Risk assessment method and system based on vascular ultrasound image, terminal and storage medium

Abstract

The invention relates to the technical field of intelligent ultrasonic examination, and discloses a risk assessment method and system based on a vascular ultrasonic image, a terminal and a storage medium, and the method comprises the steps: obtaining ultrasonic echo data of a wearable ultrasonic patch, carrying out the reconstruction processing of the ultrasonic echo data, and obtaining the vascular ultrasonic image; acquiring clinical risk factors of a patient, analyzing the vascular ultrasound image according to a vein recognition and compressibility analysis algorithm and a texture and blood flow feature extraction algorithm to obtain vein image features, and inputting the vein image features and the clinical risk factors into the trained risk assessment model for assessment to obtain a venous thrombosis risk score; and obtaining a timestamp and historical data, determining a risk level according to the timestamp, the historical data and the venous thrombosis risk score, and sending alarm information to medical personnel. The venous thrombosis risk is dynamically assessed by constructing the risk assessment model, and accurate monitoring and graded early warning of the venous thrombosis risk of the patient are achieved.

Description

Technical Field

[0001] This invention relates to the field of intelligent ultrasound examination technology, and in particular to a risk assessment method, system, terminal, and computer-readable storage medium based on vascular ultrasound images. Background Technology

[0002] In the current healthcare system, the diagnosis of venous thromboembolism primarily relies on imaging examinations, especially color Doppler ultrasound. Traditionally, ultrasound examinations are performed by sonographers or qualified physicians, and the results depend on manual observation of venous compressibility, luminal echogenicity, and blood flow signals, generating written reports and partial image archiving. This traditional approach is suitable for single or low-frequency examinations, but for high-risk groups requiring continuous monitoring of thrombosis risk, such as patients after major orthopedic surgery, post-cardiac surgery patients, and intensive care patients, continuous manual ultrasound examinations are impractical due to both manpower and equipment limitations.

[0003] Existing methods for vascular ultrasound thrombosis monitoring include: traditional ultrasound examination methods based on manual interpretation, remote monitoring methods based on physiological signal acquisition, and machine learning-assisted diagnostic methods based on static image analysis. These methods are limited by their inability to process continuous ultrasound data, lack of automated risk assessment capabilities, and insufficient centralized management of multiple devices. Consequently, they struggle to achieve real-time monitoring and early warning of venous thrombosis. This means that even when continuous, multi-site thrombosis risk assessment is needed for high-risk populations, intermittent manual examinations remain essential, making accurate early warning of thrombosis risk impossible. Therefore, existing vascular ultrasound image risk assessment technologies require further improvement and optimization. Summary of the Invention

[0004] The main objective of this invention is to provide a risk assessment method, system, terminal, and computer-readable storage medium based on vascular ultrasound images, aiming to solve the problems that existing vascular ultrasound thrombosis monitoring methods are unable to achieve real-time monitoring and early warning of venous thrombosis, and cannot accurately predict thrombosis risk.

[0005] To achieve the above objectives, the present invention provides a risk assessment method based on vascular ultrasound images, the risk assessment method based on vascular ultrasound images comprising the following steps: Acquire ultrasound echo data from a wearable ultrasound patch, and reconstruct the ultrasound echo data using an image reconstruction algorithm to obtain a vascular ultrasound image. The patient's clinical risk factors are obtained, and the vascular ultrasound image is analyzed according to the vein identification and compressibility analysis algorithm and the texture and blood flow feature extraction algorithm to obtain vein image features. The vein image features and the clinical risk factors are input into the trained risk assessment model for evaluation to obtain a venous thrombosis risk score. Obtain the current timestamp and historical data, determine the risk level based on the timestamp, the historical data and the venous thrombosis risk score, and send an alarm message to medical staff.

[0006] Optionally, the risk assessment method based on vascular ultrasound images, wherein acquiring ultrasound echo data from a wearable ultrasound patch and reconstructing the ultrasound echo data using an image reconstruction algorithm to obtain a vascular ultrasound image specifically includes: Acquire ultrasound echo data from wearable ultrasound patches; Gain compensation is performed on the ultrasonic echo data using an image reconstruction algorithm to obtain multiple echo amplitude distributions. The amplitude distributions of multiple echoes are reconstructed using an interpolation reconstruction algorithm or a beamforming algorithm to obtain a vascular ultrasound image.

[0007] Optionally, the risk assessment method based on vascular ultrasound images, wherein reconstructing the amplitude distributions of multiple echoes using an interpolation reconstruction algorithm or a beamforming algorithm to obtain a vascular ultrasound image specifically includes: The amplitude distributions of multiple echoes are reconstructed using an interpolation reconstruction algorithm or a beamforming algorithm to obtain a two-dimensional grayscale image; A one-dimensional time depth map is determined based on the time axis and the distribution of multiple echo amplitudes. The two-dimensional grayscale image and the one-dimensional time-depth map are mapped to obtain a vascular ultrasound image.

[0008] Optionally, the risk assessment method based on vascular ultrasound images, wherein obtaining the patient's clinical risk factors, analyzing the vascular ultrasound image according to a vein identification and compressibility analysis algorithm and a texture and blood flow feature extraction algorithm to obtain vein image features, and inputting the vein image features and the clinical risk factors into a trained risk assessment model for evaluation to obtain a venous thrombosis risk score, specifically includes: Obtain the patient's clinical risk factors; The vascular ultrasound image is identified using a vein recognition and compressibility analysis algorithm to obtain the lumen geometric parameters; The texture and blood flow features of the lumen geometry parameters are extracted using a texture and blood flow feature extraction algorithm to obtain venous image features; The venous imaging features and the clinical risk factors are input into the trained risk assessment model for evaluation to obtain a venous thrombosis risk score. The risk assessment model includes any one of the following: logistic regression model, random forest model, gradient boosting tree model, and neural network model; The geometric parameters of the lumen include cross-sectional area, diameter, rate of change of cross-sectional area, and diameter shrinkage ratio.

[0009] Optionally, the risk assessment method based on vascular ultrasound images, wherein the step of inputting the venous imaging features and the clinical risk factors into a trained risk assessment model for evaluation to obtain a venous thrombosis risk score, specifically includes: Obtain a historical pathology dataset, and divide the historical pathology dataset into a training set and a validation set; The training set is input into the initial risk assessment model for training to obtain the trained risk assessment model. The venous imaging features and the clinical risk factors are input into the trained risk assessment model for evaluation to obtain a venous thrombosis risk score. The validation set is used to fine-tune the structure and parameters of the trained risk assessment model.

[0010] Optionally, in the aforementioned risk assessment method based on vascular ultrasound images, the blood flow characteristics include peak blood flow velocity, average blood flow velocity, blood flow direction, spectral envelope morphology, and time proportion. The step of extracting texture and blood flow features from the lumen geometric parameters using a texture and blood flow feature extraction algorithm to obtain venous image features specifically includes: Based on the gray-level co-occurrence matrix, the texture features of the lumen geometric parameters are calculated according to the texture and blood flow feature extraction algorithm to obtain the first target feature; Feature extraction is performed on the peak blood flow velocity, the average blood flow velocity, the blood flow direction, the spectral envelope morphology, and the time ratio to obtain the second target feature. The vein image features are then determined based on the first target feature and the second target feature.

[0011] Optionally, the risk assessment method based on vascular ultrasound images, wherein obtaining the current timestamp and historical data, determining the risk level based on the timestamp, the historical data, and the venous thrombosis risk score, and sending alarm information to medical staff, specifically includes: Get the current timestamp and historical data; Based on the trend analysis module, the historical data is compared with the timestamp, the venous thrombosis risk score, and the clinical risk factors to obtain the time trend; If the venous thrombosis risk score exceeds a preset threshold or the time trend is upward, then the risk level is mapped according to the venous thrombosis risk score, and an alarm message is generated according to the risk level to remind medical staff. The risk levels include low risk, medium risk, and high risk.

[0012] Furthermore, to achieve the above objectives, the present invention also provides a risk assessment system based on vascular ultrasound images, wherein the risk assessment system based on vascular ultrasound images: An ultrasound image processing module is used to acquire ultrasound echo data from a wearable ultrasound patch, and to reconstruct the ultrasound echo data according to an image reconstruction algorithm to obtain a vascular ultrasound image. The thrombosis risk assessment module is used to acquire the patient's clinical risk factors, analyze the vascular ultrasound image according to the vein recognition and compressibility analysis algorithm and the texture and blood flow feature extraction algorithm to obtain vein image features, and input the vein image features and the clinical risk factors into the trained risk assessment model for evaluation to obtain a venous thrombosis risk score. The risk warning module is used to obtain the current timestamp and historical data, determine the risk level based on the timestamp, the historical data and the venous thrombosis risk score, and send alarm information to medical staff.

[0013] Furthermore, to achieve the above objectives, the present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a risk assessment program based on vascular ultrasound images, and when the risk assessment program based on vascular ultrasound images is executed by a processor, it implements the steps of the risk assessment method based on vascular ultrasound images as described above.

[0014] In this invention, ultrasound echo data from a wearable ultrasound patch is acquired, and the ultrasound echo data is reconstructed using an image reconstruction algorithm to obtain a vascular ultrasound image. The patient's clinical risk factors are acquired, and the vascular ultrasound image is analyzed using a vein recognition and compressibility analysis algorithm and a texture and blood flow feature extraction algorithm to obtain vein image features. These vein image features and the clinical risk factors are input into a trained risk assessment model for evaluation to obtain a venous thrombosis risk score. The current timestamp and historical data are acquired, and the risk level is determined based on the timestamp, historical data, and venous thrombosis risk score. An alarm message is then sent to medical personnel. This invention dynamically assesses the risk of venous thrombosis through image reconstruction, vein feature extraction, and a risk assessment model, achieving accurate monitoring and graded early warning of the patient's venous thrombosis risk. Attached Figure Description

[0015] Figure 1 This is a flowchart of a preferred embodiment of the risk assessment method based on vascular ultrasound images of the present invention; Figure 2 This is a structural diagram of a preferred embodiment of the risk assessment system based on vascular ultrasound images of the present invention; Figure 3This is a structural diagram of a preferred embodiment of the terminal of the device of the present invention. Detailed Implementation

[0016] To make the objectives, technical solutions, and advantages of this invention clearer and more explicit, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0017] Existing methods for vascular ultrasound thrombosis monitoring include traditional ultrasound examination methods based on manual interpretation, remote monitoring methods based on physiological signal acquisition, and machine learning-assisted diagnostic methods based on static image analysis. These methods are limited by their inability to process continuous ultrasound data, lack of automated risk assessment capabilities, and insufficient centralized management of multiple devices. They struggle to achieve real-time monitoring and early warning of venous thrombosis, resulting in reliance on intermittent manual examinations when continuous, multi-site thrombosis risk assessment is needed for high-risk populations. This makes accurate early warning of thrombosis risk impossible. Therefore, a risk assessment method based on vascular ultrasound images is needed. This method should dynamically assess venous thrombosis risk through image reconstruction, venous feature extraction, and risk assessment models, enabling accurate monitoring and tiered early warning of venous thrombosis risk in patients, thus avoiding the problem of inaccurate early warning of thrombosis risk.

[0018] The risk assessment method based on vascular ultrasound images described in the preferred embodiment of the present invention, such as... Figure 1 As shown, the risk assessment method based on vascular ultrasound images includes the following steps: Step S10: Obtain the current timestamp and historical data, determine the risk level based on the timestamp, the historical data and the venous thrombosis risk score, and send an alarm message to medical staff.

[0019] Step S10 includes: Step S11: Acquire ultrasound echo data collected by the wearable ultrasound patch; Step S12: Perform gain compensation on the ultrasonic echo data according to the image reconstruction algorithm to obtain multiple echo amplitude distributions; Step S13: Reconstruct the amplitude distribution of multiple echoes according to an interpolation reconstruction algorithm or a beamforming algorithm to obtain a vascular ultrasound image.

[0020] Specifically, the ultrasound echo data acquired by the wearable ultrasound patch is obtained (the terminal device receives a complete set of ultrasound echo data frames from the front-end transmitter and classifies and caches them through the data receiving and caching module). The ultrasound echo data is then subjected to gain compensation according to the image reconstruction algorithm to obtain multiple echo amplitude distributions (the image reconstruction module performs time gain compensation, bandpass filtering, and envelope detection on the echo signals of each channel based on the transducer array geometric parameters and acquisition timing to calculate the echo amplitude distributions at different depths). The multiple echo amplitude distributions are then reconstructed according to the interpolation reconstruction algorithm or beamforming algorithm to obtain a vascular ultrasound image.

[0021] As an example, each set of data frames contains at least one multi-channel echo signal (ultrasound echo data) collected by the transducer array within a monitoring cycle, along with patient number, patch number, patch location code, and timestamp information. The terminal device first assigns the data to the corresponding patient according to the patient number, and then distinguishes different limbs or different monitoring sites according to the patch number and location code, ensuring that subsequent processing can analyze "a specific vein of a specific patient".

[0022] Further, the multiple echo amplitude distributions are reconstructed using an interpolation reconstruction algorithm or a beamforming algorithm to obtain a two-dimensional grayscale image (using beamforming or interpolation reconstruction algorithms, multiple echo lines along the array direction are combined into a two-dimensional grayscale B-type image (two-dimensional grayscale image)); based on the time axis, a one-dimensional time depth map is determined according to the multiple echo amplitude distributions (a one-dimensional time depth map is constructed along the time axis, which is used to reflect the echo structure of the local vein and surrounding tissue at the monitoring time); the two-dimensional grayscale image and the one-dimensional time depth map are mapped to obtain a vascular ultrasound image.

[0023] In this embodiment, the patch identification information is based on the identification chip and binding process set in the ultra-thin vascular ultrasound signal acquisition patch and its connected transmitter as described in Patent 1. It is used to distinguish multiple patches and their installation positions between the same patient or different patients. The patch identification information may include at least: (1) the patch's own unique serial number or identification code, used to distinguish different physical patches; (2) the transmitter number electrically connected to the patch, used to identify the specific channel of the data source; (3) the monitoring site code preset or registered by the patch, such as "the inner side of the left calf", "the back side of the right calf", "the front side of the left thigh", etc., used to clarify which position of the patch in Patent 1 is actually attached to the thigh, calf or other limb; (4) optional side information and direction information, such as left / right side, proximal / distal direction, used to intuitively display the spatial orientation of the patch on the limb on the terminal interface; (5) the patch's model specifications, production batch and usage status mark recorded by the identification chip, etc., used for service life management and quality traceability.

[0024] Step S20: Obtain the patient's clinical risk factors, analyze the vascular ultrasound image according to the vein identification and compressibility analysis algorithm and the texture and blood flow feature extraction algorithm to obtain vein image features, input the vein image features and the clinical risk factors into the trained risk assessment model for evaluation, and obtain a venous thrombosis risk score.

[0025] Step S20 includes: Step S21: Obtain the patient's clinical risk factors; Step S22: The vascular ultrasound image is identified according to the vein identification and compressibility analysis algorithm to obtain the lumen geometric parameters; Step S23: Extract the texture and blood flow features of the lumen geometric parameters according to the texture and blood flow feature extraction algorithm to obtain the vein image features; Step S24: Input the venous imaging features and the clinical risk factors into the trained risk assessment model for evaluation to obtain a venous thrombosis risk score.

[0026] Specifically, the patient's clinical risk factors (patient's age, gender, type of surgery, length of hospital stay, history of thrombosis, tumor or pregnancy status, and use of anticoagulants) are obtained. The vascular ultrasound image is then identified using a vein recognition and compressibility analysis algorithm to obtain luminal geometric parameters (identifying the target vein lumen in the reconstructed image and calculating the luminal geometric parameters). Texture and blood flow features of the luminal geometric parameters are extracted using a texture and blood flow feature extraction algorithm to obtain venous image features (acquiring texture and blood flow features within and around the lumen). These venous image features and the clinical risk factors are then input into a trained risk assessment model for evaluation to obtain a venous thrombosis risk score. The risk assessment model includes any one of logistic regression, random forest, gradient boosting tree, and neural network models. The luminal geometric parameters include cross-sectional area, diameter, rate of change of cross-sectional area, and diameter contraction ratio.

[0027] In this embodiment, venous imaging features include venous compressibility indicators, luminal geometric changes, echo intensity and texture features, and hemodynamic features. Patient clinical risk factors include: age, sex, body mass index, type and duration of surgery, duration of bed rest or immobilization, presence of malignancy, history of venous thrombosis, peripheral phlebitis or infection, presence of a central venous catheter, and whether anticoagulants are currently in use or recently discontinued. The risk assessment model can be a rule-based scoring model or a machine learning model trained from scratch, such as logistic regression, random forest, gradient boosting tree, or neural network models.

[0028] Step S24 includes: Step S241: Obtain historical pathology dataset, and divide the historical pathology dataset to obtain training set and validation set; Step S242: Input the training set into the initial risk assessment model for training to obtain the trained risk assessment model; Step S243: Input the venous imaging features and the clinical risk factors into the trained risk assessment model for evaluation to obtain a venous thrombosis risk score.

[0029] Specifically, a historical pathology dataset is acquired, and the dataset is divided into a training set and a validation set. The training set is then input into an initial risk assessment model for training, resulting in a post-trained risk assessment model. (The historical pathology dataset is divided into a training set and a validation set. Supervised learning methods are used to fit the model parameters in the training set, so that the risk score output by the post-trained risk assessment model can distinguish between cases with and without thrombosis as much as possible.) The venous imaging features and the clinical risk factors are input into the post-trained risk assessment model for evaluation, resulting in a venous thrombosis risk score. The validation set is used to fine-tune the structure and parameters of the post-trained risk assessment model (the model's sensitivity, specificity, and area under the receiver operating characteristic curve are evaluated in the validation set to fine-tune the model's structure and parameters).

[0030] In this embodiment, a certain number of historical pathology datasets are collected. Each record in the historical pathology dataset includes continuously monitored imaging features, corresponding clinical risk factors at the time point, and annotations of the presence or absence of venous thrombosis provided by standard examinations (such as routine color Doppler ultrasound or imaging follow-up). Secondly, the trained risk assessment model is divided into a training set and a validation set. In the training set, supervised learning methods are used to fit the model parameters so that the risk score output by the trained risk assessment model can distinguish between cases with and without thrombosis as much as possible. Thirdly, the sensitivity, specificity, and area under the receiver operating characteristic curve of the model are evaluated in the validation set to optimize the model structure and parameters. Finally, the model whose performance meets the preset requirements is solidified into the terminal device, and it is periodically retrained or calibrated based on the accumulated new case data during actual operation.

[0031] Step S23 includes: Step S231: Based on the gray-level co-occurrence matrix, perform feature calculation on the texture features of the lumen geometric parameters according to the texture and blood flow feature extraction algorithm to obtain the first target feature; Step S232: Extract features from the peak blood flow velocity, the average blood flow velocity, the blood flow direction, the spectral envelope morphology, and the time ratio to obtain the second target feature, and determine the vein image features based on the first target feature and the second target feature.

[0032] Specifically, based on the gray-level co-occurrence matrix, the texture features of the lumen geometric parameters are calculated according to the texture and blood flow feature extraction algorithm to obtain the first target feature. The peak blood flow velocity, the average blood flow velocity, the blood flow direction, the spectral envelope morphology, and the time ratio are extracted to obtain the second target feature. The vein image features are determined according to the first target feature and the second target feature. The blood flow features include peak blood flow velocity, average blood flow velocity, blood flow direction, spectral envelope morphology, and time ratio.

[0033] In this embodiment, texture features may include: the average, standard deviation, maximum, and minimum grayscale values ​​of the region of interest within the lumen; statistical quantities such as contrast, homogeneity, energy, entropy, and correlation calculated based on the grayscale co-occurrence matrix; and short-run emphasis and long-run emphasis based on the grayscale run matrix, used to characterize the differences in roughness and uniformity between the hyperechoic mass in the thrombus area and the normal congested venous lumen. Blood flow features may include, but are not limited to: blood flow features measured in Doppler mode, including peak blood flow velocity, average blood flow velocity, blood flow direction, spectral envelope morphology, the time proportion of recurring low-velocity or no-flow states, and the rate of change of blood flow velocity between resting and light-pressure states, used to reflect whether there is a significant slowdown or obstruction in local hemodynamics.

[0034] Step S30: Obtain the current timestamp and historical data, determine the risk level based on the timestamp, the historical data and the venous thrombosis risk score, and send an alarm message to medical staff.

[0035] Step S30 includes: Step S31: Obtain the current timestamp and historical data; Step S32: Based on the trend analysis module, compare the historical data with the timestamp, the venous thrombosis risk score, and the clinical risk factors to obtain the time trend; Step S33: If the venous thrombosis risk score exceeds a preset threshold or the time trend is upward, then the risk level is mapped according to the venous thrombosis risk score, and an alarm message is generated according to the risk level to remind medical staff.

[0036] Specifically, the system acquires the current timestamp and historical data (storing information such as risk scores, key imaging features, and venous compressibility and geometric morphology indicators in historical data storage, and associating them with previous records in chronological order, so that the historical data and trend analysis module can generate venous compressibility change curves, long-term change curves of texture and blood flow characteristics, and time trend graphs of risk scores, providing a basis for doctors to review the evolution of the disease, evaluate the effectiveness of interventions, and optimize risk assessment models). Based on the trend analysis module, the historical data is compared with the timestamp, the venous thrombosis risk score, and the clinical risk factors to obtain the time trend. If the venous thrombosis risk score exceeds a preset threshold or the time trend is upward, a risk level is mapped according to the venous thrombosis risk score, and an alarm message is generated to remind medical staff based on the risk level, wherein the risk level includes low risk, medium risk, and high risk.

[0037] In this embodiment, venous compressibility indicators, echo intensity characteristics, texture characteristics, and blood flow characteristics are used as image inputs. These are combined with clinical risk factors such as the patient's age, gender, type of surgery, length of hospital stay, history of thrombosis, tumor or pregnancy status, and use of anticoagulants. The data are then input into a rule-based model or a machine learning model to obtain a venous thrombosis risk score. The venous thrombosis risk score can be a probability value in the range of 0 to 1, or a standardized score of 0 to 100. For example, it can be a "current thrombosis probability score" indicating the presence of venous thrombosis, or a "predicted risk score" indicating the risk of new or progressive thrombosis occurring within a preset time window (e.g., the next 24 to 72 hours). The scores can be further mapped to low-risk, medium-risk, and high-risk levels. When any score exceeds a preset threshold or shows a continuous upward trend within a certain time window, the module automatically generates an alarm message and pushes it to relevant medical staff.

[0038] For example, the terminal device compares the venous thrombosis risk score with a preset threshold. If the score exceeds the threshold or shows a significant upward trend within a certain period, an alarm is generated through the risk assessment and alarm module. The alarm time, relevant imaging feature indicators, and the current patch position are recorded, and medical staff are notified via interface display or push notification. The upward trend can be determined by calculating the moving average, slope, or difference between two adjacent scores of the risk score over a preset time window (e.g., the most recent 24 or 48 hours). For example, if the venous thrombosis risk score of the three most recent monitoring tests increases from 0.3, 0.5, to 0.7, and the moving average exceeds the high-risk threshold, the system can determine that there is a significant upward trend and issue an early warning even if the single absolute value has not yet reached an extremely high level.

[0039] Furthermore, such as Figure 2As shown, based on the above-described risk assessment method based on vascular ultrasound images, the present invention also provides a risk assessment system based on vascular ultrasound images, wherein the risk assessment system based on vascular ultrasound images includes: The ultrasound image processing module 51 is used to acquire ultrasound echo data of the wearable ultrasound patch, and to reconstruct the ultrasound echo data according to the image reconstruction algorithm to obtain a vascular ultrasound image. The thrombosis risk assessment module 52 is used to acquire the patient's clinical risk factors, analyze the vascular ultrasound image according to the vein identification and compressibility analysis algorithm and the texture and blood flow feature extraction algorithm to obtain vein image features, and input the vein image features and the clinical risk factors into the trained risk assessment model for evaluation to obtain a venous thrombosis risk score. The risk warning module 53 is used to obtain the current timestamp and historical data, determine the risk level based on the timestamp, the historical data and the venous thrombosis risk score, and send alarm information to medical staff.

[0040] Furthermore, such as Figure 3 As shown, based on the above-mentioned risk assessment method and system based on vascular ultrasound images, the present invention also provides a terminal, which includes a processor 10, a memory 20 and a display 30. Figure 3 Only some of the terminal components are shown; however, it should be understood that it is not required to implement all of the components shown, and more or fewer components may be implemented instead.

[0041] In some embodiments, the memory 20 may be an internal storage unit of the terminal, such as a hard disk or memory. In other embodiments, the memory 20 may be an external storage device of the terminal, such as a plug-in hard disk, smart media card (SMC), secure digital card (SD), flash card, etc. Further, the memory 20 may include both internal and external storage devices. The memory 20 is used to store application software and various types of data installed on the terminal, such as program code installed on the terminal. The memory 20 can also be used to temporarily store data that has been output or will be output. In one embodiment, the memory 20 stores a risk assessment program 40 based on vascular ultrasound images, which can be executed by the processor 10 to implement the risk assessment method based on vascular ultrasound images in this application.

[0042] In some embodiments, the processor 10 may be a central processing unit (CPU), a microprocessor, or other data processing chip, used to run program code stored in the memory 20 or process data, such as executing the risk assessment method based on vascular ultrasound images.

[0043] In some embodiments, the display 30 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an OLED (Organic Light-Emitting Diode) touchscreen. The display 30 is used to display information on the terminal and to display a visual user interface. The terminals communicate with each other via a system bus.

[0044] In one embodiment, when the processor 10 executes the risk assessment program 40 based on vascular ultrasound images in the memory 20, the following steps are performed: Acquire ultrasound echo data from a wearable ultrasound patch, and reconstruct the ultrasound echo data using an image reconstruction algorithm to obtain a vascular ultrasound image. The patient's clinical risk factors are obtained, and the vascular ultrasound image is analyzed according to the vein identification and compressibility analysis algorithm and the texture and blood flow feature extraction algorithm to obtain vein image features. The vein image features and the clinical risk factors are input into the trained risk assessment model for evaluation to obtain a venous thrombosis risk score. Obtain the current timestamp and historical data, determine the risk level based on the timestamp, the historical data and the venous thrombosis risk score, and send an alarm message to medical staff.

[0045] Specifically, the step of acquiring ultrasound echo data from a wearable ultrasound patch and reconstructing the ultrasound echo data using an image reconstruction algorithm to obtain a vascular ultrasound image includes: Acquire ultrasound echo data from wearable ultrasound patches; Gain compensation is performed on the ultrasonic echo data using an image reconstruction algorithm to obtain multiple echo amplitude distributions. The amplitude distributions of multiple echoes are reconstructed using an interpolation reconstruction algorithm or a beamforming algorithm to obtain a vascular ultrasound image.

[0046] The step of reconstructing the amplitude distributions of multiple echoes using an interpolation reconstruction algorithm or a beamforming algorithm to obtain a vascular ultrasound image specifically includes: The amplitude distributions of multiple echoes are reconstructed using an interpolation reconstruction algorithm or a beamforming algorithm to obtain a two-dimensional grayscale image; A one-dimensional time depth map is determined based on the time axis and the distribution of multiple echo amplitudes. The two-dimensional grayscale image and the one-dimensional time-depth map are mapped to obtain a vascular ultrasound image.

[0047] Specifically, the process of acquiring the patient's clinical risk factors, analyzing the vascular ultrasound image using a vein identification and compressibility analysis algorithm and a texture and blood flow feature extraction algorithm to obtain vein image features, and inputting the vein image features and the clinical risk factors into a trained risk assessment model for evaluation to obtain a venous thrombosis risk score includes: Obtain the patient's clinical risk factors; The vascular ultrasound image is identified using a vein recognition and compressibility analysis algorithm to obtain the lumen geometric parameters; The texture and blood flow features of the lumen geometry parameters are extracted using a texture and blood flow feature extraction algorithm to obtain venous image features; The venous imaging features and the clinical risk factors are input into a trained risk assessment model for evaluation to obtain a venous thrombosis risk score. The risk assessment model includes any one of the following: logistic regression model, random forest model, gradient boosting tree model, and neural network model; The geometric parameters of the lumen include cross-sectional area, diameter, rate of change of cross-sectional area, and diameter shrinkage ratio.

[0048] Specifically, the step of inputting the venous imaging features and the clinical risk factors into a trained risk assessment model for evaluation to obtain a venous thrombosis risk score includes: Obtain a historical pathology dataset, and divide the historical pathology dataset into a training set and a validation set; The training set is input into the initial risk assessment model for training to obtain the trained risk assessment model. The venous imaging features and the clinical risk factors are input into the trained risk assessment model for evaluation to obtain a venous thrombosis risk score. The validation set is used to fine-tune the structure and parameters of the trained risk assessment model.

[0049] The blood flow characteristics include peak blood flow velocity, average blood flow velocity, blood flow direction, spectral envelope morphology, and time ratio. The step of extracting texture and blood flow features from the lumen geometric parameters using a texture and blood flow feature extraction algorithm to obtain venous image features specifically includes: Based on the gray-level co-occurrence matrix, the texture features of the lumen geometric parameters are calculated according to the texture and blood flow feature extraction algorithm to obtain the first target feature; Feature extraction is performed on the peak blood flow velocity, the average blood flow velocity, the blood flow direction, the spectral envelope morphology, and the time ratio to obtain the second target feature. The vein image features are then determined based on the first target feature and the second target feature.

[0050] The process of obtaining the current timestamp and historical data, determining the risk level based on the timestamp, the historical data, and the venous thrombosis risk score, and sending an alarm message to medical staff specifically includes: Get the current timestamp and historical data; Based on the trend analysis module, the historical data is compared with the timestamp, the venous thrombosis risk score, and the clinical risk factors to obtain the time trend; If the venous thrombosis risk score exceeds a preset threshold or the time trend is upward, then the risk level is mapped according to the venous thrombosis risk score, and an alarm message is generated according to the risk level to remind medical staff. The risk levels include low risk, medium risk, and high risk.

[0051] The present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a risk assessment program based on vascular ultrasound images, and the risk assessment program based on vascular ultrasound images, when executed by a processor, implements the steps of the risk assessment method based on vascular ultrasound images as described above.

[0052] In summary, this invention provides a risk assessment method, system, terminal, and storage medium based on vascular ultrasound images. The method includes: acquiring ultrasound echo data from a wearable ultrasound patch; reconstructing the ultrasound echo data using an image reconstruction algorithm to obtain a vascular ultrasound image; acquiring the patient's clinical risk factors; analyzing the vascular ultrasound image using a vein recognition and compressibility analysis algorithm and a texture and blood flow feature extraction algorithm to obtain vein image features; inputting the vein image features and the clinical risk factors into a trained risk assessment model for evaluation to obtain a venous thrombosis risk score; acquiring the current timestamp and historical data; determining the risk level based on the timestamp, the historical data, and the venous thrombosis risk score; and sending an alarm message to medical personnel. This invention dynamically assesses the risk of venous thrombosis through image reconstruction, vein feature extraction, and a risk assessment model, achieving accurate monitoring and graded early warning of the patient's venous thrombosis risk.

[0053] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal system that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or terminal system. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or terminal system that includes that element.

[0054] Of course, those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware (such as a processor, controller, etc.). The program can be stored in a computer-readable storage medium, and when executed, it can include the processes described in the above method embodiments. The computer-readable storage medium can be a memory, magnetic disk, optical disk, etc.

[0055] It should be understood that the application of the present invention is not limited to the examples above. Those skilled in the art can make improvements or modifications based on the above description, and all such improvements and modifications should fall within the protection scope of the appended claims.

Claims

1. A risk assessment method based on vascular ultrasound images, characterized in that, The risk assessment method based on vascular ultrasound images includes: Acquire ultrasound echo data from a wearable ultrasound patch, and reconstruct the ultrasound echo data using an image reconstruction algorithm to obtain a vascular ultrasound image. The patient's clinical risk factors are obtained, and the vascular ultrasound image is analyzed according to the vein identification and compressibility analysis algorithm and the texture and blood flow feature extraction algorithm to obtain vein image features. The vein image features and the clinical risk factors are input into the trained risk assessment model for evaluation to obtain a venous thrombosis risk score. Obtain the current timestamp and historical data, determine the risk level based on the timestamp, the historical data and the venous thrombosis risk score, and send an alarm message to medical staff.

2. The risk assessment method based on vascular ultrasound images according to claim 1, characterized in that, The process of acquiring ultrasound echo data from a wearable ultrasound patch and reconstructing the ultrasound echo data using an image reconstruction algorithm to obtain a vascular ultrasound image specifically includes: Acquire ultrasound echo data from wearable ultrasound patches; Gain compensation is performed on the ultrasonic echo data using an image reconstruction algorithm to obtain multiple echo amplitude distributions. The amplitude distributions of multiple echoes are reconstructed using an interpolation reconstruction algorithm or a beamforming algorithm to obtain a vascular ultrasound image.

3. The risk assessment method based on vascular ultrasound images according to claim 2, characterized in that, The process of reconstructing the amplitude distributions of multiple echoes using an interpolation reconstruction algorithm or a beamforming algorithm to obtain a vascular ultrasound image specifically includes: The amplitude distributions of multiple echoes are reconstructed using an interpolation reconstruction algorithm or a beamforming algorithm to obtain a two-dimensional grayscale image; A one-dimensional time depth map is determined based on the time axis and the distribution of multiple echo amplitudes. The two-dimensional grayscale image and the one-dimensional time-depth map are mapped to obtain a vascular ultrasound image.

4. The risk assessment method based on vascular ultrasound images according to claim 1, characterized in that, The process involves acquiring the patient's clinical risk factors, analyzing the vascular ultrasound image using a vein identification and compressibility analysis algorithm and a texture and blood flow feature extraction algorithm to obtain vein image features, and inputting these vein image features and the clinical risk factors into a trained risk assessment model for evaluation to obtain a venous thrombosis risk score. Specifically, this includes: Obtain the patient's clinical risk factors; The vascular ultrasound image is identified using a vein recognition and compressibility analysis algorithm to obtain the lumen geometric parameters; The texture and blood flow features of the lumen geometry parameters are extracted using a texture and blood flow feature extraction algorithm to obtain venous image features; The venous imaging features and the clinical risk factors are input into a trained risk assessment model for evaluation to obtain a venous thrombosis risk score. The risk assessment model includes any one of the following: logistic regression model, random forest model, gradient boosting tree model, and neural network model; The geometric parameters of the lumen include cross-sectional area, diameter, rate of change of cross-sectional area, and diameter shrinkage ratio.

5. The risk assessment method based on vascular ultrasound images according to claim 4, characterized in that, The step of inputting the venous imaging features and the clinical risk factors into a trained risk assessment model for evaluation to obtain a venous thrombosis risk score specifically includes: Obtain a historical pathology dataset, and divide the historical pathology dataset into a training set and a validation set; The training set is input into the initial risk assessment model for training to obtain the trained risk assessment model. The venous imaging features and the clinical risk factors are input into the trained risk assessment model for evaluation to obtain a venous thrombosis risk score. The validation set is used to fine-tune the structure and parameters of the trained risk assessment model.

6. The risk assessment method based on vascular ultrasound images according to claim 4, characterized in that, The blood flow characteristics include peak blood flow velocity, average blood flow velocity, blood flow direction, spectral envelope morphology, and time proportion; The step of extracting texture and blood flow features from the lumen geometric parameters using a texture and blood flow feature extraction algorithm to obtain venous image features specifically includes: Based on the gray-level co-occurrence matrix, the texture features of the lumen geometric parameters are calculated according to the texture and blood flow feature extraction algorithm to obtain the first target feature; Feature extraction is performed on the peak blood flow velocity, the average blood flow velocity, the blood flow direction, the spectral envelope morphology, and the time ratio to obtain the second target feature. The vein image features are then determined based on the first target feature and the second target feature.

7. The risk assessment method based on vascular ultrasound images according to claim 1, characterized in that, The process of obtaining the current timestamp and historical data, determining the risk level based on the timestamp, the historical data, and the venous thrombosis risk score, and sending alarm information to medical staff specifically includes: Get the current timestamp and historical data; Based on the trend analysis module, the historical data is compared with the timestamp, the venous thrombosis risk score, and the clinical risk factors to obtain the time trend; If the venous thrombosis risk score exceeds a preset threshold or the time trend is upward, then the risk level is mapped according to the venous thrombosis risk score, and an alarm message is generated according to the risk level to remind medical staff. The risk levels include low risk, medium risk, and high risk.

8. A risk assessment system based on vascular ultrasound images, characterized in that, The risk assessment system based on vascular ultrasound images includes: An ultrasound image processing module is used to acquire ultrasound echo data from a wearable ultrasound patch, and to reconstruct the ultrasound echo data according to an image reconstruction algorithm to obtain a vascular ultrasound image. The thrombosis risk assessment module is used to acquire the patient's clinical risk factors, analyze the vascular ultrasound image according to the vein recognition and compressibility analysis algorithm and the texture and blood flow feature extraction algorithm to obtain vein image features, and input the vein image features and the clinical risk factors into the trained risk assessment model for evaluation to obtain a venous thrombosis risk score. The risk warning module is used to obtain the current timestamp and historical data, determine the risk level based on the timestamp, the historical data and the venous thrombosis risk score, and send alarm information to medical staff.

9. A terminal, characterized in that, The terminal includes: a memory, a processor, and a risk assessment program based on vascular ultrasound images stored in the memory and executable on the processor, wherein when the risk assessment program based on vascular ultrasound images is executed by the processor, it implements the steps of the risk assessment method based on vascular ultrasound images as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a risk assessment program based on vascular ultrasound images, which, when executed by a processor, implements the steps of the risk assessment method based on vascular ultrasound images as described in any one of claims 1-7.

Images