Early warning method for preventing urinary catheterization based on real-time pressure feedback

Abstract

The invention relates to an early warning method for preventing urinary retention catheterization based on real-time pressure feedback, and belongs to the technical field of medical intelligent monitoring. In order to solve the problem, the invention provides the early warning method for urinary retention prevention catheterization based on real-time pressure feedback, a bladder internal pressure monitoring sensor is arranged on the basis of a catheter body, and pressure gradient changes of different areas of the bladder wall are captured in real time by detecting a plurality of sensing nodes at the front end of the catheter; pressure feedback is carried out when the urinary retention bladder capacity is critical; catheterization is started or ended according to a pressure feedback result; reverse pressure is generated immediately after urethral catheterization is finished, and urine is prevented from flowing back; a bladder state classification algorithm based on LSTM is adopted to analyze pressure-urination data of a patient within 72 hours, a'pressure-urination frequency 'periodic model is automatically generated, and urinary retention and detrusor myasthenia caused by transitional catheterization are prevented in a multiple early warning mode.

Description

TECHNICAL FIELD

[0001] The application relates to a real-time pressure feedback-based early warning method for preventing urinary retention catheterization, and belongs to the technical field of medical intelligent monitoring. BACKGROUND

[0002] After the surgery of the urology department, especially after the radical prostatectomy, the patient usually indwells a catheter for 7-14 days, which serves the purpose of draining urine and compressing hemostasis. The catheter belongs to the catheter instrument in the field of urology medical instruments. The catheter currently used after the radical prostatectomy mainly comprises a catheter main body, a gas injection balloon and a drainage connection outlet, wherein the catheter main body comprises a main drainage cavity, an auxiliary cavity, a tip and a side hole, a catheter scale and a length; the drainage connection outlet comprises a drainage port connected with a urine bag and an auxiliary cavity gas injection port used for fixing the catheter. The catheter can help the patient after the radical prostatectomy to relieve the temporary dysuria and reduce the anastomotic tension to promote healing.

[0003] Clinical data shows that the incidence of urinary retention of the postoperative patient is as high as 15%-30%, and the incidence of bladder dysfunction of the patient with long-term indwelling catheter is about 40%. If the intravesical pressure continuously exceeds 40cmH2O, it will lead to insufficient blood perfusion of the bladder wall, cause the bladder mucosa ischemia and ulcer, and even cause irreversible renal function damage. The traditional catheter is only passive drainage, and cannot sense the intravesical pressure change and predict the postoperative bladder function, so the postoperative patient is prone to urinary retention due to the weakness of the bladder detrusor or urethral spasm. The clinical observation or ultrasonic examination by the medical staff is delayed, which may cause bladder injury or infection. At present, how to dynamically detect the intravesical pressure and automatically drain according to the intravesical pressure change is very important, so as to train the bladder function during the catheterization, reduce the burden of the medical staff and prevent the occurrence of urinary retention.

[0004] At present, the non-drug method, drug treatment, invasive operation and special evaluation method are mainly adopted for the urinary retention after the catheter is pulled out after the surgery, but each method has defects to some extent.

[0005] The non-drug method includes behavior induction methods such as listening to flowing water sound, warm water flushing of the perineum, hot compress of the lower abdomen (water temperature 60-70℃) and the like; physical therapy such as low-frequency electric pulse bladder treatment instrument, infrared or spectrum instrument irradiation of the bladder area; body position adjustment to help the patient to sit up or get out of bed to adopt the habitual posture to urinate; bladder function training such as the training of the bladder filling sensation by the time clamping of the catheter.

[0006] The drug treatment includes the following aspects: cholinesterase inhibitors such as neostigmine acupoint injection at Futanli; alpha receptor blockers such as tamsulosin or alfuzosin; intravesical drug perfusion such as 2% lidocaine+ofloxacin perfusion, which has the functions of relieving spasm and resisting infection; and defecation drugs such as Kaiselu (40ml) inserted into the anus to promote the micturition reflex through the rectal stimulation.

[0007] Invasive operation: clean intermittent catheterization (CIC) is the international recommended gold standard, which can reduce the risk of infection. Indwelling catheterization is used for patients with PVR>500ml or acute urinary retention. Urethral obstruction is used for patients with urethral obstruction.

[0008] However, the above method has the following defects:

[0009] First, warm water flushing perineum may increase the risk of falls in elderly patients with weak balance ability; patients after prostate cancer surgery have surgical wounds in the lower abdomen, and wound pain makes it impossible to get out of bed, and the supine urination posture does not conform to human physiology, which reduces the success rate of urination; treatment instrument, infrared and other strong dependence on machines, improper placement of electrodes can cause skin burns; repeated timing of catheter clamping may induce bladder spasm and increase nursing burden;

[0010] Second, part of the drug treatment, prostate cancer patients are generally older, and have more underlying diseases, and the use of multiple drugs may cause drug antagonism. The problem of rectal administration of Kaiselu, Kaiselu is usually administered by family members, which may ignore the measurement problem, and there is no individualized Kaiselu administration standard, and long-term rectal stimulation may also affect defecation.

[0011] Third, although clean intermittent catheterization (CIC) is marked as the gold standard, it is easy to lead to non-standard operation during operation, which may increase the infection rate.

[0012] Therefore, the existing catheterization device only realizes the catheterization function after urinary retention occurs, and cannot effectively prevent urinary retention, and is prone to infection risk and operational damage.

[0013] In summary, the traditional catheter causes the trigone area of the bladder to be compressed by the air bag during long-term indwelling, stimulates abnormal contraction of the detrusor muscle, lacks real-time monitoring of intravesical pressure, cannot dynamically adjust the drainage strategy, ignores the physiological rhythm of the bladder in the passive drainage mode, and leads to functional degradation. SUMMARY

[0014] To solve the above problems, the application provides a real-time pressure feedback-based early warning method for preventing urinary retention and catheterization, a bladder internal pressure monitoring sensor is arranged on the basis of a catheter body, a plurality of sensing nodes at the front end of the catheter are detected to capture the pressure gradient changes of different regions of the bladder wall in real time, and the bladder internal pressure value is calculated, and pressure feedback is performed when the urinary retention bladder capacity critical value is reached; the catheterization is started or ended according to the pressure feedback result; a magnetic suspension micro electromagnetic valve is used to generate reverse pressure immediately after the catheterization is ended to prevent urine backflow, and a one-way valve at the end of the catheter body is used to form a double anti-backflow structure; a bladder state classification algorithm based on LSTM is used to analyze the pressure-urination data of the patient within 72 hours to automatically generate a “pressure-urination frequency” periodic model, which is transmitted to a medical system to prevent urinary retention and detrusor weakness caused by transitional catheterization in a multiple early warning manner.

[0015] The catheter body has a three-cavity structure, including a urine drainage channel, a pressure sensing channel and a water injection channel for the air bag, the urine drainage channel is used to guide the urine in the bladder out of the body, the pressure sensing channel is directly communicated with the inside of the bladder, and the water injection channel for the air bag is used to be fixed in the bladder.

[0016] Further, the bladder internal pressure monitoring sensor adopts a flexible piezoresistive sensor array based on a PDMS substrate, the gradient substrate prepared by the layering curing technology processing can form a modulus gradient from the sensor side to the tissue contact side, and can meet the support stiffness of the sensor circuit (avoid signal distortion) and the flexible fitting requirement of the tissue contact side, reducing the interface sliding caused by mechanical mismatch.

[0017] Further, the PDMS substrate is a PDMS film prepared by a spin coating technology, which reduces the bending stiffness, so that the sensor can be more easily curled, folded and passed through the narrow catheter, and more compliantly fit the complex curved surface in the bladder without breaking or significantly affecting the measured bladder internal pressure, eight sensing nodes are annularly distributed on the PDMS substrate and located on the water injection air bag pipe wall at the front end of the catheter body, each node integrates a pressure-temperature dual parameter detection unit and can capture the pressure gradient changes of different regions of the bladder wall in real time; when the water injection air bag is inflated, the PDMS substrate unfolds the curled / folded sensor array and pushes it to the bladder wall, that is, the annularly distributed sensing nodes of the water injection air bag are unfolded and fall on the inner surface of the bladder, so that the pressure at the bottom of the bladder is maximally captured, so that the sensing nodes can play the pressure sensing function and be used for monitoring the pressure condition of the position of the bladder wall.

[0018] Further, the flexible piezoresistive sensor array comprises a plurality of flexible piezoresistive sensors arranged in a matrix array, the flexible piezoresistive sensors capable of changing the resistance of the PDMS substrate with the pressure or mechanical stress (strain) received, the resistance value change having a nonlinear relationship with the pressure in the bladder, showing that as the pressure in the bladder increases, the resistance value decreases; by measuring the change of the resistance, the pressure received by the PDMS substrate is calculated, and thus the pressure value in the bladder is obtained, and pressure feedback is performed when the bladder capacity of urinary retention reaches the critical value.

[0019] Further, when the pressure detected by the flexible piezoresistive sensor is ≥45 cmH2O, the catheterization is started, and when the detected pressure drops to 15 cmH2O, the catheterization is ended to prevent overfilling of the bladder; a magnetic suspension micro electromagnetic valve with a response time <50 ms is used to generate a 0.2 cmH2O reverse pressure immediately after the catheterization is completed to prevent urine backflow, and a one-way valve at the end of the catheter body is used to form a double anti-backflow structure.

[0020] Further, a bladder state classification algorithm based on LSTM is used to analyze the pressure-urination data of the patient within 72 hours, specifically including:

[0021] Setting the relationship between time series and numerical observations;

[0022] Using LSTM to perform time series analysis on bladder pressure to automatically identify the current state in the bladder (empty, full, and urinating) and capture the time series nature of the pressure signal;

[0023] The bladder pressure data is collected in real time by the PDMS substrate piezoresistive sensor array, and when the flexible piezoresistive sensor array is attached to the inner wall of the bladder, it will produce fluctuations in resistance or voltage signals with changes in pressure. The flexible piezoresistive sensor outputs time series signals at a fixed frequency (such as 10 Hz), and the time series signal at each time point contains multi-channel data corresponding to different positions of the sensor array, forming a multi-dimensional time series (such as a matrix dimension of [time point x channel number]).

[0024] Further, LSTM is used to perform time series analysis on bladder pressure, including:

[0025] Divide the continuous pressure data into sequences of fixed length and integrate them into multi-dimensional time series;

[0026] Decide to discard historical pressure change information (such as ignore the stable baseline) through the forget gate;

[0027] Filter important features (such as sudden pressure peaks) in the current pressure value as current information through the input gate, fuse historical change information and current information, and form new memory data (such as record a rising trend);

[0028] The hidden state is generated by the output gate and passed to the next time step (e.g., outputting the current pressure fluctuation feature);

[0029] After outputting the current pressure fluctuation, the bladder pressure data is analyzed in real time by the LSTM model;

[0030] The real-time collected pressure data is input into the model to judge the current bladder state and predict the future bladder pressure trend within a certain period of time to identify the occurrence of urination events;

[0031] By continuously inputting real-time data within 72 hours, the corresponding pressure urination data is continuously output, including the change of pressure value over time, the urination time, and the duration of each urination. The model output results need to be compared and evaluated with the actual situation. According to the evaluation results, the model parameters are further adjusted or the model structure is optimized to improve the prediction accuracy. Finally, reliable data of pressure urination within 72 hours is obtained.

[0032] Further, an automatic "pressure-urination frequency" cycle model is generated, including a prediction-update cycle of the patient's 72-hour pressure-urination data by Kalman filtering, fusion of bladder pressure monitoring sensor observation values and system dynamics model to obtain the optimal estimated volume; local time sequence features are extracted on the time axis by 1D-CNN; a curve is fitted by LSTM to integrate pressure and estimated volume data points, and further prompts are given according to the curve; 1D-CNN and LSTM are combined to resist interference and visualize features.

[0033] Further, the multiple warnings include:

[0034] Acute warning: when the pressure is continuously > 40 cmH2O, it indicates that the bladder is overfilled or the detrusor muscle is abnormally contracted, which may cause urinary retention or bladder damage;

[0035] Chronic warning: when the long-term pressure is > 20 cmH2O, it indicates neurogenic bladder dysfunction, which may cause complications such as hydronephrosis;

[0036] The pressure sequence is tracked by the memory unit, and when the predicted value of multiple consecutive times (e.g., 5-10 seconds) exceeds the pressure threshold, the multiple warnings are triggered.

[0037] Further, further prompts are given according to the curve, specifically including:

[0038] When the curve slope in the LSTM output curve increases sharply, ΔP / ΔV > 15 cmH2O / 100 ml, it indicates urethral obstruction or detrusor-sphincter coordination disorder; if the curve slope drops sharply and maintains low pressure, it indicates urinary incontinence or bladder contraction dysfunction;

[0039] When the smooth rising curve of the filling period and the peak shape of the urination period (LSTM classified as "non-urination period pressure fluctuation" + amplitude > 40 cmH2O) indicate that bladder overactivity may occur; frequent irregular contractions (such as more than 3 times per minute of small amplitude pressure fluctuations) are often seen in interstitial cystitis with sustained high pressure plateau (pressure maintained > 30 cmH2O and no micturition reflex)

[0040] When the pressure is sustained > 40 cmH2O and the LSTM urination period classification is delayed when the volume is high, it indicates that there may be outlet obstruction.

[0041] Advantages of the present application:

[0042] 1. Multi-dimensional pressure sensing and real-time monitoring mechanism: The sensing method uses an implantable or body surface near-field coupled flexible piezoresistive sensor array to directly measure the pressure change in the bladder or the corresponding area on the body surface, with an accuracy of ± 2 cmH2O (better than traditional finger pressure palpation); combined with a flexible sensing material PDMS base, which is attached to the surface of the bladder, the filling state of the bladder is inferred by the change in pressure distribution.

[0043] 2. Based on individual differences of patients (such as age, cause, bladder volume), an individualized pressure threshold model is established by LSTM-based bladder state algorithm to distinguish between "normal filling pressure" and "urinary retention warning pressure"; at the same time, combined with historical pressure data (such as 72-hour cycle model in user demand), the correlation between urination frequency and pressure change is analyzed to predict the best catheterization time and reduce the number of catheterizations.

[0044] 3. Closed-loop catheterization execution system: pressure feedback driven automatic operation: when the pressure exceeds the threshold, a magnetic suspension micro electromagnetic valve is used to control the automatic opening of the catheterization channel, avoiding the delay of manual operation; combined with an anti-reflux design (one-way valve), to prevent infection caused by urine reflux; according to the pressure drop rate, the catheterization flow rate is dynamically adjusted (such as reducing the flow rate when the pressure drops rapidly to avoid hematuria caused by sudden decompression of the bladder), improving safety.

[0045] 4. Hardware: modular design is adopted, integrating pressure sensors, microcontrollers (FPGA), and wireless transmission modules (433 MHz medical frequency band) into a wearable body surface module with a thickness of ≤8 mm and a weight of ≤20 g. BRIEF DESCRIPTION OF DRAWINGS

[0046] Figure 1 An information filtering diagram in the information processing flow of the LSTM algorithm in an embodiment of the present application.

[0047] Figure 2 A new information retention diagram in the information processing flow of the LSTM algorithm in an embodiment of the present application.

[0048] Figure 3It is a prediction data basis graph in the information processing flow of the LSTM algorithm in an embodiment of the present application.

[0049] Figure 4 It is a prediction data basis graph in the information processing flow of the LSTM algorithm in an embodiment of the present application. DETAILED DESCRIPTION

[0050] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are some of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work fall within the protection scope of the present application.

[0051] In the present application, unless specifically defined and limited otherwise, the terms "connected", "connected", "fixed" should be understood in a broad sense, for example, can be fixedly connected, can also be detachably connected, or integrated; can be mechanically connected, can also be electrically connected; can be directly connected, can also be indirectly connected through an intermediate medium, can be the internal communication of two elements or the interaction relationship between two elements. For those skilled in the art, the specific meaning of the above terms in the present application can be understood according to the specific circumstances.

[0052] In the present application, unless specifically defined and limited otherwise, the first feature "on" or "below" the second feature can include that the first and second features are in direct contact, or can include that the first and second features are not in direct contact but are in contact through another feature between them. Moreover, the first feature "on", "above" and "above" the second feature includes that the first feature is directly above and obliquely above the second feature, or only indicates that the horizontal height of the first feature is higher than that of the second feature. The first feature "below", "below" and "below" the second feature includes that the first feature is directly below and obliquely below the second feature, or only indicates that the horizontal height of the first feature is less than that of the second feature.

[0053] To solve the above problems, the application provides a real-time pressure feedback-based early warning method for preventing urinary retention and catheterization, which is based on a bladder internal pressure monitoring sensor arranged on a catheter body, detects a plurality of sensing nodes at the front end of the catheter, captures pressure gradient changes of different regions of the bladder wall in real time, calculates the bladder internal pressure value, and performs pressure feedback when the urinary retention bladder capacity critical value is reached; starts catheterization or ends catheterization according to the pressure feedback result; adopts a magnetic suspension micro electromagnetic valve to generate reverse pressure immediately after ending catheterization to prevent urine backflow, and cooperates with a one-way valve at the end of the catheter body to form a double anti-backflow structure; adopts a bladder state classification algorithm based on LSTM to analyze the pressure-urination data of a patient within 72 hours, automatically generates a "pressure-urination frequency" cycle model, and transmits it to a medical system to prevent urinary retention and detrusor weakness caused by transitional catheterization in a multiple early warning manner.

[0054] The early warning method forms a closed-loop system framework by sensing based on the bladder internal pressure monitoring sensor, decision-making based on the LSTM-based bladder state classification algorithm, and executing catheterization.

[0055] In some embodiments, the LSTM refers to an LSTM algorithm, including data preprocessing (denoising, normalization), periodic feature extraction (such as night / day pressure fluctuation rules), threshold dynamic adjustment (detrusor contraction strength judgment based on ΔP / Δt slope); according to the maximum safe pressure (Pmax) of the patient's bladder and the individual detrusor function parameters, a multi-level early warning threshold is set, wherein P1=0.7×Pmax, P2=0.9×Pmax, and when the pressure exceeds P2, catheterization is automatically triggered.

[0056] The catheter body has a three-cavity structure, including a urine drainage channel, a pressure sensing channel, and a gas bag water injection channel, the urine drainage channel is used to guide the urine in the bladder out of the body, the pressure sensing channel is directly communicated with the inside of the bladder, and the gas bag water injection channel is used to be fixed in the bladder.

[0057] The inner diameter of the urine drainage channel is made of medical silica gel which is soft and biocompatible, ensuring smooth drainage of urine and minimizing irritation to the urethra. This not only reduces the discomfort of the patient, but also has good durability and corrosion resistance, allowing long-term indwelling in the body. Preferably, the front end of the urinary catheter is designed in a smooth streamline shape, and the surface is specially treated to reduce resistance when inserted into the urethra, reducing the patient's pain. At the front end near the bladder, multiple side holes are provided, which are evenly distributed to effectively prevent the influence of urine drainage caused by catheter blockage. At the same time, the size and number of side holes are strictly calculated and tested to ensure smooth urine outflow and prevent bladder mucosa from being sucked into the side holes, causing damage. The surface of the catheter is covered with a PH-responsive hydrogel, which has a lubricating effect at normal pH, bacterial colonization when urine is stored, and triggers an "infection warning" when the acidic environment pH>7.5, releasing chitosan to reduce the rate of bacterial infection.

[0058] Further, the intravesical pressure monitoring sensor adopts a flexible piezoresistive sensor array with a PDMS substrate, which is a gradient substrate prepared by a layered curing technology. The elastic modulus gradient can be formed from the sensor side to the tissue contact side, while meeting the support stiffness of the sensor circuit (avoiding signal distortion) and the flexible fitting requirements of the tissue contact side, reducing the interface sliding caused by mechanical mismatch.

[0059] Further, the PDMS substrate is a PDMS film prepared by spin coating technology, which reduces the bending stiffness, making the sensor easier to curl, fold and pass through the narrow urinary tube, and more compliantly fit the complex curved surface in the bladder without breaking or significantly affecting the measured intravesical pressure. Eight sensing nodes are distributed in a ring shape on the PDMS substrate and located on the wall of the water injection balloon at the front end of the urinary catheter. Each node integrates a pressure-temperature dual parameter detection unit and can capture the pressure gradient changes of different regions of the bladder wall in real time. When the water injection balloon is inflated, the PDMS substrate unfolds the curled / folded sensor array and pushes it towards the bladder wall, i.e. the ring-shaped sensing nodes on the water injection balloon are unfolded and fall on the inner surface of the bladder, maximizing the pressure on the bladder bottom, so that the sensing nodes can perform pressure sensing function to monitor the pressure at the location of the bladder wall.

[0060] Further, the flexible piezoresistive sensor array includes a plurality of flexible piezoresistive sensors arranged in a matrix array. The resistance of the flexible piezoresistive sensor changes with the pressure or mechanical stress (strain) on the PDMS substrate. The resistance value changes nonlinearly with the pressure in the bladder, i.e. the resistance value decreases as the intravesical pressure increases. By measuring the change of resistance, the pressure on the PDMS substrate can be calculated, and thus the intravesical pressure value can be obtained, and pressure feedback can be performed when the urinary retention bladder capacity reaches the critical value.

[0061] Further, when the flexible piezoresistive sensor detects a pressure ≥ 45 cmH2O, the catheterization is started, and when the detected pressure drops to 15 cmH2O, the catheterization is ended, preventing overfilling of the bladder; a magnetic levitation micro electromagnetic valve with a response time < 50 ms generates a 0.2 cmH2O reverse pressure immediately after the catheterization is completed, preventing urine backflow, and a one-way valve at the end of the catheter body forms a double backflow prevention structure.

[0062] In some embodiments, the flexible piezoresistive sensor array is integrated with a catheter, for example, the flexible piezoresistive sensor is built into the top end of the catheter, which can monitor the pressure changes in real time during catheterization and prevent over-drainage.

[0063] Further, an LSTM-based bladder state classification algorithm is used to analyze the pressure-urination data of the patient within 72 hours, which specifically includes:

[0064] Setting the relationship between time series and numerical observations;

[0065] Using LSTM to perform time series analysis on bladder pressure, thereby automatically identifying the current state in the bladder (empty, full, urinating), and capturing the time series nature of the pressure signal;

[0066] The bladder pressure data is collected in real time by the PDMS-based piezoresistive sensor array. When the flexible piezoresistive sensor array is attached to the inner wall of the bladder, it will generate fluctuations in resistance or voltage signals as the pressure changes. The flexible piezoresistive sensor outputs time series signals at a fixed frequency (e.g., 10 Hz). Each time point's time series signal contains multi-channel data corresponding to different positions of the sensor array, forming a multi-dimensional time series (e.g., matrix dimension [time point x channel number]).

[0067] Further, LSTM is used to perform time series analysis on bladder pressure, including:

[0068] Divide the continuous pressure data into sequences of fixed length and integrate them into multi-dimensional time series;

[0069] Decide to discard historical pressure change information (e.g., ignore the stable baseline) through the forget gate:

[0070] f t =α(W f [h t-1 ,x t +b f ])

[0071] Filter important features (e.g., sudden pressure peaks) in the current pressure value as current information through the input gate, fuse historical change information and current information, and form new memory data (e.g., record a rising trend):

[0072] i t = a(W i · [h t-1 , x t ] + b i );

[0073] C t = tanh(W c · [h t-1 , x t ] + b c );

[0074]

[0075] Generate hidden state through output gate, pass to next time step (such as output current pressure fluctuation feature):

[0076] O t = a(W o · [h t-1 , x t ] + b o );

[0077] h t = O t x tanh(B t );

[0078] After outputting the current pressure fluctuation, analyze the bladder pressure data in real time through the LSTM model;

[0079] Input the real-time collected pressure data into the model, judge the current bladder state, predict the bladder pressure change trend in the future period of time, and identify the occurrence of urination event;

[0080] By continuously inputting real-time data within 72 hours, continuously output corresponding pressure urination data, including pressure value change over time, urination time, and each urination duration, and then compare and evaluate the model output results with the actual situation, further adjust the model parameters or optimize the model structure to improve the prediction accuracy, and finally obtain reliable data of pressure urination within 72 hours.

[0081] Further, automatically generate a "pressure-urination frequency" cycle model, including a prediction-update cycle of the patient's 72-hour pressure-urination data by Kalman filtering, fusion of bladder pressure monitoring sensor observation values and system dynamics model to obtain the optimal estimated volume; extract local time sequence features on the time axis by 1D-CNN; fit a curve graph by LSTM integrating pressure and estimated volume data points, and give further prompts according to the curve graph; combination of 1D-CNN and LSTM can resist interference and visualize features.

[0082] Further, the multiple early warnings include:

[0083] Acute early warning, when pressure persists > 40 cmH2O, suggesting overfilling of the bladder or abnormal contraction of the detrusor muscle, which may lead to urinary retention or bladder injury;

[0084] Chronic early warning, when long-term pressure > 20 cmH2O, suggesting neurogenic bladder dysfunction, which may lead to complications such as hydronephrosis;

[0085] By tracking the pressure sequence with the memory unit, when the predicted value of multiple consecutive times (such as 5-10 seconds) exceeds the pressure threshold, the multiple early warnings are triggered.

[0086] Further, further prompts are given according to the graph, specifically including:

[0087] When the curve slope in the graph output by the LSTM increases sharply, ΔP / ΔV > 15 cmH2O / 100 ml, suggesting urethral obstruction or detrusor-sphincter dyssynergia; if the curve slope drops sharply and maintains a low pressure, it suggests urinary incontinence or bladder contraction dysfunction;

[0088] When the smooth rising curve in the filling period and the sharp peak shape in the voiding period (LSTM classified as "non-voiding period pressure fluctuation" + amplitude > 40 cmH2O), it suggests that bladder overactivity may occur; frequent irregular contractions (such as > 3 times per minute of small amplitude pressure fluctuations) are often seen in interstitial cystitis with sustained high pressure plateau (pressure maintained > 30 cmH2O and no micturition reflex)

[0089] When the pressure persists > 40 cmH2O at high capacity and the LSTM voiding period classification is delayed, it suggests possible outlet obstruction.

[0090] In some embodiments, the present application is based on a medical data encryption transmission protocol in the 433 MHz frequency band, including pressure data frame format (frame header + pressure value + timestamp + CRC check) and anti-interference mechanism (frequency hopping spread spectrum FHSS). The data transmission of the present application can be transmitted to the body surface module (a modular device attached to the surface of the human body (such as skin, clothing, etc.), used for data acquisition, signal processing or wireless transmission function) through near field coupling (wireless energy or data transmission technology of near field interaction of electromagnetic field), and then transmitted wirelessly to the medical system through a medical dedicated frequency band (such as 433 MHz, 2.4 GHz), to display the bladder pressure curve in real time, assisting the treatment of urinary incontinence or neurogenic bladder patients. Bladder pressure monitoring sensor → near field coupling transmission (in vivo → body surface) → body surface module (signal processing + wireless transmission) → medical dedicated frequency band (433 MHz / 2.4 GHz) → medical system (reception + analysis).

[0091] Although the present application has been disclosed in its preferred embodiments with reference to the accompanying drawings, it is not intended to limit the present application thereto, and various modifications and alterations can be made thereto by those skilled in the art without departing from the spirit and scope of the present application, and the scope of protection of the present application should be defined by the appended claims.

Claims

1. A real-time pressure feedback based early warning method for preventing urinary retention and catheterization, characterized in that, The bladder pressure monitoring sensor is arranged based on the catheter body, the pressure gradient changes of different regions of the bladder wall are captured in real time by detecting a plurality of sensing nodes at the front end of the catheter, and the bladder pressure value is calculated, and pressure feedback is performed when the urinary retention bladder capacity critical value is reached; the catheterization is started or the catheterization is ended according to the pressure feedback result; a magnetic suspension micro electromagnetic valve is used to generate reverse pressure immediately after the catheterization is ended to prevent urine backflow, and a one-way valve at the end of the catheter body is used to form a double anti-backflow structure; a bladder state classification algorithm based on LSTM is used to analyze the pressure-urination data of a patient within 72 hours to automatically generate a 'pressure-urination frequency' cycle model, which is transmitted to a medical system to prevent urinary retention and detrusor weakness caused by transitional catheterization in a multiple warning manner; The catheter body has a three-cavity structure, including a urine drainage channel, a pressure sensing channel and a gasbag water injection channel, the urine drainage channel is used to guide the urine in the bladder out of the body, the pressure sensing channel is directly communicated with the inside of the bladder, and the gasbag water injection channel is used to be fixed in the bladder.

2. The early warning method of claim 1, wherein, The bladder pressure monitoring sensor adopts a flexible piezoresistive sensor array based on a PDMS substrate, the PDMS substrate is a gradient substrate prepared by a layered curing technology, and the elastic modulus gradient can be formed from the sensor side to the tissue contact side, while the support stiffness of the sensor circuit and the flexible fitting requirement of the tissue contact side can be met, and the interface sliding caused by mechanical mismatch is reduced.

3. The early warning method of claim 2, wherein, The PDMS substrate is a PDMS film prepared by a spin coating technology, eight sensing nodes are annularly distributed on the PDMS substrate and located on the water injection gasbag pipe wall at the front end of the catheter body, each node integrates a pressure-temperature double parameter detection unit and can capture the pressure gradient changes of different regions of the bladder wall in real time; when the water injection gasbag is inflated by water injection, the PDMS substrate unfolds the crimped / folded sensor array and pushes it to the bladder wall, that is, the annularly distributed sensing nodes of the water injection gasbag are unfolded and fall on the inner surface of the bladder, so that the pressure on the bladder bottom is maximally captured, and the sensing nodes can play the pressure sensing function to monitor the pressure condition of the position of the bladder wall.

4. The early warning method of claim 3, wherein, The flexible piezoresistive sensor array includes a plurality of flexible piezoresistive sensors arranged in a matrix array, the resistance of the flexible piezoresistive sensor changes with the pressure or mechanical stress, the resistance value change has a nonlinear relationship with the pressure in the bladder, and the resistance value decreases with the increase of the pressure in the bladder; the pressure on the PDMS substrate is calculated by measuring the change of the resistance, so that the bladder pressure value is obtained, and pressure feedback is performed when the urinary retention bladder capacity critical value is reached.

5. The early warning method of claim 4, wherein, When the pressure detected by the flexible piezoresistive sensor is greater than or equal to 45cmH2O, the catheterization is started, and when the detected pressure is reduced to 15cmH2O, the catheterization is ended; a magnetic suspension micro electromagnetic valve with a response time less than 50ms is used to generate a 0.2cmH2O reverse pressure immediately after the catheterization is ended to prevent urine backflow, and a one-way valve at the end of the catheter body is used to form a double anti-backflow structure.

6. The early warning method of claim 5, wherein, The bladder state classification algorithm based on LSTM is used to analyze the patient's pressure-urination data within 72 hours, including: Setting the relationship between time series and numerical observations; Using LSTM for time series analysis of bladder pressure to automatically identify the current state in the bladder and capture the time series nature of the pressure signal; Bladder pressure data is collected in real time by a PDMS-based piezoresistive sensor array. When the flexible piezoresistive sensor array is attached to the inner wall of the bladder, it will produce fluctuations in resistance or voltage signals as the pressure changes. The flexible piezoresistive sensor outputs time series signals at a fixed frequency. Each time point's time series signal contains multi-channel data to correspond to different positions of the sensor array, forming a multi-dimensional time series.

7. The early warning method of claim 6, wherein, Using LSTM for time series analysis of bladder pressure, including: Divide the continuous pressure data into fixed-length sequences and integrate them into multi-dimensional time series; Decide to discard historical pressure change information through the forget gate; Filter important features in the current pressure value as current information through the input gate, fuse historical change information and current information, and form new memory data; Generate hidden states through the output gate and pass them to the next time step; Real-time analysis of bladder pressure data through the LSTM model after outputting the current pressure fluctuation; Input the real-time collected pressure data into the model to judge the current bladder state and predict the bladder pressure change trend in the future period to identify the occurrence of urination events; By continuously inputting real-time data within 72 hours, the corresponding pressure-urination data is continuously output, including pressure value changes over time, urination time, and duration of each urination. The model output results need to be compared and evaluated with the actual situation. Based on the evaluation results, further adjust the model parameters or optimize the model structure to improve the prediction accuracy. Ultimately, reliable data of pressure-urination within 72 hours is obtained.

8. The early warning method of claim 7, wherein, Automatically generate a "pressure-urination frequency" cycle model, including a prediction-update cycle of patients' 72-hour pressure-urination data by Kalman filtering, which fuses the bladder pressure monitoring sensor observations and the system dynamics model to obtain the optimal estimated volume. Extract local time series features on the time axis by 1D-CNN. Fit a curve graph by LSTM integrating pressure and estimated volume data points, and give further prompts based on the curve graph.

9. The early warning method of claim 8, wherein, The multiple warnings include: Acute warning: when the pressure remains > 40 cmH2O, it indicates that the bladder is overfilled or the detrusor muscle is abnormally contracted, which may lead to urinary retention or bladder damage. Chronic warning: when the long-term pressure > 20 cmH2O, it indicates neurogenic bladder dysfunction, which may cause complications such as hydronephrosis. Track the pressure sequence through the memory unit, and trigger multiple warnings when the predicted value of multiple consecutive times exceeds the pressure threshold.

10. The early warning method of claim 8, wherein, Further prompts based on the curve graph, including: When the curve slope in the LSTM output curve graph increases sharply, ΔP / ΔV > 15 cmH2O / 100 ml, it indicates urethral obstruction or detrusor-sphincter dyssynergia. If the curve slope drops sharply and remains low, it indicates urinary incontinence or bladder contraction dysfunction. When the smooth rising curve in filling phase and the sharp peak in voiding phase, it may be a warning of overactive bladder. When the pressure is sustained >40 cmH2O in high volume and LSTM voiding phase classification is delayed, it may be an outlet obstruction.

Images