Flap blood supply monitoring and early warning system and method based on wearing technology

Abstract

The invention belongs to the related technical field of medical health monitoring, and particularly relates to a flap blood supply monitoring and early warning system and method based on a wearing technology, and the system specifically comprises a wearable blood supply monitor of a client and a cloud real-time flap state monitoring platform of a medical end. The wearable blood supply monitor comprises a front-end flexible sensor used for detecting flap state indexes and communicating with a rear-end host in real time; and the rear-end host is used for sensor data processing and visualization, and is in real-time communication with the front-end flexible sensor and the cloud real-time flap state monitoring platform. And the cloud real-time skin flap state monitoring platform analyzes client data in real time, adopts a current machine learning model for classification to divide the skin flap blood supply state, and obtains a final skin flap blood supply state classification result through a voting mechanism. Compared with a traditional blood supply monitoring mode, microcirculation changes of the skin flap can be monitored in real time, the blood supply state of the skin flap can be intelligently judged, and the blood supply monitoring efficiency, accuracy and reliability are effectively improved.

Description

Technical Field

[0001] The present invention relates to the field of medical health monitoring technology, and in particular to a flap blood supply monitoring and early warning system and method based on wearable technology. Background Art

[0002] Free flap transplantation is a common method for tumor resection or post-traumatic reconstruction. The operation requires the removal of a vascularized tissue flap from the donor area of ​​the body, which is then transplanted to the recipient area, and the flap vascular pedicle is anastomosed with the recipient area blood vessels to re-establish the flap blood supply. The challenge of this operation lies in the delicate microanastomosis of the arteriovenous vessels. Once a vascular crisis such as persistent spasm or thrombosis occurs in the anastomosed blood vessels after surgery, it will lead to ischemic necrosis of the transplanted flap and failure of the operation. Existing evidence and expert opinions indicate that early detection of possible vascular crises and timely surgical exploration after surgery will significantly increase the flap salvage rate. Since vascular crises often occur within 24 to 72 hours after surgery, the changes in flap blood supply need to be closely monitored in the early stage after free flap transplantation.

[0003] At present, there is no standardized method for postoperative blood flow monitoring of free flap transplantation in clinical practice. The clinical team mainly relies on manual regular observation of the color, swelling, and capillary response of the flap to make judgments. Although this assessment is simple, it cannot achieve continuous monitoring, and the judgment is subjective and relies on the team's experience, which may delay the discovery and treatment of vascular crisis. In recent years, several systems have been developed to continuously monitor the blood flow of flaps after surgery by observing relevant indicators of flap microcirculation. Monitoring the change of flap temperature is one of the first technologies used to evaluate the microcirculation of flaps. Clinical studies have also shown that changes in skin temperature are a more sensitive and effective way to judge the microcirculation of flaps. However, it should be pointed out that when choosing temperature as an observation indicator of flap blood flow, it is necessary to consider the influence of variables such as ambient temperature, flap coverage, and the position of the temperature sensor on the flap on the results. In general, because this method is simple and economical, it can still be used in combination with other monitoring methods to improve the accuracy of monitoring.

[0004] With the development of biochemical probes, implantable sensors such as microdialysis or tissue oxygen partial pressure have shown good application prospects, especially for blood flow monitoring of buried flaps, such as the scheme of Chinese patent application 201721412758.1. Microdialysis probes can measure metabolites including lactate, glucose, glycerol, and pyruvate. Although biochemical sensors can continuously evaluate microcirculatory biochemical indicators and are highly sensitive to changes in their concentrations, there is still a lack of detailed research on the reliability of biochemical probes for blood flow monitoring of different flaps, trauma to patients, and costs, which hinders their use in routine clinical practice. At present, physical monitoring methods such as Doppler have been developed in implantable and handheld devices. Implantable Doppler has been shown to be effective in detecting flap vascular crisis, especially for observation of buried flaps and muscle transplants. Laser Doppler is a handheld device mainly used for observation of non-buried flaps. Its operation is relatively delicate, and it is difficult to maintain a fixed penetration depth because it is sensitive to movement. Summary of the invention

[0005] The purpose of the present invention is to overcome the defects of the above-mentioned prior art and provide a flap blood supply monitoring and early warning system and method based on wearable technology.

[0006] The purpose of the present invention can be achieved by the following technical solutions:

[0007] As a first aspect of the present invention, a flap blood flow monitoring and early warning system based on wearable technology is provided, which specifically includes a wearable blood flow monitor on the client side and a cloud-based real-time flap status monitoring platform on the medical side.

[0008] The wearable blood flow monitor comprises: a front-end flexible sensor for detecting skin flap status indicators and communicating with a back-end host in real time; a back-end host for sensor data processing and visualization, communicating with the front-end flexible sensor and a cloud-based real-time skin flap status monitoring platform in real time;

[0009] The cloud-based real-time flap status monitoring platform analyzes client data in real time, uses the machine learning model currently used for classification to divide the flap blood supply status, and obtains the final flap blood supply status classification result by a voting mechanism.

[0010] As a preferred technical solution, the front-end flexible sensor includes:

[0011] Skin flap temperature monitoring sensor, used to detect the surface temperature of the skin flap after surgery;

[0012] Tissue metabolism monitoring sensors to detect lactate and glucose levels in flap tissue after surgery;

[0013] Tissue oxygen monitoring sensors for detecting tissue oxygen saturation, oxyhemoglobin levels, and deoxyhemoglobin levels;

[0014] Data transmission module, used for real-time communication with the back-end host;

[0015] The desensitized skin-friendly medical silicone membrane is used to attach each sensor to the surface of the flap.

[0016] As a preferred technical solution, the tissue metabolism monitoring sensor includes an improved glucose sensor for detecting the glucose and lactic acid content of the flap tissue;

[0017] The tissue oxygen content monitoring sensor is composed of a photoelectric detector, a red LED, a green LED and an IR LED.

[0018] As a preferred technical solution, the cloud-based real-time flap status monitoring platform includes an early warning module:

[0019] The early warning module sets the threshold of the flap status indicator according to different patients and flap types; when the monitored flap status exceeds the threshold, or the analysis result of the cloud-based real-time flap status monitoring platform is not normal, the cloud-based real-time flap status monitoring platform will send an alarm message to medical staff, and the client will also display a warning.

[0020] As a second aspect of the present invention, a flap blood supply monitoring and early warning method is provided, wherein the method adopts the flap blood supply monitoring and early warning system as described above, and the steps include:

[0021] The front-end flexible sensor collects blood circulation status indicator data, including flap temperature, tissue metabolism, and tissue oxygen content;

[0022] Preprocessing and feature extraction of the acquired blood circulation status index data;

[0023] The extracted features are input into a variety of pre-trained machine learning models for classification to obtain the classification probabilities of various skin flap blood supply states;

[0024] A voting mechanism is used for the classification probability of each machine learning model to obtain the final classification result of the flap blood supply status.

[0025] As a preferred technical solution, the tissue metabolism data includes: glucose and lactic acid, and the tissue oxygen content data includes: tissue oxygen saturation, oxygenated hemoglobin level and deoxygenated hemoglobin level.

[0026] As a preferred technical solution, the extracted features include: the change rate and standard deviation of each blood circulation status index and the ratio of glucose / lactate to the ratio of oxygenated hemoglobin level and deoxygenated hemoglobin level.

[0027] As a preferred technical solution, the blood supply status of the flap includes:

[0028] Normal state:

[0029] Skin flap temperature: within the set temperature range;

[0030] Tissue glucose levels: Similar to the patient's blood glucose levels, within the normal glucose level range;

[0031] Lactate level in tissues: low, within the normal range of lactate level;

[0032] Tissue oxygen saturation: Keep within the normal range of tissue oxygen saturation;

[0033] Tissue oxygenated hemoglobin level: higher than the set tissue oxygenated hemoglobin level threshold;

[0034] Tissue deoxyhemoglobin level: below the set tissue deoxyhemoglobin level threshold;

[0035] Suspicious status:

[0036] Flap temperature: slightly low or high, but within the set temperature range, and the temperature fluctuation is less than the threshold;

[0037] Glucose levels in tissues: slightly decreased or increased, but not exceeding the upper and lower limits of the normal glucose level range;

[0038] Tissue lactate levels: slightly elevated but not exceeding the upper limit of the normal lactate level range;

[0039] Tissue oxygen saturation: slightly decreased or increased, with the change amplitude less than the threshold;

[0040] Tissue oxygenated hemoglobin level: slightly decreased, the decrease is less than the threshold;

[0041] Tissue deoxyhemoglobin level: slightly increased, the increase is less than the threshold;

[0042] Abnormal status:

[0043] Skin flap temperature: obviously too low or too high, and the temperature is outside the set temperature range;

[0044] Glucose level in tissues: significant decrease or increase, decrease or increase greater than the threshold;

[0045] Lactate levels in tissues: significantly increased, with the increase greater than the threshold;

[0046] Tissue oxygen saturation: significantly decreased, with the decrease greater than the threshold;

[0047] Tissue oxygenated hemoglobin level: significantly decreased, with the decrease greater than the threshold;

[0048] Tissue deoxyhemoglobin level: significantly increased, the increase is greater than the threshold.

[0049] As a preferred technical solution, the multiple classification machine learning algorithms include: K nearest neighbor algorithm, support vector machine, random forest and Adaboost;

[0050] The input of the machine learning model is the extracted feature vector, and the output is the category probability of the blood supply status of each flap.

[0051] As an optimal technical solution, the method sets the threshold of the flap status indicator according to different patients and flap types. When the monitored flap parameters exceed the threshold, or the analysis results of the detection show that the flap is not in a normal state, an early warning message will be sent to medical staff.

[0052] Compared with the prior art, the present invention has the following beneficial effects:

[0053] 1) The design of the intelligent blood flow monitoring and early warning system after free flap transplantation provided by the present invention is mainly based on near-infrared spectroscopy (NIRS) technology. It integrates multiple sensors and can simultaneously and rapidly and continuously measure the changes in the concentrations of multiple components such as skin temperature, oxygenated hemoglobin and deoxygenated hemoglobin in tissues, lactic acid, and glucose, thereby improving the accuracy of flap blood flow monitoring. Compared with traditional blood flow monitoring methods, the present invention can effectively improve the efficiency, accuracy, and reliability of blood flow monitoring by observing the microcirculation parameters of the flap in real time to determine its blood flow status.

[0054] 2) The blood circulation monitor provided by the present invention is based on wearable technology, which can monitor and record various parameter data of blood circulation after free flap transplantation in real time, and upload the data to the cloud. The machine data learning model of the cloud platform will intelligently analyze and classify the data and send early warning information to medical staff. At the same time, the cloud also provides medical staff with thresholds that can be set by themselves to adapt to different patient conditions.

[0055] 3) The present invention monitors tissue oxygen content based on multispectral spatial frequency domain imaging (SFDI) technology, which can detect oxygen saturation changes secondary to partial blood flow obstruction earlier.

[0056] 4) Among the features extracted and input into the machine learning model for classification in the present invention, the lactate / glucose ratio can be used to predict partial or complete flap necrosis; and the detection of the ratio of oxyhemoglobin to deoxyhemoglobin can detect flap abnormalities before clinical suspicion of thrombosis. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 This is a workflow diagram of intelligent blood flow monitoring and early warning after free flap transplantation of the present invention;

[0058] Figure 2 A model diagram of a blood circulation monitor of a client in a certain implementation case of the present invention;

[0059] Figure 3 A sensor layout diagram of a front-end flexible sensor in a certain implementation case of the present invention;

[0060] The numbers in the figure are as follows: 1. Flexible sensor, 2. Flexible sensor storage place, 3. Temperature sensor, 4. Tissue metabolism sensor, 5. Photodetector, 6. Red LED, 7. Green LED, 8. IR LED. DETAILED DESCRIPTION

[0061] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. This embodiment is implemented based on the technical solution of the present invention, and provides a detailed implementation method and specific operation process, but the protection scope of the present invention is not limited to the following embodiments.

[0062] Example 1

[0063] As one of the embodiments of the present invention, this embodiment provides an intelligent blood flow monitoring and early warning system for free flap transplantation based on wearable technology. The present invention obtains the microcirculation parameters of the postoperative flap to monitor the blood flow status of the flap in real time, which can effectively improve the efficiency, accuracy and reliability of flap blood flow monitoring.

[0064] The intelligent blood flow monitoring and early warning system after free flap transplantation proposed by the present invention comprises: a wearable blood flow monitor as a client and a cloud-based real-time flap status monitoring platform.

[0065] The wearable blood circulation monitor includes: a front-end flexible sensor for detecting skin flap status indicators and communicating with a back-end host in real time; and a back-end host for sensor data processing and visualization and real-time communication with a cloud platform.

[0066] The front-end flexible sensors include: a flap temperature monitoring sensor for detecting the surface temperature of the flap after surgery; a tissue metabolism monitoring sensor for detecting the levels of lactate and glucose in the flap tissue after surgery; a tissue oxygen content monitoring sensor for detecting tissue oxygen saturation (StO2), oxygenated hemoglobin level (HbO2) and deoxygenated hemoglobin level (HHb); a data transmission module for real-time communication with the back-end host; and a desensitized skin-friendly medical silicone membrane for attaching the sensor to the flap surface.

[0067] The backend host includes: a data display screen, which is used for data visualization of the client host; a rechargeable battery, which is used to power the client host and sensors; a client mainboard, which is used for processing and visualization of sensor data, and real-time communication with the cloud platform; and a convenient external housing, which is used to fix and protect each device.

[0068] Among them, the sensor data processing of the client mainboard includes temperature and tissue metabolism (glucose, lactate) data registration, LED adjustment for detecting tissue oxygen content (StO2, HbO2, HHb), data preprocessing and feature extraction. Data preprocessing includes: outlier removal, which is used to remove extreme abnormal monitoring values ​​caused by factors such as environment and usage, noise filtering, which is used to remove the noise of the monitoring sensor itself, and standardization, which is used to standardize different features to the same scale.

[0069] The present invention extracts the following features from the collected feature indicators:

[0070] Change rate: Calculate the change in detection values ​​between adjacent time points to evaluate the fluctuation of the corresponding indicators;

[0071] Standard Deviation: Assess the volatility of the monitoring indicator over a period of time;

[0072] Glucose / lactate ratio: reflects metabolic status. When hypoxia occurs, lactate levels increase and the ratio decreases.

[0073] The ratio of HbO2 and HHb: reflects the balance of oxygen delivery and consumption.

[0074] Follow-up studies have confirmed that a high lactate / glucose ratio can be used to predict partial or complete flap necrosis. The detection of the ratio of HbO2 to HHb can detect flap abnormalities before clinical suspicion of thrombosis, and can detect changes in oxygen saturation secondary to partial blood flow obstruction earlier.

[0075] After the backend host processes the sensor data and visualizes the blood circulation status indicators, it sends the indicator data synchronously to the cloud platform.

[0076] The cloud-based real-time flap status monitoring platform includes: a visualization module for drawing a change curve chart of blood circulation status indicators and a corresponding interactive interface; an intelligent analysis module for real-time analysis of client data; and an early warning module for sending early warnings in a timely manner.

[0077] Among them, the characteristic of the intelligent analysis module is that the data analysis in the cloud adopts the machine learning model currently used for classification, specifically K nearest neighbor algorithm, support vector machine, random forest and Adaboost. The input of the machine learning model is the feature vectors extracted above, and the output is the classification probability of each flap status category. These models are pre-trained in the data set collected in advance by the hospital. According to the real-time status indicators of the patient's postoperative flap, the machine learning model divides the blood supply status of the flap into one of three states (normal, suspicious and abnormal). The results of each model are obtained by a voting mechanism to obtain the final classification result. Afterwards, the data analysis module synchronizes the results to the client. The specific quantitative description of the three flap states (normal, suspicious and abnormal) is as follows:

[0078] 1. Description of normal status:

[0079] Flap temperature: close to the patient's body temperature, usually between 36.5-37.5℃;

[0080] Tissue glucose level: Similar to the patient's blood sugar level, usually between 70-140 mg / dL, with specific values ​​determined by the physician;

[0081] Lactate levels in tissues: low, usually between 0.5-2.2 mmol / L;

[0082] Tissue oxygen saturation (StO2): maintained between 70% and 90%;

[0083] Tissue oxygenated hemoglobin level (HbO2): higher, reflecting adequate oxygen supply;

[0084] Tissue deoxyhemoglobin level (HHb): Low, reflecting normal oxygen consumption.

[0085] 2. Description of suspicious status:

[0086] Skin flap temperature: slightly low or high, with little temperature fluctuation;

[0087] Glucose levels in tissues: slightly decreased or increased, but not exceeding the upper and lower limits of the normal range;

[0088] Tissue lactate levels: slightly elevated but not exceeding the upper limit of normal;

[0089] Tissue oxygen saturation (StO2): slightly decreased or increased, but the change was not large;

[0090] Tissue oxygenated hemoglobin level (HbO2): slightly decreased;

[0091] Tissue deoxyhemoglobin level (HHb): slightly elevated.

[0092] 3. Abnormal status description:

[0093] Skin flap temperature: significantly low or high, indicating severe reduction in blood flow or inflammation / infection;

[0094] Glucose levels in tissues: A significant decrease or increase may indicate ischemia or inflammation;

[0095] Lactate levels in tissues: significantly elevated, indicating severe hypoxia and anaerobic metabolism;

[0096] Tissue oxygen saturation (StO2): significantly decreased, indicating insufficient local oxygen supply;

[0097] Tissue oxygenated hemoglobin level (HbO2): significantly decreased, reflecting insufficient oxygen delivery;

[0098] Tissue deoxyhemoglobin level (HHb): significantly increased, indicating severe local hypoxia.

[0099] The three states are qualitatively described as:

[0100] 1. Normal state: After surgery, the flap blood flow is normal, tissue metabolism is balanced, and there is no sign of ischemia or inflammation;

[0101] 2. Suspicious status: The flap blood flow may be slightly reduced, but not to the extent of severe ischemia, or there may be a mild inflammatory reaction in the early postoperative period;

[0102] 3. Abnormal conditions: The blood flow to the flap is significantly reduced, the tissue is ischemic and hypoxic, the flap is infected or severely inflamed, the temperature is significantly increased, and the metabolism is abnormal.

[0103] The main characteristics of the four machine learning algorithms are:

[0104] 1. K-nearest neighbor algorithm (KNN): Based on the voting mechanism of neighbor samples, the flaps are classified into normal, suspicious or abnormal states according to the similarity of their status indicators. The distance calculation uses Euclidean distance;

[0105] 2. Support Vector Machine (SVM): It realizes classification by mapping data into high-dimensional space and finding the optimal hyperplane. It is applicable to linear and nonlinear data. The kernel function is Gaussian kernel function.

[0106] 3. Random Forest (RF): integrates multiple decision trees, each of which is trained using a randomly sampled feature subset to improve classification accuracy through voting or averaging;

[0107] 4. Adaboost (AB): Iteratively trains multiple weak classifiers and adjusts sample weights based on the results of the previous iteration to improve the accuracy of the model.

[0108] These models are trained with a large amount of labeled training data in the pre-training stage to classify and evaluate the blood supply status of the flap in real-time monitoring. For machine learning model i, its accuracy on category j is ACC ij , whose output on category j is P ij , with weight λ ij Our weighted voting mechanism works as follows:

[0109] P0=λ 00 P 00 +λ 10 P 10 +λ 20 P 20 +λ 30 P 30

[0110] P1=λ 01 P 01 +λ 11 P 11 +λ 21 P 21 +λ 31 P 31

[0111] P2=λ 02 P 02 +λ 12 P 12 +λ 22 P 22 +λ 32 P 32

[0112]

[0113] Among them, P0, P1, and P2 represent the probabilities of the flap being normal, abnormal, and suspicious, respectively. The final result is determined based on the maximum probability.

[0114] The main feature of the early warning module is that medical staff can set the threshold of flap status indicators according to different patients and flap types. When the monitored flap parameters exceed the threshold, or the analysis results of the cloud are not normal, the cloud will send an early warning message to the medical staff, and the client will also display the early warning.

[0115] Example 2

[0116] As another embodiment of the present invention, Figure 1 As shown, the monitoring process of the blood circulation detection and early warning system provided by the present invention as described in the above embodiment includes the following steps:

[0117] S1. Set the flap monitoring point according to the doctor's instructions. First, connect the front-end flexible sensor to the client host for communication; then, attach the front-end flexible sensor to the flap surface designated by the medical staff. If it is easy to slip, use medical bandages or tape to fix it to prevent it from falling off. After fixing, the user can turn on the monitor. The monitor starts working.

[0118] S2, the flexible sensor monitors the flap status in real time and transmits the sensor data to the client host. The client host processes the data and converts the raw sensor data into corresponding indicator data.

[0119] S3. The client host draws the real-time change curves of various indicators on the display screen.

[0120] S4. The client host calls the cloud server API to upload data. After receiving the client data, the cloud platform draws the real-time change curve of each indicator.

[0121] S5. The cloud platform conducts intelligent analysis on the client data to determine the blood supply status of the flap (normal, suspicious, and abnormal).

[0122] S6. The early warning mechanism of the cloud platform determines the final blood circulation status based on the blood circulation status indicators and the results of the intelligent analysis module. The cloud platform synchronizes the determination results to the client and sends early warning information depending on the final blood circulation status.

[0123] like Figure 2 As shown, the wearable blood circulation monitor described in the present invention specifically includes: a front-end flexible sensor for detecting skin flap status indicators and communicating with a back-end host in real time; a back-end host for sensor data processing and visualization and real-time communication with a cloud platform.

[0124] Example 3

[0125] In this implementation case, the main scheme is the same as that in the above embodiment, wherein the layout of the front-end flexible sensor is as follows: Figure 3 As shown, its main components include: GY-MCU90614 module as a temperature sensor for measuring the temperature of the flap and surrounding skin; an improved Abbott instant glucose sensor for detecting the glucose and lactate content of the flap tissue; a tissue oxygen content monitoring sensor composed of a photodetector 5, a red LED 6, a green LED 7 and an IR LED 8; a desensitized skin-friendly medical silicone membrane for fixing the sensor to provide self-adhesive properties and mechanical robustness to ensure that the sensor adheres well to the flap without causing damage and can maintain its integrity for accurate measurement; a Bluetooth module for communicating with the back-end host; and a circuit layout connected to these sensors.

[0126] The present invention adopts near infrared spectroscopy (NIRS) to non-invasively measure the metabolic changes of biological tissues. In the near infrared spectral region (mainly 650-900nm), the absorption of light by water and hemoglobin is relatively low. Therefore, this wavelength region is more suitable for monitoring changes in tissue microcirculation. The NIRS instrument is used to capture the interaction between near infrared photons and body tissues, so as to quickly measure the concentration changes of oxygenated hemoglobin (HbO2) and deoxygenated hemoglobin (HHb), lactic acid and glucose and other components present in the tissues. It is more reliable to measure tissue oxygen saturation using NIRS because it is not affected by external factors, such as whether oxygen is inhaled, blood pressure changes, and flap-related variables, such as the size and number of perforating blood vessels. In addition, NIRS can reduce the demand for professional caregivers, thereby reducing medical costs accordingly.

[0127] In this implementation example, an Arduino UNO R3 development board is used as the client host of the blood circulation monitor.

[0128] The integrated components of the development board are as follows:

[0129] 1. ATmega328P chip: This is the main chip of Arduino UNO R3, which controls the entire system and handles input / output (I / O) and communication tasks.

[0130] 2. USB interface: This interface allows Arduino UNO R3 to connect to the computer via a USB cable and act as a serial communication port.

[0131] 3. Power plug: This plug accepts external power (DC 5V-12V) and provides power to Arduino UNO R3.

[0132] 4. Power LED indicator: This indicator lights up to indicate that the external power supply is connected and the Arduino UNO R3 is running.

[0133] 5. TX / RX LED indicator: This light will flash when communicating from the serial port through the TX and RX interfaces.

[0134] 6. Digital I / O Pins: There are 14 digital input / output pins, marked as D0-D13, which are used to connect devices such as sensors and actuators.

[0135] 7. PWM pins: Six of the digital pins (D3, D5, D6, D9, D10, D11) can generate PWM signals to control the brightness of motors or LEDs.

[0136] 8. Power Pins: Arduino UNO R3 has 3 power pins, including 5V, 3.3V and GND.

[0137] 9. Analog Input Pins: There are 6 analog input pins (A0-A5) that can be connected to analog sensors.

[0138] 10. Reset button: This button can reset Arduino UNO R3 to the default state.

[0139] 11. ESP8266 wifi module: used for the development board to connect to WiFi and communicate with the cloud platform.

[0140] In this implementation case, the above-mentioned cloud-based intelligent data analysis includes four machine learning methods for classification (K-nearest neighbor algorithm, support vector machine, random forest and Adaboost). These machine learning algorithms are pre-trained on data collected in advance by the hospital. According to the real-time status of the patient's postoperative flap, each machine learning method divides the flap blood supply status into one of three states (normal, suspicious and abnormal). After that, the cloud will use a voting mechanism (the minority obeys the majority) to determine the current patient status and synchronize the result to the client.

[0141] Example 4

[0142] In this implementation case, the main scheme is the same as that in Example 1, wherein the parameter settings of each algorithm are as follows:

[0143] K nearest neighbor algorithm: distance calculation uses Euclidean distance, K is 10, and no distance weight is set;

[0144] Support vector machine algorithm: The kernel function is Gaussian kernel function, the penalty coefficient of the objective function is 10, and the OvR (oneve rest) classification mode is adopted;

[0145] Random Forest Algorithm: The number of decision trees is 50, and the Gini index is used as the metric, and the size of the feature subset is 15;

[0146] Adaboost algorithm: The basic algorithm model is the neural network MLP, the maximum number of iterations of the base learner is 100, the weight reduction coefficient of each base learner is 0.15, and the square loss function is used.

[0147] All other parameters of each algorithm except the above ones are default parameters.

[0148] In this implementation case, medical staff can set the threshold of flap status indicators according to different patients and flap types on the cloud platform. When the monitored flap status exceeds the threshold, or the analysis result of the cloud is not normal, the cloud will send an alarm message to the medical staff, and the client will also display a warning.

[0149] The preferred specific embodiments of the present invention are described in detail above. It should be understood that a person skilled in the art can make many modifications and changes based on the concept of the present invention without creative work. Therefore, any technical solution that can be obtained by a person skilled in the art through logical analysis, reasoning or limited experiments based on the concept of the present invention on the basis of the prior art should be within the scope of protection determined by the claims.

Claims

1. A flap blood supply monitoring and early warning system based on wearable technology, characterized in that: Specifically, it includes the client's wearable blood flow monitor and the cloud-based real-time flap status monitoring platform for medical staff. The wearable blood flow monitor comprises: a front-end flexible sensor for detecting skin flap status indicators and communicating with a back-end host in real time; a back-end host for sensor data processing and visualization, communicating with the front-end flexible sensor and a cloud-based real-time skin flap status monitoring platform in real time; The cloud-based real-time flap status monitoring platform analyzes client data in real time, uses the machine learning model currently used for classification to divide the flap blood supply status, and obtains the final flap blood supply status classification result by a voting mechanism.

2. According to claim 1, a flap blood supply monitoring and early warning system based on wearable technology is characterized in that: The front-end flexible sensor comprises: Skin flap temperature monitoring sensor, used to detect the surface temperature of the skin flap after surgery; Tissue metabolism monitoring sensors to detect lactate and glucose levels in flap tissue after surgery; Tissue oxygen monitoring sensors for detecting tissue oxygen saturation, oxyhemoglobin levels, and deoxyhemoglobin levels; Data transmission module, used for real-time communication with the back-end host; The desensitized skin-friendly medical silicone membrane is used to attach each sensor to the surface of the flap.

3. According to claim 2, a flap blood supply monitoring and early warning system based on wearable technology is characterized in that: The tissue metabolism monitoring sensor includes an improved glucose sensor for detecting glucose and lactate content in flap tissue; The tissue oxygen content monitoring sensor is composed of a photoelectric detector, a red LED, a green LED and an IR LED.

4. According to claim 1, a flap blood supply monitoring and early warning system based on wearable technology is characterized in that: The cloud-based real-time flap status monitoring platform includes an early warning module: The early warning module sets the threshold of the flap status indicator according to different patients and flap types; when the monitored flap status exceeds the threshold, or the analysis result of the cloud-based real-time flap status monitoring platform is not normal, the cloud-based real-time flap status monitoring platform will send an alarm message to medical staff, and the client will also display a warning.

5. A flap blood supply monitoring and early warning method, characterized in that: The method adopts the flap blood supply monitoring and early warning system as described in any one of claims 1 to 4, and the steps include: The front-end flexible sensor collects blood circulation status indicator data, including flap temperature, tissue metabolism, and tissue oxygen content; Preprocessing and feature extraction of the acquired blood circulation status index data; The extracted features are input into a variety of pre-trained machine learning models for classification to obtain the classification probabilities of various skin flap blood supply states; A voting mechanism is used for the classification probabilities of each machine learning model to obtain the final classification results of the flap blood supply status.

6. A flap blood supply monitoring and early warning method according to claim 5, characterized in that: The tissue metabolism data include: glucose and lactic acid, and the tissue oxygen content data include: tissue oxygen saturation, oxygenated hemoglobin level and deoxygenated hemoglobin level.

7. A flap blood supply monitoring and early warning method according to claim 6, characterized in that: The extracted features include: the change rate and standard deviation of each blood flow status index and the ratio of glucose / lactate to the ratio of oxygenated hemoglobin level and deoxygenated hemoglobin level.

8. A flap blood supply monitoring and early warning method according to claim 7, characterized in that: The blood supply status of the flap includes: Normal state: Skin flap temperature: within the set temperature range; Tissue glucose levels: Similar to the patient's blood glucose levels, within the normal glucose level range; Lactate level in tissues: low, within the normal range of lactate level; Tissue oxygen saturation: Keep within the normal range of tissue oxygen saturation; Tissue oxygenated hemoglobin level: higher than the set tissue oxygenated hemoglobin level threshold; Tissue deoxyhemoglobin level: below the set tissue deoxyhemoglobin level threshold; Suspicious status: Flap temperature: slightly low or high, but within the set temperature range, and the temperature fluctuation is less than the threshold; Glucose levels in tissues: slightly decreased or increased, but not exceeding the upper and lower limits of the normal glucose level range; Tissue lactate levels: slightly elevated but not exceeding the upper limit of the normal lactate level range; Tissue oxygen saturation: slightly decreased or increased, with the change amplitude less than the threshold; Tissue oxygenated hemoglobin level: slightly decreased, the decrease is less than the threshold; Tissue deoxyhemoglobin level: slightly increased, the increase is less than the threshold; Abnormal status: Skin flap temperature: obviously too low or too high, and the temperature is outside the set temperature range; Glucose level in tissues: significant decrease or increase, decrease or increase greater than the threshold; Lactate levels in tissues: significantly increased, with the increase greater than the threshold; Tissue oxygen saturation: significantly decreased, with the decrease greater than the threshold; Tissue oxygenated hemoglobin level: significantly decreased, with the decrease greater than the threshold; Tissue deoxyhemoglobin level: significantly increased, the increase is greater than the threshold.

9. A flap blood supply monitoring and early warning method according to claim 8, characterized in that: The multiple classification machine learning algorithms include: K nearest neighbor algorithm, support vector machine, random forest and Adaboost; The input of the machine learning model is the extracted feature vector, and the output is the category probability of the blood supply status of each flap.

10. The flap blood supply monitoring and early warning method according to claim 5, characterized in that: The method sets the threshold of the flap status indicator according to different patients and flap types. When the monitored flap parameter exceeds the threshold, or the analysis result of the detection shows that the flap is not in a normal state, an early warning message will be sent to the medical staff.

Images