A method for controlling laryngeal mask ventilation in outpatient anesthesia

Abstract

The present application relates to the technical field of healthcare informatics, and discloses a laryngeal mask ventilation control system suitable for outpatient anesthesia, which comprises the following steps: before core analysis, high-frequency signals are injected and monitored to confirm the integrity of the original respiratory waveform data stream; only under the premise that the data stream is complete, real-time ventilation waveform trajectories are constructed in the differential phase space, and morphological comparison is carried out with the individualized baseline template; at the same time, when a ventilator parameter adjustment event is monitored, the baseline template is automatically recalibrated, and the present application constructs a collaborative analysis framework with pre-signal source quality inspection and dynamic benchmark adaptive capacity, converts ventilation monitoring from passive response to isolated peak points to active insight into the evolution trajectory of system dynamics, so that the risk of gradual deterioration of ventilation state caused by laryngeal mask micro-leakage or secretion accumulation can be identified in the early stage of clinical treatment, and intervention time is won for anesthesiologists.

Description

Technical Field

[0001] This invention relates to a laryngeal mask ventilation control method suitable for outpatient anesthesia, belonging to the field of healthcare informatics technology. Background Technology

[0002] Currently, especially in outpatient anesthesia clinical practice, real-time monitoring of laryngeal mask ventilation status through monitoring equipment is a fundamental technical means to ensure patient safety. The mainstream approach is to continuously monitor two key physiological indicators: end-tidal carbon dioxide concentration and peak airway pressure. This approach has become a widely accepted technical consensus in the industry due to its intuitiveness and ability to provide immediate alarms for serious ventilation events.

[0003] However, in outpatient surgeries, especially in pediatric dentistry, the procedure is often accompanied by minor, unplanned displacements of the patient's head and neck or dynamic changes in oral secretions. Under the pressure of such high-frequency, low-amplitude disturbances, an inherent limitation of the aforementioned mainstream monitoring methods becomes apparent: in order to obtain stable and clear single-point peak readings, the data processing essentially involves information dimensionality reduction. The complete respiratory waveform captured by the sensor, containing rich dynamic process information, is simplified into a few isolated scalar values. The cost of this processing is that the continuous, weak air leakage caused by the laryngeal mask due to minor displacement, or the airflow disturbance caused by the slow accumulation of secretions in the airway, will not immediately cause significant changes in the end-expiratory carbon dioxide concentration or airway pressure peak in the initial stage. These core morphological information that characterizes the gradual deterioration of the ventilation system's stability are discarded as normal fluctuations during the information dimensionality reduction process.

[0004] To address these issues, an obvious approach is to tighten the thresholds of existing alarm parameters. However, in practice, this approach can lead to a sharp increase in false alarm rates due to normal physiological fluctuations, causing alarm fatigue among clinical medical staff and ultimately reducing the reliability of the entire monitoring system. The root cause lies in the fact that the problem is not due to insufficient sensitivity of the monitoring thresholds, but rather to a defect in the dimension of the monitored indicator itself. It attempts to characterize a dynamic, continuous ventilation process with a static, discrete peak result, creating an inherent contradiction. Specifically, existing technologies suffer from the following shortcomings: 1. The monitoring information is too dimensional, lacking the ability to perceive waveform morphological changes that could reflect minor laryngeal mask movement or early airway obstruction; 2. The alarm logic is based on abrupt threshold triggering, resulting in blind spots in monitoring the gradual deterioration of ventilation status and inherent delays in alarm response. Therefore, the technical problem to be solved by this invention is how to establish a new data analysis method that no longer relies on isolated peak data that has been reduced in dimensionality, but can directly extract and quantify the evolution trajectory of dynamic stability from the inherent correlation of continuous respiratory waveform data streams in real time, thereby transforming the gradual deterioration process of ventilation status from an invisible potential risk into a visualized and early warning objective indicator. Summary of the Invention

[0005] This invention provides a laryngeal mask ventilation control method suitable for outpatient anesthesia. Its main purpose is to solve the problem that existing technologies rely on single-point peak monitoring that simplifies the respiratory waveform, which cannot effectively detect and warn of the progressive deterioration of the laryngeal mask ventilation status.

[0006] To achieve the above objectives, the present invention provides a laryngeal mask ventilation control method suitable for outpatient anesthesia, comprising the following steps:

[0007] Step a: At the acquisition front end of the respiratory waveform data stream, inject a sinusoidal high-frequency signal with a predetermined frequency and amplitude, and monitor the shape of the sinusoidal high-frequency signal in real time. When the shape of the sinusoidal high-frequency signal does not show non-sinusoidal distortion, generate a signal integrity normal mark.

[0008] Step b, if the signal integrity normal marker is present, proceed to steps c through g:

[0009] Step c: Collect the respiratory waveform data stream of the expiratory phase, and perform differential processing on the respiratory waveform data stream to obtain its first derivative data stream, and then construct a real-time ventilation waveform trajectory in the differential phase space to characterize the dynamic process of each respiratory cycle.

[0010] Step d: Based on multiple real-time ventilation waveform trajectories of the same monitored subject within the baseline establishment window after anesthesia induction, establish an individualized baseline template;

[0011] Step e: Calculate the morphological characteristics of the real-time ventilation waveform trajectory compared to the individualized baseline template, and determine whether there is a progressive deterioration in the ventilation status based on the changing trend of the morphological characteristics over multiple consecutive respiratory cycles.

[0012] Step f: Continuously monitor ventilator parameters. When an adjustment event of ventilator parameters is detected, the judgment in step e is automatically paused, and the individualized baseline template is recalibrated based on the newly formed stable respiratory waveform data after adjustment.

[0013] Step g: When it is determined that there is a gradual deterioration in the ventilation status, an early warning signal is output.

[0014] Preferably, the morphological features include shape deviation. The specific steps for calculating the shape deviation are as follows: the real-time ventilation waveform trajectory and the individualized baseline template are geometrically aligned in the differential phase space, and then a dynamic time warping algorithm is used to quantify the morphological differences between the two.

[0015] Preferably, the morphological features also include a trajectory jitter index, which characterizes the degree of non-smoothness of the real-time ventilation waveform trajectory; the trajectory jitter index The calculation is performed by applying the second derivative of the real-time ventilation waveform trajectory. In a single respiratory cycle The integral operation is performed to determine the result, and the operation rules are as follows: .

[0016] Preferably, the method also performs the following steps in parallel to generate a ventilation efficiency coefficient: within each respiratory cycle, integrating the airway pressure waveform and the respiratory flow waveform to obtain an approximate amount of ventilation work; within the same respiratory cycle, integrating the carbon dioxide concentration waveform and the respiratory flow waveform to obtain an approximate amount of gas exchange; calculating the ratio of the approximate amount of gas exchange to the approximate amount of ventilation work to generate a ventilation efficiency coefficient; and outputting an early warning signal when the ventilation efficiency coefficient shows a continuous downward trend over 5 to 10 consecutive respiratory cycles.

[0017] Preferably, the method also performs the following steps in parallel to identify human-machine aggression: low-pass filtering the acquired raw airway pressure waveform to generate a baseline pressure waveform; subtracting the baseline pressure waveform from the raw airway pressure waveform to obtain a physiological high-frequency residual waveform; calculating the root mean square value of the physiological high-frequency residual waveform in each respiratory cycle; and outputting an independent warning signal indicating human-machine aggression when the root mean square value shows a continuous upward trend over 5 to 10 consecutive respiratory cycles.

[0018] Preferably, the determination of non-sinusoidal distortion in step a is achieved by performing cross-correlation calculation between the real-time acquired sinusoidal high-frequency signal waveform and the standard sinusoidal waveform template. When the cross-correlation coefficient is lower than a predetermined threshold, it is determined that non-sinusoidal distortion has occurred.

[0019] Preferably, the step of outputting an early warning signal also includes a decision arbitration mechanism: when a progressive deterioration in ventilation status is first determined, a potential abnormality marker is generated, and a physiological response window lasting 5 to 10 respiratory cycles is established from this marker; within the physiological response window, the patient's tidal volume and respiratory rate are continuously monitored; only when the change in tidal volume or respiratory rate exceeds the corresponding physiological fluctuation threshold within the physiological response window is the potential abnormality finally confirmed as a real abnormality, and the action of outputting an early warning signal is executed.

[0020] Preferably, the warning signal is a graded warning signal, and the grading rules are as follows: when the change trend of morphological features continuously exceeds the first trend threshold, a first-level warning signal is output; when the geometric shape of the real-time ventilation waveform trajectory changes from a closed trajectory to a non-closed trajectory, or when the absolute value of the morphological features exceeds the second safety threshold, a second-level warning signal is output.

[0021] Preferably, the acquisition of the respiratory waveform data stream is performed at a sampling rate of not less than 50Hz; the number of consecutive respiratory cycles is set to 5 to 10 cycles.

[0022] Preferably, after the recalibration in step f is completed, the judgment in step e is resumed only after a delay period; the delay period is set to 2 to 3 complete respiratory cycles after the individualized baseline template has been recalibrated.

[0023] Compared with the prior art, the beneficial effects of the present invention are:

[0024] 1. This invention acquires real-time data streams of carbon dioxide concentration or airway pressure waveforms during the expiratory phase and performs differential processing to obtain their first and second derivatives. Based on the original waveform data and its first derivative data, a ventilation waveform trajectory characterizing the dynamic evolution of the respiratory system in the differential phase space is constructed. This method does not analyze specific numerical points on the waveform in isolation, but transforms the complete dynamic process of each respiratory cycle into a closed trajectory with a specific geometric shape. This allows the dynamic stability of laryngeal mask ventilation, an attribute that is originally difficult to quantify, to be directly mapped to the stability of this trajectory shape. When the laryngeal mask undergoes slight displacement or secretions accumulate in the airway, these gradual changes at the physical level will directly lead to changes in the dynamic characteristics of the ventilation system, which are presented as a continuous shift in the trajectory shape or a continuous increase in the jitter index that can be identified by the algorithm. Thus, the information distortion problem caused by relying on peak threshold monitoring in the prior art no longer constitutes an obstacle requiring preprocessing under the operating mechanism of this invention.

[0025] 2. While constructing and analyzing the aforementioned ventilation waveform trajectory morphology, this invention also performs parallel integration calculations on the airway pressure waveform and respiratory flow waveform to obtain an approximate amount of ventilation work, and integrates the carbon dioxide concentration waveform and respiratory flow waveform to obtain an approximate amount of gas exchange. A ventilation efficiency coefficient is generated by calculating the ratio between these two. This design allows the invention to simultaneously possess two independent analytical dimensions. The morphological changes in the ventilation waveform trajectory reflect structural and smoothness issues in the ventilation process, such as changes in airway resistance or compliance, while the trend changes in the ventilation efficiency coefficient reflect functional and effectiveness issues in the ventilation process, such as mismatch between work and gas exchange caused by minor air leaks. The information from these two dimensions corroborates each other and together constitutes a ventilation status assessment system, avoiding potential misjudgments that may occur when relying solely on a single morphological analysis.

[0026] 3. In processing the airway pressure waveform, this invention first obtains the baseline pressure waveform through low-pass filtering, then subtracts it from the original pressure waveform to obtain a physiological high-frequency residual waveform, and calculates the root mean square value of the residual waveform. This step reconstructs high-frequency noise information, which is traditionally considered interference, into an independent indicator characterizing the patient's autonomous physiological resistance state. Simultaneously, when the ventilation waveform trajectory analysis module captures an instantaneous abnormality in the trajectory and generates a potential abnormality marker, this invention establishes a physiological response window. Within this window, the patient's tidal volume or respiratory rate is continuously monitored. Only when the instantaneous abnormality in the trajectory is accompanied by a change in tidal volume or respiratory rate is it finally confirmed as a real abnormality and an early warning is triggered. This logic of combining high-frequency residual analysis with macroscopic physiological parameter verification allows the system to conduct an internal decision arbitration based on multi-source information regarding the source and clinical significance of the abnormality before triggering an early warning, thus improving the reliability of the early warning results. Attached Figure Description

[0027] Figure 1 This is a schematic diagram of the closed-loop adaptive ventilation monitoring framework of the present invention;

[0028] Figure 2 This is a comparison chart showing the early sensitivity verification of the trajectory jitter index of this invention to progressive changes in airway resistance.

[0029] Figure 3 This is a schematic diagram of the hardware configuration of the ventilation monitoring system of the present invention;

[0030] Figure 4 This is a timing diagram of the adaptive recalibration interaction for adjusting the baseline template response parameters of this invention. Detailed Implementation

[0031] To make the objectives, technical solutions, and advantages of the present invention clearer, the technical solutions of the present invention will be described in detail below. Obviously, the described embodiments are some embodiments of the present invention, but not all embodiments. It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of this application can be combined with each other.

[0032] The present invention provides a laryngeal mask ventilation control method suitable for outpatient anesthesia. Its overall architecture includes a pre-construction signal integrity self-diagnosis module, a core ventilation waveform differential trajectory analysis module, two parallel ventilation status assessment modules, and a post-construction decision arbitration and early warning output module. The pre-construction signal integrity self-diagnosis module is responsible for quality inspection of the acquired raw respiratory waveform data stream and outputting the complete respiratory waveform data stream that passes the inspection to the core ventilation waveform differential trajectory analysis module. This core module performs differential processing on the data stream to construct phase space trajectories, establish individualized baseline templates, and compare morphological features to determine the structural stability of the ventilation status. The two parallel ventilation status assessment modules supplement the data stream analysis from the dimensions of ventilation efficiency and human-machine asynchrony, respectively. Finally, all analysis results are sent to the post-construction decision arbitration and early warning output module, which integrates multi-dimensional information for final judgment and outputs corresponding early warning signals when a real anomaly is confirmed.

[0033] In outpatient surgical environments, the operation of high-frequency electrical equipment such as electrosurgical units often leads to non-physiological high-frequency artifacts coupling into the signal lines of anesthesia monitors. This signal contamination directly distorts the original respiratory waveform, causing misjudgments in subsequent analyses. To address this challenge, the method of this invention is configured to initiate a signal integrity self-diagnosis procedure before performing any core analysis. Specifically, in the analog circuit section at the front end of the respiratory waveform data stream acquisition, a sinusoidal high-frequency signal with a predetermined frequency and amplitude is periodically injected into the monitoring signal via a low-power oscillator. For example, a 100kHz sine wave with an amplitude far below the normal fluctuation range of physiological signals. The data processing unit monitors the morphology of this sinusoidal high-frequency signal in parallel and in real time. By performing cross-correlation calculations between the real-time acquired sinusoidal high-frequency signal waveform segments and a standard sinusoidal waveform template, the integrity of its morphology is quantified. When high-frequency noise sources such as electrosurgical units are involved, they interfere with and distort the injected sine waveform, causing a decrease in its cross-correlation coefficient. Therefore, a threshold for judging the cross-correlation coefficient can be set. This threshold is determined by recording 100 cycles of signal in an interference-free environment and calculating their average cross-correlation coefficient, taking 95% of the average as the threshold. For example, if the average is 0.99, the threshold is set to 0.94. When the real-time calculated cross-correlation coefficient is lower than this threshold, the system generates a signal integrity anomaly flag and temporarily suspends all analysis and judgment steps based on this data stream. Only when the cross-correlation coefficient consistently exceeds this threshold is a signal integrity normal flag generated, and the original respiratory waveform data stream is determined to be a reliable source, thus allowing subsequent analysis modules to intervene. Through this mechanism of active detection and quality verification at the source end, the problem of false alarms in subsequent analysis caused by environmental electrical noise is avoided.

[0034] Traditional ventilation monitoring methods rely on threshold monitoring of single-point scalar values ​​such as end-expiratory carbon dioxide or peak airway pressure. This approach reduces the dimensionality of a continuous waveform containing dynamic information, thus failing to detect the gradual deterioration of ventilation caused by minute laryngeal mask movement or slow accumulation of airway secretions. To address this challenge, the core of this invention lies in performing a differential trajectory feature analysis of the ventilation waveform. First, through a standard data interface, the carbon dioxide concentration waveform during the patient's expiratory phase is captured synchronously in real time at a sampling rate of at least 50Hz. or airway pressure waveform The algorithm first obtains the raw data stream; then, for the expiratory phase data within each respiratory cycle, it calculates the first derivative data stream in real time. or This represents the instantaneous rate of change of gas concentration or pressure; then, in a two-dimensional coordinate system, the original waveform data is used... Using the vertical axis as the ordinate, and its first derivative data... Using the horizontal axis, a real-time ventilation waveform trajectory representing the dynamic process of the respiratory cycle is plotted. Under stable ventilation conditions, this trajectory presents as a closed curve with stable and repeatable shape. This mathematical transformation process maps a dynamic process in the time domain to a geometric shape in a differential phase space, so that the dynamic stability of the ventilation system is transformed into the morphological stability of a geometric object that can be analyzed and quantified.

[0035] Because different monitored subjects have different physiological characteristics, the normal morphology of ventilation waveform trajectories also varies. Using a uniform template for comparison would reduce the accuracy of judgment. Therefore, the method of this invention is configured to establish an individualized baseline template, and its establishment procedure is deterministic. After anesthesia induction, the system automatically enters a baseline establishment window period. During this window period, for example, 30 respiratory cycles are continuously monitored. The system automatically removes the data from the 5 cycles with the greatest morphological differences to eliminate possible interference such as body movement in the initial stage. Then, the real-time ventilation waveform trajectories of the remaining 25 stable respiratory cycles are averaged to generate an individualized baseline template that can represent the current stable ventilation status of the specific monitored subject. This template is stored as a benchmark for subsequent real-time comparisons. Through this method based on the patient's own data... The baseline establishment mechanism adapts to individual differences among patients. Furthermore, during monitoring, manual adjustments to ventilator parameters by anesthesiologists, such as changes in tidal volume or respiratory rate, cause benign changes in the waveform trajectory, which are not necessarily signs of ventilation deterioration. To avoid misjudging such operations as abnormal, the method of this invention continuously monitors ventilator parameters. When a parameter adjustment event is detected, such as a change in tidal volume setting exceeding 5%, the morphological judgment based on the old template is automatically paused. After the parameter adjustment is completed and the new respiratory waveform stabilizes, for example, after 2 to 3 complete respiratory cycles as a delay period, the aforementioned individualized baseline template establishment procedure is automatically re-executed to recalibrate the baseline template. This dynamic baseline adaptive mechanism improves the applicability of the system in real clinical operating environments.

[0036] After establishing the baseline template, the key lies in accurately quantifying the morphological differences between the subsequent real-time trajectory and the template to identify progressive degradation. To this end, the method of this invention employs multi-dimensional morphological features, including shape deviation and trajectory jitter index, for comprehensive judgment. For shape deviation, the calculation steps are as follows: first, the real-time ventilation waveform trajectory and the individualized baseline template are geometrically aligned in differential phase space to eliminate errors caused by translation and rotation; then, a dynamic time warping algorithm is used to quantify the nonlinear morphological differences between the two, yielding a dimensionless deviation value. The smaller this value, the closer the morphology. For the trajectory jitter index, it aims to quantify the non-smoothness of the trajectory, which is usually related to the weak airflow disturbances caused by the accumulation of airway secretions. Its calculation is performed using the second derivative of the real-time ventilation waveform trajectory. In a single respiratory cycle The integral operation is performed to determine the result, and the operation rules are as follows: ,in, This represents the acceleration of the change in gas concentration. A smooth exhalation process has a gradual change in acceleration. The smaller the value, the larger the value. By analyzing the changing trends of shape deviation and trajectory jitter index over multiple consecutive respiratory cycles, such as 5 to 10 cycles, it can be determined that there is a progressive deterioration in ventilation status when any indicator shows a continuous upward trend. This judgment logic based on the continuous evolution trend of geometric shape enables the system to warn of early risks that are ignored by traditional peak monitoring methods.

[0037] In some cases, a small but persistent laryngeal mask air leak may not immediately cause a significant change in the waveform morphology, but it can directly disrupt the match between the work done during ventilation and the actual gas exchange effect, leading to a decrease in ventilation efficiency. To compensate for this potential blind spot of purely morphological analysis, the method of this invention performs another set of analytical procedures in parallel to generate a ventilation efficiency coefficient. Specifically, within each respiratory cycle, the airway pressure waveform is analyzed... With respiratory flow waveform Integrating the product of the two values ​​yields an approximate amount of ventilation work performed by the ventilator during that cycle; simultaneously, within the same respiratory cycle, the carbon dioxide concentration waveform is analyzed. With respiratory flow waveform The product of the two components is integrated to obtain the total amount of carbon dioxide actually expelled from the lungs during that cycle, serving as an approximation of gas exchange. Subsequently, the approximation of gas exchange is used as the numerator, and the approximation of ventilation work is used as the denominator to calculate the ratio, generating a dynamic ventilation efficiency coefficient. This coefficient reflects the gas exchange effect achieved per unit of respiratory work. When this ventilation efficiency coefficient shows a continuous downward trend over 5 to 10 consecutive respiratory cycles, even if the waveform morphology does not change significantly, the system will output a warning signal, indicating that there may be functional ventilation efficiency degradation. This adds a dimension based on energy information conversion efficiency, orthogonal to morphological analysis, to the assessment of ventilation status.

[0038] Furthermore, when a patient's anesthesia depth is insufficient or they are about to awaken, their laryngeal muscles may experience involuntary, slight tension or mild coughing. This patient-ventilator asynchrony superimposed transient, high-frequency, low-amplitude physiological artifacts onto the smooth airway pressure waveform. These artifacts are risk signals that occur earlier than changes in ventilation patterns. To capture this information, the method of this invention also executes a set of procedures in parallel to identify patient-ventilator asynchrony. The steps are as follows: first, the acquired raw airway pressure waveform... A low-pass filter is applied, with its cutoff frequency set according to the normal respiratory rate range, for example, 2Hz, to filter out high-frequency components, thereby obtaining a smooth baseline pressure waveform that represents the ventilator's baseline work. Subsequently, the basic pressure waveform was... From the original airway pressure waveform Subtracting point by point yields a physiological high-frequency residual waveform. The residual waveform mainly contains information generated by human-ventilator asynchrony. Next, the root mean square value of this physiological high-frequency residual waveform in each respiratory cycle is calculated. This root mean square value directly quantifies the intensity of human-ventilator asynchrony in that cycle. When the root mean square value shows a continuous upward trend over 5 to 10 consecutive respiratory cycles, the system outputs a warning signal indicating the state of human-ventilator asynchrony, independent of the ventilation status warning. This mechanism transforms high-frequency residual information, which is traditionally regarded as noise, into a clinical indicator that can be used to assess the depth of anesthesia.

[0039] Considering that occasional physiological fluctuations such as unconscious swallowing by the patient within a single cycle may lead to momentary abnormalities in the trajectory pattern, directly triggering an alarm would reduce the clinical reliability of the system. To address this issue, the warning signal output step of the method of this invention includes a decision arbitration mechanism. The procedure is as follows: when any of the above analysis modules first determines that there is a progressive deterioration or abnormality in the ventilation status, the system does not immediately output an alarm, but first generates an internal potential abnormality marker, and establishes a physiological response window lasting 5 to 10 respiratory cycles starting from this time point; during this window period, the system continuously monitors the patient's tidal volume. and respiratory rate These two physiological parameters are based on the logic that a real and clinically significant ventilation problem will trigger a compensatory response from the body or ventilator within a certain timeframe, namely, adjusting tidal volume or respiratory rate to maintain minute ventilation. An arbitration rule is established accordingly: if the monitored tidal volume within the entire physiological response window... Changes in volume and respiratory rate The changes in these values ​​did not exceed their respective physiological fluctuation thresholds. For example, if the tidal volume change was less than 10% of the baseline and the respiratory rate change was less than 2 breaths per minute, the potential abnormality was determined to be a clinically insignificant, occasional fluctuation, and the marker was automatically removed without triggering any warning. Only when changes were detected within the physiological response window... or Only when the change exceeds a preset physiological fluctuation threshold is the potential abnormality finally confirmed as a real abnormality, and subsequent warning signal output actions are executed; this arbitration mechanism improves the reliability of the final warning result by introducing a time-delay-based secondary verification of physiological parameters; the output warning signal can be graded, for example, when the change trend of morphological features continuously exceeds the first trend threshold, a first-level visual warning is output; while when the geometric shape of the real-time ventilation waveform trajectory changes from a stable closed trajectory to a non-closed trajectory, or the absolute value of the morphological feature exceeds the preset second safety threshold, a second-level audible and visual alarm is output.

[0040] Example 1: This example is a specific operational instance of the general technical solution described in the preceding detailed implementation in a specific clinical scenario. In an outpatient dental surgery in a pediatric department, a 6-year-old child, after receiving general anesthesia, was mechanically ventilated via a laryngeal mask airway. Initially, all routine monitoring indicators, including peak end-tidal carbon dioxide concentration and blood oxygen saturation, were within the normal range. Following the procedures outlined in the detailed implementation, the system automatically collected and established an individualized baseline template representing the child's current ventilation status during the stable ventilation phase after successful anesthesia induction. Approximately 15 minutes into the surgery… Due to intraoral suctioning and passive displacement of the child's head and neck, a subtle, continuous shift occurred in the laryngeal mask airway position. This shift did not yet cause large-scale air leakage, so the peak end-tidal carbon dioxide concentration remained within the normal range. However, at this point, the ventilation waveform differential trajectory analysis module began to capture the gradual changes in the dynamic characteristics of the ventilation system, manifested as real-time ventilation waveform trajectories over multiple respiratory cycles. Compared to the individualized baseline template, the shape deviation value showed a continuous, small upward trend. Simultaneously, due to the increased airflow disturbance caused by the slight shift, the trajectory jitter index increased. The calculation results also began to exceed its baseline stability range.

[0041] In this process, there is an inherent technical contradiction between monitoring sensitivity and false alarm rate. Simply relying on small changes in shape deviation or trajectory jitter index may not be sufficient to constitute an early warning condition due to instantaneous signal fluctuations. The method of this invention provides a solution to this problem through a parallel-running ventilation efficiency coefficient evaluation module. Within the same time window when shape deviation and trajectory jitter index begin to deteriorate, the ventilation efficiency coefficient value calculated by this module also shows a continuous downward trend over multiple respiratory cycles due to mismatch between ventilation work and gas exchange caused by minor air leakage. At this time, the structural stability degradation information output by the core ventilation waveform differential trajectory analysis module and the functional efficiency reduction information output by the ventilation efficiency coefficient module constitute two independent dimensions of information corroboration, enabling the system to determine that the ventilation status is gradually deteriorating. Correspondingly, the system does not immediately trigger a high-level audible and visual alarm, but first activates a decision arbitration mechanism, generates an internal potential abnormality marker, and opens a physiological response window that lasts for 8 respiratory cycles. In the 6th respiratory cycle of this window, the system monitors the child's tidal volume. A decrease exceeding the preset physiological fluctuation threshold occurred. This change in physiological parameters ultimately confirmed the clinical significance of the aforementioned potential abnormality. The system then output a first-level visual warning signal. Upon receiving this warning, the anesthesiologist slightly adjusted the child's head position and confirmed the laryngeal mask airway position to restore a proper fit. After intervention, the shape deviation, trajectory jitter index, and ventilation efficiency coefficient all returned to baseline levels within several respiratory cycles. A potential ventilation risk that could not be detected by conventional monitoring methods was addressed in advance. The operation of this process is not about finding an optimal single monitoring indicator, but rather about constructing an analytical framework composed of multiple orthogonal dimensions of information such as waveform geometry, system dynamics smoothness, and energy conversion efficiency. This transforms ventilation monitoring from a passive response to isolated peak points to an insight into the dynamic evolution trajectory of the system. The original problem of the sensitivity and specificity of monitoring thresholds was transformed into a problem of multi-source information collaborative verification.

[0042] Example 2: To objectively verify the performance of the method of the present invention in identifying progressive deterioration of ventilation, a comparative experiment based on a high-fidelity respiratory simulator was conducted. The purpose of this experiment was to quantify the difference in warning time between the method of the present invention and a conventional method based on the peak threshold of end-tidal carbon dioxide concentration. The experimental platform consisted of a respiratory simulator capable of simulating different lung compliance and airway resistance, a standard anesthesia ventilator, and a data acquisition and processing system. The data acquisition system included carbon dioxide airway pressure and respiratory flow sensors, and the sampling rate of their data output was set to 100Hz. This sampling rate setting was determined under the technical consideration of balancing data fidelity and processing load. A higher sampling rate can be used to calculate the second derivative of the waveform to obtain the trajectory jitter index. It provides more accurate source data, and the 100Hz setting not only meets this requirement, but its data processing capacity is also within the capabilities of current mainstream embedded processors.

[0043] The experiment set up two parallel monitoring groups: a control group, which used conventional monitoring methods and triggered an alarm only when the peak end-tidal carbon dioxide concentration output by the respiratory simulator decreased by more than 15% from the baseline stable value; and an experimental group, which ran the complete ventilation control method of this invention. When any morphological feature or ventilation efficiency coefficient showed a deterioration trend for five consecutive respiratory cycles, and this was confirmed by a decision arbitration mechanism, an early warning was issued. The experiment included two simulation scenarios. Scenario 1 simulated a small leak in the laryngeal mask airway, which linearly created a leak in the breathing circuit from 0% to 5% tidal volume within 60 seconds through a precision needle valve controlled by a stepper motor. Scenario 2 simulated the accumulation of airway secretions, which gradually increased airway resistance and induced airflow turbulence by uniformly injecting glycerol solution into the simulated trachea through a micro-pump.

[0044] In the micro-leakage test of Scenario 1, plotting the warning states of the two methods against time reveals that the control group's warning state remained untriggered for an extended period after the test began, only triggering an alarm at the 52nd second when the leakage reached approximately 4.5% and the peak carbon dioxide concentration decreased by 16%. On the same timeline, the ventilation efficiency coefficient of the experimental group began to show a identifiable linear decline at the 12th second after the test started, directly reflecting the mismatch between ventilation work and gas exchange caused by leakage. This trend persisted until the 18th second, meeting the warning conditions and triggering an alarm, 34 seconds earlier than the control group. In the secretion accumulation test of Scenario 2, the peak carbon dioxide concentration of the control group only decreased by a maximum of 6% until the end of the test, never reaching the 15% alarm threshold, thus no alarm was triggered. Meanwhile, the trajectory jitter index of the experimental group... As a quantitative indicator of airflow turbulence, the index began to rise continuously at 21 seconds after the start of the experiment and triggered an early warning at 28 seconds. Experimental data shows that the ventilation efficiency coefficient in the method of this invention has early sensitivity to minor air leaks, while the trajectory jitter index... It can identify airflow dynamic changes caused by the accumulation of secretions that do not yet significantly affect the peak concentration. Experimental results confirm that the method of the present invention, by analyzing the inherent dynamic characteristics of the respiratory waveform data stream, can provide earlier warnings when the ventilation status gradually deteriorates, compared with conventional methods that rely solely on a single peak threshold, thus allowing for a more sufficient time window for clinical intervention.

[0045] Example 3: This example combines Figures 1 to 4 This describes a laryngeal mask ventilation control method suitable for outpatient anesthesia, such as... Figure 1As shown in the figure, the process begins with the acquisition of the raw respiratory waveform data stream. This data stream first enters the signal integrity self-diagnosis module, which ensures the quality of the signal source by injecting and monitoring high-frequency signals. The data stream that passes the quality check is then sent to the core ventilation waveform differential trajectory analysis module. This module constructs a phase space trajectory and compares its morphology to determine the structural stability of the ventilation state and performs a real-time comparison with an individualized baseline template. When a ventilator parameter adjustment event is detected, the template triggers automatic recalibration. At the same time, the data stream is also sent in parallel to two supplementary evaluation modules: a ventilation efficiency coefficient evaluation module to assess the matching degree between ventilation work and gas exchange, and a human-machine interaction identification module to analyze the high-frequency residuals of the pressure waveform to identify spontaneous breathing resistance. The output results of all three analysis modules are finally converged into the decision arbitration and early warning output module. This module establishes a physiological response window and performs secondary verification by combining parameters such as tidal volume and respiratory rate. After confirming a real abnormality, it outputs a graded early warning signal, which can be divided into a first-level visual early warning and a second-level audible and visual alarm.

[0046] like Figure 2 As shown in the figure, the horizontal axis represents time in seconds, and the left vertical axis represents the trajectory jitter index calculated by the method of this invention. The right vertical axis represents the data monitored by traditional methods. The percentage change in peak value is compared to the relative percentage change as a scene input; the figure shows that as airway resistance, representing the interference source, increases linearly over 60 seconds (represented by the dotted line), the trajectory jitter index used in this invention... The solid line also shows a clear and continuous upward trend, whereas traditional methods rely on... The peak value change, represented by the dashed line, fluctuates only slightly throughout the process and fails to effectively reflect the gradual deterioration of ventilation conditions.

[0047] like Figure 3 As shown, the core is an anesthesia monitor that serves as the core processing node. It integrates embedded analysis software as the core algorithm module, a data acquisition and driving module, a user interface and early warning display module, and built-in hardware, namely a signal injection oscillator, for signal integrity self-diagnosis. The monitor is connected to respiratory and gas sensors responsible for collecting data such as flow, pressure, and carbon dioxide through sensor signal lines. However, this signal line may be subject to environmental electromagnetic interference from sources such as high-frequency electrosurgical units. At the same time, the monitor is connected to the anesthesia ventilator through a parameter monitoring link to obtain the ventilator's set parameters. Its analysis and early warning results can also be transmitted to an optional central monitoring station via a local area network.

[0048] like Figure 4As shown, this process involves the interaction of five logical units: the ventilator, parameter monitor, analysis controller, baseline calibrator, and morphological analysis module. When the ventilator adjusts the tidal volume or respiratory rate, the parameter monitor detects the change in the parameter, for example, if the tidal volume change is greater than 5%, and triggers a parameter adjustment event. Upon receiving the event, the analysis controller immediately pauses morphological assessment and sets a delay period counter, for example, waiting for 2 to 3 respiratory cycles. In the subsequent delay period loop, the analysis controller continuously monitors the stability of the respiratory waveform under the new parameters. Once the waveform stabilizes, the recalibration procedure is initiated. The baseline calibrator then collects new stable waveform data and reconstructs the individualized baseline template. After the new baseline template is ready, the analysis controller resumes morphological assessment and instructs the morphological analysis module to use the updated baseline for subsequent comparisons. If the parameter change is less than 5%, the system continues to execute the normal monitoring process.

[0049] Example 4: To ensure the decision logic of the ventilation control method of this invention possesses engineering reproducibility and environmental adaptability, the key thresholds in its core algorithm are determined through a standardized offline calibration procedure. This example describes the specific calibration process for the physiological fluctuation threshold in the decision arbitration mechanism and the second safety threshold in the secondary warning system. The initial input to this process is a high-frequency respiratory waveform database containing at least 50 similar outpatient anesthesia history cases. Each case in the database covers the complete process from the stabilization period of anesthesia induction to the end of surgery, and the time points where adverse ventilation events occurred have been marked. The tidal volume in the calibration decision arbitration mechanism is... With respiratory rate The specific steps for determining the physiological fluctuation threshold are as follows: First, select respiratory waveform data segments from all cases during the stable ventilation phase in the database, excluding data from 5 minutes before and after the marked adverse events; then, for each selected stable data segment, calculate the tidal volume for each respiratory cycle. With respiratory rate Then, using the respiratory cycle as the unit, calculate the intervals between adjacent cycles. Percentage change and The absolute value of the change was used to construct a physiological fluctuation database containing hundreds of thousands of data points; finally, statistical distribution analysis was performed on the data in the database, and samples were taken from each database. The 95th percentile of the percentage change, and The 95th percentile of the absolute value of the change is used as the physiological fluctuation threshold for this decision-making arbitration mechanism; if calculated... If the 95th percentile of the percentage change is 9.8%, then its threshold is set at 10%. If the 95th percentile of the absolute value of the change is 1.9 times / minute, then the threshold is set to 2 times / minute.

[0050] The specific steps for determining the second safety threshold of morphological characteristics in a Level II early warning system are as follows. This threshold aims to mark a critical point where ventilation status has significantly deteriorated and immediate intervention is required. First, all case data marked as having experienced adverse ventilation events are screened from the database. For each case, a data segment of the 30 respiratory cycles prior to the occurrence of the adverse event is extracted. Then, the ventilation waveform differential trajectory analysis method of this invention is applied to calculate the shape deviation value for each respiratory cycle in this data segment. By analyzing all data prior to the adverse event, a shape deviation value is determined. In more than 85% of cases, this value is within the 10th to 30th respiratory cycles prior to the occurrence of the adverse event. The first stable and continuous breakthrough occurs between the 5th respiratory cycle; this value is then determined as the second safety threshold. If the shape deviation value of 1.82 is found to conform to this statistical law, the threshold is set to 1.8. This calibration method associates the setting of the safety threshold with clinically significant adverse event consequences. By executing the above calibration procedure, the two core judgment thresholds in the method of this invention are set based on historical data statistics. When the system is deployed in a new clinical environment or connected to different models of monitoring equipment, this procedure can be repeated to adapt the key parameters, thereby ensuring the consistency and reliability of the method in different application scenarios.

[0051] Example 5: Before the actual application of the ventilation control method, in order to ensure the accuracy of its judgment logic, a system initialization and baseline adaptation procedure needs to be executed. After the monitoring device is connected to the patient and ventilator, before the formal monitoring begins, the system first enters an environmental noise self-learning phase. In this phase, the signal integrity self-diagnosis module runs continuously for 60 seconds, collecting the injected sinusoidal high-frequency signal waveform through the breathing circuit in the current electromagnetic environment. Based on the data within this 60-second period, the system automatically calculates the average shape of the sinusoidal high-frequency signal in this specific environment and sets it as the standard sinusoidal waveform template for subsequent cross-correlation calculations. At the same time, the system also calculates the statistical standard deviation of the cross-correlation coefficient in this phase and sets the mean minus 3 times the standard deviation as the dynamic predetermined threshold for subsequent judgment of non-sinusoidal distortion.

[0052] Next, the system initializes and defines the logic for trend judgment. For the step mentioned in the specific implementation, which judges based on the changing trend of morphological features over multiple consecutive respiratory cycles, the trend is determined by performing linear regression analysis on the data point sequence of the most recent 5 to 10 respiratory cycles. For a valid progressive deterioration trend to be considered valid, two conditions must be met simultaneously: first, the absolute value of the slope of the regression line is greater than a preset minimum rate of change threshold to exclude clinically insignificant random fluctuations; second, the coefficient of determination of the regression line... A value greater than 0.75 is required to confirm that the data changes have a clear linear direction rather than discrete noise. Only when both conditions are met simultaneously will the system confirm the formation of a valid trend and initiate subsequent early warning or decision arbitration logic.

[0053] Example 6: Before activating the core monitoring logic, to ensure the adaptability and reliability of the method in specific clinical environments, the system automatically executes a self-check and initialization procedure that includes baseline quality verification and online signal fault detection. In the specific implementation, after the individualized baseline template is generated, the procedure first initiates the baseline quality verification step. The system does not immediately put the generated template into use, but continues to collect data for the next 10 respiratory cycles and compares the real-time ventilation waveform trajectory of these 10 cycles with the newly generated baseline template, calculating their respective shape deviations. Only when the statistical variance of these 10 shape deviation values ​​is less than a preset initial stability threshold is the baseline template confirmed as valid and officially activated. If the variance is greater than the threshold, it indicates that the patient's initial ventilation state has not yet stabilized, and the system will discard the generated template and restart the baseline establishment window after a delay.

[0054] Throughout the monitoring process, the system continuously performs online signal fault detection to address situations such as sensor failure or severe signal distortion. For the low-pass filter used to identify human-machine confrontation, its cutoff frequency is not a fixed value, but rather determined by the airway pressure waveform during the established individualized baseline template. The system performs a Fast Fourier Transform (FFT) to analyze the energy spectrum distribution and automatically sets the cutoff frequency above the highest frequency point containing 99% of the signal energy to achieve adaptation to individual patient respiratory characteristics. Simultaneously, the system continuously performs physiological range validity checks on the input respiratory waveform data stream; if the expiratory phase carbon dioxide concentration waveform is detected... If the value remains at zero for more than 20 consecutive seconds, or if its peak value exceeds the physiological upper limit of 100 mmHg for 5 consecutive respiratory cycles, the system will determine that the sensor or signal acquisition link has failed. At this time, the system will suspend all analysis and early warning based on differential trajectory and ventilation efficiency, and output an independent technical alarm indicating that the signal source is abnormal.

[0055] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0056] Finally, 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 laryngeal mask airway control system suitable for use in outpatient anaesthesia, characterised in that, The system comprises a signal integrity self-diagnosis module, a ventilation waveform differential trajectory analysis module, and a decision arbitration and early warning output module. Before the acquisition of the respiratory waveform data stream, a sinusoidal high-frequency signal with a predetermined frequency and amplitude is injected, and the shape of the sinusoidal high-frequency signal is monitored in real time. Only when the signal integrity normal marker exists, the expiratory phase respiratory waveform data stream is collected, and the respiratory waveform data stream is differentiated to obtain its first derivative data stream, and then a real-time ventilation waveform trajectory representing the dynamic process of each breathing cycle is constructed in the differential phase space. Based on multiple real-time ventilation waveform trajectories of the same monitored subject within the baseline establishment window after anesthesia induction, an individualized baseline template is established. The morphological features of the real-time ventilation waveform trajectory compared with the individualized baseline template are calculated, and based on the change trend of the morphological features in the continuous multiple breathing cycles, it is judged whether the ventilation state exists progressive deterioration. When it is judged that the ventilation state exists progressive deterioration, an early warning signal is output. The step of outputting the early warning signal also includes a decision arbitration mechanism: when it is first judged that the ventilation state exists progressive deterioration, a potential abnormality marker is generated, and a physiological response window of 5 to 10 continuous breathing cycles is established based on the potential abnormality marker. In the physiological response window, the tidal volume and the respiratory rate of the patient are continuously monitored; only when the change amount of the tidal volume or the change amount of the respiratory rate exceeds the corresponding physiological fluctuation threshold in the physiological response window, the potential abnormality is finally confirmed as a real abnormality, and the action of outputting the early warning signal is executed. Among them, tidal volume in the calibration decision arbitration mechanism With respiratory rate The specific steps for determining the physiological fluctuation threshold are as follows: First, select respiratory waveform data segments from all cases during the stable ventilation phase in the database, excluding data from 5 minutes before and after the marked adverse events; then, for each selected stable data segment, calculate the tidal volume for each respiratory cycle. With respiratory rate Then, using the respiratory cycle as the unit, calculate the intervals between adjacent cycles. Percentage change and The absolute value of the change was used to construct a physiological fluctuation database containing hundreds of thousands of data points; finally, statistical distribution analysis was performed on the data in the database, and samples were taken from each database. The 95th percentile of the percentage change, and The 95th percentile of the absolute value of the change is used as the physiological fluctuation threshold for this decision-making arbitration mechanism; And, an effective progressive deterioration trend is determined to be true if two conditions are met simultaneously, one, the absolute value of the slope of the regression line is greater than a preset minimum change rate threshold to exclude random fluctuations of little clinical significance; and two, the determination coefficient of the regression line is greater than 0.75 to confirm that the change in data has a clear linear directionality rather than discrete noise.

2. A laryngeal mask airway ventilation control system suitable for use in outpatient anaesthesia according to claim 1 wherein, The morphological features include shape deviation, and the calculation step of the shape deviation is specifically: the real-time ventilation waveform trajectory and the individualized baseline template are geometrically aligned in the differential phase space, and then the dynamic time warping algorithm is used to quantify the morphological difference between the two.

3. A laryngeal mask airway ventilation control system suitable for use in outpatient anaesthesia according to claim 1 wherein, The morphological feature further includes a trajectory jitter index, the trajectory jitter index representing a non-smooth degree of the real-time ventilation waveform trajectory; the trajectory jitter index is calculated by integrating a second derivative of the real-time ventilation waveform trajectory in a single respiratory cycle , and the operation rule is: .

4. The laryngeal mask airway ventilation control system for outpatient anesthesia of claim 1, wherein, The system also performs the following steps in parallel to generate the ventilation efficiency coefficient: in each breathing cycle, the airway pressure waveform and the respiratory flow waveform are integrated to obtain the approximate amount of ventilation work; in the same breathing cycle, the carbon dioxide concentration waveform and the respiratory flow waveform are integrated to obtain the approximate amount of gas exchange; the approximate amount of gas exchange and the approximate amount of ventilation work are calculated by ratio to generate the ventilation efficiency coefficient. When the ventilation efficiency coefficient presents a continuous downward trend in the continuous 5 to 10 breathing cycles, the early warning signal is also output.

5. The laryngeal mask airway ventilation control system for outpatient anesthesia of claim 1, wherein, The system also performs the following steps in parallel to identify the human-machine confrontation state: the original airway pressure waveform is low-pass filtered to generate a basic pressure waveform; the basic pressure waveform is subtracted from the original airway pressure waveform to obtain a physiological high-frequency residual waveform. The root mean square value of the physiologically high frequency residual waveform is calculated in each respiratory cycle; when the root mean square value shows a continuous upward trend in 5 to 10 consecutive respiratory cycles, an independent early warning signal indicating the state of man-machine confrontation is output.

6. The laryngeal mask airway ventilation control system for outpatient anesthesia of claim 1, wherein, For the judgment of non-sinusoidal distortion, the real-time collected sinusoidal high frequency signal waveform is cross-correlated with the standard sinusoidal waveform template; when the cross-correlation coefficient is lower than the predetermined threshold, it is determined that non-sinusoidal distortion occurs.

7. The laryngeal mask airway ventilation control system for outpatient anesthesia of claim 1, wherein, The early warning signal is a graded early warning signal, and the grading rule is: when the trend of the morphological feature continuously exceeds the first trend threshold, a first-level early warning signal is output; when the geometric shape of the real-time ventilation waveform trajectory changes from a closed trajectory to a non-closed trajectory, or the absolute value of the morphological feature exceeds the second safety threshold, a second-level early warning signal is output.

8. The laryngeal mask airway ventilation control system for outpatient anesthesia of claim 1, wherein, The acquisition of the respiratory waveform data stream is performed at a sampling rate of no less than 50Hz; the number of consecutive respiratory cycles is set to 5 to 10 cycles.

9. The laryngeal mask airway ventilation control system for outpatient anesthesia of claim 1, wherein, After the recalibration is completed, the judgment of whether the ventilation state is gradually deteriorating is resumed after a delay period; the delay period is set to 2 to 3 complete respiratory cycles after the individualized baseline template completes the recalibration.

Images