Intracranial pressure monitoring method and system based on electrical impedance tomography and medium
Through the method based on electrical impedance imaging, the cerebral blood flow dynamic parameters are extracted using the electrical impedance tomography system and the random forest algorithm, and non-invasive and harmless intracranial pressure monitoring is achieved, solving the monitoring complexity and complications in the existing technology. Risk issues provide an easy-to-use real-time monitoring tool.
Patent Information
- Application Number
- CN202411252793.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-09
- Publication Date
- 2024-11-22
- Estimated Expiration
- 2044-09-09
AI Technical Summary
The existing intracranial pressure monitoring methods are mostly invasive or non-invasive indirect methods, which have complication risks or complex operations, lack ease of use, and are difficult to achieve non-invasive and harmless long-term dynamic monitoring.
Using an electrical impedance imaging method, the extracranial boundary voltage signal is collected through the electrical impedance tomography system, and the cerebral blood flow dynamic image is reconstructed using the finite element model and the damping least squares method to extract the average perfusion speed, contraction wave amplitude, and inflow volume velocity. The angle between the descending branch and the baseline is included as characteristic parameters, and a random forest algorithm is combined with the intracranial pressure change detection model to achieve non-invasive monitoring.
The monitoring results of non-destructive judgment of intracranial pressure are achieved, and auxiliary medical staff can conduct effective monitoring, providing a non-invasive and easy-to-use intracranial pressure monitoring method, which can monitor changes in intracranial pressure in real time, improving the accuracy of diagnosis and treatment.
Smart Images

Figure CN118986315A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of intracranial pressure monitoring, and in particular to an intracranial pressure monitoring method, system and medium based on electrical impedance imaging. Background Art
[0002] Neurological critical diseases, such as craniocerebral trauma, stroke, brain tumors, etc., are often accompanied by increased intracranial pressure (ICP). Continuously elevated ICP may reduce cerebral blood perfusion, leading to ischemia and hypoxia of brain tissue. If not treated in time, it will eventually lead to severe neurological damage and brain death. Monitoring and warning of elevated ICP are crucial for the correct treatment of patients in neurocritical care and improving prognosis. Therefore, real-time monitoring of ICP to prevent insufficient cerebral blood perfusion has very important clinical value in the diagnosis and treatment of craniocerebral diseases.
[0003] Existing ICP monitoring methods are mainly divided into invasive ICP monitoring and non-invasive ICP monitoring. Invasive ICP monitoring is widely regarded as the gold standard for ICP monitoring. However, this method may cause complications such as intracranial hemorrhage and infection during the placement of the pressure sensor. Currently available non-invasive ICP monitoring technologies, such as transcranial Doppler ultrasound, optic nerve sheath diameter measurement, magnetic resonance imaging, etc., mostly obtain ICP information through indirect methods. However, these methods often require professional personnel to operate and interpret, lack ease of use, and are not suitable for long-term dynamic non-invasive monitoring. Based on this, there is an urgent need for a non-invasive, harmless, and easy-to-interpret bedside ICP monitoring method in the clinic, in order to achieve monitoring and early warning of ICP changes. Summary of the invention
[0004] To solve the above problems, the present invention provides an intracranial pressure monitoring method based on electrical impedance imaging. The method obtains cerebral blood perfusion parameters through electrical impedance imaging, and monitors intracranial pressure in real time according to the change pattern of cerebral blood perfusion parameters with intracranial pressure.
[0005] To achieve the above object, the present invention provides the following technical solutions.
[0006] The present invention provides an intracranial pressure monitoring method based on electrical impedance tomography, comprising the following steps:
[0007] The electrical impedance tomography system is used to collect the extracranial original boundary voltage signal of the monitored object, and the voltage signal of the frequency band reflecting the dynamic blood perfusion situation is extracted;
[0008] The finite element model of the electrode node distribution position that conforms to the actual electrical impedance tomography system is used to solve the voltage signal and obtain the potential on each electrode node.
[0009] The damped least square method is used to perform inverse problem calculation on the potential at each electrode node to obtain a reconstructed image reflecting the dynamic change of blood flow; wherein the reconstructed image is an image of the dynamic change of resistivity caused by the pulsating change of the congestion state of brain tissue;
[0010] Based on the reconstructed images, the average perfusion velocity, contraction wave amplitude, inflow volume velocity, and the angle between the descending branch and the baseline were extracted as characteristic parameters to characterize the changes in intracranial pressure.
[0011] The random forest algorithm is used to build an intracranial pressure increase change detection model, and the characteristic parameters that characterize the intracranial pressure change are used to construct a training set to train the intracranial pressure increase change detection model; the characteristic parameters of the monitored object are obtained, and the intracranial pressure status is monitored through the trained intracranial pressure increase change detection model.
[0012] Preferably, the step of extracting the voltage signal reflecting the frequency band of dynamic blood perfusion comprises the following steps:
[0013] The voltage signal reflecting the dynamic blood perfusion condition is extracted through a bandpass filter with a passband frequency of 1Hz-5Hz.
[0014] Preferably, the method of using the damped least square method to perform inverse problem calculation on the potential at each electrode node to obtain a reconstructed image reflecting the dynamic changes of blood flow includes the following steps:
[0015] According to the potential on each electrode node, the resistivity distribution between the current frame and the reference frame is obtained:
[0016] Δρ=(J T J+λR) -1 J T ΔV
[0017] Δρ is the change in resistivity distribution between the current frame and the reference frame, ΔV is the normalized boundary voltage change between the current frame and the reference frame, J is the Jacobian matrix, λ is the regularization parameter, and R is the regularization matrix;
[0018] The average reconstructed resistivity associated with the whole brain in the electrical impedance tomography image sequence is obtained by the following formula:
[0019]
[0020] In the formula, Δρ i represents the reconstructed resistivity of the i-th pixel in the EIT image; N represents the total number of pixels in the region of interest.
[0021] Preferably, the average perfusion rate is obtained as shown in the following formula:
[0022]
[0023] Where ARV i ARV is the average reconstructed resistivity of the electrical impedance tomography image in the i-th frame; p is the peak value of resistivity; ARV v ARV is the valley value of resistivity; p ARV v For the descending branch, T d The time interval of the descending branch; from the valley value ARV v Return to next peak ARV p For the ascending branch, T a is the time interval of the ascending branch; N is the total number of pixels in the region of interest.
[0024] Preferably, the acquisition of the contraction wave amplitude is as shown in the following formula:
[0025] H s =ARV p -ARV v
[0026] Where ARV p is the peak value of resistivity; ARV v is the valley value of resistivity.
[0027] Preferably, the inflow volume velocity is obtained as shown in the following formula:
[0028]
[0029] In the formula, H s is the contraction wave amplitude, T d is the time interval of the descending branch.
[0030] Preferably, the angle between the descending branch and the baseline is obtained as shown in the following formula:
[0031]
[0032] The baseline is defined as the peak ARV value from the resistivity p Starting from, and its direction is parallel to the x-axis; X p , Y p Peak ARV p The x and y coordinates of v , Y v ARV v The x and y coordinates of the
[0033] The present invention provides an intracranial pressure monitoring system based on electrical impedance imaging, the system comprising:
[0034] processor;
[0035] a memory having stored thereon a computer program executable on the processor;
[0036] Wherein, when the computer program is executed by the processor, the steps of the intracranial pressure monitoring method based on electrical impedance imaging are implemented.
[0037] The present invention provides a computer-readable storage medium, on which a data processing program is stored. When the data processing program is executed by a processor, the steps of the intracranial pressure monitoring method based on electrical impedance imaging are implemented.
[0038] Beneficial effects of the present invention:
[0039] The present invention proposes a method, system and medium for monitoring intracranial pressure based on electrical impedance imaging. The method collects and analyzes the variation of EIT image parameters of cerebral blood perfusion with ICP through the EIT system, and develops four EIT parameters related to cerebral blood perfusion, namely, average perfusion rate, contraction wave height, inflow volume velocity and angle between descending branch and baseline. The change of ICP is analyzed and determined based on the above parameters. According to the above conclusion, the method can non-destructively judge the monitoring results of ICP by implementing EIT images for monitoring cerebral blood perfusion, and effectively assist medical staff in monitoring. BRIEF DESCRIPTION OF THE DRAWINGS
[0040] Figure 1 is a flow chart of intracranial pressure monitoring based on electrical impedance imaging according to an embodiment of the present invention;
[0041] Figure 2 Schematic diagram of the experimental method and process of the embodiment of the present invention; wherein, Figure 2 (A) is a schematic diagram of the positions of 16 copper electrodes, indwelling needles, and ICP probes; Figure 2 (B) is the experimental scene diagram; Figure 2 (C) is the experimental time. EIT data collection started with blood injection about 10 minutes after the baseline, with an interval of about 10 minutes between blood injections, for a total of 5 blood injections. Figure 2 (D) is a schematic diagram of ARV curve and perfusion parameter extraction;
[0042] Figure 3 is the variation of the EIT signal with ICP in the embodiment of the present invention; wherein, Figure 3 (A) is the original measurement data of EIT and ICP; Figure 3 (B) is a comparison chart of ARV and ICP during blood infusion; Figure 3 (C) is the average transmission impedance and perfusion cycle EIT image; Figure 3 (D) is the reconstructed EIT basic impedance image with t1 as the reference data;
[0043] Figure 4 are the average perfusion parameters and standard deviations of all samples at different ICP levels in the embodiment of the present invention; wherein, Figure 4 (A) is the average perfusion rate; Figure 4 (B) is the height of the contraction wave; Figure 4 (C) is the inflow volume velocity; Figure 4 (D) Angle;
[0044] Figure 5 4 independent perfusion parameters and the ROC curve of the random forest model for detecting ICP increase relative to the baseline in the embodiment of the present invention; wherein, Figure 5 (A) is the perfusion parameter detection ICP increase greater than 5 mmHg; Figure 5 (B) is the perfusion parameter detection ICP increase greater than 10 mmHg; Figure 5 (C) is the ROC curve of the random forest model for detecting ICP increase greater than 5 mmHg; Figure 5 (D) is the ROC curve of the random forest model for detecting ICP increase greater than 10 mmHg;
[0045] Figure 6 is the detection efficiency of the random forest model for detecting ICP increase when the reference value is changed in the embodiment of the present invention; wherein, Figure 6 (A) is to detect ICP increase greater than 5 mmHg; Figure 6 (B) is for detecting ICP increase greater than 10 mmHg. DETAILED DESCRIPTION
[0046] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.
[0047] Example 1
[0048] The present invention is a method for monitoring intracranial pressure based on electrical impedance imaging, the specific process is as follows Figure 1 As shown, the following steps are included:
[0049] S1: The original extracranial boundary voltage signal of the monitored object is collected through the electrical impedance tomography system, and the voltage signal of the frequency band reflecting the dynamic blood perfusion situation is extracted.
[0050] Furthermore, a voltage signal reflecting the dynamic blood perfusion condition is extracted through a bandpass filter with a passband frequency of 1 Hz-5 Hz.
[0051] S2: A circular finite element model that conforms to the actual electrode distribution position is used to solve the voltage signal and obtain the potential at each electrode node.
[0052] S3: The damped least square method is used to perform inverse problem calculation on the potential at each electrode node to obtain a reconstructed image reflecting the dynamic changes of blood flow; wherein the reconstructed image is an image of the dynamic change of resistivity caused by the pulsating changes of the congestion state of brain tissue.
[0053] Specifically, according to the potential on each electrode node, the resistivity distribution between the current frame and the reference frame is obtained:
[0054] Δρ=(J T J+λR) -1 J T ΔV
[0055] Δρ is the change in resistivity distribution between the current frame and the reference frame, ΔV is the normalized boundary voltage change between the current frame and the reference frame, J is the Jacobian matrix, λ is the regularization parameter, and R is the regularization matrix;
[0056] The average reconstructed resistivity associated with the whole brain in the electrical impedance tomography image sequence is obtained by the following formula:
[0057]
[0058] In the formula, Δρ i represents the reconstructed resistivity of the i-th pixel in the EIT image; N represents the total number of pixels in the region of interest.
[0059] S4: Based on the reconstructed images, the average perfusion velocity, contraction wave amplitude, inflow volume velocity, and the angle between the descending branch and the baseline were extracted as characteristic parameters to characterize the changes in intracranial pressure.
[0060] Specifically, the average perfusion rate is obtained as shown in the following formula:
[0061]
[0062] Where ARV i ARV is the average reconstructed resistivity of the electrical impedance tomography image in the i-th frame; p is the peak value of resistivity; ARV v ARV is the valley value of resistivity; p ARV v is the descending branch, T d The time interval of the descending branch; from the valley value ARV v Return to next peak ARV p For the ascending branch, T a is the time interval of the ascending branch; N is the total number of pixels in the region of interest. A complete cerebral perfusion cycle is defined as the time from the peak ARV p Start, pass the valley ARVv , reaching the next peak ARV p process.
[0063] The acquisition of contraction wave amplitude is shown in the following formula:
[0064] H s =ARV p -ARV v
[0065] Where ARV p is the peak value of resistivity; ARV v is the valley value of resistivity.
[0066] The inflow volume velocity is obtained as shown in the following formula:
[0067]
[0068] In the formula, H s is the contraction wave amplitude, T d is the time interval of the descending branch.
[0069] Preferably, the angle between the descending branch and the baseline is obtained as shown in the following formula:
[0070]
[0071] The baseline is defined as the peak ARV value from the resistivity p Starting from, and its direction is parallel to the x-axis; X p , Y p Peak ARV p The x and y coordinates of v , Y v ARV v The x and y coordinates of
[0072] S5: A random forest algorithm is used to construct an intracranial pressure increase change detection model, and characteristic parameters that characterize intracranial pressure changes are used to construct a training set to train the intracranial pressure increase change detection model; characteristic parameters of the monitored object are obtained, and the intracranial pressure status is monitored through the trained intracranial pressure increase change detection model.
[0073] In order to verify the feasibility of the above method, the present invention provides the following experiments:
[0074] A domestic pig intracranial hypertension model was constructed, and invasive ICP monitoring was used as a control method to analyze the changes in EIT cerebral blood perfusion parameters with ICP. The purpose was to verify the ability of EIT to indirectly reflect the ICP level by monitoring the cerebral blood perfusion status by observing the correlation between ICP and the cerebral blood perfusion status reflected by EIT.
[0075] Specifically:
[0076] Six healthy pigs weighing 17.2±1.5kg were selected for the experiment. The experimental protocol and procedures were approved by the Animal Ethics Committee of the Fourth Military Medical University (Ethics Permit No.: IACUC-20241299) and followed the "Guide for the Care and Use of Laboratory Animals" published by the National Institutes of Health (National Academy Press, Washington, DC, revised in 1996). All pigs were induced anesthetized with a dose of 5 mg / kg of Shutai 50, and the pigs were intubated and maintained anesthesia with 1-3% isoflurane. The femoral artery was punctured with an indwelling needle to establish a blood collection channel. After monitoring arterial blood pressure, oxygen saturation, and end-tidal carbon dioxide in the pigs, the scalp of each pig was cut longitudinally, the skull was separated and exposed, the periosteum was scraped, and a window with a sagittal length of about 60 mm and a coronal length of about 60 mm was formed. A skull drill was used to drill holes about 15 mm to the right of the sagittal suture and about 15 mm in front of the coronal suture for placement of the ICP probe. A hole was drilled about 15 mm to the right of the sagittal suture and about 10 mm behind the coronal suture, and then a 22G indwelling needle was inserted into the cranial hole to a depth of about 18 mm. In addition, 16 sterilized 1.2 mm dental root canal posts were drilled into the edge of the skull at equal intervals, as shown in the schematic diagram. Figure 2 As shown in (A), none of them penetrated the dura mater. The above cranial holes were sealed with bone wax to reduce the possibility of cerebrospinal fluid leakage.
[0077] ICP data were collected by an ICP monitor (SOPHYSA Inc., Orsay, France) with a frame rate of 2 frames per second. Figure 2 As shown in (B).
[0078] The EIT signal was collected using a new EIT system jointly developed by UTRON Technology Co., Ltd., Hangzhou, China. The operating frequency range of the system is 10kHz-250kHz, the output current range is 10μA-1250μA, and the signal-to-noise ratio is greater than 90dB. During the experiment, the frame rate of data acquisition was set to 40 frames per second and the excitation frequency was set to 50kHz.
[0079] Experimental protocol:
[0080] The intracranial hypertension model was prepared by injecting autologous non-anticoagulated blood into the brain parenchyma of domestic pigs in batches. After connecting the electrodes to the EIT system, the EIT baseline data was collected and the ICP data collection was started at the same time. About 10 minutes later, 2 ml of blood was drawn through the indwelling needle pre-placed in the femoral artery, and then injected into the brain parenchyma area at a uniform speed within 1 minute through the indwelling needle pre-placed in the skull. This process was repeated after about 10 minutes, for a total of 5 times, and its time series is shown as follows: Figure 2 As shown in (C). Norepinephrine was used to maintain blood pressure basically unchanged during the entire operation. After the operation was completed, the animal was deeply anesthetized and killed, and the brain was removed by craniotomy to observe the location of the hematoma.
[0081] Through the above steps S1 to S4, EIT image reconstruction and perfusion parameter extraction are obtained. Figure 2 As shown in (D), Figure 2 (D) is a schematic diagram of the ARV curve and perfusion parameter extraction.
[0082] Data Analysis:
[0083] SPSS27.0 statistical software (SPSS Inc, Chicago, IL, USA) was used for statistical analysis. The significance level of statistical analysis was 0.05. The Shapiro-Wilk test was used to test the normality of the data. According to the ICP grading standard, the ICP values were divided into the following four groups: 0-15 mmHg, 15-20 mmHg, 20-40 mmHg, and greater than 40 mmHg. Then, for all samples, the paired t-test was used to compare the differences in the average perfusion parameters between adjacent groups and the differences between 0-15 mmHg and other groups. The pixel value of functional EIT (fEIT) is not a simple impedance change, but valuable functional information that affects clinical decision-making. The EIT image sequence of 5 perfusion cycles in each group was calculated, and the waveform of each pixel was extracted from it to calculate the average perfusion parameters, which were presented in the form of images to generate fEIT images.
[0084] According to the International Brain Trauma Foundation Intracranial Hypertension Treatment Intervention Criteria, the data set was divided into two intervals with an ICP value of 22 mmHg as the boundary. The Spearman correlation coefficients between the perfusion parameters and ICP in these two intervals were calculated, and the paired t test was used to evaluate the differences in the correlation coefficients between the two intervals.
[0085] To evaluate the performance of EIT parameters in detecting elevated ICP, we first determined the mean MV, H, and s , IV and A db As reference values, the normalized change rates of these parameters relative to the reference values (ΔMV, ΔH s , ΔIV and ΔA db), and the corresponding ICP change (ΔICP). 500 sample data were randomly selected for each animal, and a total of 3000 sample data were analyzed for receiver operating characteristic (ROC) curve analysis, and the effectiveness of each independent EIT parameter change in detecting ICP increases greater than 5 mmHg and 10 mmHg relative to baseline was evaluated. Finally, the area under the curve, sensitivity, and specificity were calculated based on the ROC curve.
[0086] Finally, the random forest algorithm was used to establish an ICP increase change detection model. The research dataset consists of 3000 samples of ΔMV, ΔH s , ΔIV, ΔA db and Y, ΔMV, ΔH s , ΔIV and ΔA db MV, H s , IV and A db Normalized rate of change relative to the reference value, Y = {0,1} is the label (where 0 and 1 represent ΔICP less than and greater than a certain threshold, respectively). The data set was then randomly divided into training (70%) and test (30%) data sets, and the random forest model was trained using the training set. With the help of the test set, the detection performance of the model was comprehensively evaluated by calculating AUC, sensitivity, and specificity. The reference values selected are normal ICP values and intracranial hypertension values (between 15-40 mmHg). This reference range was chosen mainly because patients are often already in a state of intracranial hypertension when they receive initial monitoring, making it critical to explore the ability of the random forest algorithm to detect elevated ICP for different reference values.
[0087] in conclusion:
[0088] (1) EIT signal changes with ICP
[0089] like Figure 3 As shown, Figure 3 (A) is the original measurement data of EIT and ICP. The top shows the average transmission impedance from 256 channels, and the red arrow is the blood injection time point; the middle is the original data and filtered data of ICP; the bottom is the mean arterial pressure. Figure 3 (B) is a comparison chart of ARV and ICP during blood infusion. Figure 3 (C) shows the average transmission impedance and perfusion cycle EIT images. The upper part shows the average transmission impedance of the perfusion cycle corresponding to t1, t2 and t3, and the lower part shows the corresponding perfusion cycle EIT images. The image sequence interval is 25ms. The ICP corresponding to t1, t2 and t3 is approximately 10mmHg, 25mmHg and 40mmHg, respectively. Figure 3 (D) is the reconstructed EIT-based impedance image with t1 as the reference data.
[0090] exist Figure 3 In (A), the exemplary changes of EIT transmission impedance, ICP and mean arterial pressure during the whole experiment are presented. In this experiment, five blood injections were performed, and each blood injection caused the transmission impedance to rise rapidly. At the same time, ICP will slowly decrease after a rapid rise. During the whole experiment, the mean arterial pressure remained stable without obvious fluctuations. Figure 3 (B) shows the changes in ICP and ARV before and after a blood injection. It can be observed that after the blood injection, the ICP increases rapidly and then decreases slowly. At the same time, the amplitude of ARV decreases rapidly and then gradually increases slowly. In order to study the effect of ICP changes on the EIT images of blood perfusion, the time points corresponding to three different ICP levels of the EIT images were selected, namely Figure 3 In (B), t1, t2, and t3, each of these time points represents a cardiac cycle. Using the starting point of the systolic period as the reference frame, differential imaging of the cardiac cycle at these three time points was performed, and the results are shown in Figure 1. Figure 3 As shown in (C). By observing the dynamic cerebral blood flow perfusion EIT images at selected time points, it can be seen that as ICP gradually increases, cerebral perfusion in the edge area of the image gradually decreases. After removing the pulsating signal in EIT using a low-pass filter with an upper cutoff frequency of 0.1 Hz, the basic impedance is determined. Figure 3 (D) shows the results of EIT-based impedance imaging at t1, t2, and t3, with t1 as the reference time. Continuous EIT images show that the blue area in the lower left corner represents the blood injection area. As the number of blood injections increases, the impedance of this area gradually increases.
[0091] (2) Effect of ICP changes on EIT perfusion parameters
[0092] Figure 4 The mean perfusion parameters and standard deviations of all samples at different ICP levels are shown; Figure 4 (A) is the average perfusion rate; Figure 4 (B) is the height of the contraction wave; Figure 4 (C) is the inflow volume velocity; Figure 4 (D) angle.
[0093] The Shapiro-Wilk test showed that the data followed a normal distribution (p>0.05). In the study analyzing the effect of different levels of ICP on EIT parameters, the paired t-test was used to compare the average perfusion parameters between adjacent ICP levels, and the results showed differences. Specifically, when the perfusion parameters of ICP less than 15mmHg were compared with 15-20mmHg and 15-20mmHg with 20-40mmHg, the differences in the parameters were not consistent. However, when the ICP increased from 20-40mmHg to 40mmHg, the differences in all perfusion parameters showed significant differences (P<0.01). In addition, there were significant differences between the group with ICP less than 15mmHg and the group with ICP between 15-20mmHg and above 40mmHg (P<0.05). The fEIT image at the bottom of the figure also reveals a trend of gradual decrease in perfusion parameters in the peripheral area of the image as ICP increases.
[0094] Table 1 shows the average Spearman correlation coefficient and standard deviation between ICP and various perfusion parameters. It can be clearly seen from the table that the correlation in the group above 22 mmHg is significantly higher than that in the group below 22 mmHg (P<0.05). This shows that at higher ICP levels, there is a stronger correlation between ICP and cerebral blood perfusion, that is, ICP has a more significant effect on cerebral blood flow: as ICP further increases, cerebral blood flow decreases more significantly.
[0095] Table 1 Mean Spearman correlation coefficients and standard deviations between ICP and various perfusion parameters
[0096]
[0097] (3) Evaluation of the diagnostic efficacy of EIT perfusion parameters in detecting elevated ICP
[0098] Figure 5 (A) and (B) show the relationship between ΔMV and ΔH. s , ΔIV and ΔA db ROC curves for the ability of single variables to detect increases in ICP greater than 5 mmHg and 10 mmHg from baseline. The AUCs for these variables for detecting increases in ICP greater than 5 mmHg were 0.82, 0.71, 0.70, and 0.69, respectively, and for detecting increases in ICP greater than 10 mmHg were 0.85, 0.72, 0.72, and 0.71, respectively. Tables 2 and 3 list the specific values of the AUC, sensitivity, and specificity for each parameter. This suggests that these EIT perfusion parameters alone have a certain degree of accuracy in detecting increases in ICP.
[0099] Table 2 Evaluation indicators of different parameters in ICP>5 mmHg detection
[0100]
[0101] Table 3 Evaluation indicators of different parameters in ICP>5 mmHg detection
[0102]
[0103] In order to further improve the accuracy of detection, this embodiment uses the random forest method to establish a comprehensive model to detect ICP increase changes. Figure 5 (C) and (D) are ROC curves of the random forest model for detecting ICP increases relative to baseline. The ROC analysis results show that in detecting an ICP increase of 5 mmHg, the AUC, sensitivity, and specificity of the model reached 0.90, 92%, and 90%, respectively, and ΔMV, ΔH s , ΔIV and ΔA db The variable importances were 6.03, 2.81, 2.34 and 3.23, respectively. In detecting an increase in ICP of 10 mmHg, the AUC, sensitivity and specificity were 0.92%, 93% and 90%, respectively, and the variable importances were 7.4, 2.50, 2.96 and 1.76, respectively. Figure 6 (A) and (B) show the performance of the random forest model in detecting ICP elevations exceeding 5 mmHg and 10 mmHg at different reference values. Specifically, the model's AUC, sensitivity, and specificity all performed well, with values exceeding 0.9. More notably, these performance indicators also showed an upward trend as the reference value increased.
[0104] The above is an intracranial pressure monitoring method based on electrical impedance imaging provided by one embodiment of the present embodiment. Based on the same idea, the present embodiment also provides a corresponding intracranial pressure monitoring system based on electrical impedance imaging. For the specific definition of the intracranial pressure monitoring system based on electrical impedance imaging, please refer to the definition of the intracranial pressure monitoring method based on electrical impedance imaging in the above text, which will not be repeated here. Each module in the above-mentioned intracranial pressure monitoring system based on electrical impedance imaging can be implemented in whole or in part by software, hardware and a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory in the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above-mentioned modules.
[0105] This embodiment also provides a computer-readable storage medium, which stores a computer program, which can be used to execute the above Figure 1 Provided is an intracranial pressure monitoring method based on electrical impedance tomography.
[0106] Those of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory may include read-only memory (ROM), magnetic tape, floppy disk, flash memory or optical memory, etc. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).
[0107] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.
Claims
1. A method for monitoring intracranial pressure based on electrical impedance tomography. It is characterized in that The following steps are involved: The electrical impedance tomography system is used to collect the extracranial original boundary voltage signal of the monitored object, and the voltage signal of the frequency band reflecting the dynamic blood perfusion situation is extracted; The finite element model of the electrode node distribution position that conforms to the actual electrical impedance tomography system is used to solve the voltage signal and obtain the potential on each electrode node. The damped least square method is used to perform inverse problem calculation on the potential at each electrode node to obtain a reconstructed image reflecting the dynamic change of blood flow; wherein the reconstructed image is an image of the dynamic change of resistivity caused by the pulsating change of the congestion state of brain tissue; Based on the reconstructed images, the average perfusion velocity, contraction wave amplitude, inflow volume velocity, and the angle between the descending branch and the baseline were extracted as characteristic parameters to characterize the changes in intracranial pressure. The random forest algorithm is used to build an intracranial pressure increase change detection model, and the characteristic parameters that characterize the intracranial pressure change are used to construct a training set to train the intracranial pressure increase change detection model; the characteristic parameters of the monitored object are obtained, and the intracranial pressure status is monitored through the trained intracranial pressure increase change detection model.
2. The intracranial pressure monitoring method based on electrical impedance tomography according to claim 1, It is characterized in that The method of extracting the voltage signal of the frequency band reflecting the dynamic blood perfusion condition comprises the following steps: The voltage signal reflecting the dynamic blood perfusion condition is extracted through a bandpass filter with a passband frequency of 1Hz-5Hz.
3. The intracranial pressure monitoring method based on electrical impedance tomography according to claim 1, It is characterized in that The method of using the damped least square method to perform inverse problem calculation on the potential at each electrode node to obtain a reconstructed image reflecting the dynamic changes of blood flow includes the following steps: According to the potential on each electrode node, the resistivity distribution between the current frame and the reference frame is obtained: Dr=(J T J+λR) -1 J T ΔV Δρ is the change in resistivity distribution between the current frame and the reference frame, ΔV is the normalized boundary voltage change between the current frame and the reference frame, J is the Jacobian matrix, λ is the regularization parameter, and R is the regularization matrix: The average reconstructed resistivity associated with the whole brain in the electrical impedance tomography sequence is obtained by the following formula: In the formula, Δρ i represents the reconstructed resistivity of the ith pixel in the EIT image; N represents the total number of pixels in the region of interest.
4. The intracranial pressure monitoring method based on electrical impedance tomography according to claim 1, It is characterized in that The average perfusion rate is obtained as shown in the following formula: Where ARV i ARV is the average reconstructed resistivity of the electrical impedance tomography image in the i-th frame; p is the peak value of resistivity; ARV v ARV is the valley value of resistivity; p ARV v For the descending branch, T d The time interval of the descending branch; from the valley value ARV v Return to next peak ARV p For the ascending branch, T a is the time interval of the ascending branch; N is the total number of pixels in the region of interest.
5. The intracranial pressure monitoring method based on electrical impedance tomography according to claim 4, It is characterized in that The acquisition of the contraction wave amplitude is shown in the following formula: H s =ARV p -ARV v Where ARV p is the peak value of resistivity; ARV v is the valley value of resistivity.
6. The intracranial pressure monitoring method based on electrical impedance tomography according to claim 5, It is characterized in that The inflow volume velocity is obtained as shown in the following formula: In the formula, H s is the contraction wave amplitude, T d is the time interval of the descending branch.
7. The intracranial pressure monitoring method based on electrical impedance tomography according to claim 5, It is characterized in that The angle between the descending branch and the baseline is obtained as shown in the following formula: The baseline is defined as the peak ARV value from the resistivity p Starting from the x-axis, the vector whose direction is parallel to the x-axis; X p , Y p Peak ARV p The x and y coordinates of v , Y v ARV v The x and y coordinates of 8. An intracranial pressure monitoring system based on electrical impedance imaging, It is characterized in that The system comprises: processor; a memory having stored thereon a computer program executable on the processor; Wherein, when the computer program is executed by the processor, the steps of the intracranial pressure monitoring method based on electrical impedance imaging as described in any one of claims 1 to 7 are implemented.
9. A computer-readable storage medium, It is characterized in that The computer-readable storage medium stores a data processing program, and when the data processing program is executed by a processor, the steps of the intracranial pressure monitoring method based on electrical impedance imaging as described in any one of claims 1 to 7 are implemented.
Citation Information
Patent Citations
Method for structural information fused electrical impedance tomography
CN101564294A
Intracranial pressure estimation method and device
CN110916641A
Admittance differentiation ring for constituting detector of human organ lesion
CN1305776A
Admittance-type automatic test method of bilateral cerebral technogram and tester based on said method
CN1305778A
Systems, methods, software, and platforms for clinical decision support
US20220319718A1
