A high altitude pulmonary edema ultrasonic early warning system and method based on pulmonary artery pressure data

Abstract

The application discloses a kind of plateau pulmonary edema ultrasonic early warning system and method based on pulmonary artery pressure data, it is related to electric digital data processing technical field.The method is executed by acquiring immediate measured value on admission two-dimensional group check-in, and accordingly dynamic generation difference collection timing scheme, realize the collection of light and heavy symptom diversion.Core step includes: using the self-calibration model containing altitude linear scaling item and exponential blood oxygen correction item, eliminate the physiological pulmonary artery pressure increase deviation caused by low oxygen environment;With the calibrated pressure as the core, build the correlation coefficient, four early implicit warning features and variance compensation item of the correlation calculation model, output the disease matching degree score;Based on the score, perform stratified evaluation, output visual trend map and match clinical pathway.In addition, it also includes two-level diversion quality control and double prognosis and the generation of classification label.The application realizes whole process automation, improves early screening accuracy and efficiency.

Description

Technical Field

[0001] This invention relates to the field of electronic digital data processing technology, and in particular to a high-altitude pulmonary edema ultrasound early warning system and method based on pulmonary artery pressure data. Background Technology

[0002] For troops stationed at high altitudes, acute high-altitude pulmonary edema is a common acute condition that induces non-combat attrition and directly reduces the troops' ability to perform field missions. Currently, ultrasound examination is a commonly used rapid screening method at grassroots training sites. It is portable, non-invasive, and perfectly suited for mobile inspection scenarios in high-altitude environments. However, the traditional method of measuring pulmonary artery pressure based on the tricuspid regurgitation pressure gradient is prone to physiological biases in the hypoxic environment of high altitudes. Normal individuals experience compensatory increases in pulmonary artery pressure upon reaching high altitudes, and conventional testing methods cannot distinguish between this normal physiological fluctuation and abnormal pressure increases under pathological conditions, easily leading to missed diagnoses and misdiagnoses.

[0003] Meanwhile, most existing general-purpose ultrasound-assisted diagnostic systems adopt a standardized, fixed acquisition process, uniformly scheduling three full-scale examinations on days 1, 3, and 5 of admission regardless of the severity of the patient's condition. This wastes the limited manpower of primary care ultrasound departments and adds unnecessary examination burdens to patients with mild symptoms. Moreover, conventional systems can only output single numerical results and cannot link pulmonary artery pressure with lung ultrasound images and right ventricular function indicators. Many early potential high-risk warning signals cannot be automatically identified, and diagnosis is often delayed until patients develop typical signs of pulmonary edema, missing the optimal intervention window of 6 to 12 hours.

[0004] There is currently no fully automated early warning solution that is fully adapted to high-altitude scenarios, can automatically complete pulmonary artery pressure deviation correction, hierarchically adapt the acquisition process, and simultaneously mine early latent warning features. Clinically, it is impossible to directly rely on raw ultrasound data to automatically generate matching diagnosis and treatment path references. The entire process still requires a lot of manual intervention to adjust parameters and verify indicators. The diagnostic efficiency and accuracy cannot meet the actual needs of batch screening in high-altitude fields. Summary of the Invention

[0005] The purpose of this section is to outline some aspects of embodiments of the present invention and to briefly describe some preferred embodiments. Simplifications or omissions may be made in this section, as well as in the abstract and title of this application, to avoid obscuring the purpose of these documents; however, such simplifications or omissions should not be construed as limiting the scope of the invention.

[0006] It should be noted that, in response to the pain points of existing technologies, this invention introduces a high-altitude pulmonary edema ultrasound early warning system based on pulmonary artery pressure data that is fully adapted to the special scenarios of high altitudes. It realizes full-process automation from ultrasound data acquisition, automatic correction, intelligent early warning to matching diagnosis and treatment pathways, eliminating the need for repeated manual correction of data afterward and significantly improving the early identification accuracy of high-altitude pulmonary edema.

[0007] To achieve the above objectives, the technical solution provided by the present invention is as follows.

[0008] In a first aspect, the present invention provides an information processing method based on pulmonary artery pressure data, executed by a computer system for auxiliary assessment of high-altitude pulmonary edema, comprising the following steps:

[0009] S1: Obtain the measured value of pulmonary artery systolic pressure and physiological parameters of the subject immediately upon admission, and perform enrollment eligibility verification; the verification includes comparing the measured value with the pulmonary artery pressure reference baseline of healthy people at high altitude, and determining whether the tricuspid regurgitation signal meets the preset calculation conditions; if the verification fails, the subsequent data acquisition process is terminated and a verification failure prompt is generated.

[0010] S2: For subjects who pass the verification, a timing schedule for the acquisition of subsequent multimodal ultrasound data is dynamically generated based on the numerical range of the measured values ​​immediately upon admission. According to the scheduling schedule, the two-dimensional grayscale ultrasound, Doppler blood flow ultrasound, and tissue Doppler ultrasound data of the corresponding subjects are acquired and associated to form the effective sample set for the current batch. It is determined whether the sample size in the effective sample set has reached a preset threshold. If it has not, a supplementary sampling prompt is generated.

[0011] S3: Input the collected raw pulmonary artery systolic pressure into the preset self-calibration model, and perform calibration calculations using the current altitude of the residence, the real-time blood oxygen saturation of the subject, and the pulmonary artery pressure reference baseline value of healthy people at the same altitude to eliminate the deviation of physiological pulmonary artery pressure increase caused by high altitude hypoxia environment, and obtain the calibrated pulmonary artery systolic pressure.

[0012] S4: Using the calibrated pulmonary artery systolic pressure as the core input, construct a correlation calculation model between multimodal ultrasound features and disease severity; the input parameters of the correlation calculation model include the calibrated pulmonary artery systolic pressure and a variance compensation term determined based on the feature variance ratio of the current sample and the full dataset, and calculate the disease correlation matching score through this model;

[0013] S5: Perform stratified assessment calculation based on the correlation matching score, output visualized atlas data that includes at least the pulmonary artery pressure change curve, lung ultrasound score trend and right heart function recovery dynamic trajectory, and push matching clinical diagnosis and treatment path reference information based on the stratification results.

[0014] Preferably, the specific verification process described in S1 is as follows:

[0015] When the measured value falls within the preset pulmonary artery systolic pressure reference baseline range that characterizes healthy people at high altitudes, it is determined to be a non-target object and excluded; at the same time, the integrity of the tricuspid regurgitation spectrum is checked. If the pulmonary artery pressure cannot be measured due to the absence of regurgitation signal, it is determined to be an unqualified sample, and the subsequent acquisition process of the sample is terminated.

[0016] Preferably, the timing scheduling scheme for dynamically generating subsequent multimodal ultrasound data in S2 includes:

[0017] If the measured value immediately upon admission is not greater than the first threshold, such as 40 mmHg, then the subsequent collection nodes will be compressed into a single node.

[0018] If the measured value is greater than the first threshold, then retain at least three collection nodes that include the first, third and fifth days after admission, and trigger a temporary supplementary collection task on the second day.

[0019] Preferably, the self-calibration model in S3 performs the following calculations to obtain the calibrated pulmonary artery systolic pressure. :

[0020]

[0021] In the formula, To calibrate postpulmonary artery systolic pressure, These are the raw measured values ​​obtained by the tricuspid regurgitation pressure differential method. This represents the baseline pulmonary artery systolic blood pressure value for healthy individuals at the same altitude gradient. This is the current real-time altitude of the base. This is the reference elevation value for sea level. The preset altitude normalization threshold for plateau environments is used. This represents the subject's real-time arterial blood oxygen saturation percentage. For the natural constant An exponential function with base 0.

[0022] According to the method of claim 1, the characteristic of the disease correlation matching score in S4 is... The calculation formula is:

[0023]

[0024] in, Spearman's rank correlation coefficient between calibrated pulmonary artery systolic blood pressure and lung ultrasound 12-zone score, which was pre-determined based on historical datasets; It is the hyperbolic tangent activation function; Index of early hidden warning features. For the first Measured values ​​of latent features, For the first Clinical judgment thresholds corresponding to latent features; For indicator functions, satisfying The value is 1 if the condition is met, otherwise it is 0. This represents the variance explained by the current set of ultrasound features in the sample. This represents the total variance explained by the fully labeled dataset.

[0025] The early latent warning features consist of four items, and their corresponding preset clinical judgment thresholds are as follows: heart rate greater than 100 beats / min; pulmonary artery diameter greater than 25mm; E / A ratio not in the range of 0.8 to 2.1 and E / e' not less than 6; right ventricular index (RIMP) greater than 0.55.

[0026] Preferably, the step S5 involves scoring based on the correlation matching degree. Performing stratified assessment calculations specifically includes:

[0027] when When the value is in the range of 0 to 0.3, it is judged as either healthy or excluded.

[0028] when When the value is between 0.31 and 0.7, it is considered mild high-altitude pulmonary edema, corresponding to the first type of treatment pathway;

[0029] when When the value is in the range of 0.71 to 1, it is considered severe high-altitude pulmonary edema, corresponding to the second type of treatment pathway;

[0030] Furthermore, when the calibrated pulmonary artery systolic pressure is greater than the preset critical illness pressure threshold and the main pulmonary artery diameter is greater than the preset critical illness diameter threshold, it is directly calculated and marked as a critical illness warning state.

[0031] Preferably, the method further includes: S6: performing a two-level triage quality control operation on the dataset that has completed the evaluation;

[0032] The first round of verification of ultrasound image frame quality parameters identifies and archives abnormal samples, and generates deviation rectification prompts with correction guidelines.

[0033] The second round of verification indicators and logical consistency parameters excludes contradictory samples where the lung ultrasound score has returned to zero but the pulmonary artery systolic blood pressure is still higher than the normal threshold, triggering a second review process;

[0034] Valid data that has passed two levels of quality control will be imported into a standard resource library, and each sample will be labeled with two labels: the first label is a prognostic label for delayed recovery of right ventricular function, when the patient’s lung ultrasound score is zero but the right ventricular diastolic function has not recovered and the RIMP is still greater than 0.55, the patient is marked as a high-risk group for recurrence; the second label is a graded label generated based on the correlation matching score interval.

[0035] In a second aspect, the present invention provides a high-altitude pulmonary edema ultrasound early warning system based on pulmonary artery pressure data, used to implement the method described in any of the above-mentioned embodiments, comprising:

[0036] The enrollment verification and process control module is used to obtain the measured value of pulmonary artery systolic blood pressure and physiological parameters of subjects immediately upon admission, perform verification, and terminate the data collection process and generate a prompt when the verification fails.

[0037] The differentiated timing scheduling module is used to dynamically generate a timing scheduling scheme for the acquisition of subsequent multimodal ultrasound data based on the immediate measured values ​​upon admission, collect two-dimensional grayscale ultrasound, Doppler blood flow ultrasound and tissue Doppler ultrasound data, and generate a supplementary collection prompt when the effective sample size is insufficient; the self-calibration calculation module has a built-in self-calibration model as described in claim 4, which is used to correct the deviation of the original pulmonary artery systolic pressure and output the calibrated pulmonary artery systolic pressure.

[0038] The correlation calculation and evaluation module is used to construct the correlation calculation model as described in claim 5, and to calculate the disease correlation matching score by taking the calibrated pulmonary artery systolic blood pressure and early latent warning features as inputs.

[0039] The stratified assessment and information push module is used to perform stratified assessment calculations based on the correlation matching score, output visualized trend map data, and push matching clinical diagnosis and treatment path reference information.

[0040] The data quality control and labeling module is used to perform the two-level split quality control operation as described in claim 7 on the evaluation dataset, to label the valid data with the dual labels and import them into the standard resource library.

[0041] Thirdly, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method as described in any one of claims 1 to 7.

[0042] Fourthly, the present invention provides an electronic device, comprising:

[0043] Memory, used to store instructions;

[0044] A processor for executing the instructions, causing the device to perform the method as described in any one of claims 1 to 7.

[0045] Compared with existing technologies, this technology has the following advantages:

[0046] 1. By using a high-altitude adaptive dynamic self-calibration model and employing a two-dimensional calibration algorithm with an altitude linear scaling term and an exponential blood oxygenation correction term, the measurement interference caused by the physiological increase in pulmonary artery pressure under hypoxia is eliminated from the bottom layer. No manual parameter adjustment is required, which solves the problem of large measurement deviation in the traditional tricuspid regurgitation differential pressure method in high-altitude scenarios.

[0047] 2. By using a differentiated time-series scheduling engine, the 40 mmHg threshold for distinguishing between mild and severe cases, validated by 174 clinical samples, is directly anchored. The system automatically adapts the data collection node configuration to different conditions. For subjects with mild symptoms, only the single examination on the 3rd day of treatment is retained, which greatly improves the efficiency of screening at the grassroots level while ensuring the validity of follow-up data.

[0048] 3. Through a unique latent feature mapping network, it anchors four highly specific early warning indicators and integrates the strong correlation between pulmonary artery pressure and lung ultrasound score. It can trigger a high-risk warning when pulmonary artery pressure is slightly elevated and any two abnormal features are superimposed, advancing the diagnostic window by 6 to 12 hours. It truly realizes a closed-loop automatic flow from raw ultrasound data acquisition to intervention plan recommendation, effectively reducing non-combat attrition induced by acute high-altitude pulmonary edema. Attached Figure Description

[0049] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Wherein:

[0050] Figure 1 This is a flowchart illustrating the early warning processing method in an embodiment of the present invention;

[0051] Figure 2 This is a schematic diagram of the early warning processing system in an embodiment of the present invention.

[0052] Figure 3 This is a graph showing the inclusion criteria data of the healthy control group in an embodiment of the present invention;

[0053] Figure 4 This is a data chart showing the inclusion criteria and hospitalization treatment plan for the high-altitude pulmonary edema group in this embodiment of the invention;

[0054] Figure 5 This is a cardiac ultrasound data acquisition diagram from an embodiment of the present invention;

[0055] Figure 6 Lung ultrasound data acquisition diagram in an embodiment of the present invention;

[0056] Figure 7 This is a comparison chart of data collection in the high-altitude pulmonary edema group in this embodiment of the invention. Detailed Implementation

[0057] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0058] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0059] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0060] Example 1:

[0061] refer to Figure 1 As shown in the figure, this embodiment introduces a method for ultrasound-based early warning and processing of high-altitude pulmonary edema based on pulmonary artery pressure data. This method is automatically executed by a computer system and requires no manual intervention during operation. The overall process is as follows: Figure 1 As shown, the main steps include: dual-dimensional verification of enrollment eligibility (S1), differentiated time-series data collection and aggregation (S2), high-altitude adaptive dynamic self-calibration (S3), multimodal correlation calculation (S4), stratified evaluation and output (S5), and two-level data quality control (S6). Each step will be explained in detail below.

[0062] refer to Figures 3-7 As shown, S1: Obtain the measured pulmonary artery systolic pressure value and physiological parameters including at least blood oxygen saturation and heart rate of the subject upon admission, and perform a two-dimensional data verification for enrollment eligibility. The two-dimensional verification includes two aspects: first, comparing the measured pulmonary artery systolic pressure with the system's preset baseline range for pulmonary artery pressure in healthy high-altitude populations; and second, determining whether the tricuspid regurgitation signal characteristic value meets the measurement integrity requirements. If the verification fails, a deviation notification instruction containing the verification result identifier is generated, and the subsequent multimodal ultrasound data acquisition process for that subject is terminated.

[0063] In practice, the system incorporates a baseline range of pulmonary artery systolic blood pressure (PAP) for healthy individuals at high altitudes, calculated based on the inventors' previous large-sample study: 20.7 ± 6.5 mmHg. This baseline data is derived from statistical measurements of 106 healthy individuals at high altitudes and is expressed as mean ± standard deviation. Upon admission, the tricuspid regurgitation spectrum is acquired using ultrasound, and the measured PAP value is estimated using the tricuspid regurgitation pressure gradient method. The system automatically matches this measured value with the baseline range: if the measured value falls within the 20.7 ± 6.5 mmHg range, it means the subject's PAP pressure is within the normal fluctuation range for healthy individuals at high altitudes, classifying them as a non-target monitoring subject. The system directly determines that the inclusion verification fails and generates a deviation notification instruction, which includes an identifier indicating the verification result is within the healthy baseline range.

[0064] At the same time, the system also checks the integrity of the tricuspid regurgitation spectral signal. If the pulmonary artery systolic pressure cannot be calculated due to missing regurgitation signal or poor signal quality, the sample is deemed unqualified, triggering a deviation notification and terminating the acquisition. These two checks constitute a two-dimensional inspection.

[0065] Furthermore, the baseline range of 20.7±6.5 mmHg and the integrity judgment rules used here are all based on field studies conducted by the inventors on 174 patients with high-altitude pulmonary edema and 106 healthy individuals at the plateau, and are not based on general population data. Therefore, they have better specificity for the plateau environment.

[0066] S2: For verified subjects, the measured pulmonary artery systolic blood pressure at admission is used as input to dynamically determine the differentiated acquisition sequence of subsequent multimodal ultrasound data. Then, according to the aforementioned sequence, the three types of raw data from the corresponding subjects—two-dimensional grayscale ultrasound, Doppler blood flow ultrasound, and tissue Doppler ultrasound—along with the raw pulmonary artery systolic blood pressure calculated accordingly, are simultaneously collected into a standardized database. Furthermore, the system will determine in real time whether the effective sample size collected in the current batch has reached the preset minimum analysis sample size; if not, a supplementary collection reminder instruction will be generated.

[0067] The specific rules for dynamically determining the differentiated data acquisition timing scheme are derived from the boundary pattern of mild and severe illnesses derived by the inventors. Studies have shown that high-altitude pulmonary edema patients with pulmonary artery systolic blood pressure ≤40 mmHg upon admission can have their pulmonary artery pressure return to normal levels after 3 days of routine treatment; while patients with pulmonary artery systolic blood pressure >40 mmHg upon admission require 5 days of treatment to recover. Therefore, the following scheduling logic is preset in S2: if the measured pulmonary artery systolic blood pressure of the subject upon admission is ≤40 mmHg, the system automatically determines the subsequent acquisition node as a single node on the 3rd day of treatment, and the timing scheme automatically generated by the system only includes the single node on the 3rd day of treatment, avoiding over-collection of data from mild patients; if the measured value is >40 mmHg, the subsequent acquisition nodes are determined as three nodes on the 1st, 3rd, and 5th day of admission, and a temporary supplementary acquisition node is added on the 2nd day of treatment to closely track the evolution of the condition of severe patients. The 1st day mentioned here refers to the baseline acquisition immediately after admission, which can be regarded as the same examination as the data acquisition described in S1.

[0068] Once the time-series scheme is determined, the system acquires three types of raw image data at the corresponding time points from the ultrasound equipment via a data interface, along with the raw pulmonary artery systolic blood pressure values ​​calculated by the equipment or workstation. These data are then stored in a structured database and automatically associated with the subject's ID and the acquisition point. Simultaneously, the system counts the number of valid samples collected in the current batch to determine if the preset minimum sample size for analysis has been reached. This minimum sample size can be set by the researcher during system initialization, for example, to 50 cases.

[0069] If the criteria are not met, the system will automatically generate a supplementary sampling reminder and push it to the researcher's terminal, prompting that more eligible subjects need to be included.

[0070] S3: Input the collected raw pulmonary artery systolic blood pressure data into the high-altitude adaptive dynamic self-calibration model. This self-calibration model calculates and processes the raw data based on the current altitude of the residence, the real-time blood oxygen saturation value of the subject, and the baseline reference value of pulmonary artery pressure of healthy people at the same altitude gradient, in order to eliminate the physiological pulmonary artery pressure increase bias caused by the high-altitude hypoxic environment and obtain calibrated pulmonary artery systolic blood pressure data.

[0071] The specific self-calibration model is calculated using the following formula:

[0072]

[0073] In the formula, The unit for pulmonary artery systolic blood pressure after calibration is mmHg; The original pulmonary artery systolic pressure measured by the tricuspid regurgitation differential pressure method is in mmHg. This is the baseline reference pulmonary artery systolic blood pressure value in mmHg for healthy individuals at the same altitude gradient. This value is determined by the system based on the current altitude. Retrieved from a pre-built baseline database; The current real-time altitude of the base is in meters (m). This is the reference elevation value for sea level, usually taken as 0m; The preset altitude normalization threshold for plateau environments is in meters (m). In this embodiment, it is set to 5500m to map the altitude of common plateau settlements to the range of 0 to 1. The real-time arterial oxygen saturation percentage of the subject, ranging from 0 to 100; (.) is an exponential function with the natural constant e as its base.

[0074] This formula was developed by the inventor based on actual physiological observations during field research at high altitudes, without directly referencing any existing formula from a single publication. Its design concept mainly consists of two parts, the first part... A linear scaling term adapted to altitude is used to quantify the physiological increase in pulmonary artery pressure that naturally occurs in healthy individuals at different altitudes. The inventors' team collected pulmonary artery pressure data from healthy individuals at multiple training sites ranging from 2800m to 4500m in altitude and found that the baseline value does indeed increase with altitude. Therefore, by multiplying the normalized altitude difference by the baseline value at the same altitude, the amount of physiological increase that should be "subtracted" in the current environment can be calculated.

[0075] The second part is the improved index-based blood oxygen correction item. Conventional clinical experience suggests that decreased blood oxygen saturation is accompanied by increased pulmonary artery pressure, but this relationship is not linear in the high-altitude hypoxic region. The inventors discovered in their data analysis that the increase in pulmonary artery pressure was significantly faster when blood oxygen saturation decreased from 95% to 85% than when it decreased from 100% to 95%. Therefore, an exponential form with base e was chosen: the higher the blood oxygen saturation, the closer the correction factor is to 1, with almost no additional deduction from the original value; the lower the blood oxygen saturation, the larger the correction factor, and the more physiological deviations are deducted. After this correction, Only the abnormally elevated portion caused by the pathological state of high-altitude pulmonary edema is retained, and the interference from physiological pulmonary hypertension at high altitude is no longer included, providing a clean data basis for subsequent disease assessment.

[0076] In a specific application example, a subject was admitted to the hospital at an altitude of 3700m. The blood pressure was 42 mmHg, and the blood oxygen saturation (SaO2) was 85%. The system retrieved the corresponding healthy individuals at this altitude from the baseline database. It is 23 mmHg. Take 5500m, and H0 as 0m. Calculate the altitude adaptation term: 23×(3700-0) / 5500≈15.5mmHg.

[0077] Blood oxygen correction term: exp(1-85 / 100)=exp(0.15)≈1.16.

[0078] The product of the two is approximately 18.0 mmHg. =42-18.0=24.0mmHg. The calibrated value is approximately 27.2mmHg higher than the upper limit of the healthy baseline range but lower than 40mmHg, indicating that the patient has pathologically significant pulmonary artery pressure elevation, which is a state requiring attention but not yet severe. Without this calibration, the original value of 42mmHg might have been directly judged as severe, easily leading to overtreatment.

[0079] S4: Using calibrated pulmonary artery systolic blood pressure data as the core input variable, a pre-constructed correlation calculation model is executed to determine the correlation between multimodal ultrasound features and the severity of high-altitude pulmonary edema. This correlation calculation model uses the Spearman rank correlation coefficient between the calibrated pulmonary artery systolic blood pressure and lung ultrasound score as a weighting factor, integrates the activation values ​​of early latent warning features based on cardiac ultrasound indicators, and incorporates a variance compensation term for the dataset. The calculated correlation matching score is then output. .

[0080] The specific formula for calculating the correlation is as follows:

[0081]

[0082] in, The Spearman rank correlation coefficient is the pre-determined calibrated pulmonary artery systolic blood pressure and lung ultrasound 12-zone score based on historical datasets. The inventors previously analyzed immediate admission data from 174 patients with high-altitude pulmonary edema, obtaining a correlation coefficient of 0.839 between pulmonary artery systolic blood pressure and pulmonary edema score. Therefore, in this model, The value was set to a fixed weight of 0.839 to characterize the inherent strong correlation between pulmonary artery pressure and the severity of pulmonary edema. The Spearman rank correlation coefficient is a well-known statistical indicator, and its standard calculation formula is... ,in For each pair of observations, the rank difference For sample size, please refer to general nonparametric statistics textbooks.

[0083] The hyperbolic tangent activation function is expressed mathematically as follows:

[0084]

[0085] It is a commonly used nonlinear function in neural networks, and here it is used to compress the activation values ​​of latent features to (0, 1). Interval. The index is for early latent warning features, k=1,2,3,4, which correspond to four items: heart rate, pulmonary artery diameter, right ventricular diastolic function index combination, and right ventricular index, respectively. For the first Measured values ​​of latent features, For the first The clinical judgment threshold corresponding to the latent feature. I(. For indicator functions, when The value is 1 if the condition is met, and 0 otherwise. The four early latent warning features and their thresholds are as follows: heart rate > 100 beats / min, pulmonary artery diameter > 25mm, E / A ratio not in the range of 0.8 to 2.1 and E / e' ≥ 6, and right ventricular index RIMP > 0.55.

[0086] This represents the variance explained by the current set of ultrasound features in the sample. This represents the total variance explained by the fully labeled dataset. The ratio of the two is... As a variance compensation term, it is added to the association score to correct the estimation bias that may occur when the feature dimension of a single sample is insufficient, so that the score can simultaneously absorb the distribution information of the entire dataset.

[0087] During model computation, the system first extracts the current subjects' data from the standardized database. For the lung ultrasound 12-zone score Ls, a multiplicative weight of 0.839 was directly used. Next, the latent feature activation terms were calculated: the measured values ​​of each of the four features were compared with the threshold, summed, divided by 4, and then... Function activation; finally, add the variance compensation ratio of the current sample. Theoretically, the value range is between 0 and 1. The higher the value, the stronger the correlation between the subject's multimodal ultrasound characteristics and the severity of high-altitude pulmonary edema, and the more severe the condition.

[0088] It should be noted that the Spearman rank correlation coefficient is a well-known statistical concept, and the hyperbolic tangent function is also a publicly available mathematical function. However, the combination of the indicator function and the variance compensation term was designed by the inventor based on the characteristics of the data in this project. The integration of these three elements into a unified correlation metric is not yet fully disclosed in existing literature.

[0089] S5: Based on the correlation matching score It executes the preset stratified assessment logic, outputs structured stratified assessment report data, and simultaneously generates visualized atlas data consisting of pulmonary artery pressure change curves, lung ultrasound score trends, and right heart function recovery dynamic trajectories; based on the stratification results, it matches and outputs corresponding clinical diagnosis and treatment pathway reference information.

[0090] The specific logic of tiered evaluation is as follows: when When the value is in the 0-0.3 range, a first-level warning is generated, indicating that the subject's current ultrasound findings have a very low correlation with high-altitude pulmonary edema, and the subject is basically in a healthy or well-recovered state, and routine follow-up can be arranged; when... When the 0.31–0.7 range is reached, a second-level warning is generated, corresponding to mild high-altitude pulmonary edema. The system automatically outputs a 3-day routine treatment plan reference, such as recommending routine measures like nasal cannula oxygen therapy and oral nifedipine. When the temperature is in the range of 0.71 to 1, a third-level warning is generated, corresponding to severe high-altitude pulmonary edema. The system automatically outputs reference information for a treatment path that extends to 5 days, including enhanced oxygen therapy, and recommends transferring the patient to a medical unit equipped with a hyperbaric oxygen chamber as soon as possible.

[0091] Furthermore, the stratified assessment logic also includes an important correction rule, which is based on the severe characteristics discovered by the inventors during their research: when the calibrated pulmonary artery systolic blood pressure is >40 mmHg and the aortic pulmonary artery diameter is >25 mm, regardless of... Regardless of the calculated score, the system directly generates the highest warning level, Level 3. This is because in early clinical observations, almost all patients meeting both of these criteria were severely ill, even if some individual cases were due to timing of data collection... The risk is currently too low, but the actual clinical risk remains high. This correction rule can effectively prevent the scoring model from underestimating the risk in boundary cases.

[0092] After the assessment is completed, the system not only outputs a stratified report, but also automatically plots time trend curves based on pulmonary artery pressure, lung ultrasound scores, and right ventricular function indicators at each collection time point, forming a visual atlas. These atlases allow clinicians to see the patient's recovery speed and trajectory at a glance, without the need for manual recording and plotting.

[0093] S6: Perform two-level triage data quality control on the full dataset that has completed the S5 assessment. The first round of quality control verifies the quality of ultrasound image frames, identifies samples with unclear spectra or measurement deviations exceeding the preset range, archives them in the abnormal sample library, and generates a deviation rectification notice with correction guidelines; the second round of quality control verifies the logical consistency between indicators, excludes logically contradictory samples where the lung ultrasound score has returned to zero but the pulmonary artery systolic pressure is still higher than the normal threshold, and triggers a second review process for the contradictory samples. The valid data that has passed the two-level quality control is imported into the standard resource library, and a dual classification label is added to each valid data record: the first label is the prognostic label for delayed recovery of right ventricular function. When the patient's lung ultrasound score is zero but the right ventricular diastolic function has not recovered and the RIMP is still >0.55, he / she is marked as a high-risk group for recurrence;

[0094] The second label is either a first-level label or a second-level label based on the correlation matching score interval, corresponding to mild cases cured in 3 days, severe cases cured in 5 days, etc.

[0095] The two-tiered quality control design here is entirely geared towards ensuring the authenticity and usability of high-altitude pulmonary edema data. Under high-altitude field conditions, ultrasound examinations may produce suboptimal images due to operator inexperience, equipment movement, or poor patient coordination. The first round of quality control identifies these samples and provides corrective guidance, such as adjusting the probe angle and re-acquiring the tricuspid regurgitation spectrum, allowing on-site personnel to provide timely remediation. The second round of quality control addresses potential data entry errors or algorithmic margins. For example, some patients may have completely absorbed pulmonary edema, but their pulmonary artery pressure remains high due to other cardiopulmonary diseases. Logically, this should not be simply considered "unresolved pulmonary edema" but requires manual verification. This two-tiered screening ensures the reliability of the data ultimately entering the standard resource library.

[0096] The addition of dual labels stems directly from the inventors' core early findings: in some patients, after their lung ultrasound scores were reset to zero, the recovery of right ventricular function, especially RIMP, was significantly delayed, and these patients had a higher risk of relapse later on; while the mild and severe disease grading labels based on correlation score intervals provide high-quality labeled samples for future model iterations and recommendations for precision treatment plans. This entire quality control and labeling process enables the system to accumulate high-quality structured data assets during continuous operation.

[0097] Example 2

[0098] refer to Figure 2 As shown, this embodiment provides a high-altitude pulmonary edema ultrasound early warning system based on pulmonary artery pressure data, used to implement the method described in Embodiment 1. The system includes an enrollment verification and access control module, a differentiated time-series scheduling module, a self-calibration calculation module, a correlation assessment module, a stratified assessment and clinical push module, and a data post-processing and quality control module.

[0099] The enrollment verification and access control module is used to obtain the measured pulmonary artery systolic pressure and physiological parameters of subjects immediately upon admission, perform two-dimensional data verification, and generate a deviation notification command and terminate the subsequent data acquisition process when the verification fails. This module embeds a baseline pulmonary artery pressure database for healthy individuals at high altitudes and a tricuspid regurgitation signal integrity detection submodule.

[0100] The differentiated timing scheduling module is used to dynamically determine the acquisition timing scheme of subsequent multimodal ultrasound data based on the measured value of pulmonary artery systolic blood pressure immediately upon admission, schedule and collect two-dimensional grayscale ultrasound, Doppler blood flow ultrasound and tissue Doppler ultrasound data, and generate supplementary collection reminder instructions when the effective sample volume is insufficient.

[0101] The self-calibration calculation module has a built-in high-altitude adaptive dynamic self-calibration model, which is used to correct the deviation of the raw pulmonary artery systolic blood pressure data and output calibrated pulmonary artery systolic blood pressure data. The calculation logic of this module is exactly as shown in the formula in Example 1.

[0102] The correlation assessment module is used to execute a pre-built correlation calculation model between multimodal ultrasound features and disease severity. It takes the calibrated pulmonary artery systolic blood pressure data and early latent warning features as inputs, calculates and outputs the disease correlation matching score.

[0103] The stratified assessment and clinical recommendation module is used to perform stratified assessment based on the correlation matching score, output structured assessment reports and visualized trend map data, and match and output corresponding clinical diagnosis and treatment path reference information.

[0104] The data post-processing and quality control module is used to perform two-level triage quality control on the evaluated dataset, add prognostic and grading labels to the valid data that passes the quality control, and import the labeled data into the standard resource library.

[0105] The connections and collaborative mechanisms between the modules described above have been explained in detail in the method steps of Embodiment 1, and will not be repeated here. The system can be implemented using a general computer server architecture. The ultrasound equipment sends data to the server via a standard medical digital image transmission protocol, and the server has computer programs installed to implement the functions of each module.

[0106] Example 3

[0107] refer to Figures 1-7 As shown, this embodiment provides an electronic device, including a memory and a processor. The memory stores a computer program that can be called by the processor. The processor executes the high-altitude pulmonary edema ultrasound early warning processing method based on pulmonary artery pressure data as described in Embodiment 1 by calling the computer program.

[0108] This electronic device can be a standalone server or workstation, or a computing unit embedded in an ultrasound diagnostic instrument. Its hardware configuration may vary depending on the scenario, but it should include at least one processor and at least one memory, as well as necessary network interfaces and input / output interfaces to complete data reception, processing, and feedback.

[0109] Example 4:

[0110] This embodiment provides a computer-readable storage medium storing an erasable and rewritable computer program. When the computer program is run on a computer device, it causes the computer device to perform the method described in Embodiment 1.

[0111] The computer-readable storage medium can be any available medium such as a read-only memory, random access memory, optical disk, hard disk, solid-state drive, or optical data storage device.

[0112] The above provides a detailed description of the high-altitude pulmonary edema ultrasound early warning system and method based on pulmonary artery pressure data provided by the present invention. It should be noted that the specific parameters and thresholds described in the specific embodiments, such as the baseline range of 20.7±6.5 mmHg, the cutoff value of 40 mmHg, and the correlation coefficient of 0.839, are all derived from real research data accumulated by the inventors in the high-altitude field. They should not be construed as the present invention having to be completely bound by these specific values. After mastering the core inventive concept of the present invention, those skilled in the art can make appropriate adjustments or optimizations to the above parameters based on new data accumulated in actual application scenarios, but these adjustments should still fall within the protection scope of the present invention.

[0113] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed in this invention can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0114] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0115] In the description of this specification, references to terms such as "an embodiment," "example," "specific example," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

Claims

1. An information processing method based on pulmonary artery pressure data, executed by a computer system for auxiliary assessment of high-altitude pulmonary edema, characterized in that, Includes the following steps: S1: Obtain the measured value of pulmonary artery systolic pressure and physiological parameters of the subject immediately upon admission, and perform enrollment eligibility verification; the verification includes comparing the measured value with the pulmonary artery pressure reference baseline of healthy people at high altitude, and determining whether the tricuspid regurgitation signal meets the preset calculation conditions; If the verification fails, the subsequent data collection process will be terminated and a verification failure message will be generated. S2: For subjects who pass the verification, a timing schedule for the acquisition of subsequent multimodal ultrasound data is dynamically generated based on the numerical range of the measured values ​​immediately upon admission. According to the scheduling schedule, the two-dimensional grayscale ultrasound, Doppler blood flow ultrasound, and tissue Doppler ultrasound data of the corresponding subjects are acquired and associated to form the effective sample set for the current batch. It is determined whether the sample size in the effective sample set has reached a preset threshold. If it has not, a supplementary sampling prompt is generated. S3: Input the collected raw pulmonary artery systolic pressure into the preset self-calibration model, and perform calibration calculations using the current altitude of the residence, the real-time blood oxygen saturation of the subject, and the pulmonary artery pressure reference baseline value of healthy people at the same altitude to eliminate the deviation of physiological pulmonary artery pressure increase caused by high altitude hypoxia environment, and obtain the calibrated pulmonary artery systolic pressure. S4: Using the calibrated pulmonary artery systolic pressure as the core input, construct a correlation calculation model between multimodal ultrasound features and disease severity; the input parameters of the correlation calculation model include the calibrated pulmonary artery systolic pressure and a variance compensation term determined based on the feature variance ratio of the current sample and the full dataset, and calculate the disease correlation matching score through this model; S5: Perform stratified assessment calculation based on the correlation matching score, output visualized atlas data that includes at least the pulmonary artery pressure change curve, lung ultrasound score trend and right heart function recovery dynamic trajectory, and push matching clinical diagnosis and treatment path reference information based on the stratification results.

2. The method according to claim 1, characterized in that, The specific verification process described in S1 is as follows: When the measured value falls within the preset pulmonary artery systolic pressure reference baseline range that characterizes healthy people at high altitudes, it is determined to be a non-target object and excluded; at the same time, the integrity of the tricuspid regurgitation spectrum is checked. If the pulmonary artery pressure cannot be calculated due to the absence of regurgitation signal, it is determined to be an unqualified sample, and the subsequent collection process of the sample is terminated.

3. The method according to claim 1, characterized in that, The dynamic generation of subsequent multimodal ultrasound data acquisition timing scheduling scheme described in S2 includes: If the measured value immediately upon admission is not greater than the first threshold, for example, 40 mmHg, then the subsequent acquisition nodes are compressed into a single node; If the measured value is greater than the first threshold, then at least three acquisition nodes covering the first, third, and fifth days after admission are retained, and a temporary supplementary acquisition task is triggered on the second day.

4. The method according to claim 1, characterized in that, The self-calibration model described in S3 performs the following calculations to obtain the calibrated pulmonary artery systolic pressure. : In the formula, To calibrate postpulmonary artery systolic pressure, These are the raw measured values ​​obtained by the tricuspid regurgitation pressure differential method. This represents the baseline pulmonary artery systolic blood pressure value for healthy individuals at the same altitude gradient. This is the current real-time altitude of the base. This is the reference elevation value for sea level. The preset altitude normalization threshold for plateau environments is used. This represents the subject's real-time arterial blood oxygen saturation percentage. For the natural constant An exponential function with base 0.

5. The method according to claim 1, characterized in that, The disease-related matching score mentioned in S4 The calculation formula is: in, Spearman's rank correlation coefficient between calibrated pulmonary artery systolic blood pressure and lung ultrasound 12-zone score, which was pre-determined based on historical datasets; It is the hyperbolic tangent activation function; Index of early hidden warning features. For the first Measured values ​​of latent features, For the first Clinical judgment thresholds corresponding to latent features; For indicator functions, satisfying The value is 1 if the condition is met, otherwise it is 0. This represents the variance explained by the current set of ultrasound features in the sample. The total variance explained by the fully labeled dataset; The early latent warning features consist of 4 items, and their corresponding preset clinical judgment thresholds are as follows: heart rate greater than 100 beats / min; pulmonary artery diameter greater than 25mm; E / A ratio not in the range of 0.8 to 2.1 and E / e' not less than 6; right ventricular index (RIMP) greater than 0.

55.

6. The method according to claim 5, characterized in that, S5 describes scoring based on the correlation matching degree. Performing stratified evaluation calculations specifically includes: When When the value is in the range of 0 to 0.3, it is judged as either healthy or excluded; when... When the value is between 0.31 and 0.7, it is considered mild high-altitude pulmonary edema, corresponding to the first type of treatment pathway; when When the value is in the range of 0.71 to 1, it is considered severe high-altitude pulmonary edema, corresponding to the second type of treatment pathway; Furthermore, when the calibrated pulmonary artery systolic pressure is greater than the preset critical illness pressure threshold and the main pulmonary artery diameter is greater than the preset critical illness diameter threshold, it is directly calculated and marked as a critical illness warning state.

7. The method according to claim 1, characterized in that, The method further includes: S6: Performing a two-level triage quality control operation on the completed assessment dataset; The first round verifies the ultrasound image frame quality parameters, identifies and archives abnormal samples, and generates deviation rectification prompts with correction guidelines; The second round verifies the logical consistency parameters of the indicators, excludes contradictory samples where the lung ultrasound score has returned to zero but the pulmonary artery systolic pressure is still higher than the normal threshold, and triggers a secondary review process; The valid data that has passed the two-level quality control is imported into the standard resource library, and each sample is labeled with dual labels: The first label is a prognostic label for delayed recovery of right ventricular function, when the patient's lung ultrasound score is zero but the right ventricular diastolic function has not recovered and the RIMP is still greater than 0.55, the patient is marked as a high-risk group for recurrence; The second label is a graded label generated based on the correlation matching score interval.

8. A high-altitude pulmonary edema ultrasound early warning system based on pulmonary artery pressure data, used to implement the method as described in any one of claims 1 to 7, characterized in that, include: The enrollment verification and process control module is used to obtain the measured value of pulmonary artery systolic blood pressure and physiological parameters of subjects immediately upon admission, perform verification, and terminate the data collection process and generate a prompt when the verification fails. The differentiated timing scheduling module is used to dynamically generate a timing scheduling scheme for the acquisition of subsequent multimodal ultrasound data based on the immediate measured values ​​upon admission, collect two-dimensional grayscale ultrasound, Doppler blood flow ultrasound, and tissue Doppler ultrasound data, and generate a supplementary collection prompt when the effective sample size is insufficient; the self-calibration calculation module has a built-in self-calibration model as described in claim 4, which is used to correct the deviation of the original pulmonary artery systolic pressure and output the calibrated pulmonary artery systolic pressure; the correlation calculation and evaluation module is used to construct the correlation calculation model as described in claim 5, and calculate the disease correlation matching score with the calibrated pulmonary artery systolic pressure and early latent warning features as input; The stratified assessment and information push module is used to perform stratified assessment calculations based on the correlation matching score, output visualized trend map data, and push matching clinical diagnosis and treatment path reference information. The data quality control and labeling module is used to perform the two-level split quality control operation as described in claim 7 on the evaluation dataset, to label the valid data with the dual labels and import them into the standard resource library.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1 to 7.

10. An electronic device, characterized in that, Includes: memory, used to store instructions; A processor for executing the instructions, causing the device to perform the method as described in any one of claims 1 to 7.

Images