A method for establishing a prostate PET-MRI imaging lesion grading system

Abstract

The present invention provides a method for establishing a prostate PET-MRI image lesion grading system, comprising: acquiring a prostate PET-MRI image, projecting the prostate PET-MRI image into a preset area, obtaining original tissue information of the prostate PET-MRI image, establishing a three-dimensional tumor pedigree map based on the original tissue information, obtaining pathological features corresponding to the prostate PET-MRI image, marking lesion information in the three-dimensional tumor pedigree map, obtaining lesion features corresponding to the prostate PET-MRI image, determining the pathological grade corresponding to the prostate PET-MRI image based on the pathological features, and inputting the lesion features corresponding to the prostate PET-MRI image into a corresponding pathological grade storage domain in a preset grading system for storage. Doctors can compare the patient's PET-MRI image with known pathological grades and stages, thereby more accurately determining the malignancy and spread of the tumor, providing a basis for individualized treatment and prognosis assessment.

Description

Technical Field

[0001] The present invention relates to the technical field of establishing an equipment lesion system, and in particular to a method for establishing a prostate PET-MRI imaging lesion grading system. Background Art

[0002] Prostate PET-MRI plays an important role in monitoring prostate cancer recurrence and metastasis. PET-MRI provides more sensitive and accurate images, helping to determine the location and extent of lesions. This is crucial for developing treatment strategies, assessing treatment effectiveness, and monitoring disease progression. Prostate PET-MRI imaging plays a crucial role in the diagnosis and treatment of prostate cancer and related diseases. Combining the strengths of MRI and PET, it provides high-resolution information on anatomical structure and metabolic activity, facilitating early diagnosis, grading and staging, recurrence monitoring, and treatment guidance. Prostate PET-MRI allows physicians to more accurately assess the disease, develop personalized treatment plans, and improve patient survival and quality of life. The importance of prostate PET-MRI in assessing the grade and stage of prostate cancer is widely recognized. This combined technique provides more comprehensive and accurate information, crucial for determining the malignancy and extent of tumor spread, thus playing a vital role in treatment planning and patient prognosis. Prostate cancer grading assesses the degree of malignancy based on tumor cell atypia and proliferative activity.

[0003] Currently, doctors use PET-MRI images based on the characteristics of the gold standard of pathological morphology to preliminarily determine the grade of prostate tumors. However, this method places extremely high demands on doctors, who need years of experience to accurately determine the grade of the lesion. It is obviously highly random and requires a lot of analysis time in actual operation.

[0004] Therefore, the present invention provides a method for establishing a prostate PET-MRI imaging lesion grading system. Summary of the Invention

[0005] The present invention provides a method for establishing a prostate PET-MRI image lesion grading system. This system, through training and validation on a large number of cases, builds a model to identify and quantify PET-MRI image features associated with pathological morphology. This system allows doctors to compare a patient's PET-MRI images with known pathological grades and stages to more accurately determine the malignancy and extent of tumor spread, providing a basis for personalized treatment and prognostic assessment.

[0006] The present invention provides a method for establishing a prostate PET-MRI imaging lesion grading system, comprising:

[0007] Step 1: Acquire a prostate PET-MRI image, project the prostate PET-MRI image into a preset area, and obtain original tissue information of the prostate PET-MRI image;

[0008] Step 2: establishing a three-dimensional tumor lineage map based on the original tissue information to obtain pathological features corresponding to the prostate PET-MRI image;

[0009] Step 3: Marking lesion information in the three-dimensional tumor lineage map to obtain lesion features corresponding to the prostate PET-MRI image;

[0010] Step 4: Determine the pathological grade corresponding to the prostate PET-MRI image based on the pathological features, and input the lesion features corresponding to the prostate PET-MRI image into a corresponding pathological grade storage domain in a preset grading system for storage.

[0011] In one practicable manner,

[0012] The step 1 comprises:

[0013] Step 11: Acquire a prostate PET-MRI image and image information corresponding to the prostate PET-MRI image;

[0014] Step 12: parsing the image information to obtain a first generation time of the corresponding prostate PET-MRI image, obtaining a second generation time corresponding to each existing prostate PET-MRI image in the preset grading system, and sorting the second generation times from near to far, extracting the third generation time that ranks first in the sorting;

[0015] Step 13: determining whether the prostate PET-MRI image is acquired for the first time based on the first generation time and the third generation time;

[0016] Step 14: When the prostate PET-MRI image is acquired for the first time, the prostate PET-MRI image is projected into a preset area to obtain original tissue information of the prostate PET-MRI image.

[0017] In one practicable manner,

[0018] The step 2 comprises:

[0019] Step 21: drawing a first angle original tissue in a first preset coordinate system according to the original tissue information, and drawing a second angle original tissue in a second preset coordinate system, and establishing a three-dimensional tumor lineage map based on the first angle original tissue and the second angle original tissue;

[0020] Step 22: Mark the tissue name corresponding to each original tissue in the three-dimensional tumor lineage map, obtain the three-dimensional contour corresponding to each original tissue, and establish a tissue contour corresponding list based on the tissue name;

[0021] Step 23: Obtain the sample tissue contour corresponding to each original tissue, mark the sample tissue contour in the tissue contour corresponding list, and establish a comparison list;

[0022] Step 24: Analyze the comparison list to obtain the pathological features corresponding to each original tissue, obtain all the pathological features, and establish the pathological features corresponding to the prostate PET-MRI image.

[0023] In one practicable manner,

[0024] The step 3 comprises:

[0025] Step 31: Obtain the abnormal tissue position corresponding to each original tissue according to the pathological characteristics, obtain the abnormal contour corresponding to each abnormal tissue position, and establish the lesion contour sub-feature according to the abnormal contour;

[0026] Step 32: Mark each abnormal contour in the three-dimensional tumor pedigree map, obtain a three-dimensional abnormal image corresponding to each abnormal contour on the three-dimensional tumor pedigree map, and establish a lesion area sub-feature based on the image area corresponding to the three-dimensional abnormal image;

[0027] Step 33: Marking the contour distances between different abnormal contours on the three-dimensional tumor lineage map, and establishing lesion distribution sub-features based on the contour distances;

[0028] Step 34: Obtain the lesion feature corresponding to the prostate PET-MRI image according to the lesion contour sub-feature, the lesion area sub-feature, and the lesion distribution sub-feature.

[0029] In one practicable manner,

[0030] The step 4 comprises:

[0031] Step 41: Compare the pathological features with each preset pathological grade sample, obtain the similarity between the pathological features and each preset pathological grade sample, and establish a similarity statistical list;

[0032] Step 42: extracting the preset pathological grade sample having the highest similarity to the pathological feature according to the similarity statistical list, and determining the pathological grade corresponding to the prostate PET-MRI image;

[0033] Step 43: searching for a corresponding pathology grade storage domain in the preset grading system, and inputting the lesion characteristics into the pathology grade storage domain for storage.

[0034] In one practicable manner,

[0035] The step 24 includes:

[0036] Step 241: Input the lesion features corresponding to each original tissue into the bottom layer, middle layer, and high layer of the preset DeepLab v2 (ResNet-34) model for analysis to obtain corresponding bottom layer analysis results, middle layer analysis results, and high layer analysis results;

[0037] Step 242: Fusing the bottom-level analysis results, the middle-level analysis results, and the high-level analysis results in pairs to obtain corresponding first, second, and third fusion results, and obtaining fusion features corresponding to different original tissues in each fusion result;

[0038] Step 243: Acquire the first fusion feature, the second fusion feature, and the third fusion feature corresponding to the same original tissue to establish lesion information corresponding to each original tissue, and establish a lesion feature corresponding to each original tissue based on the lesion information and the tissue properties of the corresponding original tissue;

[0039] Step 244: Mark each lesion feature on the prostate PET-MRI image, obtain the feature interference between different lesion features, and combine the lesion features to obtain the pathological features corresponding to the prostate PET-MRI image.

[0040] In one practicable manner,

[0041] The step 31 includes:

[0042] Step 311: Mapping the pathological features to a preset single-plane lesion area to obtain a mapping result;

[0043] Step 312: Obtain a plurality of mapping profiles included in the mapping result, traverse the mapping profiles using each tissue sample, obtain the tissue name corresponding to each mapping profile, and search for the health parameter corresponding to each original tissue based on the tissue name;

[0044] Step 313: Draw a healthy contour corresponding to each original tissue according to the health parameter, compare the healthy contour with the corresponding mapping contour, and obtain the tissue abnormality position corresponding to each original tissue;

[0045] Step 314: Mark the abnormal contour corresponding to each original tissue at the abnormal tissue position, obtain contour differences and contour similarities between each abnormal contour and the corresponding healthy contour, and establish lesion contour sub-features.

[0046] In one practicable manner,

[0047] Also includes:

[0048] After inputting the lesion features into the pathology grade storage domain for storage, obtaining the lesion quantity corresponding to each existing feature in the pathology storage domain;

[0049] The existing features are reordered according to the order of low to high morbidity.

[0050] In one practicable manner,

[0051] Also includes:

[0052] After inputting the lesion features corresponding to the prostate PET-MRI image into the corresponding pathology grade storage domain in the preset grading system for storage, statistics are collected on the existing lesion features in each pathology grade storage domain in the grading system to establish a feature statistics matrix;

[0053] Obtain the original tissue information corresponding to each existing lesion feature, perform a health score on each original tissue, and generate a corresponding score report;

[0054] According to the element position of each current lesion feature in the feature statistics matrix, each scoring report is input into the corresponding storage position in the preset grading system for storage.

[0055] In one practicable manner,

[0056] Also includes:

[0057] According to the feature statistical matrix, each prostate PET-MRI image corresponding to the existing lesion feature is input into the storage location for storage.

[0058] The beneficial effects that can be achieved by the present invention are: in order to establish a complete prostate PET-MRI image lesion grading system and to match the current disease development during actual use, the acquired moral prostate PET-MRI image is projected to obtain its original tissue information, and then a three-dimensional tumor spectrum map can be established based on the original tissue information, thereby obtaining the lesion characteristics of the prostate PET-MRI image, so that the pathological grade corresponding to the prostate PET-MRI image can be determined, and the lesion characteristic data of the prostate PET-MRI image is stored in the corresponding pathological grade storage domain, thereby updating the prostate PET-MRI image lesion grading system. In this way, a grading system including different lesion characteristics can be established, and doctors can input the prostate PET-MRI images of existing patients into the system to determine the lesion grade, reducing the time for doctors to make subjective judgments and improving the efficiency of diagnosis, which is of great significance for determining the choice of treatment plan, the feasibility of surgery and predicting the patient's prognosis.

[0059] Other features and advantages of the present invention will be described in the following description, and in part will become apparent from the description, or will be understood by practicing the present invention. The purposes and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description, claims, and drawings.

[0060] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0062] Figure 1 Schematic diagram of the workflow of a method for establishing a prostate PET-MRI imaging lesion grading system according to an embodiment of the present invention;

[0063] Figure 2 This is a schematic diagram of the workflow of step 2 of a method for establishing a prostate PET-MRI image lesion grading system in an embodiment of the present invention;

[0064] Figure 3 2 is a schematic diagram of the workflow of step 3 of a method for establishing a prostate PET-MRI image lesion grading system in an embodiment of the present invention. DETAILED DESCRIPTION

[0065] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0066] Example 1

[0067] This embodiment provides a method for establishing a prostate PET-MRI imaging lesion grading system. Figure 1 Shown, including:

[0068] Step 1: Acquire a prostate PET-MRI image, project the prostate PET-MRI image into a preset area, and obtain original tissue information of the prostate PET-MRI image;

[0069] Step 2: establishing a three-dimensional tumor lineage map based on the original tissue information to obtain pathological features corresponding to the prostate PET-MRI image;

[0070] Step 3: Marking lesion information in the three-dimensional tumor lineage map to obtain lesion features corresponding to the prostate PET-MRI image;

[0071] Step 4: Determine the pathological grade corresponding to the prostate PET-MRI image based on the pathological features, and input the lesion features corresponding to the prostate PET-MRI image into a corresponding pathological grade storage domain in a preset grading system for storage.

[0072] In this example, PET-MRI imaging refers to a fusion of positron emission tomography (PET) and magnetic resonance imaging (MRI);

[0073] In this example, the preset area represents a binary image area;

[0074] In this example, the original tissue information represents the constituent tissues of the prostate;

[0075] In this example, the three-dimensional tumor spectrum diagram represents a development spectrum of prostate tumors expressed in three dimensions;

[0076] In this example, the pathological characteristics represent characteristics of the disease suffered by the prostate;

[0077] In this example, the lesion information indicates the extent and area of ​​the lesion of each original tissue in the prostate;

[0078] In this example, the lesion characteristics represent the characteristics of the lesion that occurs after the original tissue becomes diseased;

[0079] In this example, the pathological grade indicates the grade of disease in the prostate;

[0080] In this example, the preset grading system represents a system for grading lesions, which includes a number of pathology fan storage domains;

[0081] In this example, prostate PET-MRI assessment has important application value in the grading and staging of prostate cancer. By combining information from PET and MRI, doctors can more accurately assess the patient's condition and provide more precise grading and staging results.

[0082] In this example, establishing a prostate PET-MRI image lesion grading system is a complex and critical process. It requires building a multi-channel PET-MRI-based prostate tumor grading system to automate tumor grading of PET-MRI prostate images. First, a large image database containing PET-MRI images of prostate cancer patients with known pathological grades and staging results must be established.

[0083] The working principle and beneficial effects of the above technical solution are as follows: In order to establish a complete prostate PET-MRI image lesion grading system and to adapt it to the current disease development during actual use, the acquired moral prostate PET-MRI image is projected to obtain its original tissue information, and then a three-dimensional tumor lineage map can be established based on the original tissue information, thereby obtaining the lesion characteristics of the prostate PET-MRI image, so that the pathological grade corresponding to the prostate PET-MRI image can be determined, and the lesion characteristic data of the prostate PET-MRI image is stored in the corresponding pathological grade storage domain, thereby updating the prostate PET-MRI image lesion grading system. In this way, a grading system including different lesion characteristics can be established. Doctors can input the prostate PET-MRI images of existing patients into the system to determine the lesion grade, reducing the time for doctors to make subjective judgments and improving the efficiency of diagnosis, which is of great significance for determining the choice of treatment plan, the feasibility of surgery and predicting the patient's prognosis.

[0084] Example 2

[0085] Based on Example 1, the method for establishing a prostate PET-MRI imaging lesion grading system, step 1, comprises:

[0086] Step 11: Acquire a prostate PET-MRI image and image information corresponding to the prostate PET-MRI image;

[0087] Step 12: parsing the image information to obtain a first generation time of the corresponding prostate PET-MRI image, obtaining a second generation time corresponding to each existing prostate PET-MRI image in the preset grading system, and sorting the second generation times from near to far, extracting the third generation time that ranks first in the sorting;

[0088] Step 13: determining whether the prostate PET-MRI image is acquired for the first time based on the first generation time and the third generation time;

[0089] Step 14: When the prostate PET-MRI image is acquired for the first time, the prostate PET-MRI image is projected into a preset area to obtain original tissue information of the prostate PET-MRI image.

[0090] In this example, the image information represents information generated when generating the prostate PET-MRI image;

[0091] In this example, the first generation time represents the generation time of the prostate PET-MRI image;

[0092] In this example, the second generation time represents the generation time of each existing prostate PET-MRI image in the preset grading system;

[0093] In this example, the third generation time represents the generation time corresponding to the latest existing prostate PET-MRI image in the preset grading system;

[0094] In this example, "first", "second", and "third" are only used to distinguish three different generation moments and do not have the function of sorting or comparison;

[0095] In this example, when the prostate PET-MRI image is acquired for the first time, it means that the lesion features corresponding to the prostate PET-MRI image have not been stored, so they need to be updated to the preset grading system.

[0096] The working principle and beneficial effects of the above technical solution are as follows: In order to update the preset grading system, a prostate PET-MRI image and its corresponding image information are obtained, and it is determined whether the prostate PET-MRI image has been stored based on the generation time of the prostate PET-MRI image, and then it is determined whether it needs to be updated. This not only improves the update efficiency, but also avoids repeated updates.

[0097] Example 3

[0098] Based on Example 1, the method for establishing a prostate PET-MRI imaging lesion grading system is as follows: Figure 2 As shown, the step 2 includes:

[0099] Step 21: drawing a first angle original tissue in a first preset coordinate system according to the original tissue information, and drawing a second angle original tissue in a second preset coordinate system, and establishing a three-dimensional tumor lineage map based on the first angle original tissue and the second angle original tissue;

[0100] Step 22: Mark the tissue name corresponding to each original tissue in the three-dimensional tumor lineage map, obtain the three-dimensional contour corresponding to each original tissue, and establish a tissue contour corresponding list based on the tissue name;

[0101] Step 23: Obtain the sample tissue contour corresponding to each original tissue, mark the sample tissue contour in the tissue contour corresponding list, and establish a comparison list;

[0102] Step 24: Analyze the comparison list to obtain the pathological features corresponding to each original tissue, obtain all the pathological features, and establish the pathological features corresponding to the prostate PET-MRI image.

[0103] In this example, the first preset coordinate system represents the coordinate system in the XY direction, and the second preset coordinate system represents the coordinate system in the YZ direction;

[0104] In this example, the first angle original tissue represents the original tissue presented in the first preset coordinate system, and the second angle original tissue represents the original tissue presented in the second preset coordinate system;

[0105] In this example, the tissue name corresponds one-to-one with the original tissue and the three-dimensional contour;

[0106] In this example, the tissue profile corresponding list includes the 3D profile corresponding to each original tissue. In order to distinguish different 3D profiles, the tissue name is matched with the 3D profile.

[0107] In this example, the sample tissue contour represents the contour corresponding to the preset, healthy original tissue;

[0108] In this example, the comparison list contains comparison information between a tissue profile and a corresponding sample tissue profile;

[0109] In this example, the lesion feature is established by the abnormal points between the tissue contour and the corresponding sample tissue contour.

[0110] The working principle and beneficial effects of the above technical solution are as follows: first, original tissues at different angles are drawn in different coordinate systems to establish a three-dimensional tumor spectrum map, and then the tissue name corresponding to each original tissue is marked in the three-dimensional tumor spectrum map and the tissue contour corresponding to each original tissue is obtained, thereby establishing a list corresponding to the tissue contour, and then the tissue contour is compared with the corresponding sample tissue contour to establish a comparison list. Through the comparison list, the lesion characteristics corresponding to each original tissue can be intuitively obtained, thereby establishing the pathological characteristics corresponding to the prostate PET-MRI image. In this way, the original tissue can be analyzed from multiple angles, and the tissue information of different original tissues in the prostate PET-MRI image can be judged by contour analysis and comparison, so that accurate pathological characteristics can be obtained.

[0111] Example 4

[0112] Based on Example 1, the method for establishing a prostate PET-MRI imaging lesion grading system is as follows: Figure 3 As shown, the step 3 includes:

[0113] Step 31: Obtain the abnormal tissue position corresponding to each original tissue according to the pathological characteristics, obtain the abnormal contour corresponding to each abnormal tissue position, and establish the lesion contour sub-feature according to the abnormal contour;

[0114] Step 32: Mark each abnormal contour in the three-dimensional tumor pedigree map, obtain a three-dimensional abnormal image corresponding to each abnormal contour on the three-dimensional tumor pedigree map, and establish a lesion area sub-feature based on the image area corresponding to the three-dimensional abnormal image;

[0115] Step 33: Marking the contour distances between different abnormal contours on the three-dimensional tumor lineage map, and establishing lesion distribution sub-features based on the contour distances;

[0116] Step 34: Obtain the lesion feature corresponding to the prostate PET-MRI image according to the lesion contour sub-feature, the lesion area sub-feature, and the lesion distribution sub-feature.

[0117] In this instance, the abnormal tissue location indicates the location where the abnormality occurs in the original tissue;

[0118] In this example, the abnormal contour represents the contour that the original tissue presents when it becomes abnormal;

[0119] In this example, the lesion contour sub-feature represents the contour feature presented when the original tissue is lesioned;

[0120] In this example, the lesion area sub-feature represents the feature of the lesion area of ​​the original tissue;

[0121] In this example, the lesion distribution sub-feature represents a distribution feature of the locations where lesions occur on an original tissue.

[0122] The working principle of the above technical solution is as follows: In order to further analyze the lesion characteristics of prostate PET-MRI images, the lesion ethics sub-features, lesion area sub-features and lesion distribution sub-features of each original tissue are determined based on the pathological characteristics. Based on these three characteristics, the lesion characteristics are further refined from three different angles, thereby obtaining accurate lesion characteristics.

[0123] Example 5

[0124] Based on Example 1, the method for establishing a prostate PET-MRI imaging lesion grading system, step 4, includes:

[0125] Step 41: Compare the pathological features with each preset pathological grade sample, obtain the similarity between the pathological features and each preset pathological grade sample, and establish a similarity statistical list;

[0126] Step 42: extracting the preset pathological grade sample having the highest similarity to the pathological feature according to the similarity statistical list, and determining the pathological grade corresponding to the prostate PET-MRI image;

[0127] Step 43: searching for a corresponding pathology grade storage domain in the preset grading system, and inputting the lesion characteristics into the pathology grade storage domain for storage.

[0128] In this example, the similarity statistics list includes the similarities between a pathological feature and samples of different preset pathological grades.

[0129] The working principle and beneficial effects of the above technical solution are as follows: In order to store the lesion characteristics in the storage location of the corresponding level, the pathological characteristics are first compared with the preset pathological level samples, and the pathological level corresponding to the prostate PET-MRI image is determined based on the comparison results. Finally, the lesion characteristics are stored in the corresponding pathological level storage domain. In this way, the lesion characteristics of different pathological levels can be classified and stored, which is convenient for doctors to quickly find the corresponding pathological level when using the system.

[0130] Example 6

[0131] Based on Example 3, the method for establishing a prostate PET-MRI imaging lesion grading system, step 24 includes:

[0132] Step 241: Input the lesion features corresponding to each original tissue into the bottom layer, middle layer, and high layer of the preset DeepLab v2 (ResNet-34) model for analysis to obtain corresponding bottom layer analysis results, middle layer analysis results, and high layer analysis results;

[0133] Step 242: Fusing the bottom-level analysis results, the middle-level analysis results, and the high-level analysis results in pairs to obtain corresponding first, second, and third fusion results, and obtaining fusion features corresponding to different original tissues in each fusion result;

[0134] Step 243: Acquire the first fusion feature, the second fusion feature, and the third fusion feature corresponding to the same original tissue to establish lesion information corresponding to each original tissue, and establish a lesion feature corresponding to each original tissue based on the lesion information and the tissue properties of the corresponding original tissue;

[0135] Step 244: Mark each lesion feature on the prostate PET-MRI image, obtain the feature interference between different lesion features, and combine the lesion features to obtain the pathological features corresponding to the prostate PET-MRI image.

[0136] In this example, the DeepLab v2 (ResNet-34) model is a deep learning model based on dilated convolution and CRF of residual neural networks.

[0137] In this example, the DeepLab v2 (ResNet-34) model contains three training layers: bottom, middle, and top layers.

[0138] In this example, the bottom-level analysis results, middle-level analysis results, and high-level analysis results are the analysis results of the bottom-level, middle-level, and high-level outputs of the DeepLab v2 (ResNet-34) model, respectively.

[0139] In this example, fusing the bottom-level analysis results, the middle-level analysis results, and the high-level analysis results in pairs includes: fusing the bottom-level analysis results with the middle-level analysis results to generate a first fusion result, fusing the bottom-level analysis results with the high-level analysis results to generate a second fusion result, and fusing the high-level analysis results with the middle-level analysis results to generate a third fusion result;

[0140] In this example, the fused features represent the features presented after fusing the analysis results of two different layers;

[0141] In this instance, organizational attributes represent characteristics unique to the original organization;

[0142] In this example, the feature interference between different lesion features indicates that a lesion in an original tissue may have a certain diffusion or interference effect on its associated original tissue, so the feature interference is established.

[0143] The working principle and beneficial effects of the above technical solution: In order to further determine the pathological characteristics corresponding to the prostate PET-MRI images, the lesion characteristics corresponding to each original tissue are input into the preset model for training at different levels. Through the skip-layer fusion method, the model can simultaneously learn the bottom-level, middle-level and high-level features, thereby obtaining the corresponding layer analysis results. The analysis results of different layers are then fused to determine the characteristics of different tissues in different cross-sections, so that the feature interference between different lesion features can be known, thereby establishing a comprehensive pathological feature. In this process, the model is used to analyze the features of different levels, and then after fusion, the features of each tissue at different levels can be obtained, thereby establishing a comprehensive pathological feature.

[0144] Example 7

[0145] Based on Example 4, the method for establishing a prostate PET-MRI imaging lesion grading system, step 31 includes:

[0146] Step 311: Mapping the pathological features to a preset single-plane lesion area to obtain a mapping result;

[0147] Step 312: Obtain a plurality of mapping profiles included in the mapping result, traverse the mapping profiles using each tissue sample, obtain the tissue name corresponding to each mapping profile, and search for the health parameter corresponding to each original tissue based on the tissue name;

[0148] Step 313: Draw a healthy contour corresponding to each original tissue according to the health parameter, compare the healthy contour with the corresponding mapping contour, and obtain the tissue abnormality position corresponding to each original tissue;

[0149] Step 314: Mark the abnormal contour corresponding to each original tissue at the abnormal tissue position, obtain contour differences and contour similarities between each abnormal contour and the corresponding healthy contour, and establish lesion contour sub-features.

[0150] In this example, the layer lesion area refers to an area that can display the lesion outline of pathological features at different levels;

[0151] In this example, there is a one-to-one correspondence between tissue samples, mapping profiles, and tissue names;

[0152] In this example, the health parameters represent the parameters that the original tissue exhibits when it is in a healthy state;

[0153] In this example, the healthy contour represents the contour of the original tissue in a healthy state;

[0154] In this example, the contour difference points represent the differences between the abnormal contour and the healthy contour, and the contour similarity points represent the similarities between the abnormal contour and the healthy contour.

[0155] The working principle and beneficial effects of the above technical solution are as follows: the pathological features are mapped to the single-plane lesion area through mapping, and then the tissue sample is used to traverse the mapping contours in the mapping results to determine the tissue name of each mapping contour, and the health parameters corresponding to each original tissue are found. A healthy contour is established based on the health parameters, and the healthy contour domain is compared with the mapping contour, so that the tissue abnormality position corresponding to each original tissue can be obtained, and the lesion contour sub-feature is established based on the differences and similarities between the mapping contour and the healthy contour.

[0156] Example 8

[0157] Based on Example 5, the method for establishing a prostate PET-MRI imaging lesion grading system further includes:

[0158] After inputting the lesion features into the pathology grade storage domain for storage, obtaining the lesion quantity corresponding to each existing feature in the pathology storage domain;

[0159] The existing features are reordered according to the order of low to high morbidity.

[0160] The working principle of the above technical solution is: by sorting the lesion features in the same pathology storage domain in order of lesion variables from low to high, the lesion features corresponding to different lesion variables can be observed at a glance.

[0161] Example 9

[0162] Based on Example 1, the method for establishing a prostate PET-MRI imaging lesion grading system further includes:

[0163] After inputting the lesion features corresponding to the prostate PET-MRI image into the corresponding pathology grade storage domain in the preset grading system for storage, statistics are collected on the existing lesion features in each pathology grade storage domain in the grading system to establish a feature statistics matrix;

[0164] Obtain the original tissue information corresponding to each existing lesion feature, perform a health score on each original tissue, and generate a corresponding score report;

[0165] According to the element position of each current lesion feature in the feature statistics matrix, each scoring report is input into the corresponding storage position in the preset grading system for storage.

[0166] In this example, the feature statistics matrix represents the pathological features stored in the storage domains of different pathological grades in the grading system in a matrix manner;

[0167] In this example, the existing lesion features refer to lesion features that have been stored in the pathology fan storage domain.

[0168] The working principle and beneficial effects of the above technical solution: In order to facilitate doctors to quickly determine the patient's condition when using the system, and to improve the system, each existing lesion feature in the system is given a health score, and a score report is generated for each original tissue under different lesion features. Doctors can make a preliminary judgment on the patient's condition based on the score report.

[0169] Example 10

[0170] Based on Example 9, the method for establishing a prostate PET-MRI imaging lesion grading system further includes:

[0171] According to the feature statistical matrix, each prostate PET-MRI image corresponding to the existing lesion feature is input into the storage location for storage.

[0172] The working principle and beneficial effects of the above technical solution are as follows: each prostate PET-MRI image is input into the storage location for storage and can be viewed by the doctor at any time. When necessary, the prostate PET-MRI image is compared with the patient's actual condition to further determine the patient's condition.

[0173] Obviously, those skilled in the art may make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if such changes and modifications fall within the scope of the claims and their equivalents, the present invention is intended to include such changes and modifications.

Claims

1. A method for establishing a prostate PET-MRI image lesion grading system, characterized in that: include: Step 1: Acquire a prostate PET-MRI image, project the prostate PET-MRI image into a preset area, and obtain original tissue information of the prostate PET-MRI image; Step 2: establishing a three-dimensional tumor lineage map based on the original tissue information to obtain pathological features corresponding to the prostate PET-MRI image; Step 3: Marking lesion information in the three-dimensional tumor lineage map to obtain lesion features corresponding to the prostate PET-MRI image; Step 4: Determine the pathological grade corresponding to the prostate PET-MRI image based on the pathological features, and input the lesion features corresponding to the prostate PET-MRI image into a corresponding pathological grade storage field in a preset grading system for storage; The step 2 comprises: Step 21: drawing a first angle original tissue in a first preset coordinate system according to the original tissue information, and drawing a second angle original tissue in a second preset coordinate system, and establishing a three-dimensional tumor lineage map based on the first angle original tissue and the second angle original tissue; Step 22: Mark the tissue name corresponding to each original tissue in the three-dimensional tumor lineage map, obtain the three-dimensional contour corresponding to each original tissue, and establish a tissue contour corresponding list based on the tissue name; Step 23: Obtain the sample tissue contour corresponding to each original tissue, mark the sample tissue contour in the tissue contour corresponding list, and establish a comparison list; Step 24: parsing the comparison list to obtain the pathological features corresponding to each original tissue, obtaining all the pathological features to establish the pathological features corresponding to the prostate PET-MRI image; The step 3 comprises: Step 31: Obtain the abnormal tissue position corresponding to each original tissue according to the pathological characteristics, obtain the abnormal contour corresponding to each abnormal tissue position, and establish the lesion contour sub-feature according to the abnormal contour; Step 32: Mark each abnormal contour in the three-dimensional tumor pedigree map, obtain a three-dimensional abnormal image corresponding to each abnormal contour on the three-dimensional tumor pedigree map, and establish a lesion area sub-feature based on the image area corresponding to the three-dimensional abnormal image; Step 33: Marking the contour distances between different abnormal contours on the three-dimensional tumor lineage map, and establishing lesion distribution sub-features based on the contour distances; Step 34: Obtain the lesion feature corresponding to the prostate PET-MRI image according to the lesion contour sub-feature, the lesion area sub-feature, and the lesion distribution sub-feature.

2. The method for establishing a prostate PET-MRI image lesion grading system according to claim 1, wherein: The step 1 comprises: Step 11: Acquire a prostate PET-MRI image and image information corresponding to the prostate PET-MRI image; Step 12: parsing the image information to obtain a first generation time of the corresponding prostate PET-MRI image, obtaining a second generation time corresponding to each existing prostate PET-MRI image in the preset grading system, and sorting the second generation times from near to far, extracting the third generation time that ranks first in the sorting; Step 13: determining whether the prostate PET-MRI image is acquired for the first time based on the first generation time and the third generation time; Step 14: When the prostate PET-MRI image is acquired for the first time, the prostate PET-MRI image is projected into a preset area to obtain original tissue information of the prostate PET-MRI image.

3. The method for establishing a prostate PET-MRI image lesion grading system according to claim 1, wherein: The step 4 comprises: Step 41: Compare the pathological features with each preset pathological grade sample, obtain the similarity between the pathological features and each preset pathological grade sample, and establish a similarity statistical list; Step 42: extracting the preset pathological grade sample having the highest similarity to the pathological feature according to the similarity statistical list, and determining the pathological grade corresponding to the prostate PET-MRI image; Step 43: searching for a corresponding pathology grade storage domain in the preset grading system, and inputting the lesion characteristics into the pathology grade storage domain for storage.

4. The method for establishing a prostate PET-MRI image lesion grading system according to claim 1, wherein: The step 24 includes: Step 241: Inputting the lesion features corresponding to each original tissue into the bottom layer, middle layer, and high layer of the preset DeepLab v2 model for analysis, and obtaining corresponding bottom layer analysis results, middle layer analysis results, and high layer analysis results; Step 242: Fusing the bottom-level analysis results, the middle-level analysis results, and the high-level analysis results in pairs to obtain corresponding first, second, and third fusion results, and obtaining fusion features corresponding to different original tissues in each fusion result; Step 243: Acquire the first fusion feature, the second fusion feature, and the third fusion feature corresponding to the same original tissue to establish lesion information corresponding to each original tissue, and establish a lesion feature corresponding to each original tissue based on the lesion information and the tissue properties of the corresponding original tissue; Step 244: Mark each lesion feature on the prostate PET-MRI image, obtain the feature interference between different lesion features, and combine the lesion features to obtain the pathological features corresponding to the prostate PET-MRI image.

5. The method for establishing a prostate PET-MRI image lesion grading system according to claim 1, wherein: The step 31 includes: Step 311: Mapping the pathological features to a preset single-plane lesion area to obtain a mapping result; Step 312: Obtain a plurality of mapping profiles included in the mapping result, traverse the mapping profiles using each tissue sample, obtain the tissue name corresponding to each mapping profile, and search for the health parameter corresponding to each original tissue based on the tissue name; Step 313: Draw a healthy contour corresponding to each original tissue according to the health parameter, compare the healthy contour with the corresponding mapping contour, and obtain the tissue abnormality position corresponding to each original tissue; Step 314: Mark the abnormal contour corresponding to each original tissue at the abnormal tissue position, obtain contour differences and contour similarities between each abnormal contour and the corresponding healthy contour, and establish lesion contour sub-features.

6. The method for establishing a prostate PET-MRI image lesion grading system according to claim 3, wherein: Also includes: After inputting the lesion feature into the pathology grade storage domain for storage, obtaining the lesion quantity corresponding to each existing feature in the pathology grade storage domain; The existing features are reordered according to the order of low to high morbidity.

7. The method for establishing a prostate PET-MRI image lesion grading system according to claim 1, wherein: Also includes: After inputting the lesion features corresponding to the prostate PET-MRI image into the corresponding pathology grade storage domain in the preset grading system for storage, statistics are collected on the existing lesion features in each pathology grade storage domain in the grading system to establish a feature statistics matrix; Obtain the original tissue information corresponding to each existing lesion feature, perform a health score on each original tissue, and generate a corresponding score report; According to the element position of each existing lesion feature in the feature statistics matrix, each scoring report is input into the corresponding storage position in the preset grading system for storage.

8. The method for establishing a prostate PET-MRI image lesion grading system according to claim 7, wherein: Also includes: According to the feature statistical matrix, each prostate PET-MRI image corresponding to the existing lesion feature is input into the storage location for storage.

Images