Ventilator airway mucus dynamic analysis and ciliary motility visualization assay device

Abstract

The application discloses a device for dynamic analysis of airway mucus and visualization measurement of cilia activity for a breathing machine, which comprises a breathing machine and a tracheal intubation device for acquiring time series data of patient's vital signs; the tracheal intubation device comprises a tracheal intubation catheter, one end of the tracheal intubation catheter is a CLE microscope probe outlet end, a CLE microscope probe inlet end is opened on the wall of the tracheal intubation catheter and communicates with the inside of the tracheal intubation catheter, a CLE microscope probe guide wire channel is reserved in the tracheal intubation catheter for passing the CLE microscope probe, and a fiber image in the airway is acquired through the CLE microscope probe; a laser ranging sensor for acquiring the chest fluctuation of the patient and a control module for executing a monitoring method according to the fiber image, the time series data of the patient's vital signs and the chest fluctuation are arranged on the breathing machine. The application first provides a method for determining mucus generation through dynamic changes of mucus and cilia activity, and the early warning accuracy is high.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of airway feature analysis, and relates to a dynamic analysis and visualization determination device for airway mucus and cilia activity of a breathing machine. BACKGROUND

[0002] The mucus-cilia clearance system is the first line of defense against various inhaled pathogenic particles in the respiratory tract. When the respiratory function of a patient is severely damaged or cannot guarantee independent breathing, artificial tracheal intubation is usually required. During the use of the breathing machine intubation, the mucus characteristics of the airway need to be monitored at all times. When the accumulated mucus is more, negative pressure cleaning, commonly known as "sputum suction", needs to be taken in time to avoid the risk of suffocation caused by high-concentration mucus emboli.

[0003] Mucus production needs to last a long process, and in many cases, the patient's own conditions differ. The regular cleaning process cannot completely adapt to the patient's mucus production process, so real-time monitoring is needed during intubation. The prior art usually only infers the airway state through indirect sign parameters (heart rate, breathing depth, pressure, etc.), but ignores the more intuitive cilia activity and dynamic changes of mucus. Relying only on sign parameters can make it difficult to discover mucus accumulation or cilia damage in time, thereby reducing the accuracy of early warning and even delaying treatment in severe cases.

[0004] Therefore, it is of great practical significance to develop a method capable of determining mucus production through dynamic changes of mucus and cilia activity. SUMMARY

[0005] Since the prior art has the above defects, the present application provides a method capable of determining mucus production through dynamic changes of mucus and cilia activity, specifically a dynamic analysis and visualization determination device for airway mucus and cilia activity of a breathing machine, which can reliably warn of mucus production and overcome the defects that only inferring the airway state through indirect sign parameters (heart rate, breathing depth, pressure, etc.) but ignoring the more intuitive cilia activity and dynamic changes of mucus, relying only on sign parameters can make it difficult to discover mucus accumulation or cilia damage in time, thereby reducing the accuracy of early warning and even delaying treatment in severe cases.

[0006] In order to achieve the above purpose, the present application provides the following technical solutions:

[0007] The dynamic analysis and visualization determination device for airway mucus and cilia activity of a breathing machine comprises a breathing machine and a tracheal intubation device.

[0008] The tracheal intubation device comprises a tracheal intubation catheter, one end of the tracheal intubation catheter is provided with a tracheal intubation connector, the other end is sleeved with a cuff, and the end of the tracheal intubation catheter is provided with a CLE microscope probe outlet end; the tracheal intubation catheter is provided with a CLE microscope probe inlet end which is in communication with the inside of the tracheal intubation catheter; the cuff is connected with an inflation tube in the tracheal intubation catheter, and the inflation tube is arranged to pass through the middle of the tracheal intubation catheter; the inflation tube is sequentially connected with an indicator balloon and a one-way valve away from the cuff; the tracheal intubation catheter is provided with a CLE microscope probe guide wire channel for passing the CLE microscope probe; and the fiber image in the airway can be obtained through the CLE microscope probe.

[0009] The ventilator is used to obtain the time sequence data of the patient's physical parameters, the physical parameters including heart rate and breathing depth; the ventilator is provided with a laser ranging sensor and a control module; the laser ranging sensor is used to obtain the chest fluctuation amount of the patient; the control module is connected with the laser ranging sensor and the CLE microscope probe signal; the control module is used to execute a monitoring method for monitoring the airway mucus accumulation and cilia activity of the patient according to the fiber image, the time sequence data of the patient's physical parameters and the chest fluctuation amount; the monitoring method comprises:

[0010] During the tracheal intubation of the patient, the continuous fiber image in the airway of the patient, the time sequence data of the physical parameters and the chest fluctuation amount are obtained;

[0011] The time sequence data of the physical parameters is automatically divided into period data segments corresponding to each physical parameter according to the types of the physical parameters; under each physical parameter, the similarity between the positions of the time points of any two fiber images in the period data segments and the consistency of the changes between the period data segments are analyzed to obtain the state similarity value between the two fiber images; the state similarity values between the fiber images under two physical parameters are adjusted and fused based on the numerical characteristics of the chest fluctuation amount, which are used for clustering analysis of the fiber images and screening to obtain a target cluster;

[0012] In the target cluster, the similarity degree of the morphological change characteristics of the fiber images over time is analyzed, and is combined with the number characteristics of the fiber images in the target cluster to determine an abnormal index; the abnormal index is used for early warning.

[0013] The airway mucus dynamic analysis and cilia activity visualized measuring device for a respirator provided by the present application solves the problem of inaccurate monitoring and early warning of mucus generation by combining software and hardware, and airway mucus accumulation and cilia dysfunction are dynamic processes, the continuous fiber images in the airway are obtained by the CLE microscope probe, and the time sequence data of the sign parameters and the chest fluctuation volume are collected, the fiber images provide airway information, and the sign parameters and the chest fluctuation volume reflect the physiological state, because the heartbeat and respiration of the patient change periodically, the distance between the airway surface and the microscope probe is a dynamic process, and the probe needs to maintain the same distance from the airway tube surface for imaging, and the fiber images in the same physiological state are comparable, so the fiber images need to be clustered according to the change characteristics of the sign parameters, the fiber image time is mapped to the corresponding physiological state stage by period division of the time sequence data of the sign parameters, the chest fluctuation volume is used to analyze and fuse the time position similarity between the fiber images under the sign parameters in the period data segment and the change consistency of the corresponding period data segment, for clustering analysis of the fiber images, and the target cluster is selected, the fiber images in the target cluster have more consistent physiological characteristics and are more consistent in influence, and have higher comparability, under normal circumstances, mucus gradually moves towards the throat, which is a dynamic process, therefore, in the target cluster, the similarity degree of the shape change characteristics of the fiber images over time is analyzed, and the abnormal index is determined by combining the number characteristics of the fiber images, for representing the accumulation degree of mucus and the cilia activity, the cilia activity and the dynamic change of mucus are effectively combined with the sign parameters, and early warning based on the abnormal index can greatly improve the early warning accuracy.

[0014] As a preferred technical solution:

[0015] The airway mucus dynamic analysis and cilia activity visualized measuring device as described above, the method for obtaining the state similarity value comprises:

[0016] Under each sign parameter, the similarity degree between the time positions of the time points of any two fiber images in the period data segment is analyzed in time sequence, to obtain the first state similarity factor of the two fiber images;

[0017] Under each sign parameter, the change consistency between the period data segments to which the time points of any two fiber images belong is analyzed, to obtain the second state similarity factor of the two fiber images;

[0018] The product of the first state similarity factor and the second state similarity factor between any two fiber images is normalized as a state similarity value between the two fiber images under each sign parameter.

[0019] The first state similarity factor acquisition method of the ventilator airway mucus dynamic analysis and cilia activity visualization measurement device includes:

[0020] For any fiber image, the period data section to which the time of the fiber image belongs in the sign parameter time sequence data is taken as a target data section corresponding to the fiber image under each sign parameter.

[0021] The length of the target data section is taken as a period length, and the ratio of the difference between the time of the fiber image and the starting time of the target data section to the period length is taken as a relative position feature value of the fiber image in the target data section.

[0022] The absolute value of the difference between the relative position feature values of any two fiber images is negatively correlated and normalized as a first state similarity factor between the two fiber images under each sign parameter.

[0023] The second state similarity factor acquisition method of the ventilator airway mucus dynamic analysis and cilia activity visualization measurement device includes:

[0024] The DTW value of the target data sections corresponding to any two fiber images is negatively correlated and normalized as a second state similarity factor between the two fiber images under each sign parameter.

[0025] The target cluster acquisition method of the ventilator airway mucus dynamic analysis and cilia activity visualization measurement device includes:

[0026] The chest fluctuation amount includes a respiratory fluctuation and an apnea fluctuation, the respiratory fluctuation is the maximum distance of the chest fluctuation caused by respiration during the tracheal intubation of the patient, and the apnea fluctuation is the maximum distance of the chest fluctuation during apnea.

[0027] The ratio of the apnea fluctuation to the respiratory fluctuation is taken as a heartbeat influence ratio.

[0028] For any two fiber images, the state similarity values of the two fiber images under two sign parameters are adjusted and fused by using the heartbeat influence ratio, so as to determine a comprehensive state similarity index between the two fiber images.

[0029] In all fiber images, each fiber image is sequentially taken as a clustering center in time sequence, and in all the remaining fiber images, the fiber images with a comprehensive state similarity index greater than a preset similarity threshold are classified into a class with the clustering center, thereby obtaining a corresponding clustering cluster when each fiber image is taken as a clustering center.

[0030] In all clustering clusters, a clustering cluster with a number of fiber images greater than a preset number threshold is taken as a target clustering cluster.

[0031] The airway mucus dynamic analysis and cilia activity visualization measuring device for respirator as described above, the method for obtaining the comprehensive state similarity index comprises:

[0032] For any two fiber images, the product of the heartbeat influence proportion and the state similarity value of the two fiber images at the heart rate is taken as a similarity parameter, and the normalized value of the sum of the similarity parameter and the state similarity value of the two fiber images at the depth of respiration is taken as the comprehensive state similarity index of the two fiber images.

[0033] The airway mucus dynamic analysis and cilia activity visualization measuring device for respirator as described above, the method for obtaining the abnormal index comprises:

[0034] In the target clustering cluster, the similarity degree of the morphological change characteristics of the fiber images over time is analyzed to determine an abnormal trend degree value.

[0035] In all target clustering clusters, the normalized value of the number of fiber images in each target clustering cluster is taken as the number weight of each target clustering cluster, the product of the number weight of each target clustering cluster and the abnormal trend degree value is taken as a weighted abnormal factor, and the normalized value of the mean value of the weighted abnormal factors of all target clustering clusters is taken as the abnormal index.

[0036] The airway mucus dynamic analysis and cilia activity visualization measuring device for respirator as described above, the method for obtaining the abnormal trend degree value comprises:

[0037] In each target clustering cluster, Otsu threshold segmentation is performed on each fiber image, a region with a small gray value is taken as a foreground region, and the number of pixel points in the foreground region is taken as a mucus performance characteristic value in each fiber image.

[0038] In each target clustering cluster, all fiber images are arranged in time sequence to obtain a sorting sequence, in the sorting sequence, the ratio of the absolute value of the difference of the mucus performance characteristic values to the absolute value of the time difference between each adjacent two fiber images is calculated to obtain a unit difference performance value, and all unit difference performance values between the fiber images are arranged in order in the sorting sequence to obtain a change sequence.

[0039] In the change sequence, a first-order difference sequence of the unit difference performance value is calculated, the mean value of the absolute values of all values in the first-order difference sequence is negatively correlated and normalized, and the value after the normalization is taken as a first abnormal trend factor, and the product of the number of negative values in the first-order difference sequence and the first abnormal trend factor is normalized, and the value after the normalization is taken as the abnormal trend degree value.

[0040] The airway mucus dynamic analysis and cilia activity visualization measuring device for a respirator as described above, the early warning based on the abnormal index, comprising:

[0041] When the abnormal index is greater than a preset abnormal threshold, a mucus accumulation abnormality early warning is performed, otherwise, no early warning is performed.

[0042] The airway mucus dynamic analysis and cilia activity visualization measuring device for a respirator as described above, the periodic data segment acquisition method comprising:

[0043] The periodic division of the time series data of the vital signs parameters is performed based on the LSTM neural network one by one according to the types of the vital signs parameters, so as to obtain all the periodic data segments corresponding to all the types of the vital signs parameters.

[0044] The above technical solutions are only one possible technical solution of the present application, and the protection scope of the present application is not limited to this, and the person skilled in the art can reasonably adjust the specific design according to the actual needs.

[0045] The above-mentioned application has the following advantages or beneficial effects:

[0046] (1) The airway mucus dynamic analysis and cilia activity visualization measuring device for a respirator of the present application solves the problem of inaccurate monitoring and early warning of the mucus generation process by using a combination of software and hardware.

[0047] (2) The airway mucus dynamic analysis and cilia activity visualization measuring device for a respirator of the present application provides a method for determining mucus generation through the dynamic changes of mucus and cilia activity for the first time in the field, which has high early warning accuracy and good reliability.

[0048] (3) The airway mucus dynamic analysis and cilia activity visualization measuring device for a respirator of the present application can obtain the accumulation degree of sputum in real time, intuitively reflect the abnormality of mucus and cilia activity, reduce the cost of high-frequency periodic manual screening, and has good application prospect. BRIEF DESCRIPTION OF DRAWINGS

[0049] The present application, its features, shapes and advantages will become more apparent from the following detailed description of non-limiting embodiments, made with reference to the accompanying drawings. In all the drawings, the same references indicate the same parts. The drawings are not necessarily drawn to scale, the emphasis being on illustrating the principles of the present application.

[0050] Figure 1 Structure diagram of the tracheal intubation device in the present application;

[0051] Figure 2 Structure diagram of the control module in the present application;

[0052] Figure 3 Method flow chart of the monitoring method involved in the present application;

[0053] Figure 4 Fiber image collected in the present application;

[0054] Figure 5 Method flow chart of the state similarity value acquisition method involved in the present application;

[0055] Figure 6 Method flow chart of the target cluster acquisition method involved in the present application;

[0056] Figure 7 Image diagram of the image after threshold segmentation of the fiber image in the present application;

[0057] Wherein, 1 - tracheal intubation connector, 2 - tracheal intubation catheter, 3 - one-way valve, 4 - indicating balloon, 5 - CLE microscope probe inlet end, 6 - inflation tube, 7 - cuff, 8 - CLE microscope probe outlet end, 200 - processor, 201 - memory, 202 - bus, 203 - communication interface. DETAILED DESCRIPTION

[0058] The present application will be further described below in conjunction with the accompanying drawings and specific embodiments, but not as a limitation of the present application.

[0059] Example 1

[0060] A dynamic analysis and ciliary activity visualization measuring device for airway mucus of a breathing machine, comprising a breathing machine and a tracheal intubation device;

[0061] The tracheal intubation device is as follows Figure 1As shown, including tracheal intubation connector 1; tracheal intubation catheter 2; one-way valve 3; indicating balloon 4; CLE microscope probe inlet end 5; inflation tube 6; cuff 7; CLE microscope probe outlet end 8; wherein the tracheal intubation connector 1 connects the breathing machine and the tracheal intubation catheter 2, ensuring seamless docking of the gas passage, the tracheal intubation catheter 2 is the main channel, used to transport gas from the connector to the patient's trachea, and also provides operating space for the suction tube; one-way valve 3 controls the direction of gas flow; indicating balloon 4 directly shows the pressure in cuff 7, assisting medical staff in determining whether the inflation amount is appropriate; inflation tube 6 connects indicating balloon 4 and tracheal intubation catheter 2; cuff 7 fixes the position of tracheal intubation catheter 2, while closing the gap; in the present application, a CLE microscope guide wire channel (not shown in the figure, specifically, the inner mirror is not less than 1mm, and the standard tracheal intubation tube diameter is not less than 8mm) is reserved in the tracheal intubation catheter 2, used for passing through the CLE microscope probe, there is a gap between the reserved CLE microscope guide wire channel and the air guide channel in the tracheal intubation catheter 2, and they do not directly contact, and both are sterile environments, the CLE microscope probe enters through the CLE microscope probe inlet end 5, passes through the reserved CLE microscope guide wire channel, and protrudes out of the CLE microscope probe outlet end 8, the CLE microscope probe can obtain the fiber image in the airway;

[0062] When the patient needs tracheal mucus dynamic monitoring and analysis, the tracheal intubation device is inserted into the human body through the ordinary tracheal nose or mouth, and then the CLE microscope probe is input to the proximal end of the tracheal tube through the CLE microscope guide wire channel, and the CLE microscope probe inlet end 5 is locked to avoid bacterial infection and gas leakage, the CLE microscope probe irradiates the tissue surface by laser beam (usually 488nm wavelength fluorescence), needs to use high-speed imaging mode, receives and detects fluorescence, and obtains the fiber image in the patient's airway; the breathing machine can obtain the time sequence data of the patient's sign parameters, and the breathing machine is also provided with a laser ranging sensor and a control module, the laser ranging sensor is used to obtain the chest fluctuation amount of the patient, and the control module can realize the monitoring method for monitoring the airway mucus accumulation and cilia activity of the patient according to the above-mentioned various data.

[0063] The control module is, for example Figure 2As shown, at least includes a memory and a processor, including a processor 200, memory 201, bus 202 and communication interface 203, the processor 200, communication interface 203 and memory 201 are connected through bus 202;Wherein, the memory 201 can include a high-speed random access memory, the bus 202 can be ISA bus, PCI bus or EISA bus, etc., the processor 200 can be an integrated circuit chip, with signal processing capacity;Memory 201 stores at least one instruction, at least one program, code set or instruction set, at least one instruction, at least one program, code set or instruction set is loaded and executed by the processor to realize the step in the monitoring method for monitoring the airway mucus accumulation and the cilia activity of the patient.

[0064] The monitoring method comprises the following steps: Figure 3 As shown in the figure, the method comprises the following steps:

[0065] Step S1: in the process of tracheal intubation of the patient, the continuous fiber image, the sign parameter time series data and the chest fluctuation amount in the airway of the patient are acquired, wherein the sign parameter includes heart rate and breathing depth.

[0066] In the process of tracheal intubation of the patient, when the trachea is intubated to the appropriate position, the continuous fiber image of the tracheal surface is acquired by the CLE microscope probe, and under the irradiation condition of the spontaneous laser of the CLE microscope probe, because the cells will produce cell metabolites (NADH, FAD, etc.), and there are collagen and elastin, and these are the main substances of reflected fluorescence, so on the fiber image, it presents high light reflection;Because the respiratory fibers of the muscle are usually transverse, longitudinal or reticular structure, and the cell gap is lack of cell composition, so the image presents “reticular” or “wavy” performance, Figure 4 That is, the schematic diagram of the fiber image collected by the present application.

[0067] At the same time, the sign parameter time series data and the chest fluctuation amount in the process of tracheal intubation of the patient are acquired by the ventilator, the type of sign parameter includes heart rate and breathing depth, and the chest fluctuation amount includes respiratory fluctuation degree (the maximum distance amount of chest fluctuation caused by respiration in the process of tracheal intubation of the patient) and apnea fluctuation degree (the maximum distance amount of chest fluctuation when apnea, such as at the end of expiration and at the end of inspiration).

[0068] It should be noted that the collection period of the above various data needs to be the same, and the length is set to 3 minutes;The acquisition frequency of fiber image and sign parameter time series data should be consistent, and the acquisition frequency should not be less than 30Hz, and the present application is set to 50Hz.

[0069] It should be noted that the collection and acquisition of personal information data in the present application are all authorized by relevant users, and the process does not violate relevant laws and regulations and does not violate public order and good customs.

[0070] Step S2: automatically dividing the time sequence data of the vital sign parameters one by one according to the types of the vital sign parameters to obtain period data segments; under each vital sign parameter, analyzing the similarity between the positions of the time points of any two fiber images in the period data segments and the consistency of changes between the period data segments in which the time points are located to obtain the state similarity value between the two fiber images under each vital sign parameter; and adjusting and fusing the state similarity values between the fiber images under two vital sign parameters based on the numerical characteristics of the chest fluctuation to cluster the fiber images and screen the target cluster.

[0071] Since the heartbeat and respiration of the human body will produce certain physical movements, such as the ventricular contraction and diastole of the heartbeat which will produce certain vibration effects on the trachea, and the respiration will also move the trachea along with the movement of the diaphragm and the expansion of the lungs, and in some scenarios, such as ICU, general anesthesia surgery, etc., the patient is in a coma state during tracheal intubation, or in long-term bedridden patients with respiratory failure, the patient's own influence on the trachea is generally limited, so the main factors affecting the change of the distance between the trachea and the probe are respiration and heartbeat. Therefore, due to the influence of the patient's heartbeat and respiration, the distance between the trachea surface and the probe is a dynamic change process, and the CLE microscope probe needs to maintain the same distance from the trachea surface for imaging, and only then can the irregular shadows obtained have comparability. Therefore, the fiber images generated at different times in the continuous fiber images cannot be directly compared, so it is necessary to cluster the fiber images based on the change characteristics of the patient's vital sign parameters, so that the fiber images at different times in the same cluster can be regarded as imaging results of the probe and the trachea surface at the same distance, thereby realizing dynamic evaluation of mucus and evaluation of cilia activity.

[0072] The physiological activities (respiration, heartbeat) of the human body have natural periodicity, and if the continuous time sequence data is directly analyzed, the pattern differences within the period may be ignored, and different stages have different effects on the chest pressure, which also leads to different morphological differences of the fiber images. Therefore, first, the vital sign parameter time sequence data is automatically divided into period data segments according to the types of the vital sign parameters.

[0073] Specifically, the method for obtaining the period data segments comprises:

[0074] The period division extracts the dynamic characteristics of the vital sign parameter time sequence data, otherwise the time sequence data of different periods may be confused in the subsequent process, further causing a decrease in comparability.

[0075] The time series data of the physical parameters is divided into cycles one by one according to the types of the physical parameters based on the LSTM neural network, so that all cycle data segments corresponding to all physical parameters are obtained, that is, the time series data of the physical parameters is taken as the input of the LSTM neural network one by one according to the types of the physical parameters, and the node data of the divided cycle is output, and then the time series data of the physical parameters can be disconnected based on the node data, so that all cycle data segments are obtained.

[0076] It should be noted that the training process of the LSTM neural network is a known technology, and the specific process will not be repeated here.

[0077] The fiber images generated at different times cannot be directly compared because the influence of different physiological states on the trachea is different, but under the same breathing and heartbeat state, it can be considered that the influence on the trachea is the same, and in general, the same breathing and heartbeat state means the same position in the breathing and heartbeat cycle, and the changes in the physical parameters should also be consistent, so under each physical parameter, the similarity between the positions of the time of any two fiber images in the cycle data segment is analyzed, and the change consistency between the cycle data segments where the time is located, to obtain the state similarity value between the two fiber images under each physical parameter, which is used to represent the comparability between the two fiber images.

[0078] Specifically, the method for obtaining the state similarity value is as shown in Figure 5 , which includes the following steps:

[0079] Step S501: Under each physical parameter, the similarity between the positions of the time of any two fiber images in the cycle data segment is analyzed in time series, to obtain a first state similarity factor of the two fiber images.

[0080] Under each physical parameter, for any fiber image, the cycle data segment to which the time of the fiber image belongs in the physical parameter time series data is taken as the target data segment corresponding to the fiber image. For example, the time of a fiber image is 5, and the time 5 belongs to the second cycle data segment in the heart rate time series data, so the second cycle data segment in the heart rate time series data is the target data segment corresponding to the fiber image under the heart rate physical parameter.

[0081] The length of the target data segment is taken as the cycle length, the difference between the time of the fiber image and the starting time of the target data segment is calculated, and then the ratio of the difference to the cycle length is taken as the relative position feature value of the fiber image in the target data segment. The relative position feature value eliminates the influence of possible differences in cycle length and reflects the relative position of the time of the fiber image in the target data segment.

[0082] Finally, under each sign parameter, the absolute value of the difference between the relative position feature values of any two fiber images is calculated. The smaller the absolute value of the difference, the more similar the relative positions of the two fiber images in the respective target data segments, and thus the more likely they are in the same physiological state stage. Therefore, the absolute value of the difference is negatively correlated and normalized to correct the logical relationship and obtain the first state similarity factor of the two fiber images under each sign parameter. The larger the first state similarity factor, the more similar the stages of the two fiber images, and thus the more similar the effects of the trachea under the sign parameter, and the higher the comparability of the fiber images. The negative correlation and normalization processing can be performed using the formula exp(-x), where exp() represents the exponential function with the natural constant e as the base, and x represents the independent variable.

[0083] Step S502: Under each sign parameter, the degree of change consistency between the time periods to which any two fiber images belong is analyzed to obtain the second state similarity factor of the two fiber images.

[0084] Only based on the comparison of the positions of the time points in the period in step S501, the trend of the data values in the period data segment is ignored, and DTW can measure the similarity in shape by nonlinearly aligning two time series. Therefore, under each sign parameter, for any two fiber images, the DTW value of the target data segments corresponding to the two fiber images is calculated. The smaller the DTW value, the more similar the dynamic change patterns of the target data segments corresponding to the two fiber images. Therefore, the DTW value is negatively correlated and normalized to correct the logical relationship and obtain the second state similarity factor of the two fiber images under each sign parameter. The larger the second state similarity factor, the more similar the stages of the two fiber images, and thus the more similar the effects of the trachea under the sign parameter, and the higher the comparability of the fiber images. The negative correlation and normalization processing can be performed using the formula exp(-x), where exp() represents the exponential function with the natural constant e as the base, and x represents the independent variable.

[0085] It should be noted that the process of obtaining the DTW value is a known technology, and the specific process is not described here.

[0086] Step S503: Under each sign parameter, the first state similarity factor and the second state similarity factor between any two fiber images are combined to obtain the state similarity value between the two fiber images.

[0087] Based on the logic in the foregoing step S501 and step S502, it is known that, under each sign parameter, the first state similarity factor and the second state similarity factor between the fiber images are positively correlated with the state similarity degree between the fiber images, so in the present application, under each sign parameter, the value of the product of the first state similarity factor and the second state similarity factor between any two fiber images after normalization is taken as the state similarity value between the two fiber images under each sign parameter, and the greater the state similarity value, the higher the comparability of the two fiber images and the closer the physiological state of the trachea.

[0088] Based on the foregoing steps, the state similarity value of any two fiber images under each sign parameter can be obtained. Since the amplitude of the chest movement caused by respiration is generally significantly greater than that caused by heartbeat, the greater the expansion distance of the chest, the higher the degree of influence on the trachea, and since the expansion of the chest can be quantified by the chest fluctuation amount, the chest fluctuation amount includes the respiratory fluctuation (the maximum distance of the chest fluctuation caused by respiration during the tracheal intubation of the patient) and the apnea fluctuation (the maximum distance of the chest fluctuation during apnea); therefore, in the present application, the state similarity values between the fiber images under two sign parameters are adjusted and fused based on the numerical characteristics of the chest fluctuation amount, an index capable of comprehensively representing the similarity degree of the fiber images is obtained, and the fiber images are subjected to cluster analysis based on the index to obtain a target cluster, so that the degree of influence change of the trachea in the fiber images in the target cluster remains consistent.

[0089] Specifically, the method for obtaining the target cluster is as shown in Figure 6 and includes the following steps:

[0090] Step S601: determining the heartbeat influence ratio based on the numerical characteristics of the respiratory fluctuation and the apnea fluctuation.

[0091] During apnea (such as at the end of expiration and at the end of inspiration), the chest displacement is mainly caused by heartbeat (such as chest wall vibration caused by heartbeats), so the respiratory fluctuation corresponding to the respiration process reflects the dominant effect of normal respiration on the airway state, and the apnea fluctuation reflects the independent influence of heartbeat. The ratio of the two can quantify the influence ratio of heartbeat on the features in the fiber image, so the ratio of the apnea fluctuation to the respiratory fluctuation is taken as the heartbeat influence ratio, and the value of the heartbeat influence ratio is theoretically always less than 1, and the closer to 1, the greater the influence degree of the apnea fluctuation.

[0092] Step S602: For any two fiber images, the state similarity values of the two fiber images under the two vital sign parameters are adjusted and fused by using the heartbeat influence proportion, so as to determine the comprehensive state similarity index between the two fiber images.

[0093] For any two fiber images, the product of the heartbeat influence proportion and the state similarity value of the two fiber images under the heart rate is taken as the similarity parameter, that is, the heartbeat influence proportion is taken as the influence weight of the heart rate on the fiber image. In the present application, the influence weight of the respiration on the fiber image is set to 1, so that the value of the sum of the normalized values of the similarity parameter and the state similarity value of the two fiber images under the depth of respiration is taken as the comprehensive state similarity index of the two fiber images. At this time, the comprehensive state similarity index fuses the state characteristics of the vital sign parameters between the fiber images and the influence degree of the vital sign parameters on the fiber image, so that the comprehensive similarity degree between the two fiber images can be more comprehensively reflected. The greater the value is, the more similar the comprehensive state is, and the higher the comparability of the fiber images is. The normalization is a technical means familiar to those skilled in the art, and the selection of the normalization function can be linear normalization or standard normalization, and the specific normalization method is not limited herein.

[0094] Step S603: In all fiber images, clustering analysis is performed on all fiber images based on the comprehensive state similarity index between the fiber images, to obtain clustering clusters.

[0095] In all fiber images, each fiber image is sequentially taken as a clustering center in time sequence, and in the remaining all fiber images, the fiber images with a comprehensive state similarity index greater than a preset similarity threshold value are classified into a class with the clustering center, so as to obtain the corresponding clustering cluster when each fiber image is taken as the clustering center. At this time, each fiber image corresponds to a clustering cluster, and the fiber images in the clustering cluster have relatively consistent characteristics.

[0096] It should be noted that in the present application, the preset similarity threshold value is set to 0.8, and the specific value can be adjusted according to the implementation scene, which is not limited herein.

[0097] Step S604: In all clustering clusters, a target clustering cluster is selected based on the number characteristics of the fiber images in the clustering cluster.

[0098] In all clustering clusters, the clustering cluster with a number of fiber images greater than a preset number threshold value is taken as the target clustering cluster.

[0099] It should be noted that in the present application, the preset number threshold value is one fourth of the total number of fiber images, and the specific value can be adjusted according to the implementation scene, which is not limited herein.

[0100] Step S3: In the target cluster, the similarity of the morphological change characteristics of the fiber images over time is analyzed, and combined with the number characteristics of the fiber images in the target cluster, an abnormality index is determined; and an early warning is made based on the abnormality index.

[0101] The above steps obtain the fiber images at different times after clustering. Under the same cluster, the changes in the influence on the bronchus can be regarded as the same, so the differences between the microscope probe and the bronchus surface are relatively close, and thus the state of mucus and cilia can be directly determined by analyzing the differences between the fiber images.

[0102] Since the cilia movement presents high-frequency periodicity under normal circumstances, the mucus is continuously transmitted from the deep part of the trachea to the throat. The mucus has a small gray value relative to the fiber composed of cells under the action of fluorescence because it lacks related proteins, and the mucus mostly has the characteristics of blockiness and irregular shape. Therefore, when similar shadows appear in the fiber images, it is highly probable that the mucus has passed through. Moreover, in the process of the mucus moving toward the throat along the cilia, the shadows should be less similar over time, which indicates that the mucus is gradually discharged to the throat and belongs to the normal situation. Conversely, if the suspected shadow of the mucus is similar over a long period of time, it indicates that the vitality of the cilia is abnormal, and the accumulation of mucus is high. Therefore, in the present application, the similarity of the morphological change characteristics of the fiber images over time is analyzed in the target cluster, and the abnormality index is determined by combining the number characteristics of the fiber images in the target cluster, to reflect the accumulation of mucus in the trachea and the vitality of the cilia.

[0103] Specifically, the method for obtaining the abnormality index comprises:

[0104] In the target cluster, the similarity of the morphological change characteristics of the fiber images over time is analyzed, and the abnormality trend degree value is determined:

[0105] Since the mucus region lacks fluorescent proteins, the gray value is significantly lower than the fiber tissue composed of cells, so in the target cluster, Otsu threshold segmentation is performed on each fiber image, the region with a small gray value is regarded as a foreground region, and the number of pixel points in the foreground region is obtained as a mucus performance characteristic value in each fiber image. Otsu threshold segmentation is a known technology, and the specific process is not described here. Figure 7 The image obtained after threshold segmentation of the fiber image in the present application.

[0106] Since the changes of mucus and cilia belong to dynamic change process, in each target clustering cluster, all fiber images are arranged in time sequence to obtain a sorting sequence, in the sorting sequence, the ratio of the absolute value of the difference of mucus performance characteristic value and the absolute value of time difference between each adjacent two fiber images is calculated to obtain a unit difference performance value, the unit difference performance value reflects the change characteristics of mucus performance between adjacent two fiber images in unit time, the greater the value, the more obvious the change characteristics, which implies greater cilia activity.

[0107] Under normal cilia movement, mucus should move to the throat with time, thereby causing rapid change of mucus distribution between adjacent images, therefore, the unit difference performance values between all fiber images are arranged in order in the sorting sequence to obtain a change sequence; and in the change sequence, a first-order difference sequence of the unit difference performance values is calculated, the absolute value of the value in the first-order difference sequence can reflect whether the change of mucus in the fiber image is similar, the smaller the value, the more similar the change, and then abnormality may occur, mucus is prone to accumulate, therefore, the value obtained by negatively correlating and normalizing the mean value of the absolute values of all values in the first-order difference sequence is taken as a first abnormal trend factor. The negatively correlating and normalizing processing here can adopt the formula exp(-x), wherein exp() represents an exponential function with natural constant e as the base, and x represents an independent variable.

[0108] And the positive and negative of the value in the first-order difference sequence can also reflect the change of cilia activity in the fiber image at adjacent time, a negative value indicates weakened cilia activity, abnormality occurs, and may cause accumulation of mucus, therefore, the value obtained by normalizing the product of the number of negative values in the first-order difference sequence and the first abnormal trend factor is taken as an abnormal trend degree value, based on the foregoing analysis, the greater the abnormal trend degree value, the poorer the physiological condition of the current patient reflected by the fiber images in the target clustering cluster, and the higher the possibility of mucus accumulation and cilia activity reduction. Normalization is a technology familiar to those skilled in the art, and the selection of the normalization function can be linear normalization or standard normalization, and the specific normalization method is not limited here.

[0109] Finally, in all target clustering clusters, the normalized value of the number of fiber images in each target clustering cluster is taken as the quantity weight of each target clustering cluster. The greater the quantity weight, the more the number of fiber images in the target clustering cluster, and the greater the reference. Therefore, the product of the quantity weight of each target clustering cluster and the abnormal trend degree value is taken as a weighted abnormal factor, and the normalized value of the mean of the weighted abnormal factors of all target clustering clusters is taken as an abnormal index. The greater the abnormal index, the greater the degree of mucus accumulation and the greater the abnormal degree of cilia activity in the current tracheal intubation process of the patient. The normalization is a technology known to those skilled in the art, and the selection of the normalization function can be linear normalization or standard normalization, and the specific normalization method is not limited herein.

[0110] After obtaining the abnormal index of the patient in the tracheal intubation process, the index can be used for abnormal early warning.

[0111] Specifically, the early warning based on the abnormal index includes:

[0112] When the abnormal index is greater than a preset abnormal threshold, mucus accumulation abnormal early warning is needed; otherwise, no early warning is needed.

[0113] It should be noted that in the present application, the preset abnormal threshold is 0.5, and the specific value can be adjusted according to the implementation scene, which is not limited herein.

[0114] In summary, the airway mucus accumulation and cilia dysfunction are dynamic processes, so the continuous fiber images in the airway are obtained by the CLE microscope probe, and the time series data of the sign parameters and the chest fluctuation volume are collected, the fiber images provide the airway information, and the sign parameters and the chest fluctuation volume reflect the physiological state. Because the patient's heartbeat and breathing change periodically, the distance between the airway surface and the microscope probe is a dynamic process, and the probe needs to maintain the same distance from the airway tube surface for imaging, and the fiber images in the same physiological state are comparable, so the fiber images need to be clustered according to the change characteristics of the patient's sign parameters: the time of the fiber images is mapped to the corresponding physiological state stage by period division of the time series data of the sign parameters, and because the heartbeat and breathing have different effects on the expansion of the chest, the chest fluctuation volume is used to analyze and fuse the time position similarity between the fiber images under the sign parameters in the period data segment and the change consistency of the corresponding period data segment, for clustering analysis of the fiber images, and the target cluster is selected, at this time, the fiber images in the target cluster have more consistent physiological characteristics and are more consistent in the impact, and the comparability is higher. Under normal circumstances, mucus gradually moves cilia towards the throat, which is a dynamic process, so finally, in the target cluster, the similarity degree of the shape change characteristics of the fiber images over time is analyzed, and the number characteristics of the fiber images are combined to determine the abnormal index, which is used to represent the accumulation degree of mucus and the activity of cilia, and effectively combines the ciliary activity and the dynamic changes of mucus with the sign parameters, so the early warning based on the abnormal index can greatly improve the early warning accuracy.

[0115] Those skilled in the art should understand that those skilled in the art can make changes in combination with the prior art and the above embodiments, which are not described here. Such changes do not affect the essential content of the present application, and are not described here.

[0116] The preferred embodiments of the present application are described above. It should be understood that the present application is not limited to the above specific embodiments, and the devices and structures not fully described should be understood as implemented in the ordinary way in the art; any person skilled in the art can make many possible changes and modifications to the technical solutions of the present application, or modify them as equivalent embodiments with equivalent changes, without departing from the scope of the technical solutions of the present application, which does not affect the essential content of the present application. Therefore, any simple modification, equivalent change and modification of the above embodiments made according to the technical essence of the present application, without departing from the technical solutions of the present application, all still belong to the scope of protection of the technical solutions of the present application.

Claims

1. A device for visualizing and measuring airway mucus dynamics and ciliary activity in a ventilator, characterized by, The application relates to a ventilator and a tracheal intubation device. The tracheal intubation device comprises a tracheal intubation catheter, one end of the tracheal intubation catheter is provided with a tracheal intubation connector, the other end is sleeved with a cuff, and the end of the tracheal intubation catheter is provided with a CLE microscope probe outlet end; the tracheal intubation catheter is provided with a CLE microscope probe inlet end which is communicated with the inside of the tracheal intubation catheter; the cuff is connected with an inflation tube in the tracheal intubation catheter, and the inflation tube is arranged to pass through the middle of the tracheal intubation catheter; the inflation tube is sequentially connected with an indicating balloon and a one-way valve at the end away from the cuff; the tracheal intubation catheter is provided with a CLE microscope probe guide wire channel reserved for the CLE microscope probe; and the fiber image in the airway can be obtained through the CLE microscope probe. The ventilator is used for obtaining time sequence data of patient's physical parameters, the physical parameters include heart rate and breathing depth, a laser ranging sensor and a control module are arranged on the ventilator, the laser ranging sensor is used for obtaining the chest fluctuation amount of the patient, the control module is connected with the laser ranging sensor and the CLE microscope probe signal, the control module is used for executing a monitoring method for monitoring the airway mucus accumulation and cilia activity of the patient according to the fiber image, the time sequence data of the patient's physical parameters and the chest fluctuation amount, and the monitoring method comprises the following steps: During the tracheal intubation of the patient, the continuous fiber image in the airway of the patient, the time sequence data of the physical parameters and the chest fluctuation amount are obtained; The time sequence data of the physical parameters is automatically divided into period data segments according to the types of the physical parameters; under each type of the physical parameters, the similarity between the time positions of any two fiber images in the period data segments and the consistency of the changes between the period data segments are analyzed to obtain the state similarity value between the two fiber images; the state similarity values between the fiber images under two types of the physical parameters are adjusted and fused based on the numerical characteristics of the chest fluctuation amount, which are used for clustering analysis of the fiber images and screening of a target cluster. In the target cluster, the similarity degree of the morphological change characteristics of the fiber images over time is analyzed, and is combined with the quantity characteristics of the fiber images in the target cluster to determine an abnormal index; the abnormal index is used for early warning.

2. The apparatus for airway mucus dynamic analysis and ciliary motility visualization measurement for a breathing machine according to claim 1, characterized in that, The state similarity value is obtained by the following method: Under each type of the physical parameters, the similarity between the time positions of any two fiber images in the period data segments is analyzed to obtain the first state similarity factor of the two fiber images; Under each type of the physical parameters, the consistency of the changes between the period data segments of any two fiber images is analyzed to obtain the second state similarity factor of the two fiber images; Under each type of the physical parameters, the product of the first state similarity factor and the second state similarity factor between any two fiber images is normalized to obtain the state similarity value between the two fiber images.

3. The apparatus for airway mucus dynamic analysis and ciliary motility visualization measurement for a breathing machine according to claim 2, wherein, The first state similarity factor is obtained by the following method: For any fiber image, a period data segment to which a time point of the fiber image belongs in the body sign parameter time sequence data is taken as a target data segment corresponding to the fiber image; A difference value between the time point of the fiber image and a starting time point of the target data segment is divided by a length of the target data segment, and a ratio is taken as a relative position feature value of the fiber image in the target data segment; For any two fiber images, an absolute value of a difference between relative position feature values of the two fiber images is negatively correlated and normalized, and a value after the negative correlation and the normalization is taken as a first state similarity factor of the two fiber images under each body sign parameter.

4. The apparatus for airway mucus dynamic analysis and ciliary motility visualization measurement for a breathing machine according to claim 3, wherein, The second state similarity factor includes: For any two fiber images, a DTW value of target data segments corresponding to the two fiber images is negatively correlated and normalized, and a value after the negative correlation and the normalization is taken as a second state similarity factor of the two fiber images under each body sign parameter.

5. The apparatus for airway mucus dynamic analysis and ciliary motility visualization measurement for a breathing machine according to claim 1, wherein, The target clustering cluster includes: The chest fluctuation includes a respiratory fluctuation and an apnea fluctuation, the respiratory fluctuation is a maximum distance of chest fluctuation caused by respiration during a tracheal intubation process of the patient, and the apnea fluctuation is a maximum distance of chest fluctuation during apnea; A ratio of the apnea fluctuation to the respiratory fluctuation is taken as a heartbeat influence ratio; For any two fiber images, the heartbeat influence ratio is used to adjust and fuse state similarity values of the two fiber images under two body sign parameters, so as to determine a comprehensive state similarity index between the two fiber images; In all fiber images, each fiber image is sequentially taken as a clustering center in time sequence, and fiber images with a comprehensive state similarity index greater than a preset similarity threshold value are classified into a same class with the clustering center in all remaining fiber images, so as to obtain a clustering cluster corresponding to the fiber image as the clustering center; In all clustering clusters, a clustering cluster with a number of fiber images greater than a preset number threshold value is taken as a target clustering cluster.

6. The apparatus for airway mucus dynamic analysis and ciliary motility visualization measurement for a breathing machine according to claim 5, wherein, The comprehensive state similarity index includes: For any two fiber images, a product of the heartbeat influence ratio and a state similarity value of the two fiber images under a heart rate is taken as a similarity parameter, and a normalized value of a sum of the similarity parameter and a state similarity value of the two fiber images under a respiratory depth is taken as a comprehensive state similarity index of the two fiber images.

7. The apparatus for airway mucus dynamic analysis and visualization of ciliary motility determination of claim 1, wherein, The abnormal index includes: In the target clustering cluster, a similarity degree of a morphological change feature of the fiber images over time is analyzed, and an abnormal trend degree value is determined; In all target clustering clusters, a number of fiber images in each target clustering cluster is normalized, and a value after the normalization is taken as a number weight of each target clustering cluster, a product of the number weight of each target clustering cluster and the abnormal trend degree value is taken as a weighted abnormal factor, and a mean value of the weighted abnormal factors of all target clustering clusters is normalized, and a value after the normalization is taken as the abnormal index.

8. The apparatus for airway mucus dynamic analysis and ciliary motility visualization measurement for a breathing machine according to claim 7, characterized in that, The abnormal trend degree value includes: In each target clustering cluster, Otsu threshold segmentation is performed on each fiber image, a region with a small gray value is taken as a foreground region, and a number of pixel points in the foreground region is taken as a mucus performance characteristic value in each fiber image; In each target clustering cluster, all fiber images are arranged in time sequence to obtain a sorting sequence, in the sorting sequence, a ratio of an absolute value of a difference between mucus performance characteristic values of each adjacent two fiber images to an absolute value of a time difference is calculated to obtain a unit difference performance value, all unit difference performance values between the fiber images are arranged in order in the sorting sequence to obtain a change sequence; In the change sequence, a first-order difference sequence of the unit difference performance value is calculated, a value obtained by negatively correlating and normalizing a mean value of absolute values of all values in the first-order difference sequence is taken as a first abnormal trend factor, and a value obtained by normalizing a product of a number of negative values in the first-order difference sequence and the first abnormal trend factor is taken as the abnormal trend degree value.

9. The apparatus for airway mucus dynamic analysis and visualization of ciliary motility determination of claim 1, wherein, The early warning based on the abnormal index includes: When the abnormal index is greater than a preset abnormal threshold, a mucus accumulation abnormal early warning is performed; otherwise, no early warning is performed.

10. The apparatus for airway mucus dynamic analysis and visualization of ciliary motility determination of claim 1, wherein, The acquisition method of the periodic data segment includes: Periodic division is performed on the time sequence data of the physical parameters according to the types of the physical parameters based on the LSTM neural network, so as to obtain all periodic data segments corresponding to all types of physical parameters.

Images