Intelligent anesthesia breathing safety monitoring system and method

Abstract

The invention discloses an intelligent anesthesia breathing safety monitoring system and method, and the system comprises a sensing module which is used for collecting the expiration metabolism data, alveolar structure data, respiratory mechanics data, body temperature data and anesthesia depth data of a target object in a non-invasive manner; the multi-source data fusion and risk assessment module is used for carrying out bidirectional calibration on the expiration metabolism data and the alveolar structure data so as to carry out localized real-time calculation and output a risk assessment result; the closed-loop regulation and control module is used for evaluating a result according to the risk; the data management module is used for storing monitoring data, evaluation results and regulation and control records; according to the system, by integrating multi-dimensional noninvasive sensing modules such as expiration metabolism, alveolar structure, respiratory mechanics, body temperature and anesthesia depth, the system breaks through the limitation that traditional monitoring only pays attention to single or few life parameters, and a panoramic view of the physiological state of a patient is constructed.

Description

Technical Field

[0001] This invention belongs to the field of medical device technology, and in particular relates to an intelligent anesthesia respiratory safety monitoring system and method. Background Technology

[0002] Currently, maintaining respiratory and circulatory stability is the core of anesthesia management, especially for the particularly vulnerable neonate population, during surgical anesthesia. Widely used anesthesia safety monitoring technologies primarily rely on continuous monitoring and threshold alarms of multiple vital signs. These technologies typically include: circulatory function assessment based on ECG, oxygen saturation (SpO2), and non-invasive blood pressure monitoring; ventilation function assessment through monitoring respiratory rate, end-tidal carbon dioxide (EtCO2) waveforms, and partial pressures; and monitoring of inhaled and exhaled anesthetic drug concentrations using anesthetic gas analyzers. Furthermore, some advanced monitoring devices can integrate respiratory mechanics monitoring (such as airway pressure and flow rate) to roughly assess lung compliance and resistance, or use tools such as the bispectral index (BIS) to quantify the depth of anesthesia. These monitoring methods provide anesthesiologists with fundamental information to assess the patient's physiological state, forming a crucial line of defense for modern anesthesia safety. The technical approach essentially involves physically or informationally integrating multiple single-parameter monitoring modules and issuing audiovisual alarms when parameters exceed preset safety ranges, thereby prompting medical staff to intervene manually.

[0003] While existing technologies form the foundation of anesthesia safety, their inherent limitations are particularly pronounced when dealing with complex physiological states such as neonates. These limitations manifest in three main ways, which are the core issues this invention aims to address. First, at the risk perception level, existing technologies largely operate on a "single-dimensional threshold alarm" model, with each parameter judged independently, lacking effective deep fusion of multi-source information and early correlation analysis. For example, a decrease in SpO2 or abnormal EtCO2 is already a consequence of respiratory depression, not a precursor; the system cannot extract risk information from earlier biomarkers such as expiratory metabolite profiles and dynamic changes in alveolar microstructure, resulting in severely delayed warnings and an inability to distinguish the root causes of risk (e.g., metabolic, mechanical, or central). Second, at the decision-making and intervention level, existing systems are completely "open-loop," only providing alarms. All diagnostic and treatment adjustments rely entirely on the anesthesiologist's immediate experience and manual operation. In neonatal anesthesia, where the condition changes rapidly and multiple devices need to be managed simultaneously, this model struggles to guarantee response speed, operational accuracy, and consistency, posing a risk of adverse events due to human fatigue or delayed judgment. Finally, regarding individualized adaptation, the alarm thresholds of existing monitoring parameters are mostly general settings, making it difficult to flexibly adapt to the significant individual physiological differences in newborns due to vast variations in age, weight, and organ function maturity. Furthermore, they cannot achieve dynamic risk prediction based on individual characteristics. Therefore, there is an urgent clinical need for a new type of safety monitoring system capable of multi-dimensional non-invasive sensing, intelligent prospective risk assessment, and integrated with treatment equipment to form a personalized, precise, closed-loop control system. Summary of the Invention

[0004] In order to overcome the above-mentioned defects of the prior art, the present invention provides an intelligent anesthesia respiratory safety monitoring system and method, which solves the problems of delayed early warning, reliance on manual intervention and inability to achieve individualized and accurate risk prediction in the prior art anesthesia monitoring system.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] An intelligent anesthesia respiratory safety monitoring system and method, comprising:

[0007] The sensing module is used to collect expiratory metabolic data, alveolar structural data, respiratory mechanics data, body temperature data, and anesthesia depth data of the target subject in a non-invasive manner.

[0008] The multi-source data fusion and risk assessment module is connected to the sensing module and is used to perform bidirectional calibration of the exhaled metabolic data and alveolar structure data. The fused and calibrated multi-dimensional data and the individual physiological parameters of the target object are input into a preset respiratory depression risk prediction model to calculate and output the risk assessment results in a localized real-time manner.

[0009] The closed-loop control module is connected to the multi-source data fusion and risk assessment module, and is used to generate control instructions and automatically adjust the operating parameters of the linked anesthesia breathing equipment and anesthetic drug infusion equipment based on the risk assessment results.

[0010] The data management module is used to store monitoring data, evaluation results and control records, and provides a visual interface for real-time display;

[0011] The power module is used to supply power to the system.

[0012] Preferably, the sensing module includes:

[0013] The metabolic monitoring unit uses a mass spectrometry sensor integrated into the breathing circuit to collect the concentrations of anesthetic drug metabolic intermediates and endogenous respiratory markers in exhaled breath in real time. The anesthetic drug metabolic intermediates include propofol glucuronide and inorganic fluoride products of inhaled anesthetic drugs, and the endogenous respiratory markers include acetone, isoprene and methyl lactate.

[0014] The alveolar monitoring unit uses a flexible ultrasound probe that is applied to the projection area of ​​the thoracic lung lobes to collect and process dynamic image data to quantify the percentage of alveolar ventilation area and the local alveolar expansion non-uniformity index.

[0015] The respiratory mechanics monitoring unit uses impedance sensors attached to the body surface to collect data on airway resistance and lung compliance.

[0016] The body temperature monitoring unit and the anesthesia depth monitoring unit are used to collect body temperature data and calculate the anesthesia depth index, respectively.

[0017] Preferably, the bidirectional calibration includes:

[0018] A positive calibration process for the effective early warning threshold of metabolite concentration using alveolar ventilation area;

[0019] A reverse calibration process that uses the changing trends of specific metabolic markers to verify the reliability of alveolar data;

[0020] The respiratory depression risk prediction model is a machine learning model trained based on the random forest algorithm, used to predict the probability of metabolic latent respiratory depression occurring within the next 8-12 minutes.

[0021] Preferably, the risk assessment results include three risk levels—low, medium, and high—based on predicted probabilities, and risk types including metabolic inhibition, alveolar damage, and central nervous system depression.

[0022] Preferably, the closed-loop control module executes a differentiated control strategy:

[0023] When the risk type is alveolar injury, the generated instructions prioritize adjusting the positive end-expiratory pressure and peak inspiratory pressure of the anesthesia breathing equipment.

[0024] When the risk type is metabolic inhibition, the generated instructions prioritize adjusting the infusion rate or inhalation concentration of the anesthetic drug infusion device and increasing the inhaled oxygen concentration.

[0025] When the risk type is central nervous system depression, the generated instruction prioritizes reducing the supply of anesthetic drugs.

[0026] When the risk level is medium or high, a local audible and visual alarm is triggered and a warning message is simultaneously sent to the remote monitoring terminal.

[0027] Preferably, the automatic adjustment of operating parameters is a gradual adjustment process: the adjustment range of a single parameter does not exceed 20% of the set base value, and after adjustment, wait for 1-3 data update cycles to assess changes in risk trends.

[0028] Preferably, the data management module includes:

[0029] The local storage unit uses non-volatile memory to continuously store raw data, processing data, and event logs for at least 72 hours, and supports data export.

[0030] The interactive display unit uses a touch screen to provide visual displays in the form of trend curves, digital dashboards, and color-coded risk indicator charts.

[0031] Preferably, the power module includes a rechargeable lithium battery and a DC power adapter interface, supporting uninterrupted power supply switching; the multi-source data fusion and risk assessment module has a built-in low-power edge computing chip for realizing the localized real-time computing.

[0032] Preferably, an intelligent anesthesia respiratory safety monitoring method includes the following steps:

[0033] S1. System initialization and calibration: Input the individual physiological parameters of the target object, deploy and calibrate each non-invasive sensor, and establish a communication connection with the external anesthesia equipment;

[0034] S2. Multi-dimensional data synchronous acquisition: Real-time parallel acquisition of expiratory metabolism, alveolar structure, respiratory mechanics, body temperature and depth of anesthesia data;

[0035] S3. Data Fusion and Risk Calculation: Preprocess and bidirectionally calibrate multi-source data, and input the fused data and individual parameters into the risk prediction model to calculate the real-time risk level and type.

[0036] S4. Intelligent closed-loop control: Automatically generates and executes progressive adjustment instructions for respiratory and anesthetic drug infusion parameters based on risk results, and triggers an alarm when the warning threshold is reached;

[0037] S5. Full Data Recording and Output: Continuously stores all data, events, and operation records, and generates anesthesia safety reports.

[0038] Preferably, the specific method of the bidirectional calibration is as follows: a regression model with alveolar ventilation area as the independent variable and metabolite concentration threshold as the dependent variable is established for positive calibration; at the same time, the slope of change of specific metabolic markers is calculated, and when it exceeds the set range, the alveolar data of the same time period is marked as a state to be verified and reverse calibration is performed; the parameter adjustment range of each progressive adjustment command does not exceed 20% of the base value, and after adjustment, 1-3 data update cycles are waited for reassessment of risk.

[0039] The technical effects and advantages of the intelligent anesthesia respiratory safety monitoring system and method of the present invention are as follows:

[0040] 1. This invention integrates multi-dimensional non-invasive sensing modules, including those for expiratory metabolism, alveolar structure, respiratory mechanics, body temperature, and depth of anesthesia. This system overcomes the limitations of traditional monitoring, which focuses on only a single or a few vital parameters, and constructs a panoramic view of the patient's physiological state. More importantly, through multi-source data fusion and a built-in machine learning risk prediction model, the system can perform early identification and probability prediction of different types of respiratory depression risks, such as metabolic, alveolar damage, and central nervous system depression. This achieves a fundamental shift from "post-event alarm" to "pre-event warning," buying valuable time for clinical intervention.

[0041] 2. The invention's closed-loop control module can automatically execute differentiated control strategies based on the specific risk type and level output by the risk assessment module. This system can intelligently link anesthesia and respiratory equipment with anesthetic drug infusion equipment, prioritizing adjustments to ventilation pressure parameters for alveolar damage risks and prioritizing adjustments to anesthesia depth and oxygenation for metabolic or central nervous system depression risks, achieving precise intervention tailored to specific needs. This automated closed-loop process replaces the traditional model that relies entirely on the experience and judgment of medical staff and manual operation, significantly improving the timeliness, accuracy, and consistency of intervention, and effectively reducing the risks associated with human error and response delays.

[0042] 3. Addressing the challenges of significant individual differences and rapid changes in the condition of newborns, this invention significantly improves the accuracy and robustness of individualized risk assessment by incorporating individual physiological parameters (such as age, weight, and liver and kidney function) into the risk assessment model and employing a two-way calibration mechanism using metabolic and alveolar data. Its progressive control logic ensures smooth and safe treatment adjustments. The overall solution integrates scattered data, complex judgments, and precise operations into a collaborative intelligent system, greatly enhancing the ability to protect the lives of patients, especially newborns, in critical and complex anesthesia scenarios. Attached Figure Description

[0043] Figure 1 This is a flowchart of an intelligent anesthesia respiratory safety monitoring system and method proposed in this invention. Detailed Implementation

[0044] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0045] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include," "contain," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "includes..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0046] refer to Figure 1 This invention provides an intelligent anesthesia respiratory safety monitoring system and method, aiming to solve the problems of delayed early warning, reliance on manual intervention, and inability to provide individualized control in existing technologies. The system includes a sensing module, a multi-source data fusion and risk assessment module, a closed-loop control module, a data management module, and a power supply module. The sensing module simultaneously collects multi-dimensional data on the target's expiratory metabolism, alveolar structure, respiratory mechanics, body temperature, and depth of anesthesia in a non-invasive manner. The multi-source data fusion and risk assessment module performs bidirectional calibration of metabolic and alveolar structure data and performs localized real-time calculations based on individual physiological parameters and machine learning risk prediction models, outputting the risk level and type. The closed-loop control module automatically generates and executes differentiated control instructions based on the risk assessment results, adjusting the operating parameters of the anesthesia respiratory equipment and anesthetic drug infusion equipment in a coordinated manner to achieve progressive and precise intervention. This method fully implements the above process. This invention achieves a leap from passive alarm to active early warning, from manual judgment to intelligent decision-making, and from general monitoring to individualized closed-loop control, significantly improving the level of anesthesia respiratory safety management.

[0047] Example 1

[0048] Early warning and closed-loop regulation of metabolic inhibition risk in neonatal laparoscopic surgery:

[0049] Purpose of implementation:

[0050] This embodiment aims to demonstrate how the system can achieve early identification, accurate judgment, and automatic intervention of metabolic respiratory depression risk through non-invasive multi-dimensional monitoring in typical neonatal surgeries, thereby verifying the overall workflow and core value of the system.

[0051] Implementation System:

[0052] The systems used in the implementation include:

[0053] Sensing module: Used for non-invasive collection of expiratory metabolic data, alveolar structure data, respiratory mechanics data, body temperature data, and anesthesia depth data of children.

[0054] Multi-source data fusion and risk assessment module: Connected to the sensing module, it is used to perform bidirectional calibration of expiratory metabolic data and alveolar structure data, and input the fused and calibrated multi-dimensional data and the child's individual physiological parameters (age in months, weight) into a preset respiratory depression risk prediction model for localized real-time calculation and output of risk assessment results.

[0055] Closed-loop control module: Connected to the risk assessment module, it is used to generate control instructions based on the risk assessment results and automatically adjust the operating parameters of the linked anesthesia breathing equipment and anesthetic drug infusion equipment.

[0056] Data management module and power supply module: used for data storage, display and system power supply.

[0057] Implementation steps:

[0058] The implementation process of this embodiment follows the following steps:

[0059] S1: System Initialization and Calibration: Input the age and weight information of a 15-day-old newborn weighing 3.2kg. Deploy and calibrate the mass spectrometer sensor integrated into the anesthesia breathing circuit, the flexible ultrasound probe applied to the chest, the respiratory impedance sensor, the body temperature sensor, and the BIS sensor. Establish communication connections between the system and the anesthesia machine and the propofol target-controlled infusion pump, and complete parameter calibration.

[0060] S2: Multi-dimensional data synchronous acquisition: After the start of the operation, the system collects the following data in real time and in parallel: concentrations of propofol glucuronide (12 ng / L) and acetone (0.8 ppm) in exhaled air provided by the mass spectrometer every 1.5 respiratory cycles; percentage of alveolar ventilation area in the right lower lobe (65%) and expansion non-uniformity index (0.25) provided by the ultrasound probe; airway resistance (35 cmH2O / L / s) and lung compliance (0.8 mL / cmH2O) provided by the impedance sensor; axillary temperature (36.8℃) provided by the body temperature sensor; and anesthesia depth index (45) provided by the BIS sensor.

[0061] S3: Data Fusion and Risk Calculation: The system performs preprocessing on the above multi-source data, including time alignment and filtering. Then, a two-way calibration is performed: based on the current alveolar ventilation area (65%), the warning threshold for acetone concentration is dynamically corrected from 1.0 ppm to 0.85 ppm. After calibration, all data and the child's physiological parameters are input into the risk prediction model. The model outputs the following results: Risk Level "Intermediate", Risk Type "Metabolic Inhibition".

[0062] S4: Intelligent Closed-Loop Control: Based on the "intermediate, metabolically suppressive" risk level, the system automatically executes control strategies. First, it sends a command to the target-controlled infusion pump to reduce the propofol infusion rate by 20%. Simultaneously, it sends a command to the anesthesia machine to increase the inhaled oxygen concentration by 10%. After the control commands are issued, the system enters monitoring mode, awaiting data updates.

[0063] S5: Full Data Recording and Output: Throughout the entire surgical procedure, the system continuously stores all raw data, intermediate model results, risk warning records, and equipment control commands. Post-operatively, an anesthesia safety report is automatically generated, summarizing monitored and intervention events.

[0064] Implementation results:

[0065] Approximately 3 minutes after the control command was executed, the system monitored a decrease in propofol metabolite concentration, and the BIS value rebounded to 55. The risk level reassessed by the model decreased to "low." This intervention successfully provided an early warning and automatically corrected potential respiratory depression risks approximately 10 minutes before routine indicators such as blood oxygen saturation became abnormal, thus preventing adverse events. The complete timeline recorded by the data management module provided detailed evidence for postoperative review.

[0066] Example 2

[0067] Identification of latent risks based on metabolic biomarker trends:

[0068] Purpose of implementation:

[0069] The system demonstrates how it can utilize the continuous trends in expiratory metabolic markers to specifically identify the risk of latent metabolic disorders caused by factors such as immature liver function, even before alveolar and mechanical parameters have significantly deteriorated.

[0070] Implementation System:

[0071] The same system as described in Example 1.

[0072] Implementation steps:

[0073] S1: Initialize the system for a premature infant with elevated liver enzymes.

[0074] S2: The system continuously monitored the concentration of the endogenous marker methyl lactate during the operation, which gradually increased from 0.5 ppm to 1.2 ppm within 15 minutes.

[0075] S3: At this point, alveolar ventilation area (70%) and lung compliance data are stable, and the positive calibration in the two-way calibration did not trigger threshold correction. However, based on the combined trend of increasing metabolites and the baseline liver and kidney function of the child, the risk prediction model determined that there was a "metabolic suppression type, low risk".

[0076] S4: The system performs preventative micro-adjustments, reducing the anesthetic infusion rate by 5% and displaying a notification to monitor metabolic status.

[0077] S5: Record the risk event and the intervention.

[0078] Implementation results:

[0079] Subsequent blood gas analysis confirmed that the child had mild metabolic acidosis. The system's early warning based on metabolic fingerprinting enabled clinicians to intervene in advance, validating its specificity and sensitivity to metabolic risks.

[0080] Example 3

[0081] Purpose of implementation:

[0082] The effectiveness of the system in monitoring alveolar structure in real time via non-invasive ultrasound and automatically implementing lung-protective ventilation strategies when the risk of alveolar overinflation is identified was validated.

[0083] Implementation System:

[0084] The same system as described in Example 1.

[0085] Implementation steps:

[0086] S1: Deploy the system for a newborn requiring one-lung ventilation, focusing on calibrating the thoracic ultrasound probe array.

[0087] S2-S3: Intraoperatively, the system monitored a rapid increase in the alveolar ventilation area percentage of the ventilated lung to 85%, and the expansion unevenness index worsened to 0.45, indicating a risk of alveolar overinflation. Respiratory compliance decreased simultaneously. The model classified it as "alveolar injury type, high risk".

[0088] S4: The system immediately activates the differentiated control strategy, prioritizing the sending of instructions to the anesthesia machine to gradually reduce positive end-expiratory pressure (PEEP) and peak inspiratory pressure.

[0089] S5: Record changes in alveolar dynamics and regulatory parameters.

[0090] Implementation results:

[0091] After adjustment, ultrasound monitoring showed that the alveolar ventilation area returned to 78%, the non-uniformity index improved, and the high-risk alarm was lifted. The system achieved precise and automatic lung-protective ventilation guided by visualized alveolar structures, potentially reducing the risk of ventilator-associated lung injury.

[0092] Example 4

[0093] Identification of central respiratory depression and adjustment of anesthesia depth:

[0094] Purpose of implementation:

[0095] The system demonstrates how it integrates anesthesia depth data, identifies central respiratory depression caused by excessive anesthesia, and implements targeted control strategies to avoid misintervention.

[0096] Implementation System:

[0097] The same system as described in Example 1.

[0098] Implementation steps:

[0099] S1-S2: During the operation, the system monitored that the BIS value was consistently below 40 (deep anesthesia state), but there were no abnormalities in expiratory metabolic markers and alveolar ultrasound data.

[0100] S3: The model makes a comprehensive judgment, excludes metabolic or lung injury factors, and outputs "central depression type, intermediate risk".

[0101] S4: The system executes the corresponding strategy, prioritizing the sending of instructions to reduce the infusion rate of anesthetic drugs, while only making maintenance adjustments to the ventilator parameters.

[0102] S5: Record changes in anesthesia depth and regulatory response.

[0103] Implementation results:

[0104] After reducing the depth of anesthesia, the child's spontaneous breathing resumed, and the BIS value returned to the target range. This case demonstrates the advantages of multi-dimensional information fusion in the system, which can accurately identify the root cause of risk and thus implement the most appropriate intervention, avoiding inappropriate increases in respiratory support when anesthesia is too deep.

[0105] Example 5

[0106] Comprehensive decision-making and orderly regulation in scenarios with multiple concurrent risks:

[0107] Purpose of implementation:

[0108] The test system demonstrates its ability to process conflicting information from multiple sources, make comprehensive judgments, and execute orderly and safe controls in complex clinical scenarios (such as hypotension and concurrent changes in body temperature).

[0109] Implementation System:

[0110] The same system as described in Example 1.

[0111] Implementation steps:

[0112] S1: Monitor children who experience transient hypotension and hypothermia during surgery.

[0113] S2-S3: The system simultaneously captures: slowed metabolism (increased acetone), partial alveolar collapse (decreased ultrasound ventilation area), and BIS value fluctuations. A two-way calibration model helps to weight different signals. The final model outputs "composite risk (primarily metabolic suppression)" and "high-level risk".

[0114] S4: System activation strategy: First, adjust the anesthetic and oxygen concentrations to address the primary risk (metabolic suppression); then, for alveolar collapse, perform a low-level lung recruitment and set a low PEEP. All adjustments are gradual.

[0115] S5: Complete log of handling concurrency risks.

[0116] Implementation results:

[0117] Through orderly and gradual regulation, the system gradually stabilized the child's respiratory status and successfully handled the complex situation of multiple abnormal indicators, demonstrating its robustness and clinical applicability as an intelligent decision support system.

[0118] Comparative Example 1

[0119] Traditional multi-parameter monitor alarm system

[0120] Purpose of implementation:

[0121] By comparing with traditional monitoring methods, the technological advancements of this invention in terms of timely risk warning, cause specificity, and automated intervention are highlighted.

[0122] Implementation System:

[0123] The current standard operating room configuration is adopted: independent electrocardiogram and pulse oximeter, end-tidal carbon dioxide monitor, and anesthetic gas analyzer, relying on medical staff observation and manual intervention.

[0124] Implementation steps (traditional manual process):

[0125] In a scenario similar to the early metabolic inhibition in Example 1:

[0126] The monitor only displays routine vital signs, and the change in end-tidal carbon dioxide waveform may not be significant.

[0127] When blood oxygen saturation begins to slowly decrease, the monitor triggers a non-specific "low blood oxygen" alarm.

[0128] After hearing the alarm, the anesthesiologist needs to manually observe and investigate the cause (check the airway, ventilator, depth of anesthesia, etc.).

[0129] Based on their experience, doctors manually adjust the ventilator parameters or the dosage of anesthetic drugs.

[0130] Observe whether vital signs improve after intervention.

[0131] Implementation results:

[0132] The entire process suffers from severe early warning delays, typically triggering an alarm only after blood oxygen levels have dropped; the cause of the alarm is unclear, relying on doctors' experience for trial-and-error intervention; and the response speed is slow, taking several minutes from detection to effective treatment, by which time the child may have already experienced a hypoxic phase. Compared to the early warning, root cause identification, and automatic closed-loop control provided by this invention, traditional methods are significantly deficient in terms of safety, accuracy, and timeliness.

[0133] Compared to Examples 1-5 and Comparative Example 1, the intelligent anesthesia respiratory safety monitoring system of the present invention demonstrates a paradigm shift from "passive delayed alarm" to "active early warning and precise closed-loop control" compared to traditional monitoring modes. Its technological advantages are specifically reflected in three fundamental aspects. At the risk perception and identification level, the system constructs a multi-dimensional panoramic view of vital signs through the simultaneous acquisition of data from multiple non-invasive sensing modules (expiratory metabolism, alveolar ultrasound, respiratory mechanics, body temperature, and electroencephalography). Its core lies in the bidirectional calibration of metabolic data and alveolar structural data by the multi-source data fusion and risk assessment module, as well as risk prediction based on machine learning models. This enables the system to capture early, specific risk signals that traditional monitoring cannot detect. As shown in Examples 1 and 2, the system can accurately identify "metabolic inhibition" latent risks 8-12 minutes before a decrease in blood oxygen saturation through subtle trend changes in expiratory metabolic markers (such as acetone and methyl lactate); Example 3 shows that it can directly monitor the dynamic changes in alveolar structure through ultrasound to identify "alveolar damage" risks. In contrast, traditional systems in comparison have a single perception dimension (mainly relying on blood oxygen and end-tidal carbon dioxide), isolated information, and can only trigger non-specific alarms (such as "low blood oxygen") after physiological compensation imbalance, resulting in severely delayed warnings and unclear root causes.

[0134] At the decision-making and intervention level, the advantages of this invention are even more significant. The system has an intelligent closed-loop control module that can automatically execute differentiated and gradual control strategies based on the specific risk type (metabolic inhibition, alveolar damage, central nervous system depression) and level output by the risk assessment module. For example, it prioritizes adjusting the infusion of anesthetic drugs and oxygen concentration for metabolic inhibition risk (Examples 1 and 2), and prioritizes adjusting ventilator pressure parameters for alveolar damage risk (Example 3), achieving precise intervention by "treating the symptoms." This intervention is real-time, automatic, and orderly. As shown in Example 5, even in complex scenarios with multiple concurrent risks, the system can make comprehensive judgments and sequentially execute control instructions. In contrast, the traditional model in the comparison example relies entirely on manual intervention: after receiving a delayed non-specific alarm, the doctor needs to rely on personal experience to conduct time-consuming problem investigation and trial-and-error manual adjustments, resulting in slow response speed, inability to guarantee the accuracy and consistency of intervention, and the child may already be in a dangerous state during the process. In summary, the system of this invention achieves a qualitative leap in the timeliness, accuracy and automation of risk management through an integrated technical path of "non-invasive multi-dimensional perception - data fusion and intelligent prediction - automatic closed-loop control", providing a revolutionary proactive protection solution for the anesthesia and respiratory safety of newborns and critically ill patients.

[0135] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of protection of the claims.

[0136] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An intelligent anesthesia respiratory safety monitoring system and method, characterized in that, include: The sensing module is used to collect expiratory metabolic data, alveolar structural data, respiratory mechanics data, body temperature data, and anesthesia depth data of the target subject in a non-invasive manner. The multi-source data fusion and risk assessment module is connected to the sensing module and is used to perform bidirectional calibration of the exhaled metabolic data and alveolar structure data. The fused and calibrated multi-dimensional data and the individual physiological parameters of the target object are input into a preset respiratory depression risk prediction model to calculate and output the risk assessment results in a localized real-time manner. The closed-loop control module is connected to the multi-source data fusion and risk assessment module, and is used to generate control instructions and automatically adjust the operating parameters of the linked anesthesia breathing equipment and anesthetic drug infusion equipment based on the risk assessment results. The data management module is used to store monitoring data, evaluation results and control records, and provides a visual interface for real-time display; The power module is used to supply power to the system.

2. The intelligent anesthesia respiratory safety monitoring system as described in claim 1, characterized in that, The sensing module includes: The metabolic monitoring unit uses a mass spectrometry sensor integrated into the breathing circuit to collect the concentrations of anesthetic drug metabolic intermediates and endogenous respiratory markers in exhaled breath in real time. The anesthetic drug metabolic intermediates include propofol glucuronide and inorganic fluoride products of inhaled anesthetic drugs, and the endogenous respiratory markers include acetone, isoprene and methyl lactate. The alveolar monitoring unit uses a flexible ultrasound probe that is applied to the projection area of ​​the thoracic lung lobes to collect and process dynamic image data to quantify the percentage of alveolar ventilation area and the local alveolar expansion non-uniformity index. The respiratory mechanics monitoring unit uses impedance sensors attached to the body surface to collect data on airway resistance and lung compliance. The body temperature monitoring unit and the anesthesia depth monitoring unit are used to collect body temperature data and calculate the anesthesia depth index, respectively.

3. The intelligent anesthesia respiratory safety monitoring system as described in claim 1, characterized in that, The bidirectional calibration includes: A positive calibration process for the effective early warning threshold of metabolite concentration using alveolar ventilation area; A reverse calibration process that uses the changing trends of specific metabolic markers to verify the reliability of alveolar data; The respiratory depression risk prediction model is a machine learning model trained based on the random forest algorithm, used to predict the probability of metabolic latent respiratory depression occurring within the next 8-12 minutes.

4. The intelligent anesthesia respiratory safety monitoring system as described in claim 1, characterized in that, The risk assessment results include three risk levels—low, medium, and high—based on predicted probabilities, as well as risk types including metabolic inhibition, alveolar damage, and central nervous system depression.

5. The intelligent anesthesia respiratory safety monitoring system as described in claim 4, characterized in that, The closed-loop control module executes differentiated control strategies: When the risk type is alveolar injury, the generated instructions prioritize adjusting the positive end-expiratory pressure and peak inspiratory pressure of the anesthesia breathing equipment. When the risk type is metabolic inhibition, the generated instructions prioritize adjusting the infusion rate or inhalation concentration of the anesthetic drug infusion device and increasing the inhaled oxygen concentration. When the risk type is central nervous system depression, the generated instruction prioritizes reducing the supply of anesthetic drugs. When the risk level is medium or high, a local audible and visual alarm is triggered and a warning message is simultaneously sent to the remote monitoring terminal.

6. The intelligent anesthesia respiratory safety monitoring system as described in claim 5, characterized in that, The automatic adjustment of operating parameters is a gradual adjustment process: the adjustment range of a single parameter does not exceed 20% of the set base value, and after adjustment, wait for 1-3 data update cycles to assess changes in risk trends.

7. The intelligent anesthesia respiratory safety monitoring system as described in claim 1, characterized in that, The data management module includes: The local storage unit uses non-volatile memory to continuously store raw data, processing data, and event logs for at least 72 hours, and supports data export. The interactive display unit uses a touch screen to provide visual displays in the form of trend curves, digital dashboards, and color-coded risk indicator charts.

8. The intelligent anesthesia respiratory safety monitoring system as described in claim 1, characterized in that, The power module includes a rechargeable lithium battery and a DC power adapter interface, supporting uninterrupted power supply switching; the multi-source data fusion and risk assessment module has a built-in low-power edge computing chip for realizing the localized real-time computing.

9. A method for intelligent monitoring of anesthesia respiratory safety, characterized in that, The system applied to any one of claims 1 to 8 comprises the following steps: S1. System initialization and calibration: Input the individual physiological parameters of the target object, deploy and calibrate each non-invasive sensor, and establish a communication connection with the external anesthesia equipment; S2. Multi-dimensional data synchronous acquisition: Real-time parallel acquisition of expiratory metabolism, alveolar structure, respiratory mechanics, body temperature and depth of anesthesia data; S3. Data Fusion and Risk Calculation: Preprocess and bidirectionally calibrate multi-source data, and input the fused data and individual parameters into the risk prediction model to calculate the real-time risk level and type. S4. Intelligent closed-loop control: Automatically generates and executes progressive adjustment instructions for respiratory and anesthetic drug infusion parameters based on risk results, and triggers an alarm when the warning threshold is reached; S5. Full Data Recording and Output: Continuously stores all data, events, and operation records, and generates anesthesia safety reports.

10. The intelligent anesthesia respiratory safety monitoring method as described in claim 1, characterized in that, The specific method of the bidirectional calibration is as follows: a regression model with alveolar ventilation area as the independent variable and metabolite concentration threshold as the dependent variable is established for positive calibration; at the same time, the slope of change of specific metabolic markers is calculated, and when it exceeds the set range, the alveolar data of the same time period is marked as a state to be verified and reverse calibration is performed; the parameter adjustment range of each progressive adjustment command does not exceed 20% of the base value, and after adjustment, 1-3 data update cycles are waited for reassessment of risk.

Images