Portal vein hypertension screening diagnosis and treatment auxiliary system and method based on artificial intelligence and ultrasound

Abstract

The invention provides a portal hypertension screening diagnosis and treatment auxiliary system and method based on artificial intelligence and ultrasound. The system comprises an ultrasonic imaging system, an intelligent image recognition system, a secondary ultrasonic imaging radiography system and an intelligent reporting and analyzing system. Artificial intelligence and ultrasound are combined, ultrasound imaging is a non-invasive detection mode, the system has extremely high safety, is based on non-ionizing radiation, is lower in risk compared with X-rays and other imaging systems using ionizing radiation, is combined with a diagnosis and treatment auxiliary system provided by artificial intelligence, is high in automation degree, can automatically grab and recognize related information and obtain parameters, and is high in diagnosis and treatment efficiency. The operation process is simple and friendly; the method is based on an ultrasonic system, has low requirements on equipment, does not have high-value consumable requirements, and is low in inspection cost. Compared with an existing method for representing portal vein pressure through hepatic vein pressure gradient measurement, the method has the advantages of being noninvasive, easy to operate and low in price, and is suitable for wide clinical application and popularization.

Description

[Technical field]

[0001] The present invention relates to the field of diagnosis and treatment assistance systems, and in particular to an artificial intelligence and ultrasound-based portal hypertension screening diagnosis and treatment assistance system and method. [Background technology]

[0002] Digestive tract diseases are common diseases, among which cirrhosis is a common digestive tract disease in my country and is the terminal stage of many liver diseases. Portal hypertension is one of the most important complications in the development of cirrhosis, which will further deteriorate liver function and even lead to the death of patients. Cirrhosis increases the blood pressure in the liver, and it is difficult for portal vein blood to flow back to the inferior vena cava through the liver, which can easily cause the portal vein pressure to increase, causing the communication branch between the portal and systemic veins to open, resulting in a large amount of portal vein blood, directly entering the systemic circulation without passing through the liver, further leading to decreased liver function. The esophageal gastric venous communication branch is the most important communication branch between the portal and systemic veins. Its rupture can lead to acute gastrointestinal bleeding, which is an important cause of death in patients with cirrhosis.

[0003] Portal vein pressure is closely related to the prognosis of patients with cirrhosis. Portal hypertension indicates the tendency of liver function deterioration and the risk of gastrointestinal bleeding. Therefore, early identification of patients with portal hypertension in cirrhosis must be strengthened clinically, and early diagnosis and treatment must be achieved to improve the prognosis of liver patients. Accurately measuring the portal vein pressure of patients with cirrhosis is the key to assessing the patient's condition and taking timely intervention measures.

[0004] Direct measurement of portal vein pressure is traumatic, risky, and easily affected by intra-abdominal pressure. Currently, most clinical tests use blood tests (inflammatory indicators, liver fibrosis markers) and imaging tests (spleen volume, ascites, portal vein width) to indirectly show portal vein pressure. However, they are all qualitative indicators of portal vein pressure and cannot quantify portal vein pressure levels, resulting in low screening accuracy, late detection, and inability to accurately quantify the degree of liver cirrhosis, which brings risks to subsequent treatment.

[0005] Non-invasive assessment of portal pressure has become an important part of medical research, and HVPG has gradually become the gold standard for pressure detection. The method for detecting portal pressure is to measure the hepatic venous pressure gradient (HVPG). Hepatic venous pressure gradient detection requires puncture and catheterization of the internal jugular vein or femoral vein. After catheterization, the hepatic venous wedge pressure (WHVP) and free hepatic venous pressure (FHVP) are measured respectively. The hepatic venous wedge pressure represents the hepatic sinusoidal pressure and can indirectly reflect the portal pressure when portal hypertension is caused by sinusoidal reasons. The free hepatic venous pressure can represent the pressure of the intra-abdominal vena cava. The difference between the two is the hepatic venous pressure gradient, which reflects the portal pressure and excludes the influence of intra-abdominal pressure. Under normal circumstances, the range of HVPG is 3-5 mmHg. If HVPG is greater than 5 mmHg, it indicates the presence of portal hypertension.

[0006] Although HVPG has the advantage of accurate measurement, it still has the following shortcomings:

[0007] 1. Hepatic venous pressure gradient detection is an invasive test with high technical difficulty and high cost, and cannot be promoted in ordinary primary hospitals.

[0008] 2. Hepatic venous pressure gradient testing is an invasive examination with high risks. The test needs to be performed through deep vein catheterization. Deep vein catheterization may cause complications such as bleeding, hematoma, thrombosis, and catheter-related infection. In addition, if the internal jugular vein catheter is used, the operation may also damage lung tissue and cause pneumothorax, subcutaneous emphysema, or hemothorax.

[0009] 3. The hepatic venous pressure gradient detection operation is technically difficult and requires special pressure sensors and other instruments. The operation steps are complicated and many domestic hospitals are currently unable to carry out this operation.

[0010] 4. At the same time, pressure testing requires the use of consumables such as compliant balloon catheters, and the testing cost is relatively high.

[0011] To solve the above problems, Chinese patent CN107945878B proposes a method for constructing a hepatic venous pressure gradient calculation model based on radiomics. This method calculates the hepatic venous pressure gradient value based on the extraction results of the patient's first liver portal layer image features and the spleen portal layer image features and the hepatic venous pressure gradient calculation model based on radiomics, which solves the problems of high invasive risk and high operational difficulty. The concept of rHVPG was proposed for the first time, and important data was obtained in its methodology construction and model verification, which is conducive to exploring and establishing a new technology for safe, non-invasive and accurate quantitative portal pressure calculation. Chinese patent CN109528196B proposes a non-invasive assessment method for hepatic venous pressure gradient based on multimodal imaging and empirical knowledge, which uses a convolutional neural network to extract features from multimodal medical images and obtain HVPG estimates based on multimodal images, uses a deep neural network to perform regression analysis on HVPG-related empirical knowledge parameters and obtains HVPG estimates based on empirical knowledge, fuses the HVPG estimates based on multimodal imaging and empirical knowledge obtained in the above steps to obtain fused HVPG estimates, and uses an optimization algorithm to jointly train the above-mentioned convolutional neural network and deep neural network. After the training is completed, they can be used to predict HVPG to obtain a quantitative estimate of HVPG based on multimodal imaging and empirical knowledge.

[0012] Although the above two methods have conducted certain innovative research based on HVPG, these methods have certain radiation to patients, and the algorithm model only estimates the portal pressure. There is no prognostic risk grading and no initial screening procedure, which leads to excessive use of medical resources and high prices. [Summary of the invention]

[0013] The purpose of the present invention is to solve the problems in the prior art and to provide a portal hypertension screening and diagnosis and treatment auxiliary system and method based on artificial intelligence and ultrasound, which can reduce radiation, perform prognostic risk grading, alleviate the phenomenon of excessive use of medical resources, and has a low price.

[0014] To achieve the above objectives, the present invention proposes an artificial intelligence and ultrasound-based portal hypertension screening and diagnosis and treatment auxiliary system, comprising:

[0015] Ultrasound imaging system: Use ultrasound equipment to image the portal vein, splenic vein, and spleen in two dimensions and Doppler;

[0016] Intelligent image recognition system: used to identify and evaluate the sections displayed by ultrasound during ultrasound scanning, and to obtain the data required for portal vein pressure measurement after identifying the ultrasound sections and images that can be used for data acquisition;

[0017] Secondary ultrasound imaging and angiography system: For subjects suspected of portal hypertension in the initial screening, further measurement of ultrasound elasticity imaging and ultrasound angiography related data is performed, and the image is analyzed and measured in combination with the intelligent image recognition system;

[0018] Intelligent reporting and analysis system: Combines the patient's clinical information with the analytical measurement data of the intelligent image recognition system to assess the patient's portal vein pressure and grade the prognostic risk.

[0019] Preferably, the intelligent image recognition system comprises:

[0020] Image preprocessing module: preprocess the ultrasound images to be identified to reduce the interference factors in the images to the algorithm;

[0021] Image structure recognition module: performs structural recognition on the preprocessed ultrasound image, and measures and collects data on the identified target structure;

[0022] Result delivery module: Integrates the data collected in the image structure recognition module and delivers it to the intelligent reporting and analysis system.

[0023] Preferably, the intelligent reporting and analysis system comprises:

[0024] Portal vein pressure assessment algorithm module: analyzes the collected ultrasound data and assesses the patient's portal vein pressure;

[0025] Prognosis analysis module for patients with portal hypertension: Based on the input patient clinical information and portal pressure assessment results, make corresponding prognosis assessments and issue reports.

[0026] To achieve the above purpose, the present invention also proposes an artificial intelligence and ultrasound-based portal hypertension screening and diagnosis and treatment auxiliary method, which comprises the following steps in sequence:

[0027] a) Ultrasound imaging: Ultrasound equipment is used to obtain two-dimensional and Doppler images of the portal vein, splenic vein, and spleen;

[0028] b) Intelligent image recognition: During the ultrasound scan, the sections displayed by the ultrasound are identified and evaluated. After the ultrasound sections and images that can be used for data acquisition are identified, the data required for portal vein pressure measurement are acquired;

[0029] c) Determine whether portal hypertension is suspected, if yes, proceed to step d); if no, proceed to step f);

[0030] d) Secondary ultrasound imaging: For subjects suspected of portal hypertension in the initial screening, further measurement of ultrasound elastography and ultrasound contrast-enhanced data is performed, and the images are analyzed and measured in combination with an intelligent image recognition system;

[0031] e) Input patient clinical information: Input patient clinical information into the intelligent reporting and analysis system;

[0032] f) Intelligent reporting and analysis: Combine the patient's clinical information with the analytical measurement data of the intelligent image recognition system to assess the patient's portal vein pressure;

[0033] g) Issue a diagnosis report and prognosis analysis: Issue a diagnosis report based on the patient's portal vein pressure assessment data, and conduct a graded analysis of the prognosis risk.

[0034] Preferably, the a) ultrasound imaging comprises: turning on the intelligent image recognition system, firstly performing a preliminary screening of the subject's portal vein pressure, the sections displayed by the ultrasound comprising: right subcostal section, right intercostal oblique section, left intercostal oblique section, and subxiphoid section; after identifying the appropriate section according to the prompt of the image recognition system, freezing the image, and in grayscale mode, the system automatically measures the portal vein inner diameter and spleen thickness; in Doppler mode, the system identifies and records the portal vein blood flow velocity and blood flow direction; performing a preliminary screening of patients with portal hypertension according to the system's built-in portal vein pressure discrimination criteria; for patients judged to have no manifestations of portal vein hypertension, a relevant report is issued and the examination is terminated; for patients judged to have suspected portal vein hypertension, further examination of portal vein hypertension is prompted.

[0035] Preferably, the d) secondary ultrasound imaging includes further examination of patients suspected of portal hypertension, including liver and spleen elasticity examination and ultrasound contrast examination, and the sections involved in the further examination include right subcostal section, right intercostal oblique section, left intercostal oblique section, and subxiphoid section; after selecting the corresponding section, the intelligent image recognition system continuously records the section image and collects the required information in different modes, elasticity mode: collects liver and spleen hardness values; ultrasound contrast mode: collects the hepatic vein arrival time and hepatic artery arrival time of the contrast agent.

[0036] Preferably, said e) inputting the patient's clinical information includes the patient's Child liver function grade, and forming a structured report by analyzing the patient's clinical information and data obtained by the portal vein identification system, the report content includes the patient's portal vein pressure assessment, and the risk grade of complications.

[0037] Preferably, the intelligent image recognition comprises an image preprocessing module, an image structure recognition module and a result transfer module in sequence; the image preprocessing comprises denoising, enhancement and edge detection preprocessing of the collected ultrasound image; the image structure recognition comprises analyzing the ultrasound image using a deep learning artificial intelligence algorithm, automatically identifying relevant anatomical structures and extracting relevant features.

[0038] Beneficial effects of the present invention: The present invention combines artificial intelligence with ultrasound. Ultrasound imaging is a non-invasive detection method with extremely high safety. It is based on non-ionizing radiation and has lower risks compared to X-rays and other imaging systems that use ionizing radiation. Combined with the diagnosis and treatment assistance system provided by artificial intelligence, it has a high degree of automation and can automatically capture and identify relevant information and obtain parameters. The operation process is simple and friendly. The present invention is based on an ultrasound system, has low equipment requirements, no high-value consumables requirements, and low inspection costs. Compared with the existing method of characterizing portal vein pressure by measuring the hepatic vein pressure gradient, the present invention has the advantages of being non-invasive, easy to operate, and low-priced, and is suitable for widespread clinical application and promotion.

[0039] The features and advantages of the present invention will be described in detail through embodiments in conjunction with the accompanying drawings.

Brief Description of the Drawings

[0040] Figure 1 It is a schematic diagram of the portal hypertension screening and diagnosis and treatment auxiliary system based on artificial intelligence and ultrasound of the present invention;

[0041] Figure 2 It is a schematic diagram of an intelligent image recognition system in the portal hypertension screening and diagnosis and treatment auxiliary system based on artificial intelligence and ultrasound of the present invention;

[0042] Figure 3 It is a schematic diagram of the intelligent reporting and analysis system in the portal hypertension screening and diagnosis and treatment auxiliary system based on artificial intelligence and ultrasound of the present invention. [Specific implementation method]

[0043] See also Figure 1 , an artificial intelligence and ultrasound-based portal hypertension screening and diagnosis and treatment auxiliary system, including:

[0044] Ultrasound imaging system: Use ultrasound equipment to image the portal vein, splenic vein, and spleen in two dimensions and Doppler;

[0045] Intelligent image recognition system: used to identify and evaluate the sections displayed by ultrasound during ultrasound scanning, and to obtain the data required for portal vein pressure measurement after identifying the ultrasound sections and images that can be used for data acquisition;

[0046] Secondary ultrasound imaging and angiography system: For subjects suspected of portal hypertension in the initial screening, further measurement of ultrasound elasticity imaging and ultrasound angiography related data is performed, and the image is analyzed and measured in combination with the intelligent image recognition system;

[0047] Intelligent reporting and analysis system: Combines the patient's clinical information with the analytical measurement data of the intelligent image recognition system to assess the patient's portal vein pressure and grade the prognostic risk.

[0048] See also Figure 2The intelligent image recognition system includes: an image preprocessing module: preprocessing the ultrasound image to be recognized to reduce the interference factors in the image to the algorithm; an image structure recognition module: performing structural recognition on the preprocessed ultrasound image, and measuring and collecting data on the recognized target structure; a result transmission module: integrating the data collected in the image structure recognition module and transmitting it to the intelligent reporting and analysis system.

[0049] See also Figure 3 The intelligent reporting and analysis system includes: portal vein pressure assessment algorithm module: analyzes the collected ultrasound data to assess the patient's portal vein pressure; portal hypertension patient prognosis analysis module: makes corresponding prognosis assessment and issues a report based on the input patient clinical information and portal vein pressure assessment results.

[0050] See also Figure 1 The screening and diagnosis and treatment auxiliary method of portal hypertension based on artificial intelligence and ultrasound includes the following steps in sequence:

[0051] a) Ultrasound imaging: Ultrasound equipment is used to take two-dimensional and Doppler images of the portal vein, splenic vein, and spleen. Ultrasound imaging includes: turning on the intelligent image recognition system, firstly conducting a preliminary screening of the subject's portal vein pressure, and the sections displayed by ultrasound include: right subcostal section, right intercostal oblique section, left intercostal oblique section, and subxiphoid section; after identifying the appropriate section according to the prompt of the image recognition system, freeze the image, and in grayscale mode, the system automatically measures the portal vein inner diameter and spleen thickness; in Doppler mode, the system identifies and records the portal vein blood flow velocity and blood flow direction; preliminary screening of patients with portal hypertension is carried out according to the system's built-in portal vein pressure discrimination criteria; for patients who are judged to have no manifestations of portal hypertension, a relevant report is issued and the examination is completed; for patients who are judged to be suspected of portal hypertension, further examination of portal hypertension is prompted.

[0052] b) Intelligent image recognition: During the ultrasound scan, the sections displayed by the ultrasound are identified and evaluated. After the ultrasound sections and images that can be used for data acquisition are identified, the data required for portal vein pressure measurement are acquired; intelligent image recognition includes an image preprocessing module, an image structure recognition module, and a result transfer module in sequence. The image preprocessing includes denoising, enhancement, and edge detection preprocessing of the collected ultrasound images; image structure recognition includes using a deep learning artificial intelligence algorithm to analyze the ultrasound image, automatically identify relevant anatomical structures, and extract relevant features.

[0053] The intelligent image recognition system is generated based on ResNet50. First, data preprocessing is performed: by converting ultrasound images into a format suitable for ResNet50 input, cropping and standardizing the images, etc., so that the image data can be directly input into ResNet50.

[0054] Training model: Use ResNet50 as the base network and fine-tune or train it according to the specific task. Use ResNet50 as the feature extractor, then add a fully connected layer for classification and fine-tuning, and use ResNet50 as the encoder, then add a decoder for segmentation or measurement.

[0055] Model validation and tuning: Use the validation set to validate and tune the trained model to achieve the best performance on the test set.

[0056] c) Determine whether portal hypertension is suspected, if yes, proceed to step d); if no, proceed to step f);

[0057] d) Secondary ultrasound imaging and angiography: For subjects suspected of having portal hypertension in the initial screening, further measurements of ultrasound elasticity imaging and ultrasound angiography related data are performed, and the images are analyzed and measured in combination with the intelligent image recognition system; secondary ultrasound imaging and angiography include further examinations of patients suspected of portal hypertension, including liver and spleen elasticity examinations and ultrasound angiography examinations. The sections involved in the further examinations include the right subcostal section, right intercostal oblique section, left intercostal oblique section, and subxiphoid section; after selecting the corresponding section, the intelligent image recognition system continuously records the section images and collects the required information in different modes: elasticity mode: collects liver and spleen hardness values; ultrasound angiography mode: collects the hepatic vein arrival time and hepatic artery arrival time of the contrast agent.

[0058] e) Input the patient's clinical information: Input the patient's clinical information into the intelligent reporting and analysis system; input the patient's clinical information including the patient's Child liver function grade, and generate a structured report by analyzing the patient's clinical information and the data obtained by the portal vein identification system. The report content includes the patient's portal vein pressure assessment and the risk level classification of complications.

[0059] f) Intelligent reporting and analysis: Combine the patient's clinical information with the analytical measurement data of the intelligent image recognition system to evaluate the patient's portal vein pressure.

[0060] The intelligent reporting and analysis system is generated based on convolutional neural networks. During the operation, the patient's medical records, examination results and other data are collected, and the data is preprocessed, including data cleaning, feature extraction and data normalization, so that the data can be directly input into the CNN model. Then the CNN model is designed, including convolutional layers, pooling layers, fully connected layers, etc. According to the characteristics of the patient data, the appropriate model structure and hyperparameters are selected so that the model can better learn the characteristics of the incidence of complications of patients. The prepared data set is divided into training set, validation set and test set, and the CNN model is trained with the training set, and the model is tuned on the validation set to avoid problems such as overfitting and underfitting. The trained CNN model is tested with the test set, and the prediction accuracy, recall rate, precision and other indicators of the model are calculated to evaluate the prediction ability of the CNN model.

[0061] Finally, the trained model is applied to the intelligent reporting and analysis system in this patent, which is combined with the patient's clinical information and related ultrasound data to evaluate the incidence of complications of the patient and improve the diagnostic efficiency and accuracy of clinicians.

[0062] g) Issue a diagnosis report and prognosis analysis: Issue a diagnosis report based on the patient's portal vein pressure assessment data, and conduct a graded analysis of the prognosis risk.

[0063] This patented screening and diagnosis assistance system can evaluate the patient's portal vein pressure level by combining artificial intelligence with the recognition of ultrasound images. For subjects who are initially suspected of having portal hypertension, the ultrasound elastography and ultrasound angiography related data can be further measured, and the images can be analyzed and measured in combination with artificial intelligence. At the same time, combined with the patient's clinical information, an artificial intelligence prognostic analysis algorithm is constructed to analyze the patient's risk of complications. This patent analyzes ultrasound images in multiple modes. Compared with existing imaging methods such as CT and MRI for portal vein pressure detection, it is more convenient, radiation-free, and relatively inexpensive. Since this patent sets up a preliminary screening procedure for portal vein pressure, further ultrasound elastography and ultrasound angiography examinations are only performed on patients suspected of having portal hypertension, thereby reducing the excessive use of medical resources. At the same time, the measured portal vein pressure-related data, combined with the patient's clinical information, can be used to further analyze the prognosis of patients with portal hypertension through an artificial intelligence report and analysis system, and assess the risk of complications in patients. This patent is easy to operate, reduces the requirements for the operating level and experience of ultrasound physicians, and does not require expensive consumables. It is non-invasive, convenient, and easy to promote, and can also guide clinical physicians to formulate corresponding diagnosis and treatment plans.

[0064] The artificial intelligence image recognition system of the present invention can automatically identify relevant structures, extract relevant tissue features, reduce the measurement errors caused by human factors during the measurement process, and improve the accuracy and repeatability of the measurement. The portal vein system of the patient is evaluated using multiple ultrasound modes, which is not limited to the patient's portal vein pressure measurement, and also includes the prognosis analysis of the patient, which can provide more accurate guidance for clinical treatment. A more reasonable screening process is established to avoid the waste of medical resources generated during the screening of patients with portal hypertension, and for patients suspected of having portal hypertension, ultrasound elasticity and ultrasound angiography are further evaluated according to the system guidance. Non-invasive ultrasound technology is used. Compared with the traditional direct measurement method of portal vein pressure, there is no risk of infection and bleeding. Compared with CT, MRI and other examinations, it has the advantages of no radiation, low price, and can be repeated in a short period of time. Low cost, easy to promote, compared with traditional methods, the use of ultrasound + artificial intelligence to measure portal vein pressure does not require the use of expensive equipment and consumables, and at the same time, the doctor's operation skills and experience level are required to be low, which can effectively reduce the cost required for measurement, and it is also easy to promote on a large scale in grassroots hospitals.

[0065] Compared with the existing method of characterizing portal vein pressure by measuring the hepatic vein pressure gradient, the present invention has the advantages of being non-invasive, easy to operate and low-cost, and is suitable for wide clinical application and promotion. Ultrasound imaging is a non-invasive detection method with an excellent safety record and is one of the preferred clinical examination methods. Ultrasound detection is based on non-ionizing radiation, and has lower risks compared to X-rays and other imaging systems that use ionizing radiation. It is widely used in various organ systems throughout the body and prenatal examinations of the fetus, and is also very safe for pregnant women. The present invention is based on a diagnosis and treatment auxiliary system provided by artificial intelligence, with a high degree of automation, which can automatically capture and identify relevant information, obtain parameters, and the operation process is simple and friendly. At the same time, ultrasound, as one of the preferred clinical examination methods, has good popularity, which is conducive to the promotion of the present invention. The present invention is based on an ultrasound system, has low equipment requirements, no high-value consumables requirements, and low inspection costs.

[0066] The above embodiments are intended to illustrate the present invention, not to limit the present invention. Any solution that is a simple transformation of the present invention belongs to the protection scope of the present invention.

Claims

1. Artificial intelligence and ultrasound-based portal hypertension screening and diagnosis and treatment auxiliary system, characterized by: include: Ultrasound imaging system: Use ultrasound equipment to image the portal vein, splenic vein, and spleen in two dimensions and Doppler; Intelligent image recognition system: used to identify and evaluate the sections displayed by ultrasound during ultrasound scanning, and to obtain the data required for portal vein pressure measurement after identifying the ultrasound sections and images that can be used for data acquisition; Secondary ultrasound imaging and angiography system: For subjects suspected of portal hypertension in the initial screening, further measurement of ultrasound elasticity imaging and ultrasound angiography related data is performed, and the image is analyzed and measured in combination with the intelligent image recognition system; Intelligent reporting and analysis system: Combines the patient's clinical information with the analytical measurement data of the intelligent image recognition system to assess the patient's portal vein pressure and grade the prognostic risk.

2. The artificial intelligence and ultrasound-based portal hypertension screening and diagnosis and treatment auxiliary system according to claim 1, characterized in that: The intelligent image recognition system comprises: Image preprocessing module: preprocess the ultrasound images to be identified to reduce the interference factors in the images to the algorithm; Image structure recognition module: performs structural recognition on the preprocessed ultrasound image, and measures and collects data on the identified target structure; Result delivery module: Integrates the data collected in the image structure recognition module and delivers it to the intelligent reporting and analysis system.

3. The artificial intelligence and ultrasound-based portal hypertension screening and diagnosis and treatment auxiliary system according to claim 1, characterized in that: The intelligent reporting and analysis system includes: Portal vein pressure assessment algorithm module: analyzes the collected ultrasound data and assesses the patient's portal vein pressure; Prognosis analysis module for patients with portal hypertension: Based on the input patient clinical information and portal pressure assessment results, make corresponding prognosis assessments and issue reports.

4. An artificial intelligence and ultrasound-based portal hypertension screening and treatment auxiliary method, characterized by: The following steps are included in sequence: a) Ultrasound imaging: Ultrasound equipment is used to obtain two-dimensional and Doppler images of the portal vein, splenic vein, and spleen; b) Intelligent image recognition: During the ultrasound scan, the sections displayed by the ultrasound are identified and evaluated. After the ultrasound sections and images that can be used for data acquisition are identified, the data required for portal vein pressure measurement are acquired; c) Determine whether portal hypertension is suspected, if yes, proceed to step d); if no, proceed to step f); d) Secondary ultrasound imaging: For subjects suspected of portal hypertension in the initial screening, further measurement of ultrasound elastography and ultrasound contrast-enhanced data is performed, and the images are analyzed and measured in combination with an intelligent image recognition system; e) Input patient clinical information: Input patient clinical information into the intelligent reporting and analysis system; f) Intelligent reporting and analysis: Combine the patient's clinical information with the analytical measurement data of the intelligent image recognition system to assess the patient's portal vein pressure; g) Issue a diagnosis report and prognosis analysis: Issue a diagnosis report based on the patient's portal vein pressure assessment data, and conduct a graded analysis of the prognosis risk.

5. The artificial intelligence and ultrasound-based portal hypertension screening and treatment auxiliary method according to claim 4, characterized in that: The a) ultrasound imaging includes: turning on the intelligent image recognition system, firstly performing a preliminary screening of the portal vein pressure of the examinee, the sections displayed by the ultrasound include: right subcostal section, right intercostal oblique section, left intercostal oblique section, and subxiphoid section; after identifying the appropriate section according to the prompt of the image recognition system, freezing the image, and in grayscale mode, the system automatically measures the portal vein inner diameter and spleen thickness; in Doppler mode, the system identifies and records the portal vein blood flow velocity and blood flow direction; performing a preliminary screening of patients with portal hypertension according to the system's built-in portal vein pressure discrimination standard; if the patient is judged to have no manifestation of portal hypertension, a relevant report is issued and the examination is completed; if the patient is judged to have suspected portal hypertension, further examination of portal hypertension is prompted.

6. The artificial intelligence and ultrasound-based portal hypertension screening and treatment auxiliary method according to claim 5, characterized in that: The d) secondary ultrasound imaging angiography includes further examination of patients suspected of portal hypertension, including liver and spleen elasticity examination and ultrasound angiography examination. The sections involved in the further examination include right subcostal section, right intercostal oblique section, left intercostal oblique section, and subxiphoid section. After selecting the corresponding section, the intelligent image recognition system continuously records the section image and collects the required information in different modes. Elasticity mode: collects liver and spleen hardness values; ultrasound angiography mode: collects the hepatic vein arrival time and hepatic artery arrival time of the contrast agent.

7. The artificial intelligence and ultrasound-based portal hypertension screening and diagnosis and treatment auxiliary method according to claim 6, characterized in that: e) inputting the patient's clinical information including the patient's Child liver function grade, and forming a structured report by analyzing the patient's clinical information and the data obtained by the portal vein identification system. The report content includes the patient's portal vein pressure assessment and the risk grade of complications.

8. The artificial intelligence and ultrasound-based portal hypertension screening and diagnosis and treatment auxiliary method according to any one of claims 4 to 7, characterized in that: The b) intelligent image recognition includes an image preprocessing module, an image structure recognition module, and a result transmission module in sequence. The image preprocessing includes denoising, enhancement, and edge detection preprocessing of the collected ultrasound image; the image structure recognition includes using a deep learning artificial intelligence algorithm to analyze the ultrasound image, automatically identify relevant anatomical structures, and extract relevant features.

Images