A method, system and device for identifying intestinal wall gas

Abstract

The present invention provides a method, system, device, medium and program product for identifying intestinal wall gas, and a method, system, device, medium and program product for predicting neonatal necrotizing enterocolitis, which relates to the field of intelligent medical care. The present invention solves the problem that intestinal wall gas and intestinal cavity gas cannot be accurately identified when using ultrasound to identify gas accumulation, thereby affecting the diagnosis of neonatal necrotizing enterocolitis; the present invention uses machine learning to more accurately identify the relevant ultrasound manifestations of NEC, timely discover and treat it, reduce long-term complications, and improve survival rate.

Description

Technical Field

[0001] The present invention relates to the field of intelligent medical treatment, and more specifically, to a method, system, device, medium and program product for identifying intestinal wall gas. Background Art

[0002] Necrotizing enterocolitis (NEC) is a serious gastrointestinal disease in newborns, especially premature infants, with high morbidity and mortality. The clinical manifestations of NEC vary greatly, mainly manifested as abdominal distension, vomiting, diarrhea, bloody stools, shock or multiple organ failure, and surviving children may have digestive and nervous system sequelae. Active prevention, early identification and standardized treatment are the key to reducing the occurrence of NEC, increasing the survival rate of children with NEC and improving their prognosis. Therefore, how to identify and diagnose NEC children who need surgical intervention early and strive for the best surgical opportunity for children to improve prognosis is of great significance to improving the survival rate and quality of life of newborns.

[0003] However, early diagnosis of NEC in children is difficult, and ultrasound technology is highly dependent. At present, the main indicator of ultrasound diagnosis of NEC is intestinal wall gas. However, when using ultrasound for clinical identification, intestinal wall gas is easily interfered by the gas in the intestinal cavity, which makes it impossible to accurately identify whether there is intestinal wall gas. In addition, due to the non-standard ultrasound scanning section, intestinal wall gas is sometimes not accurately identified. It is possible to capture non-abnormal intestinal tubes due to misjudgment of intestinal gas, resulting in misjudgment or missed diagnosis. In addition, the method for ultrasound physicians to distinguish the intestines in clinical practice is to first continuously scan and locate the duodenum, and then observe the morphology of the sagittal and transverse sections to distinguish the small intestine or large intestine. The whole process is prone to non-standard operation, resulting in a certain section or a section of the intestine not being scanned, resulting in missed diagnosis. Summary of the invention

[0004] The present invention aims to solve at least one of the technical problems existing in the prior art. To this end, the present invention provides a method for identifying intestinal wall gas.

[0005] In a first aspect, the present application discloses a system for identifying intestinal wall gas, the system comprising:

[0006] A first acquisition module is used to acquire an ultrasonic image of a subject;

[0007] A first processing module is used to segment the ultrasound image to obtain a segmented target intestinal image;

[0008] The second processing module is used to input the target intestinal image into the gas accumulation segmentation model to obtain the result of whether there is gas accumulation in the target intestinal area;

[0009] A third processing module is used for obtaining an intestinal wall region from the target intestinal image using an intestinal wall segmentation model if there is gas accumulation in the target intestinal region;

[0010] The judgment module is used to judge whether there is a first intersection area between the target intestinal area and the intestinal wall area where gas accumulation exists, and if there is a first intersection area, the result that the subject has intestinal wall gas accumulation is obtained.

[0011] The second aspect of the present application discloses a prediction system for neonatal necrotizing enterocolitis, the system comprising:

[0012] The second acquisition module is used to acquire an ultrasonic image of the subject;

[0013] An intestinal gas classification module, used for analyzing whether the ultrasonic image output of the sample to be tested contains a classification result of intestinal gas based on an intestinal gas recognition method;

[0014] The first output module outputs a classification result that the subject is NEC when the subject has intestinal wall gas.

[0015] In some embodiments, the system further comprises:

[0016] A portal vein gas classification module, used for inputting the ultrasound image into the portal vein gas segmentation model to obtain a result of whether the subject has portal vein gas when the subject does not have intestinal wall gas;

[0017] The second output module obtains a result of suspected NEC if the subject has portal vein gas.

[0018] In some embodiments, the system further comprises: an SMA pulsation classification module, inputting the ultrasound image containing portal vein gas into the SMA pulsation classification model to obtain the SMA pulsation index and the SMA resistance index of the ultrasound image; if the SMA pulsation index and the SMA resistance index increase, obtaining a result of abnormal SMA pulsation; if the SMA pulsation index and the SMA resistance index are normal, obtaining a result of normal SMA pulsation;

[0019] In some embodiments, the system further includes: an intestinal wall thickness classification module, which is used to obtain the intestinal wall thickness in the ultrasound image based on the intestinal wall segmentation model, compare the intestinal wall thickness with a preset intestinal wall thickness, and if the intestinal wall thickness is greater than a first preset intestinal wall thickness, obtain a result that the intestinal wall thickness is thickened; if the intestinal wall thickness is less than a second preset intestinal wall thickness, obtain a result that the intestinal wall thickness is thinned; based on the result that the intestinal wall thickness is thickened or thinned, obtain a result that the intestinal wall thickness of the subject is abnormal;

[0020] In some embodiments, the system further includes: an intestinal wall blood flow signal classification module, which is used to calculate the area of ​​the intestinal wall blood flow signal based on the intestinal wall segmentation model, and compare the area of ​​the intestinal wall blood flow signal with a preset area; if the area of ​​the intestinal wall blood flow signal is greater than the preset area, a result of an increase in the area of ​​the intestinal wall blood flow signal is obtained; if the intestinal wall thickness is less than the preset area, a result of a decrease in the area of ​​the intestinal wall blood flow signal is obtained; based on the result of the increase or decrease in the area of ​​the intestinal wall blood flow signal, a result of abnormal intestinal wall blood flow signal of the subject is obtained;

[0021] In some embodiments, the system further comprises: a third output module, which classifies the subject as NEC if the subject has portal vein gas accumulation and SMA pulsation abnormalities, as well as any one of abnormal intestinal wall thickness and abnormal intestinal wall blood flow signal.

[0022] A third aspect of the present application discloses a method for identifying intestinal wall gas, the method comprising:

[0023] 101, obtaining an ultrasound image of a subject;

[0024] 102, performing segmentation processing on the ultrasound image to obtain a segmented target intestinal image;

[0025] 103, inputting the target intestinal image into a gas accumulation segmentation model to obtain a result of whether there is gas accumulation in the target intestinal region;

[0026] 104, if there is gas accumulation in the target intestinal region, obtaining an intestinal wall region from the target intestinal image using an intestinal wall segmentation model;

[0027] 105. Determine whether there is a first intersection area between the target intestinal area and the intestinal wall area where gas accumulation exists. If there is a first intersection area, a result is obtained that the subject has intestinal wall gas accumulation.

[0028] In some embodiments, the segmentation processing method includes: inputting the ultrasound image into an intestinal segmentation model for segmentation;

[0029] The target intestine includes any one or more of the following: duodenum, jejunum, ileum, colon;

[0030] The colon includes any one or more of the following: ascending colon, transverse colon, descending colon, and sigmoid colon.

[0031] In some embodiments, determining in 105 whether there is a first intersection area between the target intestinal area with gas accumulation and the intestinal wall area includes:

[0032] Obtaining a set of coordinate points of the gas accumulation area by indexing the gas accumulation area in the target intestinal image;

[0033] Obtaining a set of coordinate points of the intestinal wall region by indexing the intestinal wall region;

[0034] Calculating the minimum distance between the coordinate point set of the gas accumulation area and the coordinate point set of the intestinal wall area;

[0035] The minimum distance is compared with a preset threshold value, and a result of whether the subject has intestinal wall gas is output.

[0036] In some embodiments, after 102 and before 103, the method further includes: 1021, determining whether the target intestinal image meets the requirements of a standard intestinal image, if so, executing 103; if not, interrupting the execution; the requirements of the standard intestinal image include being able to identify key feature points of the target intestinal tract;

[0037] Optionally, if it does not meet the requirements, the following steps may be performed: 1022, comparing the target intestinal image with the standard intestinal image, obtaining and outputting the intestinal types in the target intestinal image that do not meet the standard intestinal image.

[0038] The fourth aspect of the present application discloses a method for predicting neonatal necrotizing enterocolitis, the method comprising:

[0039] Acquiring an ultrasound image of a subject;

[0040] Based on the method disclosed in the third aspect of the present application, analyzing the classification result of whether the ultrasound image output of the sample to be tested contains intestinal wall gas;

[0041] When the subject has intestinal wall gas, a classification result of the subject being NEC is output.

[0042] In some embodiments, the method further comprises: when the subject does not have intestinal wall gas, inputting the ultrasound image into a portal vein gas segmentation model to obtain a result of whether the subject has portal vein gas; if the subject has portal vein gas, obtaining a result of suspected NEC;

[0043] In some embodiments, the method further includes: inputting an ultrasound image containing portal vein gas accumulation into an SMA pulsation classification model to obtain an SMA pulsation index and an SMA resistance index of the ultrasound image; if the SMA pulsation index and the SMA resistance index increase, obtaining a result of abnormal SMA pulsation; if the SMA pulsation index and the SMA resistance index are normal, obtaining a result of normal SMA pulsation.

[0044] In some embodiments, the method further includes: obtaining the intestinal wall thickness in the ultrasound image based on the intestinal wall segmentation model, comparing the intestinal wall thickness with a preset intestinal wall thickness, and obtaining a result of thickening of the intestinal wall thickness if the intestinal wall thickness is greater than a first preset intestinal wall thickness; obtaining a result of thinning of the intestinal wall thickness if the intestinal wall thickness is less than a second preset intestinal wall thickness; and obtaining a result of abnormal intestinal wall thickness of the subject based on the result of thickening or thinning of the intestinal wall thickness;

[0045] In some embodiments, the method further includes: calculating the area of ​​the intestinal wall blood flow signal based on the intestinal wall segmentation model, and comparing the area of ​​the intestinal wall blood flow signal with a preset area; if the area of ​​the intestinal wall blood flow signal is greater than the preset area, obtaining a result that the area of ​​the intestinal wall blood flow signal increases; if the intestinal wall thickness is less than the preset area, obtaining a result that the area of ​​the intestinal wall blood flow signal decreases; based on the result of the increase or decrease in the area of ​​the intestinal wall blood flow signal, obtaining a result that the intestinal wall blood flow signal of the subject is abnormal;

[0046] If the subject has portal vein gas and SMA pulsation abnormalities, as well as any one of abnormal intestinal wall thickness and abnormal intestinal wall blood flow signal, the subject is classified as NEC.

[0047] In some embodiments, the method further comprises: judging whether the subject has single-focal NEC, multi-focal NEC, or pan-intestinal NEC based on intestinal wall gas, abnormal intestinal wall thickness, and abnormal intestinal wall blood flow signal in the subject;

[0048] The method further includes: providing a recommendation for conservative observation based on the result of the unifocal NEC; providing a recommendation for surgical treatment based on the result of the multifocal NEC;

[0049] In some embodiments, the method further includes: based on the results of the pan-intestinal NEC, such as intestinal wall thinning (intestinal wall thickness ≤ 1 mm), weakened intestinal wall blood flow signal, anechoic free peritoneal effusion, complex localized effusion (peritoneal effusion with more flocculent or punctate deposits or with septa), pneumoperitoneum, and increased intestinal wall echo, giving a recommendation for surgical treatment.

[0050] The fifth aspect of the present application discloses a computer device, which includes: a memory and a processor; the memory is used to store program instructions; the processor is used to call the program instructions, and when the program instructions are executed, it is used to execute the steps of the above method.

[0051] A sixth aspect of the present application discloses a computer-readable storage medium having a computer program stored thereon, which implements the steps of the above method when the computer program is executed by a processor.

[0052] This application has the following beneficial effects:

[0053] 1. The present application innovatively discloses a method for identifying intestinal wall gas. The method first uses an intestinal segmentation model to segment each intestinal tract in an image, and then uses the gas segmentation model to observe whether there is an intersection area between the intestinal area and the gas accumulation area, so as to determine whether there is intestinal wall gas. This method well solves the problem that intestinal wall gas and intestinal cavity gas cannot be accurately identified when using ultrasound to identify gas accumulation. The reasons why intestinal wall gas and intestinal cavity gas cannot be accurately identified include: the main indicator of NEC ultrasound diagnosis is intestinal wall gas, but ultrasound identification is difficult and easily interfered by gas in the intestinal cavity; in addition, the ultrasound scanning section is not standardized, and sometimes intestinal wall gas cannot be accurately identified, and non-target abnormal intestinal tubes may be captured due to misjudgment of gas in the intestinal cavity;

[0054] 2. This application innovatively introduces a step to determine whether an ultrasound image meets the standard image, which solves the problem that ultrasound physicians in the current clinical practice fail to scan a certain section or a certain section of the intestine due to improper operation during intestinal scanning, thereby obtaining an ultrasound image that cannot evaluate the entire intestine, leading to missed diagnosis or even misjudgment;

[0055] 3. The present application discloses a prediction method for neonatal necrotizing enterocolitis. The method, based on the identification of abnormal intestinal tubes, quantifies and classifies intestinal wall gas, portal vein gas, SMA pulsatility index, SMA resistance index, intestinal wall thickness, and intestinal wall blood flow signals: single-focal NEC, multifocal NEC, pan-intestinal NEC, and evaluates the extensiveness of intestinal lesions. This method overcomes the problem that most ultrasound indicators in existing literature are binary indicators, there are few quantitative indicators, and the ultrasound scanning section is not standardized;

[0056] Because ultrasound is highly technically dependent, the significance of this patent is to more accurately identify the relevant ultrasound manifestations of NEC through machine learning, so as to detect and treat it in time, reduce long-term complications and improve survival rate. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use 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 work.

[0058] Figure 1 is a schematic diagram of a method flow chart provided by the third aspect of an embodiment of the present invention;

[0059] Figure 2 is a schematic diagram of a method flow chart provided by the fourth aspect of an embodiment of the present invention;

[0060] Figure 3 is a schematic flow chart of a system provided by the first aspect of an embodiment of the present invention;

[0061] Figure 4 is a schematic flow chart of a system provided by the second aspect of an embodiment of the present invention;

[0062] Figure 5 It is a schematic diagram of a computer device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0063] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0064] In some of the processes described in the specification and claims of the present invention and the above-mentioned figures, multiple operations that appear in a specific order are included, but it should be clearly understood that these operations may not be executed in the order in which they appear in this article or executed in parallel. The serial numbers of the operations, such as 101, 102, etc., are only used to distinguish different operations, and the serial numbers themselves do not represent any execution order. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions of "first", "second", etc. in this article are used to distinguish different messages, devices, modules, etc., do not represent the order of precedence, and do not limit the "first" and "second" to be different types.

[0065] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0066] Figure 1 The following is a flow chart of a method for identifying intestinal wall gas provided by an embodiment of the present invention. Specifically, the method comprises the following steps:

[0067] 101: Acquire an ultrasound image of a subject;

[0068] In some embodiments, the term "subject" or "test subject" or "test sample" used herein refers to any animal (e.g., mammal), including but not limited to humans, non-human primates, rodents, etc., which will be the recipient of a specific treatment. Generally, the terms "subject" and "patient" are used interchangeably herein when referring to human subjects.

[0069] Preferably, the subject is a human.

[0070] In some embodiments, the subject is a patient undergoing clinical prognosis assessment.

[0071] In some embodiments, the method for obtaining the ultrasound image includes: the child fasts for 4 hours, lies supine, and, if necessary, lies on the left side or right side. If the condition permits, gastrointestinal decompression can be performed to reduce intestinal gas interference. A series of longitudinal, transverse and oblique scans are performed starting from the mid-upper abdomen, starting from the duodenal bulb, and observing the jejunum, ileum, and colon in turn, observing the intestinal cavity morphology, intestinal wall echo structure, intestinal peristalsis, intestinal wall blood perfusion, and intestinal wall gas accumulation; the liver is scanned between the right ribs to observe whether there is portal gas accumulation; the intestinal space and lower abdomen are scanned to observe whether there is free effusion or complex localized effusion, as well as signs such as free gas in the abdominal cavity.

[0072] 102: performing segmentation processing on the ultrasound image to obtain a segmented target intestinal image;

[0073] In some embodiments, the segmentation processing method includes: inputting the ultrasound image into an intestinal segmentation model for segmentation; optionally, the intestinal segmentation model construction method includes: obtaining an ultrasound image of a training set sample containing the duodenum, jejunum, ileum, and colon; the ultrasound image has been annotated with key features that can identify the duodenum, jejunum, ileum, and colon; the ultrasound image is input into a machine learning model to obtain a predicted segmentation result, which is compared with the key features, and the model is optimized according to the comparison result to obtain a constructed intestinal segmentation model.

[0074] Optionally, the target intestine includes any one or more of the following: duodenum, jejunum, ileum, colon;

[0075] Optionally, the colon includes any one or more of the following: ascending colon, transverse colon, descending colon, sigmoid colon; specifically, identify each key feature point of the target intestine (including but not limited to the starting point or line, the ending point or line, some key intestinal cavities, and identification points of the intestinal wall) by introducing the expert consensus method to mark the ultrasound images in the training set samples. The method for identifying the target intestine is as follows: Identify the mark of the duodenum: First, identify the pyloric canal structure. The intestine that continues with the pyloric canal is the A duodenum; in order: 1. Duodenal bulb: It is located in the upper right of the pancreatic head and is continuous with the pyloric canal (identified by the pyloric canal as a mark). 2. Descending part of the duodenum: C-shaped around the pancreatic head, located on the right side of the pancreatic head (identified by the pancreatic head as a mark). 3. The horizontal part is located in front of the left side of the abdominal aorta, between the superior mesenteric artery (SMA) and the abdominal aorta (AO) (identified by the SMA and AO as markers) 4. The horizontal part of the duodenum forms the duodenal jejunal fold after it reaches the left upper abdomen. The ascending part of the duodenum is between the horizontal part and the fold, and the distal end of the fold is the beginning of the jejunum. B Small intestine: 1. The jejunum is mostly located in the left abdomen, with thick intestinal walls and large diameters. The jejunum is an arterial arch. 2 The ileum is mostly located in the right abdomen, with thinner intestinal walls than the jejunum and thinner diameters than the jejunum. The ileum is a secondary arterial arch. C Ileocecal part: It continues with the end of the ileum and is petal-shaped. D Colon: 1. Ascending colon: It is located between the liver and kidney, with strong intestinal wall echoes and a colon bag structure. 2. Transverse colon: It is connected to the hepatic flexure of the colon and crosses the middle and upper abdomen to form the transverse colon. 3 Descending colon: The transverse colon runs between the liver and the left kidney and runs to the left lower abdomen to form the descending colon. 4. Sigmoid colon: Behind the bladder is the sigmoid colon.

[0076] In some embodiments, after 102 and before 103, the method further includes: 1021, determining whether the target intestinal image meets the requirements of a standard intestinal image, and if so, executing 103; if not, interrupting the execution; the requirements of the standard intestinal image include being able to identify key feature points of the target intestine; because some intestinal cavities are in the middle part, they must also be distinguished in combination with intestinal wall features; the key feature points of the target intestine include not only starting points or lines, end points or lines, but also some key intestinal cavities and intestinal wall identification points.

[0077] Optionally, if it does not meet the requirements, the following steps can be performed: 1022, compare the target intestinal image with the standard intestinal image, obtain and output the intestinal types in the target intestinal image that do not meet the standard intestinal image. Through comparison, it is found whether the target intestinal image includes all the intestinal tracts in the duodenum, jejunum, ileum, ascending colon, transverse colon, descending colon, and sigmoid colon. If not, the names of the missing intestinal types are output accordingly.

[0078] Specifically, when making a comparison, the target intestinal image can be a single image or a group of images. The group of images are preprocessed into a single image, and then this single image is compared with the standard intestinal image based on expert consensus. Alternatively, a group of images can be obtained, and the machine side defaults to this group of images as a task, and then they are identified separately based on the respective characteristics of the key feature points of the four intestines. If they are identified, the image meets the standard. If all four intestines are not identified, it does not meet the standard (considering that in clinical operations, a picture is usually only a certain section of a certain intestine, and most of them cannot contain information about all intestines at the same time).

[0079] 103: inputting the target intestinal image into a gas accumulation segmentation model to obtain a result of whether there is gas accumulation in the target intestinal region;

[0080] In some embodiments, the method for constructing the gas accumulation segmentation model includes: obtaining a target intestinal image and a corresponding label of a training set sample, wherein the label includes whether it contains gas accumulation; inputting the target intestinal image into a machine learning model to obtain a predicted segmentation result, comparing it with the label, optimizing the model according to the comparison result, and obtaining a constructed gas accumulation segmentation model.

[0081] In some specific embodiments, the presence of intestinal wall gas is manifested as dot-shaped or clustered gas echoes in the intestinal wall;

[0082] 104: if there is gas accumulation in the target intestinal region, obtaining an intestinal wall region from the target intestinal image using an intestinal wall segmentation model;

[0083] In some embodiments, the method for constructing the intestinal wall segmentation model includes: obtaining a target intestinal image of a training set sample and labels for the intestinal wall and / or intestinal cavity in the image; inputting the target intestinal image into a machine learning model to obtain a predicted segmentation result, comparing it with the label, optimizing the model according to the comparison result, and obtaining a constructed intestinal wall segmentation model.

[0084] 105: Determine whether there is a first intersection area between the target intestinal area with gas accumulation and the intestinal wall area, and if there is a first intersection area, obtain a result that the subject has intestinal wall gas accumulation.

[0085] In some embodiments, determining in step 105 whether there is a first intersection area between the target intestinal area with gas accumulation and the intestinal wall area includes:

[0086] Obtaining a set of coordinate points of the gas accumulation area by indexing the gas accumulation area in the target intestinal image;

[0087] Obtaining a set of coordinate points of the intestinal wall region by indexing the intestinal wall region;

[0088] Calculating the minimum distance between the coordinate point set of the gas accumulation area and the coordinate point set of the intestinal wall area;

[0089] Compare the minimum distance with a preset threshold value, and output the result of whether the subject has intestinal wall gas;

[0090] Optionally, the preset threshold is selected based on training of a training set;

[0091] Optionally, the minimum distance between the coordinate point set of the gas accumulation area and the coordinate point set of the intestinal wall area is calculated by using Euclidean distance or standardized Euclidean distance;

[0092] In some embodiments, the method used by the machine learning model includes: R-CNN, fast R-CNN, fasterR-CNN, R-FCN, YOLO, SSD (Single Shot MultiBox Detector);

[0093] Optionally, YOLO (target tracking algorithm) can be BOOSTING Tracker, MIL Tracker, CSRTTracker, MedianFlow Tracker, TLD Tracker, MOSSE Tracker, GOTURN Tracker, etc.

[0094] Optionally, YOLO (target tracking algorithm) can use generative models and discriminative models; the generative model reflects the similarity of the same category, including optical flow method, particle filter, Meanshift algorithm, Camshift algorithm, KCF, SRDCF, SRDCFdecon, stable, EBT and other algorithms; the discriminative model reflects the difference between different categories, and extracts the target model by comparing the difference between the target model and the background information, so as to obtain the target position in the current frame; including MIL, OAB, struck, MEEM, TLD, support vector machine and other algorithms;

[0095] Optionally, YOLO (target tracking algorithm) can adopt deep learning methods. Target tracking algorithms based on deep learning can be divided into: Improvement of correlation filtering algorithm: C-COT, ECO; Pre-training deep model with auxiliary image data and fine-tuning during online tracking: DLT, SO-DLT; Extracting features using CNN classification network pre-trained with existing large-scale classification data sets: FCNT; Pre-training with tracking sequence and fine-tuning during online tracking: MDNet; New idea of ​​using recursive neural network for target tracking: RTT, DeepTracking.

[0096] Figure 2The present invention provides a flow chart of a method for predicting neonatal necrotizing enterocolitis. Specifically, the method comprises the following steps:

[0097] 201: Acquire an ultrasound image of a subject;

[0098] 202: Analyze the classification result of whether the ultrasound image output of the sample to be tested contains intestinal wall gas based on the method described in the first aspect of the present application;

[0099] 203: When the subject has intestinal wall gas, outputting a classification result that the subject is NEC;

[0100] In some embodiments, the method further includes: when the subject does not have intestinal wall gas, that is, the target intestinal area with gas and the intestinal wall area do not have an intersection, the ultrasound image is input into the portal vein gas segmentation model to obtain a result of whether the subject has portal vein gas; specifically, if there are point-like strong echoes in the portal vein, the result of portal vein gas is obtained; if there are no point-like strong echoes in the portal vein, the result of no portal vein gas is obtained; the ultrasound image is an ultrasound image of the portal vein system; if there is portal vein gas, a result of suspected NEC is obtained, because if there is only portal vein gas, but intestinal wall gas is not clearly identified, it cannot be directly diagnosed as NEC, and it needs to be combined with the SMA pulsation. Portal vein gas: point-like or line-like strong echoes are distributed in the lumen of the portal trunk and branches.

[0101] If the subject has portal vein gas accumulation, a result of suspected NEC is obtained; and the ultrasound image containing portal vein gas accumulation is input into the SMA pulsation classification model to obtain the SMA pulsation index and SMA resistance index of the ultrasound image; if the SMA pulsation index and the SMA resistance index increase, a result of abnormal SMA pulsation is obtained; if the SMA pulsation index and the SMA resistance index are normal, a result of normal SMA pulsation is obtained;

[0102] In some embodiments, the method further includes: obtaining the intestinal wall thickness in the ultrasound image based on the intestinal wall segmentation model, comparing the intestinal wall thickness with a preset intestinal wall thickness, and obtaining a result of thickened intestinal wall thickness if the intestinal wall thickness is greater than a first preset intestinal wall thickness; and obtaining a result of thinned intestinal wall thickness if the intestinal wall thickness is less than a second preset intestinal wall thickness; and obtaining a result of abnormal intestinal wall thickness of the subject based on the result of thickened or thinned intestinal wall thickness; wherein an intestinal wall thickness of ≥2.7 mm indicates thickened intestinal wall thickness, and an intestinal wall thickness of ≤1 mm indicates thinned intestinal wall thickness;

[0103] In some embodiments, the method further includes: calculating the area of ​​the intestinal wall blood flow signal based on the intestinal wall segmentation model, and comparing the area of ​​the intestinal wall blood flow signal with a preset area; if the area of ​​the intestinal wall blood flow signal is greater than the preset area, obtaining a result that the area of ​​the intestinal wall blood flow signal increases, and if the intestinal wall thickness is less than the preset area, obtaining a result that the area of ​​the intestinal wall blood flow signal decreases; based on the result of the increase or decrease in the area of ​​the intestinal wall blood flow signal, obtaining a result that the intestinal wall blood flow signal of the subject is abnormal; wherein, the method for evaluating intestinal wall blood supply is: using a linear array probe, the speed range is set to 8.6 cm / s, if the blood flow signal is greater than 9 blood flow signals / cm2, it is defined as rich blood flow signals, that is, the area of ​​the intestinal wall blood flow signal increases; if no blood flow signal is detected, the speed range is reduced to 2.9 cm / s, if the blood flow signal is still not detected, it is defined as the disappearance of the blood flow signal;

[0104] Optionally, if the subject has portal vein gas and SMA pulsation abnormalities, as well as any one of abnormal intestinal wall thickness and abnormal intestinal wall blood flow signal, the subject is classified as NEC.

[0105] In some embodiments, the method further includes: judging whether the subject to be tested has single-focus NEC, multi-focus NEC, or pan-intestinal NEC based on the intestinal wall gas, abnormal intestinal wall thickness, and abnormal intestinal wall blood flow signal present in the subject to be tested; further, if the subject to be tested only has intestinal wall gas and the number is 1, or only has intestinal wall thickness abnormalities and the number is 1, or only has intestinal wall blood flow signal abnormalities, outputting the result that the subject to be tested has single-focus NEC; if only intestinal wall gas exists and the number is greater than or equal to 2, or only has intestinal wall thickness abnormalities and the number is greater than or equal to 2, or if any two or more of intestinal wall gas, abnormal intestinal wall thickness, and abnormal intestinal wall blood flow signal exist, outputting the result that the subject to be tested has single-focus NEC. The result is multi-focal NEC; if the subject has any one or more of the following abnormalities, including intestinal wall gas, abnormal intestinal wall thickness, and abnormal intestinal wall blood flow signal, and the entire intestine is present, the result of the subject being pan-intestinal NEC is output; specifically, the method for determining the number of intestinal wall gas includes: calculating and outputting the number of the first intersection areas; the specific calculation method includes: calculating the minimum distance between the coordinate point sets of two adjacent first intersection areas; comparing the minimum distance with a preset threshold, if the minimum distance is greater than or equal to the preset threshold, outputting the number of the first intersection areas as 2; if the minimum distance is less than the preset threshold, outputting the number of the first intersection areas as 1. The method also includes: calculating and outputting the number of abnormal intestinal wall thickness;

[0106] In some embodiments, the method further includes: based on the result of the unifocal NEC, giving a recommendation for conservative observation; based on the result of the multifocal NEC, giving a recommendation for surgical treatment; based on the result of the pan-intestinal NEC, such as the occurrence of intestinal wall thinning (intestinal wall thickness ≤ 1 mm), weakened intestinal wall blood flow signal, echo-free peritoneal free effusion, complex localized effusion (peritoneal effusion with more flocculent or punctate deposits or with septa), pneumoperitoneum, and increased intestinal wall echo, giving a recommendation for surgical treatment. If it is a unifocal, conservative treatment is used. Pan-focal treatment is surgical treatment. Multifocal treatment requires further identification of intestinal wall thickness and intestinal wall blood flow signals to determine whether to be conservative or surgical;

[0107] In some embodiments, the contents to be observed and measured from the ultrasound images of the subject or the training set samples include:

[0108] (1) Intestinal wall thickness: using a linear array probe, measure in the short axis cross section of the intestine: ① intestinal wall thickening: intestinal wall thickness in the short axis cross section of the intestine ≥ 2.7 mm; ② intestinal wall thinning: intestinal wall thickness ≤ 1 mm;

[0109] (2) Increased echogenicity of the intestinal wall;

[0110] (3) Intestinal wall gas: Use a linear array probe to observe the short axis cross section of the intestine. There are dot-shaped or clustered gas strong echoes in the intestinal wall (it should be noted that the gas in the intestinal cavity can change with intestinal peristalsis, breathing or abdominal pressure, but the gas in the intestinal wall does not change accordingly. This is one of the difficulties in understanding and is also the necessity of introducing artificial intelligence): ① Single-focal NEC: NEC lesions (intestinal wall gas / intestinal wall thickness / blood flow signal changes) are in one place in the intestine (small intestine and colon); ② Multifocal NEC: NEC lesions (intestinal wall gas / intestinal wall thickness / blood flow signal changes) are in multiple places (≥2) in the intestine (small intestine and colon), and normal intestinal tissue can be detected between the lesions; ③ Pan-intestinal NEC: NEC lesions are extensive (involving the entire small intestine or the entire colon)

[0111] (4) Intestinal wall blood flow signal: Using a linear array probe, in the short axis cross section of the intestinal tube, ① set the velocity range to 8.6 cm / s. If the blood flow signal is greater than 9 blood flow signals / cm2 per square centimeter, it is defined as a rich blood flow signal. If no blood flow signal is detected when the velocity range is set to 8.6 cm / s, reduce the velocity range to 2.9 cm / s. If no blood flow signal is still detected in the intestinal wall, it is defined as the disappearance of the blood flow signal.

[0112] (5) Intestinal peristalsis: Dynamic observation in the four quadrants of the abdomen: if the number of intestinal peristalsis is less than 10 times / minute, it is judged as weakened intestinal peristalsis;

[0113] (6) Intestinal morphology;

[0114] (7) Portal gas accumulation: Strong echoes are distributed in the portal vein trunk and branches in a dot-like or line-like manner, which is not related to the prognosis of the disease;

[0115] (8) Anechoic free peritoneal effusion;

[0116] (9) Complex localized effusion: Ascites is accompanied by a large number of flocculent or punctate deposits or with septa.

[0117] (10) Free gas in the abdominal cavity.

[0118] Figure 4 It is a NEC prediction flow chart based on the method of the present invention provided in an embodiment of the present invention. Specifically, intestinal stenosis is one of the complications of NEC. Children with NEC accompanied by intestinal stenosis need to undergo surgery to remove the stenotic segment, and the probability of long-term complications such as short bowel syndrome increases. That is, the quality of life of children with NEC combined with intestinal stenosis is lower than that of children with NEC without intestinal stenosis.

[0119] Figure 3 A system for identifying intestinal wall gas provided by an embodiment of the present invention comprises:

[0120] The first acquisition module 301 is used to acquire an ultrasonic image of a subject;

[0121] A first processing module 302 is used to segment the ultrasound image to obtain a segmented target intestinal image;

[0122] The second processing module 303 is used to input the target intestinal image into the gas accumulation segmentation model to obtain the result of whether there is gas accumulation in the target intestinal area;

[0123] The third processing module 304 is used to obtain the intestinal wall region from the target intestinal image using an intestinal wall segmentation model if there is gas accumulation in the target intestinal region;

[0124] The judgment module 305 is used to judge whether there is a first intersection area between the target intestinal area and the intestinal wall area where gas accumulation exists. If there is a first intersection area, the result that the subject has intestinal wall gas accumulation is obtained.

[0125] In some embodiments, the segmentation processing method in the first processing module includes: inputting the ultrasound image into an intestinal segmentation model for segmentation; the target intestine includes any one or more of the following: duodenum, jejunum, ileum, colon; the colon includes any one or more of the following: ascending colon, transverse colon, descending colon, sigmoid colon.

[0126] In some embodiments, the system further includes: a first comparison module located after the first processing module and before the second processing module, for determining whether the target intestinal image meets the requirements of the standard intestinal image, and if so, executing the second processing module, and if not, interrupting the execution;

[0127] Preferably, it further comprises a second comparison module for comparing the target intestinal image with the standard intestinal image, and obtaining and outputting specific features of the target intestinal image that do not conform to the standard intestinal image.

[0128] In some embodiments, the judgment module judges whether there is a first intersection area between the target intestinal region and the intestinal wall region where gas accumulation exists, including: obtaining a set of coordinate points of the gas accumulation region by indexing the gas accumulation region in the target intestinal image; obtaining a set of coordinate points of the intestinal wall region by indexing the intestinal wall region; calculating a minimum distance between the set of coordinate points of the gas accumulation region and the set of coordinate points of the intestinal wall region; comparing the minimum distance with a preset threshold, and outputting a result of whether the subject has intestinal wall gas accumulation.

[0129] Figure 4 A prediction system for neonatal necrotizing enterocolitis provided by an embodiment of the present invention comprises:

[0130] The second acquisition module 401 is used to acquire an ultrasonic image of the subject;

[0131] An intestinal wall gas classification module 402, used to analyze the classification result of whether the ultrasound image output of the sample to be tested contains intestinal wall gas based on the method described in the first aspect of the present application;

[0132] The first output module 403 outputs a classification result that the subject is NEC when the subject has intestinal wall gas;

[0133] In some embodiments, the system further comprises:

[0134] A portal vein gas classification module, used for inputting the ultrasound image into the portal vein gas segmentation model to obtain a result of whether the subject has portal vein gas when the subject does not have intestinal wall gas;

[0135] The second output module obtains a result of suspected NEC if the subject has portal vein gas accumulation;

[0136] In some embodiments, the system further comprises:

[0137] The SMA pulsation classification module inputs the ultrasound image containing portal vein gas accumulation into the SMA pulsation classification model to obtain the SMA pulsation index and the SMA resistance index of the ultrasound image; if the SMA pulsation index and the SMA resistance index increase, the result of abnormal SMA pulsation is obtained; if the SMA pulsation index and the SMA resistance index are normal, the result of normal SMA pulsation is obtained;

[0138] In some embodiments, the system further includes: an intestinal wall thickness classification module, which is used to obtain the intestinal wall thickness in the ultrasound image based on the intestinal wall segmentation model, compare the intestinal wall thickness with a preset intestinal wall thickness, and if the intestinal wall thickness is greater than a first preset intestinal wall thickness, obtain a result that the intestinal wall thickness is thickened; if the intestinal wall thickness is less than a second preset intestinal wall thickness, obtain a result that the intestinal wall thickness is thinned; based on the result that the intestinal wall thickness is thickened or thinned, obtain a result that the intestinal wall thickness of the subject is abnormal;

[0139] In some embodiments, the system further includes: an intestinal wall blood flow signal classification module, which is used to calculate the area of ​​the intestinal wall blood flow signal based on the intestinal wall segmentation model, and compare the area of ​​the intestinal wall blood flow signal with a preset area; if the area of ​​the intestinal wall blood flow signal is greater than the preset area, a result of an increase in the area of ​​the intestinal wall blood flow signal is obtained; if the intestinal wall thickness is less than the preset area, a result of a decrease in the area of ​​the intestinal wall blood flow signal is obtained; based on the result of the increase or decrease in the area of ​​the intestinal wall blood flow signal, a result of abnormal intestinal wall blood flow signal of the subject is obtained;

[0140] In some embodiments, the system further comprises: a third output module, which classifies the subject as NEC if the subject has portal vein gas accumulation and SMA pulsation abnormalities, as well as any one of abnormal intestinal wall thickness and abnormal intestinal wall blood flow signal.

[0141] Figure 5 A computer device is provided in an embodiment of the present invention, and the device includes: a memory and a processor; the memory is used to store program instructions; the processor is used to call program instructions, and when the program instructions are executed, it is used to execute the steps of the method described in the third aspect or the fourth aspect of the present application.

[0142] The embodiment of the present invention also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps of the method described in the third aspect or the fourth aspect of the present application are implemented.

[0143] An embodiment of the present invention further discloses a computer program product, including a computer program, which, when executed by a processor, implements the steps of the method described in the third aspect or the fourth aspect of the present application.

[0144] The validation results of this validation example show that assigning intrinsic weights to indications can moderately improve the performance of the method relative to the default setting.

[0145] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0146] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0147] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0148] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0149] A person skilled in the art may understand that all or part of the steps in the various methods of the above embodiments may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, and the storage medium may include: a read-only memory (ROM), a random access memory (RAM), a disk or an optical disk, etc.

[0150] A person skilled in the art can understand that all or part of the steps in the above-mentioned embodiment method can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable storage medium. The above-mentioned storage medium can be a read-only memory, a disk or an optical disk, etc.

[0151] The above is a detailed introduction to a computer device provided by the present invention. For a person skilled in the art, according to the concept of the embodiments of the present invention, there may be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as limiting the present invention.

Claims

1. A prediction system for neonatal necrotizing enterocolitis, characterized in that: The system comprises: The second acquisition module is used to acquire an ultrasonic image of the subject; The intestinal wall gas classification module is used to analyze whether the ultrasound image output of the subject contains the classification result of intestinal wall gas; specifically includes: a first processing module, used to segment the ultrasound image to obtain a segmented target intestinal image; a second processing module, used to input the target intestinal image into a gas segmentation model to obtain a result of whether there is gas accumulation in the target intestinal area; a third processing module, used to obtain the intestinal wall area from the target intestinal image using an intestinal wall segmentation model if there is gas accumulation in the target intestinal area; a judgment module, used to judge whether there is a first intersection area between the target intestinal area with gas accumulation and the intestinal wall area, and if there is a first intersection area, obtain a result that the subject has intestinal wall gas; A first output module, when the subject has intestinal wall gas, outputs a classification result that the subject is NEC; The system further comprises: a portal vein gas classification module, which is used to input the ultrasound image into the portal vein gas segmentation model to obtain a result of whether the subject has portal vein gas when the subject does not have intestinal wall gas; The second output module obtains a result of suspected NEC if the subject has portal vein gas.

2. The prediction system for neonatal necrotizing enterocolitis according to claim 1, characterized in that: The system also includes: an SMA pulsation classification module, which inputs the ultrasound image containing portal vein gas accumulation into the SMA pulsation classification model to obtain the SMA pulsation index and SMA resistance index of the ultrasound image; if the SMA pulsation index and the SMA resistance index increase, a result of abnormal SMA pulsation is obtained; if the SMA pulsation index and the SMA resistance index are normal, a result of normal SMA pulsation is obtained.

3. The prediction system for neonatal necrotizing enterocolitis according to claim 2, characterized in that: The system further includes: an intestinal wall thickness classification module, which is used to obtain the intestinal wall thickness in the ultrasound image based on the intestinal wall segmentation model, compare the intestinal wall thickness with a preset intestinal wall thickness, and obtain a result of thickening the intestinal wall thickness if the intestinal wall thickness is greater than a first preset intestinal wall thickness; and obtain a result of thinning the intestinal wall thickness if the intestinal wall thickness is less than a second preset intestinal wall thickness; Based on the result of thickening or thinning of the intestinal wall thickness, a result indicating that the intestinal wall thickness of the subject is abnormal is obtained.

4. The prediction system for neonatal necrotizing enterocolitis according to claim 3, characterized in that: The system also includes: an intestinal wall blood flow signal classification module, which is used to calculate the area of ​​the intestinal wall blood flow signal based on the intestinal wall segmentation model, and compare the area of ​​the intestinal wall blood flow signal with a preset area; if the area of ​​the intestinal wall blood flow signal is larger than the preset area, a result of an increase in the area of ​​the intestinal wall blood flow signal is obtained; if the intestinal wall thickness is smaller than the preset area, a result of a decrease in the area of ​​the intestinal wall blood flow signal is obtained; based on the result of the increase or decrease in the area of ​​the intestinal wall blood flow signal, a result of abnormal intestinal wall blood flow signal of the subject is obtained.

5. The prediction system for neonatal necrotizing enterocolitis according to claim 4, characterized in that: The system further comprises: a third output module, which obtains a classification result that the subject is NEC if the subject has portal vein gas accumulation and SMA pulsation abnormality, as well as any one of abnormal intestinal wall thickness and abnormal intestinal wall blood flow signal.

6. A method for predicting neonatal necrotizing enterocolitis, characterized in that: The method comprises: Acquiring an ultrasound image of a subject; Analyze whether the ultrasound image output of the subject contains the classification result of intestinal wall gas accumulation; specifically include: 102, segmenting the ultrasound image to obtain a segmented target intestinal image; 103, inputting the target intestinal image into a gas accumulation segmentation model to obtain a result of whether there is gas accumulation in the target intestinal area; 104, if there is gas accumulation in the target intestinal area, using the intestinal wall segmentation model to obtain the intestinal wall area from the target intestinal image; 105, judging whether there is a first intersection area between the target intestinal area with gas accumulation and the intestinal wall area, and if there is a first intersection area, obtaining a result that the subject has intestinal wall gas accumulation; When the subject has intestinal wall gas, outputting a classification result that the subject is NEC; The method further includes: when the subject does not have intestinal wall gas, inputting the ultrasound image into a portal vein gas segmentation model to obtain a result of whether the subject has portal vein gas; if the subject has portal vein gas, obtaining a result of suspected NEC.

7. The method for predicting neonatal necrotizing enterocolitis according to claim 6, characterized in that: The method further includes: inputting the ultrasound image containing portal vein gas accumulation into the SMA pulsation classification model to obtain the SMA pulsation index and the SMA resistance index of the ultrasound image; if the SMA pulsation index and the SMA resistance index increase, obtaining a result of abnormal SMA pulsation; if the SMA pulsation index and the SMA resistance index are normal, obtaining a result of normal SMA pulsation.

8. The method for predicting neonatal necrotizing enterocolitis according to claim 7, characterized in that: The method also includes: obtaining the intestinal wall thickness in the ultrasound image based on the intestinal wall segmentation model, comparing the intestinal wall thickness with a preset intestinal wall thickness, and if the intestinal wall thickness is greater than a first preset intestinal wall thickness, obtaining a result that the intestinal wall thickness is thickened; if the intestinal wall thickness is less than a second preset intestinal wall thickness, obtaining a result that the intestinal wall thickness is thinned; and based on the result of the thickening or thinning of the intestinal wall thickness, obtaining a result that the intestinal wall thickness of the subject is abnormal.

9. The method for predicting neonatal necrotizing enterocolitis according to claim 8, characterized in that: The method also includes: calculating the area of ​​the intestinal wall blood flow signal based on the intestinal wall segmentation model, and comparing the area of ​​the intestinal wall blood flow signal with a preset area; if the area of ​​the intestinal wall blood flow signal is larger than the preset area, obtaining a result that the area of ​​the intestinal wall blood flow signal has increased, and if the intestinal wall thickness is smaller than the preset area, obtaining a result that the area of ​​the intestinal wall blood flow signal has decreased; based on the result of the increase or decrease in the area of ​​the intestinal wall blood flow signal, obtaining a result that the intestinal wall blood flow signal of the subject is abnormal.

10. The method for predicting neonatal necrotizing enterocolitis according to claim 9, characterized in that: If the subject has portal vein gas and SMA pulsation abnormalities, as well as any one of abnormal intestinal wall thickness and abnormal intestinal wall blood flow signal, the subject is classified as NEC.

11. The method for predicting neonatal necrotizing enterocolitis according to claim 6, characterized in that: The segmentation processing method includes: inputting the ultrasound image into an intestinal segmentation model for segmentation.

12. The method for predicting neonatal necrotizing enterocolitis according to claim 6, characterized in that: The target intestine includes any one or more of the following: duodenum, jejunum, ileum, and colon.

13. The method for predicting neonatal necrotizing enterocolitis according to claim 12, characterized in that: The colon includes any one or more of the following: ascending colon, transverse colon, descending colon, and sigmoid colon.

14. The method for predicting neonatal necrotizing enterocolitis according to claim 6, characterized in that: The step 105 of determining whether there is a first intersection area between the target intestinal area with gas accumulation and the intestinal wall area includes: Obtaining a set of coordinate points of the gas accumulation area by indexing the gas accumulation area in the target intestinal image; Obtaining a set of coordinate points of the intestinal wall region by indexing the intestinal wall region; Calculating the minimum distance between the coordinate point set of the gas accumulation area and the coordinate point set of the intestinal wall area; The minimum distance is compared with a preset threshold value, and a result of whether the subject has intestinal wall gas is output.

15. The method for predicting neonatal necrotizing enterocolitis according to claim 6, characterized in that: After 102 and before 103, the method further includes: 1021, determining whether the target intestinal image meets the requirements of a standard intestinal image, and if so, executing 103; if not, interrupting the execution; the requirements of the standard intestinal image include being able to identify key feature points of the target intestine.

16. The method for predicting neonatal necrotizing enterocolitis according to claim 15, characterized in that: If it does not match, the following steps can also be performed: 1022, compare the target intestinal image with the standard intestinal image, obtain and output the intestinal types in the target intestinal image that do not match the standard intestinal image.

17. The method for predicting neonatal necrotizing enterocolitis according to claim 9, characterized in that: The method further includes: judging whether the subject has single-focus NEC, multi-focus NEC, or pan-intestinal NEC based on intestinal wall gas, abnormal intestinal wall thickness, and abnormal intestinal wall blood flow signal present in the subject.

18. The method for predicting neonatal necrotizing enterocolitis according to claim 17, characterized in that: The method further includes: providing a recommendation for conservative observation based on the result of the unifocal NEC; and providing a recommendation for surgical treatment based on the result of the multifocal NEC.

19. The method for predicting neonatal necrotizing enterocolitis according to claim 17, characterized in that: The method also includes: based on the results of the pan-intestinal NEC, such as thinning of the intestinal wall, weakened intestinal wall blood flow signal, anechoic free peritoneal effusion, complex localized effusion, pneumoperitoneum, and enhanced intestinal wall echo, giving a recommendation for surgical treatment.

20. A computer device, characterized in that: The device comprises: a memory and a processor; the memory is used to store program instructions; the processor is used to call program instructions, and when the program instructions are executed, it is used to execute the steps of the method described in any one of claims 6 to 19.

21. A computer-readable storage medium, characterized in that: A computer program is stored thereon, and when the computer program is executed by a processor, the steps of the method described in any one of claims 6 to 19 are implemented.

Images