threshold condition, the system activates the decoupling matrix to separate the composite residual into organ re-expansion degree score data pointing to organ compliance recovery and cavity air leakage activity index pointing to wound closure. The preset periodic offset threshold condition is specifically: within a set shift time window, the pressure residual characteristics... The mean of the first derivative is greater than the empirical constant. And the variance is less than the set volatility. Among them, the empirical setting constant The value range is set to Set volatility The value range is set to The specific numerical range mentioned above was obtained by statistical distribution calibration based on historical large sample data of the target object's age group and weight; The decoupling matrix constructs an augmented vector using the residuals and their rates of change. The specific operational form of this vector is as follows: in, For time variables, for Pressure residual characteristics at any given time This represents the first derivative of the pressure residual characteristic with respect to time. For decoupling matrix, Data for scoring the degree of organ re-expansion. For cavity leakage activity index, matrix elements The model was trained and calibrated offline using preoperative baseline model data; where, in the variable subscripts, Indicates re-opening. This indicates an air leak; Specifically, the elements of the decoupling matrix The dimensions are , The dimensions are ; The dimensions are , The dimensions are ; Offline training specifically includes: collecting time-series sequences of pressure residuals from patients undergoing similar surgeries in the past as input to the training set; using the degree of re-expansion assessed by corresponding clinical experts and the measured air leakage volume as true labels; and calculating the matrix elements through partial least squares regression. The decoupling matrix in this embodiment is mainly used for rapid precondition screening of slight sudden physiological noise, while the state estimation model in embodiment 4 below is used to accurately quantify the deep physical parameters of long-term series. The two are a serial complementary relationship of precondition determination and subsequent accurate inference. If the fluctuation time of the pressure residual characteristic is extremely short and does not meet the preset periodic offset threshold condition, such as being judged as a single physiological and sensor artifact or a neurological tremor, the system will directly block the decoupling process, output a normal state maintenance signal, and observe silently. For example, in the main scenario of monitoring patients after lung resection, the patient experienced a severe, irregular spasm due to wound pain during the recovery period. The system detected a sudden spike in the pressure residual, but due to its extremely short duration, it failed to meet the preset threshold condition of continuous deviation for 3 minutes. The system determined it to be physiological noise and maintained the original monitoring state. Several hours later, the system detected that the pressure residual showed a continuous and slow upward deviation over 45 minutes, meeting the periodicity condition. The system then decoupled this residual, calculated the current organ re-expansion score as low, indicating lung collapse, and the cavity leakage activity index as high, indicating persistent pleural fistula. Based on this, the system generated an intervention instruction carrying pathological causality. The purpose of this step is to filter out physiological and sensor artifacts caused by the human body's dynamic recovery curve and coughing, establish a true clinical pathological causal link, and avoid unreasonable predictions caused by the lack of mechanistic constraints in purely data-driven models.

[0019] Example 3: The target population is post-thoracic surgery patients; the baseline modeling module is specifically used for: Preoperative pulmonary function test data of patients after thoracic surgery were obtained as organ function test data. Data on the extent of surgical resection is used as data on the extent of physiological structural changes. The initial pleural surface area and the expected remaining vital capacity were used as structured boundary parameters.

[0020] The IoT data acquisition terminal includes multiple physiological parameter acquisition units for collecting body surface physiological time-series data and intracavitary real pressure data; the body surface physiological time-series data includes at least chest and abdominal displacement tension data, time-series blood oxygen saturation data, and digital monitoring drainage volume data.

[0021] The above embodiments rely on the input of basic data under ideal conditions. However, in real thoracic surgery environments, conventional methods often process vital sign data and thoracic drainage data separately, relying on subjective experience and judgment, which makes the model boundaries blurred. Therefore, this embodiment provides a multi-source cross-domain IoT data fusion and boundary anchoring mechanism. Specifically, during the initialization phase, the system forcibly inputs the patient's objective medical records after thoracic surgery, locking in preoperative pulmonary function test data and specific physiological structural changes. The system then uses a geometric mapping algorithm to reduce the dimensionality of this data, generating two rigid structured boundary parameters: the initial pleural surface area and the estimated remaining vital capacity. Specifically, the formula for calculating the estimated remaining vital capacity is: in, To estimate remaining lung capacity, The preoperative forced vital capacity was obtained from the preoperative pulmonary function test data. This represents the total number of lung segments removed. This refers to the sequence number of the lung segment that was removed. For the first Volume weighting coefficient for each resected lung segment This represents the dimensionless standardized volume proportion of the corresponding lung segment to the total lung volume; in the parameter subscript, Indicates the remainder. Indicates preoperative, Representing lung segment; initial pleural surface area The surface area was obtained by nonlinear calculation using an empirical formula and the resection ratio; specifically, the mathematical formula for nonlinear calculation is as follows: in, This represents the initial pleural surface area. The target subject's weight, in kg. The height of the target object is expressed in cm. The preceding formula is based on the DuBois empirical formula for body surface area to calculate body surface characteristics; 0.425 and 0.725 are fixed exponential constants in this empirical formula. The pleural cavity conversion factor is preferably in the range of 0.6 to 0.8. The total standardized volume of the resected lung segment. This represents the preoperative standardized total lung volume; where, in the parameter subscripts, Indicates that it has been removed. The total is represented; 0.007184 is the constant coefficient in the empirical formula for the surface area of ​​this body. Meanwhile, the IoT data acquisition terminal activates multiple physiological parameter acquisition units distributed on the hospital bed and the patient's body surface, and acquires chest and abdominal displacement tension data, time-series blood oxygen saturation data, and digital monitoring drainage data in parallel; if one of the acquisition units slips or acquires continuous zero noise, the system will use the cross-correlation of the other dimensions to perform a downgraded estimation and prompt the nurse to go to the bedside to check the physical connection. For example, in the main scenario, the system pre-programs the physiological structural changes resulting from the resection of the left lower lobectomy in the 65-year-old patient, sets the expected remaining lung capacity to 75% of the preoperative value, and uses this as a structured boundary parameter. During the postoperative evolution phase, the wearable breathing belt captures chest and abdominal displacement tension data, the pulse oximeter captures time-series blood oxygen saturation data, and the digital chest drainage system simultaneously outputs digital monitoring drainage volume data. These three high-frequency data streams serve as a joint driving force to jointly deduce the true state of the patient's current closed variable-volume physical system. The purpose of this mechanism is to break the current situation of fragmented and superficial information quantification after thoracic surgery, and to provide comprehensive fluid dynamics and aerodynamic boundary condition inputs in real time through multi-dimensional sensor networking.

[0022] Example 4: The physiological mechanism dynamic equation is a nonlinear pleural cavity aerodynamic equation, used to describe the physiological constraint relationship between the theoretical pressure in the cavity and the real-time volume and the rate of volume change; the body surface physiological time series data includes real-time volume data and the rate of volume change data; the state estimation model is a physical information neural network model or a Bayesian filtering model.

[0023] When using the state estimation model for inference calculations, the physiological state extrapolation server is also used for: The predicted values ​​of data representations are corrected by fluid dynamics laws using nonlinear pleural cavity aerodynamic equations. Real-time organ compliance parameters are extracted from the data stream containing non-target environmental noise, and continuous quantitative assessment results of organ function are generated based on the real-time organ compliance parameters.

[0024] Existing conventional deep learning models in artificial intelligence fall into the category of purely data-driven approaches. When dealing with small postoperative sample data and high-concurrency noise, they are prone to overfitting, producing erroneous predictions that violate physical principles. Therefore, this embodiment provides a gray-box model computation mechanism for correcting physical equations. Specifically, a physical information neural network model is instantiated within the physiological state deduction server. This model includes an input layer, three hidden layers, and an output layer. The input layer receives data containing real-time volumetric data. With volume change rate data The model outputs a temporal vector; each hidden layer contains 64 neurons and uses hyperbolic tangent as the activation function to ensure second-order differentiability with respect to time; the output layer outputs the predicted theoretical intrapleural pressure and implicit physical parameters; the loss function of this model incorporates equations describing nonlinear pleural cavity aerodynamics, such as... The total loss function of the physical information neural network model Data-driven loss and physical equation residual loss The composition, specifically the formula, is as follows: in, For the total loss function, This represents the total number of sampling points within the time window. The sampling point sequence number, For the first The time corresponding to each sampling point For time variables, for Real-time volume data at any given moment. for Volume change rate data at time t, for The theoretical intracavitary pressure value is always derived from aerodynamic equations. for Real intracavitary pressure data at any given time. for The predicted intracavitary pressure output by the neural network model at any given time. For physical constraint adaptive weighting coefficients, Given the initial resting pressure constant of the pleural cavity, the system automatically updates the network weights and latent organ compliance parameters through backpropagation. and airway resistance parameters The first term on the right-hand side of the total loss function is the data-driven loss. The second term is the residual loss of the physical equation. ; in, It refers to the respiratory system. Indicates airway, Indicates the pleural cavity. Indicating truth, Indicates prediction, Representing theory, To indicate the total, Representing data, Indicates physics; During the simulation, the model incorporates real-time volumetric data from the physiological time-series data of the body surface. With volume change rate data When the neural network outputs a predicted value based on the data representation, the system forces it to undergo fluid dynamics law constraint correction using the above equations; after correction, the system solves for the hidden real-time organ compliance parameters from the noisy data stream. If, during the calculation process, the predicted value of the original output of the neural network deviates significantly from the equation constraints, such as calculating the compliance of negative numbers, the system will determine that the current input data has serious distortion, immediately discard the appearance of this batch of data, and restart the estimation cycle. For example, in the main scenario, due to the patient's irregular shallow and rapid breathing caused by incision pain, the pure data-driven model may incorrectly predict extreme hypoxia in the next two hours based on blood oxygen fluctuations; however, the state estimation model in this embodiment substitutes the current volume change rate into the nonlinear pleural cavity aerodynamic equation for verification and finds that the prediction violates the dynamic balance law of elastic resistance and airway resistance; the system then corrects the predicted value and accurately extracts the organ compliance parameters that reflect the true physiological state, generating continuous quantitative assessment results of organ function; The purpose of this mechanism is to overcome the technical bottleneck of extracting real organ compliance parameters from apparent temporal fluctuations in the absence of physical constraints, and to realize the transformation from pure data-driven prediction to causal inference based on fused mechanisms.

[0025] Example 5: Intervention instructions include: Breathing regulation instructions used to guide the target subject to actively clear sputum, or equipment control instructions used to limit the equipment from increasing the oxygen concentration.

[0026] Previous anomaly detection algorithms or models only output alarms, indicating the deterioration of superficial data, but failing to provide decision support for the clinical implementation end to address the root cause. This forces medical staff to spend a lot of effort to investigate the cause again. Therefore, this embodiment provides an automated closed-loop response step with clear physiological guidance. Specifically, after receiving the decoupled implicit key parameters, the feedback control terminal maps out specific execution actions according to the built-in medical logic tree, generating active sputum expectoration and breathing regulation instructions to directly guide the target object, or sending equipment control instructions to the oxygen supply terminal to limit the oxygen concentration; the feedback control terminal establishes a communication connection with relevant medical hardware devices in the ward through the hospital's local area network or IoT gateway to realize the issuance of control instructions and the synchronization of equipment status; The built-in medical logic tree mapping includes the following specific decision rules: If the organ re-expansion degree score ( ) And the airway resistance parameters are extracted in real time ( ) The system is identified as airway obstruction by secretions, and an active expectoration breathing regulation command is generated. If the organ re-expansion degree score ( ) And the cavity leakage activity index ( ) If the cavity is determined to be continuously leaking air, a device control command is generated to limit the equipment from increasing the oxygen concentration. in, The set threshold for scoring the degree of organ re-expansion, The set threshold for abnormal airway resistance, The threshold value for the cavity leakage activity index is set; further, in this embodiment, The value is set to the standard full score state. ; The value is set to ; The value is set to Those skilled in the art will understand that when the real-time decoupled data crosses the aforementioned specific threshold, the system can clearly determine that a corresponding adverse physiological event has occurred. If the system assesses that the target subject's current organ re-expansion score is extremely low and the vital signs are too weak to independently execute the breathing regulation command for active sputum clearance, the command flow will be automatically downgraded and redirected to issue the highest level alarm for manual sputum suction and thoracoscopic assisted examination to the nursing station. For example, in the main scenario, after residual analysis, the system determines that the current decrease in blood oxygen in the patient who underwent left lower lobectomy is not due to insufficient oxygen supply, but rather due to a surge in micro-airway resistance caused by airway secretion obstruction. At this point, the system automatically generates an interpretable instruction and pushes it to the bedside terminal: since real-time lung compliance has not improved and there is a resistance residual, it is recommended to actively cough up sputum instead of increasing the oxygen concentration. After the patient or nurse follows the instruction to perform sputum clearance, the lungs re-expand, and the newly generated high-quality time-series data flows back into the system to form a new historical baseline. The purpose of this step is to expand the entire system beyond the superficial IoT function of remotely viewing data, into a non-invasive, continuously working online clinical organ function assessment system, achieving closed-loop control by using quantitative functional assessment to guide actual medical interventions.

[0027] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A remote processing system for postoperative thoracic surgery patient health data based on the Internet of Things, characterized in that, include: Baseline modeling module, IoT data acquisition terminal, physiological state projection server, and feedback control terminal; The baseline modeling module is used to acquire organ function test data and physiological structure change range data of the target object; establish a parameterized baseline model based on the organ function test data and the physiological structure change range data; and extract structured boundary parameters from the parameterized baseline model. The IoT data acquisition terminal is used to collect the surface physiological time-series data and intracavitary real pressure data of the target object in real time, and send the surface physiological time-series data and intracavitary real pressure data to the physiological state inference server. The physiological state extrapolation server is configured with a state estimation model, the loss function of which embeds physiological mechanism dynamic equations as hard constraints. The physiological state extrapolation server uses the state estimation model, taking the body surface physiological time-series data as the driving force, and combines the parameterized baseline model and the structured boundary parameters to extrapolate and calculate the theoretical pressure change data within the cavity; and performs difference fitting calculation between the actual pressure data within the cavity and the theoretical pressure change data within the cavity to extract pressure residual features. The feedback control terminal is used to generate intervention commands based on the pressure residual characteristics, and to feed back the new time series data generated after executing the intervention commands to the baseline modeling module to update the parameterized baseline model.

2. The IoT-based remote processing system for postoperative thoracic surgery patient health data according to claim 1, characterized in that, The feedback control terminal is specifically used for: The pressure residual characteristics are conditionally determined; In response to the pressure residual feature satisfying a preset periodic offset threshold condition, the pressure residual feature is decoupled and analyzed, and implicit key parameters containing physiological and pathological correlations are output. In response to the pressure residual characteristic not meeting the preset periodic offset threshold condition, a normal state maintenance signal is output.

3. The IoT-based remote processing system for postoperative thoracic surgery patient health data according to claim 2, characterized in that, The hidden key parameters include organ re-expansion degree scoring data and cavity air leakage activity index; The feedback control terminal is also used to generate the intervention instruction containing pathological causal explanation based on the extracted organ re-expansion degree score data and the cavity air leakage activity index.

4. The IoT-based remote processing system for postoperative thoracic surgery patient health data according to claim 1, characterized in that, The target group is post-thoracic surgery patients; the baseline modeling module is specifically used for: Preoperative pulmonary function test data of the patients who underwent thoracic surgery were obtained as organ function test data. The surgical resection extent data is obtained as the physiological structural change extent data; The initial pleural surface area and the expected remaining vital capacity are used as the structured boundary parameters.

5. The IoT-based remote processing system for postoperative thoracic surgery patient health data according to claim 1, characterized in that, The IoT data acquisition terminal includes multiple physiological parameter acquisition units for acquiring the body surface physiological time-series data and the intracavitary real pressure data; the body surface physiological time-series data includes at least chest and abdominal displacement tension data, time-series blood oxygen saturation data, and digital monitoring drainage volume data.

6. The IoT-based remote processing system for postoperative thoracic surgery patient health data according to claim 1, characterized in that, The physiological mechanism dynamic equation is a nonlinear pleural cavity aerodynamic equation, used to describe the physiological constraint relationship between the theoretical pressure in the cavity and the real-time volume and the rate of volume change; the body surface physiological time series data includes real-time volume data and volume change rate data; the state estimation model is a physical information neural network model or a Bayesian filtering model.

7. The IoT-based remote processing system for postoperative thoracic surgery patient health data according to claim 6, characterized in that, When the physiological state extrapolation server performs extrapolation calculations using the state estimation model, it is also used for: The predicted values ​​of the data representation are corrected by fluid dynamics laws using the nonlinear pleural cavity aerodynamic equation. Real-time organ compliance parameters are extracted from the data stream containing non-target environmental noise, and continuous quantitative assessment results of organ function are generated based on the real-time organ compliance parameters.

8. The IoT-based remote processing system for postoperative thoracic surgery patient health data according to claim 1, characterized in that, The intervention instructions include: Breathing regulation commands used to guide the target subject to actively clear sputum, or device control commands used to limit the device from increasing the oxygen concentration.

Images