Dynamic measurement method and system for ascites pressure of cirrhosis

Abstract

The invention discloses a cirrhosis ascites pressure dynamic measurement method and system, and relates to the technical field of medical auxiliary measurement. Comprising the following steps: acquiring an abdominal cavity image, ascites pressure dynamics and abdominal cavity blood flow pulsation data of a patient; extracting periodical fluctuation characteristics of the pressure data, and dynamically calibrating the image anatomical region according to the periodical fluctuation characteristics to obtain a candidate region; performing phase decorrelation analysis on the fluctuation characteristics of the region and the abdominal blood flow pulsation data to exclude a blood vessel adjacent region and obtain a target cavity region; based on this, a puncture path prediction model is constructed, and path data is output; finally, safety judgment is conducted on the path data, and when the pressure gradient of the path data is stable within the preset length range and meets the decorrelation condition, a puncture positioning point is output. The system comprises corresponding modules for executing the steps. Through multi-modal data fusion and dynamic analysis, precise planning and safety verification of puncture points are achieved, and the risk of accidentally injuring blood vessels and intestinal canals in the abdominal paracentesis is reduced.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of medical auxiliary measurement technology, in particular to a cirrhosis ascites pressure dynamic measurement method and system. BACKGROUND

[0002] Ultrasound-guided peritoneocentesis is a key technique for treating cirrhosis ascites, and its core difficulty lies in how to safely and accurately locate the ascites accumulation area and plan a puncture path in a dynamically changing peritoneal cavity environment to avoid mispuncturing the intestine and blood vessels and causing serious complications. The existing technology generally relies on static medical images (such as ultrasound and CT) for preoperative assessment, and identifies anatomical structures to determine the fluid accumulation area and generally avoid major blood vessels.

[0003] Such ultrasound and CT-based navigation methods provide basic guidance for clinical puncture and reduce the operation risk to a certain extent. However, the peritoneal cavity is a complex environment full of dynamic physiological processes, and respiratory motion will cause periodic displacement of the diaphragm, intestine and abdominal wall, while cardiovascular pulsation will generate pressure conduction in blood vessels and adjacent tissues. The existing technology fails to fully utilize these dynamic physiological information, and for ascites areas adjacent to blood vessels or cases with complex anatomical structures, it is difficult to effectively distinguish the pressure fluctuations caused by blood vessel pulsation from the benign fluctuations of the ascites area caused by fluid conduction, and to reflect the dynamic changes of the tissue structure on the puncture path in real time, resulting in the risk of damaging blood vessels or intestine when operating according to the preoperative planned path.

[0004] Therefore, the existing peritoneocentesis navigation technology lacks in-depth analysis and utilization of the dynamic characteristics of ascites pressure, and cannot achieve safe path planning for individualized peritoneocentesis based on the real-time physiological state of the patient. SUMMARY

[0005] In view of the deficiencies of the prior art, the present application provides a cirrhosis ascites pressure dynamic measurement method and system.

[0006] In order to achieve the above-mentioned purpose, the technical scheme of the present application is as follows:

[0007] In a first aspect, the present application discloses a cirrhosis ascites pressure dynamic measurement method, comprising the following steps:

[0008] Obtaining peritoneal cavity image data, ascites pressure dynamic data and peritoneal blood flow pulsation data of a patient, and extracting an anatomical region in the peritoneal cavity image data and a periodic fluctuation feature of the ascites pressure dynamic data;

[0009] Extracting a respiratory fluctuation component and a heart beat fluctuation component in the periodic fluctuation feature; filtering the respiratory fluctuation component and the heart beat fluctuation component, and obtaining an aligned time sequence relationship through time sequence alignment;

[0010] According to the aligned time sequence relationship, a phase change with respiratory motion is delineated in the abdominal cavity image data, an anatomic region with a pressure fluctuation amplitude greater than a preset fluctuation threshold and a stable fluctuation law is taken as the candidate region data, and a candidate fluctuation feature of the candidate region data is extracted;

[0011] The candidate fluctuation feature is subjected to phase decorrelation analysis with the abdominal cavity blood flow pulsation data to exclude regions similar to the blood vessel pulsation feature to obtain target cavity region data subjected to decorrelation processing;

[0012] A puncture path prediction model is constructed based on the target cavity region data, and prediction path data is output;

[0013] Safety of the prediction path data is determined, and when the prediction path data maintains a stable pressure gradient within a preset path length range and the prediction path data satisfies a decorrelation threshold condition, corresponding puncture positioning point data is determined and output.

[0014] In a second aspect, the present application discloses a liver cirrhosis ascites pressure dynamic measurement system using the above-mentioned liver cirrhosis ascites pressure dynamic measurement method, comprising:

[0015] A collection module is configured to acquire abdominal cavity image data, ascites pressure dynamic data and abdominal cavity blood flow pulsation data of a patient, and extract an anatomic region in the abdominal cavity image data and a periodic fluctuation feature of the ascites pressure dynamic data;

[0016] An alignment module is configured to extract a respiratory fluctuation component and a heart beat fluctuation component in the periodic fluctuation feature, filter the respiratory fluctuation component and the heart beat fluctuation component, and obtain an aligned time sequence relationship through time sequence alignment;

[0017] A calibration processing module is configured to delineate a phase change with respiratory motion in the abdominal cavity image data according to the aligned time sequence relationship, take an anatomic region with a pressure fluctuation amplitude greater than a preset fluctuation threshold and a stable fluctuation law as the candidate region data, and extract a candidate fluctuation feature of the candidate region data;

[0018] A decorrelation analysis module is configured to subject the candidate fluctuation feature to phase decorrelation analysis with the abdominal cavity blood flow pulsation data to exclude regions similar to the blood vessel pulsation feature to obtain target cavity region data subjected to decorrelation processing;

[0019] A path prediction module is configured to construct a puncture path prediction model based on the target cavity region data and output prediction path data;

[0020] A determination module is configured to perform safety determination on the prediction path data output by the puncture path prediction model, and when the prediction path data keeps a stable pressure gradient within a preset path length range and the prediction path data satisfies a decorrelation threshold condition, corresponding puncture positioning point data is determined and output.

[0021] Compared with the prior art, the present application has the following beneficial effects:

[0022] 1. The present application can effectively distinguish the potential liquid region changing with the respiratory motion law from the fixed solid organ tissue by dynamically calibrating the anatomical region according to the periodic fluctuation characteristics of the ascites pressure dynamic data, thereby realizing the transcendence of the traditional static image interpretation method, and the candidate region preliminarily screened is more consistent with the physiological dynamic actuality, thereby laying a reliable foundation for subsequent accurate positioning.

[0023] 2. The present application can accurately exclude the region adjacent to the blood vessel from the functional level rather than the simple morphological level by performing phase decorrelation analysis on the fluctuation characteristics of the candidate region and the abdominal cavity blood flow pulsation data, can acutely identify those dangerous regions which may not be significant on the image but have highly homologous pressure fluctuation and blood vessel pulsation, thereby improving the accuracy and safety of the selection of the target cavity region, and overcoming the limitation of single image examination.

[0024] 3. The present application realizes the transcendence of dynamic prediction and verification by constructing a puncture path prediction model and performing double safety determination on the output based on the stability of the pressure gradient and the decorrelation condition; through the decision mechanism based on multi-dimensional data fusion of the present application, the output puncture positioning point data is not only a coordinate, but also a safe channel which is fully verified and can effectively avoid dynamic physiological interference, thereby significantly improving the success rate and safety of the abdominal puncture operation as a whole. BRIEF DESCRIPTION OF DRAWINGS

[0025] The disclosure of the present application will be described with reference to the accompanying drawings. It should be understood that the drawings are only for illustrative purposes, and are not intended to limit the scope of protection of the present application. In the drawings, the same reference numerals are used to refer to the same parts. Among them:

[0026] Figure 1 is a step flow chart of the present application;

[0027] Figure 2 is a step flow chart of the working principle of the present application;

[0028] Figure 3 is a process flow chart of the decorrelation analysis of the present application;

[0029] Figure 4 is a system module connection diagram of the present application;

[0030] Figure 5 System module flowchart of the present application. DETAILED DESCRIPTION

[0031] It is easy to understand that, according to the technical solution of the present application, those skilled in the art can propose a plurality of structure modes and implementation modes which can be replaced with each other without changing the essential spirit of the present application. Therefore, the following specific embodiments and drawings are only exemplary descriptions of the technical solution of the present application, and should not be regarded as the whole or as the limitation or restriction of the technical solution of the present application.

[0032] SUMMARY

[0033] In the prior art, liver cirrhosis ascites puncture positioning is mostly dependent on static medical image navigation, which is difficult to cope with dynamic interference caused by respiratory motion and cardiovascular pulsation. When the abdominal cavity pressure changes dynamically, the traditional method is easily affected by the displacement of the intestinal tract and the pulsation of blood vessels, resulting in a decrease in positioning accuracy. The existing equipment cannot synchronously perceive the influence of physiological fluctuations on puncture safety, especially when the ascites volume is small or the distribution is complex, single image navigation will have systematic deviation, which is difficult to meet the clinical precise puncture demand.

[0034] In order to solve the above problems, the inventors found that the ascites pressure fluctuation characteristics are coupled with the respiratory phase and the blood vessel pulsation, and the dynamic compensation is realized by establishing a multi-modal data fusion model. It is found in the research process that the respiratory dominant fluctuation has high positioning sensitivity, and the heart beat dominant fluctuation has large interference but can be used for blood vessel identification, thus the idea of dynamically switching safety criterion according to fluctuation correlation is proposed. Further through clinical verification, the pressure gradient stability and decorrelation threshold are introduced into the path planning decision mechanism to form an adaptive optimization system.

[0035] Specifically, the detection system first synchronously collects the abdominal cavity image, ascites pressure dynamics and blood flow pulsation data of the patient. By analyzing the periodic fluctuation characteristics of the pressure data, the respiratory fluctuation component and the heart beat fluctuation component are separated and time-sequentially aligned. When the pressure fluctuation of the candidate region and the blood flow pulsation correlation exceeds the set threshold, the system automatically switches to the decorrelation analysis mode, uses the phase difference characteristics to exclude the blood vessel interference region, and combines the pre-established fluctuation-morphology coupling model to plan the safe path. In the continuous detection process, the system dynamically corrects the path prediction model parameters through the feedback mechanism according to the real-time pressure gradient change characteristics, forming an optimized closed loop. For the case where the correlation threshold is not exceeded, the standard path planning based on respiratory phase synchronization is continued to output the puncture positioning.

[0036] Compared with the prior art, the traditional method relies on static anatomical structure and lacks dynamic compensation mechanism, and positioning error is easy to occur under the interference of physiological fluctuation. The application fuses multi-modal real-time data, and realizes dynamic optimization of the puncture path by establishing a fluctuation-morphology coupling model. Unlike the existing static planning method, the scheme can intelligently switch the safety criterion according to the real-time physiological state, and continuously calibrate the model parameters through closed-loop feedback, thereby significantly improving the puncture safety in a complex abdominal cavity environment.

[0037] Through the above technical scheme, the positioning error problem caused by physiological fluctuation is effectively overcome, the puncture precision is improved while ensuring the operation real-time, the dynamic correlation analysis mechanism takes into account the high sensitivity of respiratory fluctuation and the blood vessel recognition advantage of heart beat fluctuation, and the model parameter self-adaptive function ensures the accuracy under individualized difference. The method provides a reliable technical means for liver cirrhosis ascites puncture, and is particularly suitable for clinical puncture scenes of complex cases such as obesity and intestinal distension.

[0038] After introducing the basic idea of the application, the embodiments of the application will be specifically introduced with reference to the drawings.

[0039] Embodiment one:

[0040] As shown in the figure, a liver cirrhosis ascites pressure dynamic measurement method comprises the following steps: Figure 1

[0041] Obtain abdominal cavity image data, ascites pressure dynamic data and abdominal cavity blood flow pulsation data of a patient, and extract an anatomical region in the abdominal cavity image data and a periodic fluctuation feature of the ascites pressure dynamic data;

[0042] Extract a respiratory fluctuation component and a heart beat fluctuation component in the periodic fluctuation feature; after filtering the respiratory fluctuation component and the heart beat fluctuation component, obtain an aligned time sequence relationship through time sequence alignment;

[0043] According to the aligned time sequence relationship, delineate a phase change with respiratory motion in the abdominal cavity image data, take an anatomical region with a pressure fluctuation amplitude greater than a preset fluctuation threshold and a stable fluctuation rule as candidate region data, and extract candidate fluctuation features of the candidate region data;

[0044] Perform phase decorrelation analysis on the candidate fluctuation features and the abdominal cavity blood flow pulsation data to exclude regions similar to the blood vessel beat feature to obtain target cavity region data after decorrelation processing;

[0045] Construct a puncture path prediction model based on the target cavity region data and output prediction path data;

[0046] ​The safety judgment is performed on the predicted path data, and when the predicted path data keeps the pressure gradient stable within a preset path length range and the predicted path data satisfies a decorrelation threshold condition, corresponding puncture positioning point data is determined and output.

[0047] As shown in Figure 2 The working principle of the present application is to obtain the abdominal cavity image data, ascites pressure dynamic data and abdominal cavity blood flow pulsation data of the patient. The abdominal cavity image data is obtained by an ultrasound, CT or MRI imaging device, which can provide intuitive anatomical information of the distribution of tissues and ascites in the patient's body. The ascites pressure dynamic data is collected in real time by an implantable or body surface pressure sensor, reflecting the periodic fluctuation changes of the patient's abdominal fluid with breathing and heart beat. The abdominal cavity blood flow pulsation data is obtained by Doppler ultrasound or other blood flow monitoring devices, which is used to characterize the pulsation of abdominal blood vessels. By simultaneously obtaining these three types of data, the patient's abdominal cavity environment can be comprehensively characterized from three dimensions of spatial structure, pressure change and blood flow signal.

[0048] Periodic fluctuation features are extracted from the ascites pressure dynamic data, and are decomposed into a breathing fluctuation component driven by breathing and a heart beat fluctuation component driven by cardiovascular pulsation. Since the two types of fluctuations have different frequency characteristics, the system filters them respectively to eliminate noise and cross interference. On this basis, the fluctuation components are synchronized with the abdominal cavity image data through time sequence alignment, so that the phase change of the pressure signal can accurately correspond to the specific frame in the image sequence, ensuring the reliable correspondence between the signal and the image in the subsequent analysis.

[0049] According to the aligned time sequence relationship, the anatomical region that is displaced by respiratory motion is outlined in the image. By setting a preset fluctuation threshold, only those regions with a pressure fluctuation amplitude greater than the preset fluctuation threshold and maintaining regular stability within a period are selected as candidate region data, and the candidate fluctuation features of these candidate regions are further extracted. This process can effectively eliminate interference signals caused by image noise or local irregular motion, ensuring that the candidate regions have high representativeness and stability. The candidate fluctuation features represent the periodic pressure change signals generated by respiratory fluctuation and heart beat fluctuation.

[0050] The preset fluctuation threshold setting method includes but is not limited to the following methods:

[0051] Based on the real-time ascites pressure dynamic data of the current patient, the amplitude distribution mean μ and standard deviation σ of the respiratory fluctuation component within the whole abdominal cavity are calculated;

[0052] According to the formula A dynamic adaptive threshold is calculated, where the coefficient K is a value preset according to clinical experience, and the preferred range is 1.5 to 2.5; an absolute lower threshold and an absolute upper threshold to prevent misjudgment due to too extreme data distribution, determined according to the noise base of the pressure sensor used, determined according to the safety maximum in historical clinical data;

[0053] The effective threshold interval of the final preset fluctuation threshold is The above calculation is dynamically performed through a time domain sliding window to ensure that the preset fluctuation threshold is adaptively adjusted according to the physiological state of the patient.

[0054] On this basis, the system performs phase decorrelation analysis on the candidate fluctuation features and the abdominal cavity blood flow pulsation data. If the fluctuation features of the candidate region and the blood vessel pulsation have a high correlation, it is determined that it is close to the blood vessel or has a risk of blood vessel interference, and it is excluded; otherwise, it is included in the target cavity region data. Through this step, the cavity candidate region avoiding blood vessel distribution is finally obtained, providing a reliable basis for subsequent planning of a safe puncture path.

[0055] The system constructs a puncture path prediction model based on the target cavity region data. The puncture path prediction model can simulate the abdominal cavity space of the patient and automatically output prediction path data to plan a puncture path from the body surface to the target cavity region. After the path is generated, the system performs safety determination on the prediction path. When the path maintains a stable pressure gradient within a preset length range, and the fluctuation features meet the decorrelation threshold condition, it is determined that the path has a low puncture risk. At this time, the system finally outputs corresponding puncture positioning point data.

[0056] Through the above steps, the present application realizes joint analysis and fusion determination of the abdominal cavity image, the dynamic characteristics of ascites pressure, and the blood flow pulsation data, and can dynamically and individually determine the optimal puncture positioning point while avoiding blood vessel and intestinal tube interference. This method not only improves the safety and accuracy of puncture operation, but also realizes intelligent prediction of the puncture path through the multi-modal combination of pressure signals and images, thereby improving the efficiency and effect of clinical treatment of cirrhosis ascites.

[0057] The present application further proposes that after filtering the respiratory fluctuation component and the heart beat fluctuation component, the specific steps of obtaining the aligned time sequence relationship through time sequence alignment include:

[0058] The frequency band of the respiratory fluctuation component is set to 0.1-0.5 Hz, and the frequency band of the heart beat fluctuation component is set to 0.8-2.0 Hz, and the corresponding fluctuation amplitude features and phase features are obtained respectively; such frequency band division can effectively separate the low-frequency pressure fluctuation caused by respiratory movement from the high-frequency pressure fluctuation caused by heart beat, avoiding mutual interference between the two in the signal processing process. For the above two components, the fluctuation amplitude features and phase features are extracted respectively, ensuring that both the energy information of the signal and the phase information of the dynamic change are retained;

[0059] The fluctuation amplitude features and phase features are filtered by band-pass filtering to obtain filtered pressure phase features; by band-pass filtering, random noise and non-target frequency band interference in the measurement environment can be effectively eliminated, ensuring that the obtained pressure phase features are more pure and stable, so as to more accurately reflect the dynamic change of ascites pressure with breathing and heart beat. The filtered pressure phase features are used for subsequent time sequence alignment analysis;

[0060] The pressure phase features are time sequence matched with the synchronously collected abdominal cavity image data to generate an aligned time sequence relationship;

[0061] In the time sequence matching process, first, the key image frames at the end of expiration and the end of inspiration are marked in the abdominal cavity image sequence, which can clearly reflect the spatial position change of the abdominal cavity tissue at different respiratory phases. At the same time, the peak points and trough points in the pressure phase features are marked to represent the extreme points of the respiratory and heart beat fluctuations. By corresponding matching of the end of expiration and the end of inspiration of the image frame with the peak and trough of the pressure phase features, accurate alignment between the image data and the pressure signal can be realized, and a unified time sequence relationship is obtained.

[0062] This processing method can ensure that the abdominal cavity image data and the ascites pressure signal are strictly synchronized in the time dimension, so that the spatial change of the anatomical region and the phase change of the pressure fluctuation correspond one by one. On this basis, the anatomical region that dynamically changes with the respiratory movement can be more accurately outlined, and further combined with the pressure amplitude threshold to screen the candidate region, avoiding the deviation of region recognition caused by the asynchronization of breathing and heart beat.

[0063] By introducing the frequency band separation, band-pass filtering and time sequence alignment processing procedures in the signal processing, the application can significantly improve the accuracy and stability of the candidate region identification, and realize the dynamic fusion of image and pressure. Not only can the real pressure distribution in the abdominal cavity be more accurately reflected, but also the positioning deviation caused by blood vessel beat or image noise can be avoided, thereby providing more reliable data support in the output process of the puncture path prediction and puncture positioning point, and finally improving the safety and controllability of the cirrhotic ascites puncture operation.

[0064] AsFigure 3 As shown, the present application further proposes that the candidate wave characteristics are subjected to phase decorrelation analysis with the abdominal cavity blood flow pulsation data, and regions similar to the blood vessel pulsation characteristics are excluded to obtain the specific steps of the decorrelation-processed target cavity region data.

[0065] The blood vessel pulsation signal is extracted from the synchronously collected abdominal cavity blood flow pulsation data, which directly reflects the pulsation law of the fluid in the blood vessel and has high phase stability.

[0066] The cross-correlation coefficient of the candidate wave characteristics and the abdominal cavity blood flow pulsation data is calculated to measure the similarity of the two groups of signals in phase and frequency; the cross-correlation analysis can reveal whether the wave of the candidate region has a strong synchronization relationship with the blood vessel pulsation, thereby providing a quantitative basis for the identification of the region adjacent to the blood vessel.

[0067] The average value of the cross-correlation coefficient in the preset time domain sliding window is calculated, and through the introduction of the sliding window, the influence of transient noise or incidental interference on the correlation calculation result can be eliminated, and the correlation determination is more stable and reliable.

[0068] After obtaining the average cross-correlation coefficient in the sliding window, the present embodiment sets a preset decorrelation threshold for distinguishing the region highly correlated with the blood vessel pulsation signal and the region far away from the blood vessel; when the average value is lower than the preset decorrelation threshold, it indicates that the similarity of the wave characteristics of the candidate region and the blood vessel pulsation signal is low, and it is determined that the candidate region data is far away from the blood vessel, and the candidate region data is taken as the target cavity region data. On the contrary, when the average value is higher than the decorrelation threshold, it is determined that the candidate region is too close to the blood vessel, and the region is excluded to avoid the puncture path being close to the blood vessel.

[0069] Specifically, first, in the prediction path data, the candidate wave characteristics of the target cavity region sampled along the path are extracted, denoted as time series signal X(t). At the same time, the pulsation signal of the blood vessel adjacent to the path is extracted, denoted as Y(t). To determine the correlation of the two, the standardized cross-correlation function is used, and its formula is as follows:

[0070]

[0071] Wherein, and are the mean values of the signals, and τ is the time delay. The value range of is [-1, 1], and the closer the absolute value is to 1, the stronger the correlation of the two signals.

[0072] In the actual determination process, a time domain sliding window W is set, and the average cross-correlation coefficient in the window is calculated :

[0073]

[0074] When , it is determined that the current path segment is de-correlated with the blood vessel pulsation signal, indicating that the predicted path is away from the blood vessel distribution area, and can be used as a candidate safe path; if , the path segment is rejected or the path is adjusted.

[0075] wherein the de-correlation threshold β is not fixed, but is adaptively adjusted according to the main frequency amplitude of the patient's blood flow pulsation characteristics:

[0076]

[0077] wherein, is a basic threshold (such as 0.2-0.3), and γ is an adjustment coefficient, used to reflect the dynamic influence of the blood vessel pulsation amplitude on the path safety. In this way, the greater the pulsation amplitude, the lower the tolerance of the system to correlation, and the more strictly the path close to the blood vessel is rejected, ensuring the safety of puncture.

[0078] Through the above technical solutions, the present application avoids the misjudgment caused by simply relying on image morphology or pressure amplitude screening, improves the accuracy of identifying the region adjacent to the blood vessel, and on the other hand, ensures that the target cavity region output finally meets the safety requirements of clinical puncture, thereby significantly reducing the risk of mispuncturing the blood vessel, and providing higher quality data input for the construction of subsequent puncture path prediction model.

[0079] The present application further proposes specific steps of constructing a puncture path prediction model based on target cavity region data and outputting prediction path data, in order to realize safer and more accurate puncture positioning, including:

[0080] Three-dimensional reconstruction is performed on the target cavity region data obtained through phase de-correlation analysis to generate a puncture path prediction model; the reconstruction process not only includes the anatomical structure information of the abdominal cavity, but also uniformly integrates the patient's skin surface, abdominal wall layer, intestinal drift region and target ascites pool into the same spatial coordinate system, thereby generating a complete puncture path prediction model, which can intuitively reflect the hierarchical relationship and dynamic characteristics of the patient's body cavity structure, and provide spatial basis for path planning;

[0081] After generating the three-dimensional space model, an initial path is planned from the patient's surface puncture point to the center of the target cavity in the puncture path prediction model; the path is usually in the form of a straight line, and can be used as a basic reference for subsequent dynamic adjustment.

[0082] ​In order to consider the influence of the respiratory motion of the patient on the position of the intestinal tract, the initial path is coupled with the respiratory motion of the patient, the drift motion of the intestinal tract relative to the initial path at different respiratory phases is simulated, the minimum distance between each point of the initial path and the drift intestinal tract is calculated in the inspiration, expiration and respiratory transition phases, and the initial path is adjusted to an optimal path with the lowest risk according to the minimum distance, and the output is prediction path data.

[0083] Specifically, based on the calculation result of the minimum distance between each point of the initial path and the drift intestinal tract, the initial path is further dynamically adjusted. When the minimum distance between some path sections and the intestinal tract is less than a safety threshold, the system automatically adjusts the path direction of the section to deviate to a position away from the drift intestinal tract. After iterative optimization, an optimal path that satisfies the minimum safety distance condition at all respiratory phases is gradually obtained, and the path is output as the final prediction path data.

[0084] Through the above technical solution, the present application realizes automatic avoidance of the risk of intestinal tract drift in the puncture path prediction in a three-dimensional space. Compared with the traditional method of relying on the experience of doctors to infer in a two-dimensional image, the present application not only can accurately depict the individualized dynamic abdominal cavity environment of the patient, but also can comprehensively consider the influence of the respiratory motion on the drift of the intestinal tract in the path planning, thereby effectively reducing the risk of damaging the intestinal canal in the puncture process and improving the safety and reliability of the selection of the puncture positioning point, and providing more scientific data support for the clinical treatment of cirrhosis ascites.

[0085] The present application further proposes that the step of constructing a puncture path prediction model based on target cavity region data further comprises:

[0086] An individualized database based on historical successful puncture cases is established, the database is constructed by a plurality of clinically verified successful puncture cases, and the database at least stores data characteristics of patients with different body characteristics, and the data characteristics include safe puncture path data and corresponding ascites pressure dynamic data. In this way, the database can reflect the corresponding relationship between the pressure environment and the path selection in the actual puncture operation of patients with different body types;

[0087] In the step of constructing the puncture path prediction model, the database is called, the body characteristics and data characteristics of the current patient are matched with the database, and the prediction path data is optimized and corrected through the matching result.

[0088] In the matching process, not only the body surface parameters (such as abdominal circumference, body mass index) of the patient are considered, but also comprehensive indexes such as abdominal wall thickness, liver volume reduction degree and ascites distribution are combined to ensure the accuracy of the matching.

[0089] After the matching is completed, the system extracts a number of reference cases closest to the current patient's body type, and takes the safe puncture path data in these cases and the corresponding ascites pressure dynamic data as the comparison standard, and combines it into the adjustment of the current predicted path. By comparing with historical data, the present application can check the stability of the predicted path and correct the predicted path if necessary. For example, when the pressure gradient performance of the current predicted path in a certain section is similar to the pressure distribution characteristics of the known high-risk path in the similar case, the system will automatically offset or re-plan the path, so as to ensure that the output predicted path meets the clinical safety requirements.

[0090] By introducing the individualized database based on historical cases, the present application further increases the empirical constraint conditions from real clinical experience on the basis of the three-dimensional geometric modeling and respiratory motion coupling of the prediction model. In this way, not only the individualized adaptability of the predicted path is improved, but also the ability to identify and avoid potential risks in a complex abdominal anatomical environment is enhanced. The final output of the predicted path data can take into account the dynamic anatomical environment of the patient and the historical experience rules, thereby significantly improving the reliability and clinical operability of the puncture path prediction.

[0091] The present application further proposes that in the safety judgment of the predicted path data output by the puncture path prediction model, the judgment method for judging whether the predicted path data maintains a stable pressure gradient within a preset path length range comprises:

[0092] Within the preset path length range, the pressure difference between adjacent sampling points on the predicted path data is calculated;

[0093] When the absolute value of the pressure difference between all consecutive sampling points is less than the pressure gradient threshold value, it is judged that the pressure gradient is stable, wherein the pressure gradient threshold value is adaptively adjusted according to the average amplitude of the ascites pressure dynamic data.

[0094] Specifically, within a preset path length range (denoted as L), the system first uniformly samples the predicted path data to obtain a series of sampling points on the path For each adjacent sampling point pair , the corresponding pressure difference value is calculated:

[0095]

[0096] Wherein, represents the ascites pressure value of the i-th sampling point on the path, represents the pressure gradient between adjacent points.

[0097] In the judgment process, the system will count the pressure difference of all adjacent sampling points, and if the following conditions are met:

[0098]

[0099] If the pressure gradient is within the preset length range, it is determined that the path is stable in the pressure gradient. In the formula, θ is the pressure gradient threshold.

[0100] The pressure gradient threshold θ is not a fixed constant, but is dynamically adjusted according to the average amplitude of the ascites pressure dynamic data of the patient The specific calculation formula is:

[0101]

[0102] Wherein, α is a proportional coefficient, which can be 0.05-0.2, used to balance the threshold setting which is too strict or too loose. This adaptive method ensures that in patients with large ascites pressure fluctuation amplitude, reasonable pressure changes are still tolerated, while in patients with small pressure fluctuation amplitude, the path stability is more stringent.

[0103] Through the above determination method, the application can quantitatively evaluate the stability of the predicted path in the pressure distribution, avoid selecting the pressure gradient mutation region as the puncture channel, and reduce the risk of puncture needle deviation or damage caused by local abdominal stress difference. Combined with the previous candidate region screening, phase decorrelation analysis, three-dimensional path prediction and individual database optimization, the safety determination step finally forms a closed loop path selection mechanism, which ensures that the output puncture positioning point not only avoids blood vessels and intestinal tract regions, but also maintains stability in the pressure distribution level, further improving the safety and clinical operability of puncture.

[0104] The application further proposes that the preset path length range is determined by the following method:

[0105] Based on the abdominal cavity image data, the total thickness of the abdominal wall tissue from the body surface puncture point to the entrance of the ascites cavity region in the target cavity region data is measured;

[0106] The preset path length range is set to the total length from the body surface puncture point, extending along the predicted path data into the body, covering the total thickness of the abdominal wall tissue and additionally increasing a safety margin distance;

[0107] Wherein, the safety margin distance is a fixed value preset according to clinical experience, or a dynamically calculated value proportional to the depth of the target ascites region.

[0108] Specifically, first, based on the abdominal cavity image data, the total thickness of the abdominal wall tissue from the body surface puncture point to the entrance of the ascites cavity in the target cavity region data is measured, denoted as:

[0109]

[0110] Wherein, wherein, d, is the thickness of the i-th layer of abdominal wall tissue along the puncture path, and n is the number of tissue layers. This thickness measurement can be obtained by image segmentation and three-dimensional reconstruction methods.

[0111] Subsequently, the preset path length range is set as:

[0112]

[0113] wherein, ΔL is the safety margin distance. This margin distance can be set in two ways:

[0114] Fixed value method: according to clinical experience, set , usually 5-15 mm, to ensure that even in the presence of individual measurement errors, the path coverage range is still reliable.

[0115] Dynamic calculation method: according to the depth of the target ascites region set the safety margin, in the form of:

[0116]

[0117] wherein, a is the proportionality coefficient (such as 0.05-0.2), which is dynamically adjusted by the proportion of the patient's ascites depth. When the ascites region is deeper, the margin is automatically increased, thereby improving the adaptability of the model to patients of different body types and different conditions.

[0118] Through the above method, the preset path length range not only covers the total thickness of the abdominal wall tissue, but also further ensures that the determination interval completely contains the possible puncture channel through the safety margin, thereby providing a unified, reasonable and adaptive boundary condition for pressure gradient smoothness determination and decorrelation threshold determination.

[0119] The present application further proposes that the liver cirrhosis ascites pressure dynamic measurement method further comprises:

[0120] Based on the construction of the puncture path prediction model based on the target cavity region data, further introduce the patient's historical puncture data, construct an individualized prediction model. The historical puncture data at least includes: puncture body surface position, puncture path depth information, ascites pressure dynamic change curve, puncture success rate and complication occurrence.

[0121] Specifically, by feature extraction and induction on the above historical data, an individualized database can be formed, and an individualized prediction model can be established based on this database by machine learning or statistical modeling methods.

[0122] In the implementation process, the XGBoost algorithm is adopted, and the training set contains 200 cases of liver cirrhosis patient puncture data to train the individualized prediction model. The individualized prediction model calculates the risk score data of the patient under different candidate puncture paths by comparing the body parameters (such as abdominal wall thickness, ascites distribution characteristics, and intestinal tube drift characteristics) of the current patient with the parameters of similar patients in the database, and is recorded as:

[0123]

[0124] wherein, represents the multi-modal feature parameters of the current patient, represents the historical puncture data set, and the function The risk score calculation is realized by modeling.

[0125] After obtaining the risk score data, the application proposes to fuse the risk score data with the aforementioned target cavity region data. Specifically, first, the spatial safety index (such as pressure gradient stability, decorrelation threshold satisfaction, optimal path and intestinal tube distance margin, etc.) of the target cavity region data is quantified to obtain the spatial safety score S. Then, the risk score data R and the spatial safety score S are weighted or nonlinearly combined according to the preset fusion strategy to obtain the comprehensive fusion result F:

[0126]

[0127] wherein, , is a weight coefficient, which can be adaptively set according to clinical statistical experience or training data, and is determined by regression analysis of 200 historical puncture data, wherein = 0.6, = 0.4, the prediction error is minimum.

[0128] When the fusion result F is lower than the preset fusion threshold (typically 0.4 ~ 0.6, which can be adjusted according to the complexity of the patient's body type), the system automatically outputs the individualized recommended puncture positioning point data. The output not only depends on the current image and pressure dynamic characteristics, but also fully combines the individual difference rules in historical clinical data, thereby establishing a more robust association between spatial constraints and clinical experience.

[0129] Through the fusion of the above individualized prediction model and the target region data, the application not only can further reduce the potential blind area risk in the puncture path selection, but also can provide more accurate recommendations for puncture of special body type patients or complex pathological states. The clinical usability and individual adaptability of the predicted path are improved, avoiding the problem of lack of universality caused by simply relying on images and pressure dynamic characteristics, significantly improving the safety and success rate of puncture operation.

[0130] This application further proposes that the method for dynamic measurement of ascites pressure in cirrhosis also includes a real-time feedback step during surgery:

[0131] During the actual puncture process, the real-time pressure data of the needle tip moving along the predicted path data is obtained in real time and recorded as a time series. At the same time, the system records the expected pressure fluctuation characteristics of the corresponding position points in the predicted path, which are recorded as This feature comes from the previously constructed puncture path prediction model, combined with the results of pressure gradient stability, decorrelation threshold determination, and individualized prediction model optimization;

[0132] The fluctuation characteristics of the real-time needle tip pressure data are compared in real time with the expected pressure fluctuation characteristics of the corresponding position point in the predicted path data. During the real-time comparison, the system calculates the deviation between the real-time needle tip pressure and the predicted pressure:

[0133]

[0134] When the deviation of the comparison result exceeds the preset tolerance δ, the system determines that the needle tip has deviated from the ideal path and there may be a risk of approaching the intestine or blood vessels. The preset tolerance δ can be ±2 mmHg to ±5 mmHg, preferably ±3 mmHg.

[0135] If the preset tolerance δ is exceeded, the system generates and outputs a guidance signal for needle tip adjustment. The guidance signal includes the direction of the needle tip adjustment. and amplitude ΔL, which is based on the magnitude of the deviation ΔP(t) and the instantaneous correlation between the real-time pressure data of the needle tip and the peritoneal blood flow pulsation data of the adjacent blood vessels Perform the calculation:

[0136]

[0137] Where, is the proportional coefficient, which controls the linear relationship between the adjustment range and the pressure deviation; The larger the value, the higher the correlation between the needle tip and the blood vessel pulsation, and the adjustment range should be reduced accordingly to avoid excessive adjustment and damage to the blood vessels. It is calculated through a predictive path model and real-time imaging or needle tip position sensor, pointing to the spatial direction with the least risk.

[0138] Through the above-mentioned technical solution, the present application's real-time intraoperative feedback mechanism can continuously verify the safety of the puncture needle path during advancement, providing immediate warning of pressure anomalies and dynamic path adjustment. This forms a closed-loop safety control system from preoperative planning to intraoperative execution. This not only reduces the risk of puncture errors and tissue damage, but also improves the puncture success rate and controllability of clinical operations, achieving dynamic protection for the patient's ascites puncture throughout the entire process.

[0139] The following is a specific embodiment of a clinical application of a liver cirrhosis ascites pressure dynamic measurement method:

[0140] A 62-year-old male patient with liver cirrhosis decompensation (Child-Pugh score B, abdominal circumference 112 cm) needs therapeutic puncture due to massive tense ascites. Clinical examination shows that the abdomen is swollen and frog-like, and the mobile dullness is positive. Traditional ultrasound examination shows that the intestinal canal is widely floating, and there is a 6.8x5.2cm liquid dark area around the superior mesenteric vein branch, but it is difficult to determine the safe puncture path by traditional image navigation. The patient takes a semi-recumbent position, and the liver cirrhosis ascites pressure dynamic measurement system of the application is used for precise puncture planning and real-time guidance.

[0141] The ultrasound probe scans the patient's abdominal cavity at a frequency of 3.5MHz, and captures the diaphragm movement and intestinal canal drift trajectory in real time. At the same time, the body surface pressure sensor array attached from the xiphoid to the pubic region monitors the abdominal wall pressure fluctuation at a sampling rate of 100Hz, and the Doppler blood flow probe synchronously tracks the pulsatile characteristics of the portal vein branch. The system realizes hardware-level synchronization of three sources of data through precise clock protocol, and the timing error is controlled within milliseconds. At the end of expiration phase, the diaphragm is lifted to the 10th intercostal space, and the pressure sensor records a trough value of 12.8mmHg; at the end of inspiration, the pressure peak value is 15.3mmHg, and the ultrasound image captures the dynamic change of the intestinal canal moving down about 2.3cm. The accurate synchronization of respiration-image-pressure lays a time-space foundation for subsequent safe region labeling.

[0142] In the region 5cm above the navel, the system identifies the liquid area that fluctuates regularly with respiration: the pressure fluctuation amplitude is greater than 8mmHg, and the coefficient of variation is stable below 0.15, which meets the characteristics of the candidate area. When the cross-correlation coefficient between the heart beat fluctuation and the portal vein blood flow spectrum of this area reaches 0.58 (significantly higher than the safety threshold of 0.35), the system immediately starts the decorrelation analysis mechanism, and the three-dimensional reconstruction shows that this place is only 1.2cm away from the blood vessel, which is automatically excluded as a high-risk area. Finally, a 4.3x3.8cm safe target area is confirmed in the left iliac fossa, and the pressure fluctuation and blood vessel pulsation correlation coefficient of this area is only 0.12, which meets the safe puncture conditions.

[0143] Based on the three-dimensional modeling of the target area, the system plans a puncture path from the skin entry point, through the 2.1 cm subcutaneous fat layer, the 1.8 cm posterior sheath of the abdominal rectus muscle, to the peritoneal layer. Considering the displacement of the intestinal tract caused by respiration, the path planning unit automatically adjusts the puncture angle by 22° to ensure a safe distance of more than 1.2 cm from the intestinal tract throughout the respiratory cycle. The system retrieves successful cases of similar patients from the historical database, and uses adaptive pressure gradient threshold to verify the safety of the path: the pressure difference in the subcutaneous layer is 0.85 mmHg / cm, the muscle layer is 1.12 mmHg / cm, and the peritoneal area is 0.43 mmHg / cm, all of which are lower than the individualized threshold of 1.26 mmHg / cm.

[0144] During the puncture process, when the needle tip travels to a depth of 3.2 cm, real-time pressure monitoring shows abnormal fluctuations. The system instantaneously compares the predicted path data and detects that the pressure deviation exceeds the safety tolerance, immediately initiating blood vessel correlation analysis. According to the spatial offset parameters calculated in real time, the system prompts "adjust 15° to the head side, advance 2 mm" by voice. After correction, the needle tip pressure fluctuation returns to normal, and finally 1100 ml of grass yellow clear ascites is safely drained. Postoperative ultrasound review confirms no blood vessel injury and intestinal penetration.

[0145] This embodiment fully presents the closed-loop protection capability of the system in complex ascites puncture. Through the dynamic coupling of multi-modal data, the three technical bottlenecks in traditional puncture are successfully solved: anatomical structure displacement caused by respiratory motion, difficulty in distinguishing blood vessels and ascites area, and lack of real-time correction during operation. The respiratory-pressure synchronization model accurately captures millimeter-level tissue displacement, the phase decorrelation algorithm effectively identifies the adjacent area of hidden blood vessels, and the real-time feedback system during operation is like a "pressure-aware navigator" for the doctor. Finally, in high-risk cases with extensive intestinal floating, the puncture process is achieved with zero complications, providing a new precision solution for the treatment of cirrhotic ascites.

[0146] Example Two:

[0147] As shown in Figure 4 A cirrhotic ascites pressure dynamic measurement system using the above-mentioned cirrhotic ascites pressure dynamic measurement method, comprising:

[0148] A collection module for acquiring abdominal cavity image data, ascites pressure dynamic data, and abdominal cavity blood flow pulsation data of a patient, and extracting an anatomical region in the abdominal cavity image data and a periodic fluctuation feature of the ascites pressure dynamic data;

[0149] An alignment module for extracting a respiratory fluctuation component and a heart beat fluctuation component in the periodic fluctuation feature; after filtering the respiratory fluctuation component and the heart beat fluctuation component, a time sequence relationship after alignment is obtained through time sequence alignment;

[0150] The calibration processing module is configured to delineate the phase change of the anatomical region with the respiratory motion in the abdominal cavity image data according to the aligned time sequence relationship, take the anatomical region with the pressure fluctuation amplitude greater than the preset fluctuation threshold and the stable fluctuation rule as candidate region data, and extract candidate fluctuation features of the candidate region data.

[0151] The decorrelation analysis module is configured to perform phase decorrelation analysis on the candidate fluctuation features and the abdominal cavity blood flow pulsation data, and exclude the regions with similar blood vessel pulsation features to obtain target abdominal cavity region data after decorrelation processing.

[0152] The path prediction module is configured to construct a puncture path prediction model based on the target abdominal cavity region data and output prediction path data.

[0153] The determination module is configured to perform safety determination on the prediction path data output by the puncture path prediction model, and determine corresponding puncture positioning point data and output when the prediction path data remains stable in the pressure gradient within the preset path length range and the prediction path data satisfies the decorrelation threshold condition.

[0154] As shown in Figure 5 The acquisition module includes a medical image acquisition unit (such as an ultrasound probe or a CT scanner), a pressure sensing unit (for monitoring the dynamic changes of ascites pressure), and a blood flow detection unit (such as a Doppler blood flow probe). The module also includes a pre-processing circuit and a memory for enhancing the resolution of the acquired image data, suppressing noise of the pressure and blood flow signals, and extracting anatomical region information in the abdominal cavity image data and periodic fluctuation features of the ascites pressure dynamic data.

[0155] The alignment module extracts the respiratory fluctuation component and the heart beat fluctuation component in the periodic fluctuation features, and performs frequency filtering processing through a band-pass filter. The module also includes a time sequence synchronization processing unit for synchronizing the respiratory, heart beat signals with the image acquisition frame to obtain the aligned time sequence relationship, ensuring the consistency of the multi-modal data in time domain and space.

[0156] The calibration processing module delineates the anatomical region with the respiratory motion change in the abdominal cavity image data according to the aligned time sequence relationship, takes the region with the pressure fluctuation amplitude greater than the preset threshold and the stable rule as candidate region data, and further extracts the candidate fluctuation features. The module is composed of an image processing chip, a feature extraction unit and a storage buffer, and is configured to generate a candidate region distribution map in real time.

[0157] The de-correlation analysis module performs phase de-correlation analysis on the candidate fluctuation features and the abdominal cavity blood flow pulsation data, calculates the cross-correlation coefficient in a preset time domain sliding window, and compares it with a de-correlation threshold value, to exclude a high-risk area similar to the blood vessel pulsation features, so as to obtain target cavity region data after de-correlation processing. The module is composed of a signal processing unit and a special computing circuit (DSP / FPGA) to realize high-speed real-time calculation.

[0158] The path prediction module constructs a puncture path prediction model based on the target cavity region data. Specifically, it includes a three-dimensional reconstruction unit, a path planning unit, and a database calling unit. The three-dimensional reconstruction unit generates a three-dimensional model containing the skin surface, the abdominal wall layer, the intestinal drift region, and the target ascites pool. The path planning unit establishes an initial path from the body surface puncture point to the center of the ascites cavity and corrects the path by simulating intestinal drift in combination with respiratory motion. The database calling unit retrieves safe path data of similar body type patients from the historical puncture case database and optimizes the predicted path.

[0159] The determination module determines the safety of the predicted path data output by the puncture path prediction model. The module includes a pressure gradient calculation unit and a path risk analysis unit. The pressure gradient calculation unit calculates the pressure difference point by point within a preset path length range and compares it with a pressure gradient threshold value to ensure that the path has a smooth pressure change from penetrating the abdominal wall to the ascites cavity entrance. The path risk analysis unit further investigates whether the path is close to a blood vessel in combination with the de-correlation threshold condition, to determine whether the final puncture path meets the safety requirements. When it is determined to be safe, the corresponding puncture positioning point data is output.

[0160] The liver cirrhosis ascites pressure dynamic measurement system also includes an intraoperative feedback module, which includes a needle tip sensing unit and a real-time comparison unit, for collecting needle tip pressure data and comparing it with the predicted path data in real time, and outputting a needle tip adjustment guide signal when the deviation exceeds a preset tolerance, to provide dynamic guidance for the doctor, with a response time from pressure anomaly detection to guide signal output ≤50 ms.

[0161] In an actual application, the doctor obtains the patient's abdominal cavity ultrasound image, real-time ascites pressure curve, and blood flow pulsation signal through the acquisition module of the system. After aligning the respiratory fluctuation features with the end-expiratory frame of the image, the calibration processing module selects the candidate regions in the abdominal cavity that meet the fluctuation rules. The de-correlation analysis module further excludes regions highly related to blood vessel pulsation signals to obtain safe target ascites cavity data. Subsequently, the path prediction module reconstructs the patient's three-dimensional abdominal cavity model, optimizes the puncture path in combination with the database, and the final determination module determines the pressure gradient and blood vessel correlation of the path to output the recommended puncture positioning point. If the path does not match the prediction during the operation, the feedback module provides real-time adjustment direction and amplitude, thereby realizing safe and reliable puncture operation.

[0162] Through the above system composition, the application not only realizes the multimodal fusion of image data, pressure dynamic data and blood flow pulsation data, but also effectively avoids the puncture needle into the high-risk areas such as blood vessels or intestinal tubes through multiple safety verifications such as candidate region screening, decorrelation analysis, individualized path prediction and pressure gradient judgment. At the same time, combined with the historical database and real-time feedback function, the system has predictability before operation and dynamic adaptability during operation, greatly improves the safety and success rate of liver cirrhosis ascites puncture operation, and has good clinical application value.

[0163] The technical scope of the present application is not limited to the content in the above description, and those skilled in the art can make various modifications and changes to the above embodiments without departing from the technical idea of the present application, and these modifications and changes should all belong to the protection scope of the present application.

Claims

1. A method for dynamic measurement of ascites pressure in cirrhosis, characterized by: The method comprises the following steps: acquiring abdominal cavity image data, ascites pressure dynamic data and abdominal cavity blood flow pulsation data of a patient, and extracting an anatomic region in the abdominal cavity image data and a periodic fluctuation feature of the ascites pressure dynamic data; extracting a respiratory fluctuation component and a heart beat fluctuation component in the periodic fluctuation feature; after filtering the respiratory fluctuation component and the heart beat fluctuation component, obtaining an aligned time sequence relationship through time sequence alignment; based on the aligned time sequence relationship, delineating a phase change with respiratory motion in the abdominal cavity image data, taking an anatomic region with a pressure fluctuation amplitude greater than a preset fluctuation threshold and a stable fluctuation law as candidate region data, and extracting a candidate fluctuation feature of the candidate region data; performing phase decorrelation analysis on the candidate fluctuation feature and the abdominal cavity blood flow pulsation data to exclude regions similar to blood vessel pulsation features to obtain target cavity region data after decorrelation processing; constructing a puncture path prediction model based on the target cavity region data and outputting prediction path data; performing safety determination on the prediction path data, and when the prediction path data maintains a stable pressure gradient within a preset path length range and the prediction path data satisfies a decorrelation threshold condition, determining corresponding puncture positioning point data and outputting the same.

2. The method of claim 1, wherein: The specific steps of filtering the respiratory fluctuation component and the heart beat fluctuation component in the method and obtaining an aligned time sequence relationship through time sequence alignment comprise: setting a frequency band of the respiratory fluctuation component to 0.1-0.5 Hz and a frequency band of the heart beat fluctuation component to 0.8-2.0 Hz, and obtaining corresponding fluctuation amplitude features and phase features respectively; filtering the fluctuation amplitude features and the phase features through band-pass filtering to obtain filtered pressure phase features; performing time sequence matching on the pressure phase features and synchronously collected abdominal cavity image data to generate an aligned time sequence relationship; wherein the time sequence matching comprises marking image frames at end-expiratory and end-inspiratory phases in the abdominal cavity image data, and aligning the image frames with corresponding wave peak and wave trough phase points in the pressure phase features.

3. The method of claim 2, wherein: The specific steps of performing phase decorrelation analysis on the candidate fluctuation feature and the abdominal cavity blood flow pulsation data to exclude regions similar to blood vessel pulsation features to obtain target cavity region data after decorrelation processing comprise: calculating a cross-correlation coefficient of the candidate fluctuation feature and the abdominal cavity blood flow pulsation data; calculating an average value of the cross-correlation coefficient within a preset time domain sliding window; when the average value is lower than a preset decorrelation threshold value, determining that the candidate region data is far away from blood vessels, and taking the candidate region data as the target cavity region data.

4. The method of claim 1, wherein: The specific steps of constructing a puncture path prediction model based on the target cavity region data and outputting prediction path data comprise: performing three-dimensional reconstruction on the target cavity region data to generate the puncture path prediction model; the puncture path prediction model is a three-dimensional space model comprising a skin surface, an abdominal wall layer, an intestinal drift region and a target ascites pool. an initial path is planned from a body surface puncture point of a current patient to a center of a target cavity in the puncture path prediction model; the initial path is coupled with respiratory motion of the patient, a drifting motion of an intestinal tract relative to the initial path at different respiratory phases is simulated, a minimum distance between each point of the initial path and the drifting intestinal tract is calculated, and the initial path is adjusted to an optimal path with a lowest risk according to the minimum distance, and output as the prediction path data.

5. The method of claim 1, wherein: The step of constructing the puncture path prediction model based on the target cavity region data further comprises: an individualized database based on historical successful puncture cases is established, and the database at least stores data features of patients with different body characteristics, the data features including safe puncture path data and corresponding ascites pressure dynamic data; in the step of constructing the puncture path prediction model, the database is called, the body characteristics of the current patient and the data features are matched with the database, and the prediction path data is optimized and corrected through the matching result.

6. The method of claim 1, wherein: In the safety judgment of the prediction path data output by the puncture path prediction model, the judgment method that the prediction path data maintains a stable pressure gradient within a preset path length range comprises: within the preset path length range, a pressure difference between adjacent sampling points on the prediction path data is calculated; when the absolute values of pressure differences between all continuous sampling points are less than a pressure gradient threshold value, it is determined that the pressure gradient is stable, wherein the pressure gradient threshold value is adaptively adjusted according to an average amplitude of the ascites pressure dynamic data.

7. The method of claim 4, wherein the method further comprises: The preset path length range is determined by the following method: based on the abdominal cavity image data, the total thickness of abdominal wall tissue from the body surface puncture point to the entrance of the ascites cavity region in the target cavity region data is measured; the preset path length range is set as a total length from the body surface puncture point, extending into the body along the prediction path data, covering the total thickness of the abdominal wall tissue and additionally increasing a safety margin distance; wherein the safety margin distance is a fixed value preset according to clinical experience, or a dynamically calculated value proportional to the depth of the target ascites region.

8. The method of claim 1, wherein: The ascites pressure dynamic measurement method for liver cirrhosis further comprises: an individualized prediction model is constructed based on historical puncture data of the patient, and risk score data output by the individualized prediction model is fused with the target cavity region data; when the risk score of the fusion result is lower than a preset fusion threshold value, individualized recommended puncture positioning point data is output.

9. The method of claim 1, wherein: The ascites pressure dynamic measurement method for liver cirrhosis further comprises an intraoperative real-time feedback step: in the actual puncture process, real-time needle tip pressure data advancing along the prediction path data is acquired in real time; fluctuation characteristics of the needle tip real-time pressure data are compared in real time with expected pressure fluctuation characteristics of corresponding position points in the prediction path data; When the deviation of the comparison result exceeds a preset tolerance, a needle tip adjustment instruction signal is generated and output, the instruction signal including a direction and magnitude of the needle tip adjustment, the magnitude being calculated based on the size of the deviation and the instantaneous correlation of the needle tip real-time pressure data and the abdominal blood flow pulsation data of the adjacent blood vessel.

10. A liver cirrhosis ascites pressure dynamic measurement system, characterized in that: The application discloses a liver cirrhosis ascites pressure dynamic measurement method, comprising the following steps: An acquisition module is configured to acquire abdominal cavity image data, ascites pressure dynamic data and abdominal blood flow pulsation data of a patient, and extract an anatomical region in the abdominal cavity image data and a periodic fluctuation feature of the ascites pressure dynamic data; An alignment module is configured to extract a respiratory fluctuation component and a heart beat fluctuation component in the periodic fluctuation feature, filter the respiratory fluctuation component and the heart beat fluctuation component, and obtain an aligned time sequence relationship through time sequence alignment; A calibration processing module is configured to draw a phase change with respiratory motion in the abdominal cavity image data according to the aligned time sequence relationship, take an anatomical region with a pressure fluctuation amplitude greater than a preset fluctuation threshold and a stable fluctuation law as candidate region data, and extract a candidate fluctuation feature of the candidate region data; A decorrelation analysis module is configured to perform phase decorrelation analysis on the candidate fluctuation feature and the abdominal blood flow pulsation data, and exclude regions similar to blood vessel pulsation features to obtain target abdominal cavity region data after decorrelation processing; A path prediction module is configured to construct a puncture path prediction model based on the target abdominal cavity region data and output prediction path data; A determination module is configured to perform safety determination on the prediction path data output by the puncture path prediction model, determine corresponding puncture positioning point data and output the corresponding puncture positioning point data when the prediction path data remains stable in a pressure gradient within a preset path length range and the prediction path data satisfies a decorrelation threshold condition.

Images