Primary liver cancer auxiliary diagnosis method and system based on iconography data

Abstract

The invention relates to the technical field of MRI image analysis, in particular to a primary liver cancer auxiliary diagnosis method and system based on iconography data, and the method comprises the steps: obtaining MRI images in each period, and obtaining the possibility that adjacent pixel points are located in the same tissue area according to the difference between the adjacent pixel points in the MRI images; obtaining a plurality of pixel point clusters; acquiring a first liver cancer factor and a second liver cancer factor of the pixel point clusters in the arterial phase MRI image according to the pixel point clusters in the different MRI images; and according to the first liver cancer factor and the second liver cancer factor, obtaining the possibility that a region corresponding to the pixel point cluster in the arterial phase MRI image is a liver cancer region, and judging whether the region corresponding to the pixel point cluster in the arterial phase MRI image is the liver cancer region or not. According to the method, the characteristics of the liver cancer area are accurately extracted by analyzing the performance of the liver cancer area in the MRI images in different periods so as to assist in detecting the liver cancer.

Description

Technical Field

[0001] The present invention relates to the technical field of MRI image analysis, and particularly to a method and system for auxiliary diagnosis of primary liver cancer based on imaging data. Background Art

[0002] Primary liver cancer (Hepatocellular Carcinoma, HCC) is one of the malignant tumors with a relatively high fatality rate globally, and early diagnosis is crucial for improving the survival rate of patients. Currently, MRI imaging is an important means for clinical diagnosis of liver cancer, and its multi-phase enhanced scans (such as arterial phase, portal vein phase, equilibrium phase, delayed phase) can dynamically reflect the blood supply characteristics of lesions. However, traditional diagnostic methods mainly rely on the empirical judgment of physicians, with high subjectivity and risks of missed diagnosis and misdiagnosis. Summary of the Invention

[0003] The present invention provides a method and system for auxiliary diagnosis of primary liver cancer based on imaging data to solve the existing problems: traditional diagnosis of primary liver cancer through MRI images relies on the empirical judgment of physicians, and there may be risks of missed diagnosis and misdiagnosis.

[0004] The method and system for auxiliary diagnosis of primary liver cancer based on imaging data of the present invention adopt the following technical solutions: An embodiment of the present invention provides a method for auxiliary diagnosis of primary liver cancer based on imaging data, and the method includes the following steps: Obtain an initial MRI image, an arterial phase MRI image, a portal vein phase MRI image, an equilibrium phase MRI image, and a delayed phase MRI image, and obtain the human body region in each MRI image; According to the pixel points within the local range of each pixel point in the human body region, obtain the blur degree of each pixel point in the human body region; according to the gray values and blur degrees of adjacent pixel points in the human body region, obtain the possibility that the adjacent pixel points in the human body region are located in the same tissue region; according to the possibility that the adjacent pixel points in the human body region are located in the same tissue region, divide the human body region into several pixel point cluster classes; According to the pixel point cluster classes in different MRI images, obtain the first abnormal degree, the second abnormal degree, the third abnormal degree, and the fourth abnormal degree of the overall pixel point cluster class in the arterial phase MRI image; according to the first abnormal degree, the second abnormal degree, the third abnormal degree, and the fourth abnormal degree of the overall pixel point cluster class in the arterial phase MRI image, obtain the first liver cancer factor and the second liver cancer factor of the pixel point cluster class in the arterial phase MRI image; Based on the first liver cancer factor and the second liver cancer factor of the pixel cluster class in the arterial-phase MRI image, obtain the possibility that the corresponding region of the pixel cluster class in the arterial-phase MRI image is a liver cancer region, and determine whether the corresponding region of the pixel cluster class in the arterial-phase MRI image is a liver cancer region.

[0005] Preferably, the method for obtaining the blur degree of each pixel point in the human body region according to the pixel points within the local range of each pixel point in the human body region includes the following specific steps: For any pixel point in any human body region, construct a -sized local window centered on the pixel point in the human body region to obtain the local window of the pixel point in the human body region; for the th pixel point in the local window of the pixel point in the human body region, if the gray values of the pixel points within the eight-neighborhood of the th pixel point are all less than the gray value of each pixel point, then mark the th pixel point as a maximum value point; For the th maximum value point in the local window of the pixel point in the human body region, connect the th maximum value point in the local window of the pixel point in the human body region with each other maximum value point to obtain a number of line segments of the th maximum value point, and use the pixel points on the line segment where the th maximum value point is located as the reference points of the th maximum value point; according to the distances between all the maximum value points in the local window of the pixel point in the human body region and the gray-scale differences of their reference points, obtain the blur degree of the pixel point in the human body region. The specific calculation formula is: ; In the formula, represents the blur degree of the pixel point in the human body region; represents the number of maximum value points in the local window of the pixel point in the human body region; represents the distance between the th and the th maximum value points in the local window of the pixel point in the human body region; represents the number of reference points of the th maximum value point in the local window of the pixel point in the human body region; represents the gray value of the th maximum value in the local window of the pixel point in the human body region; represents the the gray value of the control point corresponding to the maximum value; denotes the absolute value function; denotes the linear normalization function.

[0006] Preferably, the method for obtaining the possibility that adjacent pixel points in the human body region are located in the same tissue region according to the gray values and blurring degrees of adjacent pixel points in the human body region includes the following specific steps: For any two adjacent pixel points in any human body region, according to the gray values and blurring degrees of the two adjacent pixel points in the human body region, obtain the similarity degree of the two adjacent pixel points in the human body region. The specific calculation formula is: ; In the formula, denotes the possibility that the two adjacent pixel points in the human body region are located in the same tissue region; denotes the difference in blurring degree between the two adjacent pixel points in the human body region; denotes the difference in gray value between the two adjacent pixel points in the human body region; denotes the average value of the blurring degrees of the two adjacent pixel points in the human body region; denotes the absolute value function; denotes the exponential function with the natural constant as the base.

[0007] Preferably, the method for obtaining the first abnormal degree of the overall pixel point cluster in the arterial phase MRI image includes the following specific steps: For the human body region in any MRI image, take the average gray value of all pixel points in the human body region in the MRI image as the overall gray value of the MRI image; For the th pixel point in any pixel point cluster in the arterial phase MRI image, take all pixel points in all MRI images with the same coordinates as the th pixel point as the corresponding pixel points of the th pixel point; take the difference in gray value between the pixel point corresponding to the th pixel point in the initial MRI image and the th pixel point as the first abnormal feature of the th pixel point; Obtain the first abnormal features of all pixel points in the pixel point cluster in the arterial phase MRI image, and take the average value of the first abnormal features of all pixel points in the pixel point cluster in the arterial phase MRI image as the first abnormal degree of the overall pixel point cluster in the arterial phase MRI image; Subtract the difference between the overall gray level of the initial MRI image and the arterial-phase MRI image from the first abnormal feature of the overall pixel cluster class in the arterial-phase MRI image, and use the result as the first abnormal degree of the pixel cluster class in the arterial-phase MRI image.

[0008] Preferably, the specific method for obtaining the second abnormal degree is as follows: Take the difference in gray value between the pixel corresponding to the th pixel in the portal-phase MRI image and the pixel corresponding to the th pixel as the second abnormal feature of the th pixel; Obtain the second abnormal features of all pixels in the pixel cluster class in the arterial-phase MRI image, and take the mean of the second abnormal features of all pixels in the pixel cluster class in the arterial-phase MRI image as the second abnormal feature of the overall pixel cluster class in the arterial-phase MRI image; Subtract the difference between the overall gray level of the arterial-phase MRI image and the portal-phase MRI image from the second abnormal feature of the overall pixel cluster class in the arterial-phase MRI image, and use the result as the second abnormal degree of the pixel cluster class in the arterial-phase MRI image.

[0009] Preferably, the specific method for obtaining the third abnormal degree is as follows: Take the difference in gray value between the pixel corresponding to the th pixel in the equilibrium-phase MRI image and the pixel corresponding to the th pixel as the third abnormal feature of the th pixel; Obtain the third abnormal features of all pixels in the pixel cluster class in the arterial-phase MRI image, and take the mean of the third abnormal features of all pixels in the pixel cluster class in the arterial-phase MRI image as the third abnormal feature of the overall pixel cluster class in the arterial-phase MRI image; Subtract the difference between the overall gray level of the arterial-phase MRI image and the equilibrium-phase MRI image from the third abnormal feature of the overall pixel cluster class in the arterial-phase MRI image, and use the result as the third abnormal degree of the pixel cluster class in the arterial-phase MRI image.

[0010] Preferably, the specific method for obtaining the fourth abnormal degree is as follows: Take the difference in gray value between the pixel corresponding to the th pixel in the equilibrium-phase MRI image and the pixel corresponding to the th pixel in the delayed-phase MRI image as the fourth abnormal feature of the th pixel; Obtain the fourth abnormal feature of all the pixel points in the pixel point cluster class in the arterial phase MRI image, and use the mean value of the fourth abnormal feature of all the pixel points in the pixel point cluster class in the arterial phase MRI image as the fourth abnormal feature of the overall pixel point cluster class in the arterial phase MRI image; Subtract the difference between the overall gray levels of the equilibrium phase MRI image and the delayed phase MRI image from the fourth abnormal feature of the overall pixel point cluster class in the arterial phase MRI image, and use the result as the fourth abnormal degree of the pixel point cluster class in the arterial phase MRI image.

[0011] Preferably, the specific method for obtaining the first liver cancer factor and the second liver cancer factor of the pixel point cluster class in the arterial phase MRI image includes: For any pixel point cluster class in the arterial phase MRI image, take the absolute value of the product of the first abnormality, the second abnormal degree, and the third abnormal degree of the pixel point cluster class as the first liver cancer factor of the pixel point cluster class in the arterial phase MRI image; For any pixel point cluster class in the arterial phase MRI image, take the pixel point cluster class adjacent to the pixel point cluster class in the arterial phase MRI image as the neighborhood cluster class of the pixel point cluster class; according to the differences in the first abnormal degree, the second abnormal degree, and the third abnormal degree between the neighborhood cluster class of the pixel point cluster class and the pixel point cluster class, obtain the second liver cancer factor of the pixel point cluster class in the arterial phase MRI image, and its specific calculation formula is: ; In the formula, represents the second liver cancer factor of the pixel point cluster class in the arterial phase MRI image; represents the number of neighborhood cluster classes of the pixel point cluster class in the arterial phase MRI image; represents the difference in the first abnormal degree between the neighborhood cluster class of the pixel point cluster class and the pixel point cluster class; represents the difference in the second abnormal degree between the neighborhood cluster class of the pixel point cluster class and the pixel point cluster class; represents the difference in the third abnormal degree between the neighborhood cluster class of the pixel point cluster class and the pixel point cluster class; represents the difference in the fourth abnormal degree between the neighborhood cluster class of the pixel point cluster class and the pixel point cluster class; represents the linear normalization function.

[0012] Preferably, the specific method for obtaining the possibility that the area corresponding to the pixel point cluster class in the arterial phase MRI image is a liver cancer area based on the first liver cancer factor and the second liver cancer factor of the pixel point cluster class in the arterial phase MRI image, and determining whether the area corresponding to the pixel point cluster class in the arterial phase MRI image is a liver cancer area includes: For any pixel cluster class in the arterial-phase MRI image, normalize the product of the first liver cancer factor and the second liver cancer factor of the pixel cluster class in the arterial-phase MRI image, and use the normalization result as the possibility that the corresponding region of the pixel cluster class in the arterial-phase MRI image is a liver cancer region; Preset a possibility threshold , for any pixel cluster class in the arterial-phase MRI image, if the possibility that the corresponding region of the pixel cluster class in the arterial-phase MRI image is a liver cancer region is greater than or equal to , then the corresponding region of the pixel cluster class in the arterial-phase MRI image is a liver cancer region.

[0013] The present invention also provides an auxiliary diagnosis system for primary liver cancer based on imaging data, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of any one of the above-mentioned auxiliary diagnosis methods for primary liver cancer based on imaging data are implemented.

[0014] An embodiment of the present invention provides an auxiliary diagnosis system for primary liver cancer based on imaging data, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of any one of the above-mentioned auxiliary diagnosis methods for primary liver cancer based on imaging data are implemented.

[0015] The beneficial effect of the technical solution of the present invention is as follows: By collecting MRI images in each period, according to the differences between adjacent pixels in the MRI image, the possibility that adjacent pixels are located in the same tissue region is obtained; several pixel cluster classes are obtained in this way. Since the MRI image contains various tissues in the human body, to accurately obtain the corresponding region of liver cancer in the MRI image, the MRI image needs to be divided into several pixel cluster classes. Also, because the differences within the same tissue are small, the human body region can be divided into several pixel cluster classes by analyzing the differences between adjacent pixels in the human body region.

[0016] Moreover, since liver cancer tissues are more susceptible to contrast agents than normal tissues, and liver cancer is mainly supplied by the hepatic artery while normal liver tissues are supplied by both the portal vein and the hepatic artery, the pixels in the pixel cluster class described in the arterial-phase MRI image are used to obtain the first liver cancer factor and the second liver cancer factor of the pixel cluster class in the arterial-phase MRI image due to the differences between the pixels in other MRI images, so as to accurately evaluate the possibility that the area corresponding to the pixel cluster class in the arterial-phase MRI image is a liver cancer area. Finally, based on the possibility that the area corresponding to the pixel cluster class in the arterial-phase MRI image is a liver cancer area, it is determined whether the area corresponding to the pixel cluster class in the arterial-phase MRI image is a liver cancer area. In this application, by comparing and analyzing MRI images at different times, the characteristics of liver cancer areas in MRI images are accurately extracted, thereby improving the accuracy of auxiliary detection of liver cancer. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0018] Figure 1 It is a flowchart of the steps of the method for auxiliary diagnosis of primary liver cancer based on imaging data of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0019] In order to further elaborate on the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following will, in conjunction with the drawings and preferred embodiments, describe in detail the specific implementation manners, structures, features, and effects of the method and system for auxiliary diagnosis of primary liver cancer based on imaging data according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0020] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the technical field to which the present invention belongs.

[0021] The following will specifically describe the specific solutions of the method and system for auxiliary diagnosis of primary liver cancer based on imaging data provided by the present invention with reference to the drawings.

[0022] Please refer to Figure 1 , which shows a flowchart of the steps of the method for auxiliary diagnosis of primary liver cancer based on imaging data provided by an embodiment of the present invention. The method includes the following steps: Step S001: Obtain the initial MRI image, arterial phase MRI image, portal venous phase MRI image, equilibrium phase MRI image, and delayed phase MRI image, and obtain the human body regions in each MRI image.

[0023] It should be noted that, as an auxiliary diagnosis method for primary liver cancer based on imaging data, the specific purpose of this embodiment is to detect whether a patient has liver cancer through the patient's MRI images. Therefore, it is first necessary to collect the patient's MRI images. Since the collected MRI images contain background regions, in order to avoid the interference of the background regions, it is also necessary to remove the background parts in the MRI images.

[0024] Specifically, let the patient lie flat on the MRI scanning bed, and inject a contrast agent into the patient's body through intravenous injection, and collect the initial MRI image, arterial phase MRI image, portal venous phase MRI image, equilibrium phase MRI image, and delayed phase MRI image. Furthermore, through a semantic segmentation algorithm, remove the background parts in each MRI image to obtain the human body regions in each MRI image.

[0025] It should be noted that the initial MRI image is the MRI image collected within 20 seconds after injecting the contrast agent; the arterial phase MRI image is the MRI image collected within 20 to 30 seconds after injecting the contrast agent; the portal venous phase MRI image is the MRI image collected within 50 to 60 seconds after injecting the contrast agent; the equilibrium phase MRI image is the MRI image collected within 3 to 5 minutes after injecting the contrast agent; the delayed phase MRI image is the MRI image collected within 5 to 10 minutes after injecting the contrast agent. Since both the collection of MRI images and the semantic segmentation algorithm are well-known prior arts, they will not be elaborated in this embodiment.

[0026] Step S002: Obtain the blurring degree of each pixel point in the human body region according to the pixel points within the local range of each pixel point in the human body region; obtain the possibility that adjacent pixel points in the human body region are in the same tissue region according to the gray values and blurring degrees of adjacent pixel points in the human body region; divide the human body region into several pixel point cluster classes according to the possibility that adjacent pixel points in the human body region are in the same tissue region.

[0027] It should be noted that the human body regions in the MRI images contain various tissues in the human body, such as bones, muscles, liver, liver cancer, etc., in order to accurately obtain the regions corresponding to liver cancer in the MRI images. Since the differences between tissues in the MRI images are large, while the differences within each tissue in the MRI images are small, the human body region can be divided into several pixel point cluster classes by analyzing the differences between adjacent pixel points in the human body region, and one pixel point cluster class corresponds to one tissue region.

[0028] It should be further noted that in a clear image, sharp edges or details will form sparse and concentrated maximum points, and the gray level of surrounding pixels changes smoothly; while blurring will cause feature diffusion, the maximum points are more densely distributed. At the same time, the smoothing effect of blurring will significantly reduce the gray level of pixels in the area between the maximum points, forming a greater difference. Therefore, the denser the maximum points and the more obvious the difference, it indicates that the high-frequency details of the image are weakened and the degree of blurring is higher. Therefore, the degree of blurring of pixel points can be obtained based on this.

[0029] Specifically, for any pixel point in any human body area, a -sized local window is constructed with the pixel point in the human body area as the center, and the local window of the pixel point in the human body area is obtained. The specific value of can be set according to the actual situation by itself, and there is no hard requirement in this embodiment. In this embodiment, is taken as an example for description; for the th pixel point in the local window of the pixel point in the human body area, if the gray level values of the pixel points in the eight-neighborhood of the th pixel point are all less than the gray level value of each pixel point, then the th pixel point is recorded as a maximum point; Furthermore, for the th maximum point in the local window of the pixel point in the human body area, connect the th maximum point in the local window of the pixel point in the human body area with each other maximum point, and obtain several line segments of the th maximum point. The pixel points located on the line segment of the th maximum point are used as the comparison points of the th maximum point; according to the distances between all the maximum points in the local window of the pixel point in the human body area, and the gray level differences of their comparison points, the degree of blurring of the pixel point in the human body area is obtained. The specific calculation formula is: ; In the formula, represents the degree of blurring of the pixel point in the human body area; represents the number of maximum points in the local window of the pixel point in the human body area; represents the th and the th distance between the maximum points in the local window of the pixel point in the human body area; represents the number of comparison points of the th maximum point in the local window of the pixel point in the human body area; represents the gray value of the th maximum value within the local window of the pixel point in the human body region; represents the gray value of the th control point of the th maximum value within the local window of the pixel point in the human body region; represents the absolute value function; represents the linear normalization function, and its specific normalization range is for all pixel points in all human body regions .

[0030] It should be noted that represents the density of all maximum value points within the local window of the pixel point in the human body region; represents the difference in gray value between all pixel points located between the maximum value points and the maximum value points within the local window of the pixel point in the human body region; since the denser the distribution of the maximum value points and the greater the difference in gray value between the pixel points between the maximum value points within the local range of the pixel point, the more blurred the pixel, so the larger the value, the more blurred the pixel point in the human body region; in this embodiment, the specific calculation process of the difference is the absolute value of the difference between two values.

[0031] It should be further noted that when the blurring degrees of two adjacent pixel points are large, it indicates that the high-frequency information (such as edges and textures) in their regions has been significantly suppressed, resulting in the reduction of the originally possible gray value difference due to the smoothing effect. If the difference in their gray values is small itself, it further indicates that they may originally have similar tissue characteristics, that is, the two adjacent pixel points are more likely to be located in the same tissue region. Therefore, the possibility that two adjacent pixel points are located in the same tissue region can be obtained based on the blurring degrees and gray values of the two adjacent pixel points.

[0032] Specifically, for any two adjacent pixel points in any human body region, according to the gray values and blurring degrees of the two adjacent pixel points in the human body region, the similarity degree of the two adjacent pixel points in the human body region is obtained, and its specific calculation formula is: ; In the formula, represents the possibility that the two adjacent pixel points in the human body region are located in the same tissue region; represents the difference in blurring degree between the two adjacent pixel points in the human body region; represents the difference in gray value between the two adjacent pixel points in the human body region; represents the average value of the blurring degrees of the two adjacent pixel points in the human body region; represents the absolute value function; represents the exponential function with the natural constant as the base.

[0033] It should be noted that since the smaller the difference in the degree of blurriness and the gray value between two adjacent pixel points, and the greater the degree of blurriness of the two adjacent pixel points, the more likely the two adjacent pixel points are located in the same tissue region. Therefore the larger the value of , the more likely the two adjacent pixel points in the human body region are located in the same tissue region. That is, the human body region can be divided into several tissue regions based on this.

[0034] Specifically, for any pixel point in any human body region, obtain the similarity degree between the pixel point in the human body region and each adjacent pixel point, and classify the adjacent pixel point with the greatest similarity degree to the pixel point in the human body region into the same pixel point cluster class; Similarly, classify all pixel points in the human body region to obtain several pixel point cluster classes of the human body region.

[0035] So far, several pixel point cluster classes of the human body region are obtained.

[0036] Step S003: According to the pixel point cluster classes in different MRI images, obtain the first abnormal degree, the second abnormal degree, the third abnormal degree, and the fourth abnormal degree of the whole pixel point cluster class in the arterial phase MRI image; according to the first abnormal degree, the second abnormal degree, the third abnormal degree, and the fourth abnormal degree of the whole pixel point cluster class in the arterial phase MRI image, obtain the first liver cancer factor and the second liver cancer factor of the pixel point cluster class in the arterial phase MRI image.

[0037] It should be noted that the pixels in a pixel point cluster class in the MRI image are all pixel points in the same tissue of the patient, while the pixel points in adjacent pixel point cluster classes in the MRI image are pixel points in different tissues in the patient's body. Also, since liver cancer tissue is more susceptible to the influence of the contrast agent than normal tissue; therefore, by comparing the gray value changes of the pixel points in the pixel point cluster classes of different MRI images, the abnormal characteristics of the pixel point cluster class can be obtained, and further, according to the abnormal characteristics of the pixel point cluster class, the first liver cancer factor and the second liver cancer factor of the pixel point cluster class can be obtained for subsequent evaluation of the possibility that the region corresponding to the pixel point cluster class is a liver cancer region.

[0038] Specifically, for the human body region in any MRI image, take the average gray value of all pixel points in the human body region in the MRI image as the overall gray value of the MRI image; For the th pixel point in any pixel point cluster class in the arterial phase MRI image, for all MRI images, the coordinates are the same as the Pixels with the same pixel values as the th pixel are used as the corresponding pixels of the th pixel; the difference in gray values between the pixel corresponding to the th pixel in the initial MRI image and the th pixel is used as the first abnormal feature of the th pixel; Similarly, the first abnormal features of all pixels in the pixel cluster in the arterial phase MRI image are obtained, and the average value of the first abnormal features of all pixels in the pixel cluster in the arterial phase MRI image is used as the overall first abnormal feature of the pixel cluster in the arterial phase MRI image; Furthermore, the difference obtained by subtracting the difference in overall gray levels between the initial MRI image and the arterial phase MRI image from the overall first abnormal feature of the pixel cluster in the arterial phase MRI image is used as the first abnormal degree of the pixel cluster in the arterial phase MRI image, and its specific calculation formula is: ; In the formula, represents the first abnormal degree of the pixel cluster in the arterial phase MRI image; represents the overall first abnormal feature of the pixel cluster in the arterial phase MRI image; represents the difference in overall gray levels between the initial MRI image and the arterial phase MRI image.

[0039] The difference in gray values between the pixel corresponding to the th pixel in the portal venous phase MRI image and the th pixel is used as the second abnormal feature of the th pixel; Similarly, the second abnormal features of all pixels in the pixel cluster in the arterial phase MRI image are obtained, and the average value of the second abnormal features of all pixels in the pixel cluster in the arterial phase MRI image is used as the overall second abnormal feature of the pixel cluster in the arterial phase MRI image; Furthermore, the difference obtained by subtracting the difference in overall gray levels between the arterial phase MRI image and the portal venous phase MRI image from the overall second abnormal feature of the pixel cluster in the arterial phase MRI image is used as the second abnormal degree of the pixel cluster in the arterial phase MRI image, and its specific calculation formula is: ; In the formula, represents the second abnormal degree of the pixel cluster in the arterial phase MRI image; represents the overall second abnormal feature of the pixel cluster in the arterial phase MRI image; Indicates the difference in overall gray level between the arterial-phase MRI image and the portal-venous-phase MRI image.

[0040] Take the difference in gray value between the pixel point corresponding to the th pixel point in the equilibrium-phase MRI image and the th pixel point as the third abnormal feature of the th pixel point; Similarly, obtain the third abnormal features of all pixel points in the pixel point cluster class in the arterial-phase MRI image, and take the mean of the third abnormal features of all pixel points in the pixel point cluster class in the arterial-phase MRI image as the overall third abnormal feature of the pixel point cluster class in the arterial-phase MRI image; Furthermore, subtract the difference between the overall gray level of the arterial-phase MRI image and the equilibrium-phase MRI image from the overall third abnormal feature of the pixel point cluster class in the arterial-phase MRI image, and take the result as the third abnormal degree of the pixel point cluster class in the arterial-phase MRI image. The specific calculation formula is: ; In the formula, represents the third abnormal degree of the pixel point cluster class in the arterial-phase MRI image; represents the overall third abnormal feature of the pixel point cluster class in the arterial-phase MRI image; represents the difference in overall gray level between the arterial-phase MRI image and the equilibrium-phase MRI image.

[0041] Take the difference in gray value between the pixel point corresponding to the th pixel point in the equilibrium-phase MRI image and the pixel point corresponding to the th pixel point in the delayed-phase MRI image as the fourth abnormal feature of the th pixel point; Similarly, obtain the fourth abnormal features of all pixel points in the pixel point cluster class in the arterial-phase MRI image, and take the mean of the fourth abnormal features of all pixel points in the pixel point cluster class in the arterial-phase MRI image as the overall fourth abnormal feature of the pixel point cluster class in the arterial-phase MRI image; Furthermore, subtract the difference between the overall gray level of the equilibrium-phase MRI image and the delayed-phase MRI image from the overall fourth abnormal feature of the pixel point cluster class in the arterial-phase MRI image, and take the result as the fourth abnormal degree of the pixel point cluster class in the arterial-phase MRI image. The specific calculation formula is: ; In the formula, represents the fourth abnormal degree of the pixel point cluster class in the arterial-phase MRI image; Represents the fourth abnormal feature of the overall cluster of pixel points in the arterial phase MRI image; Represents the difference in overall gray scale between the equilibrium phase MRI image and the delayed phase MRI image.

[0042] It should be noted that the first, second, third, and fourth abnormal degrees of the cluster of pixel points in the above-mentioned arterial phase MRI image are used to measure the gray scale change degree of the pixel points in the cluster of pixel points; since liver cancer tissue is more susceptible to the influence of contrast agents than normal tissue, the greater the gray scale change degree of the pixel points in the cluster of pixel points, the more abnormal the cluster of pixel points. Since liver cancer is mainly supplied by the hepatic artery, it will show significant early enhancement in the arterial phase. Therefore, the first, second, and third abnormal degrees are obtained by comparing the differences in gray scale values between the arterial phase MRI image and other MRI images; also, since the gray scale value of pixel points gradually increases from the initial MRI image to the equilibrium phase MRI image, and the gray scale value of pixel points gradually decreases from the equilibrium phase MRI image to the delayed phase MRI image, the fourth abnormal degree is obtained by comparing the differences in gray scale values between the equilibrium phase MRI image and the delayed phase MRI image.

[0043] It should be further noted that when the first, second, third, and fourth abnormal degrees of the cluster of pixel points in the arterial phase MRI image are greater, the cluster of pixel points in the arterial phase MRI image is more likely to be a liver cancer area for the region. Therefore, the first liver cancer factor of the cluster of pixel points is obtained based on this; also, since liver cancer is mainly supplied by the hepatic artery, while normal liver tissue is supplied by both the portal vein and the hepatic artery, and the blood supply ratio is about 3:1. Therefore, if the cluster of pixel points is a liver cancer area for the region, the difference in the first abnormal degree between the cluster of pixel points and its adjacent cluster of pixel points is much greater than the differences in the second, third, and fourth abnormal degrees. Therefore, the second liver cancer factor of the cluster of pixel points in the arterial phase MRI image can be obtained based on this.

[0044] Specifically, for any cluster of pixel points in the arterial phase MRI image, the absolute value of the product of the first abnormal, second abnormal degree, and third abnormal degree of the cluster of pixel points is used as the first liver cancer factor of the cluster of pixel points in the arterial phase MRI image; Furthermore, for any cluster of pixel points in the arterial phase MRI image, the cluster of pixel points adjacent to the cluster of pixel points in the arterial phase MRI image is used as the neighborhood cluster of the cluster of pixel points; according to the differences in the first abnormal degree, second abnormal degree, and third abnormal degree between the neighborhood cluster of the cluster of pixel points and the cluster of pixel points, the second liver cancer factor of the cluster of pixel points in the arterial phase MRI image is obtained. The specific calculation formula is: ; In the formula, The second liver cancer factor representing the cluster of pixel points in the arterial-phase MRI image; The number of neighborhood cluster classes of the cluster of pixel points in the arterial-phase MRI image; The difference between the neighborhood cluster class of the cluster of pixel points and the cluster of pixel points in the first degree of abnormality; The difference between the neighborhood cluster class of the cluster of pixel points and the cluster of pixel points in the second degree of abnormality; The difference between the neighborhood cluster class of the cluster of pixel points and the cluster of pixel points in the third degree of abnormality; The difference between the neighborhood cluster class of the cluster of pixel points and the cluster of pixel points in the fourth degree of abnormality; Represents a linear normalization function, and its specific normalization range is for all cluster classes of pixel points in the arterial-phase MRI image .

[0045] It should be noted that The ratio of The larger the ratio, the greater the difference between the cluster of pixel points and its adjacent cluster of pixel points in the first degree of abnormality, compared to the differences in the second, third, and fourth degrees of abnormality. That is, the cluster of pixel points has more characteristics of a liver cancer region for the area.

[0046] Thus, the first liver cancer factor and the second liver cancer factor of all cluster classes of pixel points in the arterial-phase MRI image are obtained.

[0047] Step S004: According to the first liver cancer factor and the second liver cancer factor of the cluster of pixel points in the arterial-phase MRI image, obtain the possibility that the region corresponding to the cluster of pixel points in the arterial-phase MRI image is a liver cancer region, and determine whether the region corresponding to the cluster of pixel points in the arterial-phase MRI image is a liver cancer region.

[0048] It should be noted that after obtaining the first liver cancer factor and the second liver cancer factor of each cluster of pixel points in the arterial-phase MRI image through the above step S003, the possibility that the region corresponding to each cluster of pixel points in the arterial-phase MRI image is a liver cancer region can be quantified based on the first liver cancer factor and the second liver cancer factor of each cluster of pixel points in the arterial-phase MRI image. Further, according to the possibility that the region corresponding to each cluster of pixel points in the arterial-phase MRI image is a liver cancer region, it can be determined whether the patient has liver cancer.

[0049] Specifically, for any cluster of pixel points in the arterial-phase MRI image, normalize the product of the first liver cancer factor and the second liver cancer factor of the cluster of pixel points in the arterial-phase MRI image, and use the normalized result as the possibility that the region corresponding to the cluster of pixel points in the arterial-phase MRI image is a liver cancer region; in this implementation, the specific function for normalizing the product of the first liver cancer factor and the second liver cancer factor is function

[0050] Furthermore, a possibility threshold is preset , and the specific value of which can be set according to the actual situation and is not rigidly required in this embodiment. In this embodiment, taking as an example, for any pixel cluster class in the arterial phase MRI image, if the possibility that the area corresponding to the pixel cluster class in the arterial phase MRI image is a liver cancer area is greater than or equal to , then the area corresponding to the pixel cluster class in the arterial phase MRI image is a liver cancer area; if the possibility that the area corresponding to the pixel cluster class in the arterial phase MRI image is a liver cancer area is less than , then the area corresponding to the pixel cluster class in the arterial phase MRI image is not a liver cancer area.

[0051] This embodiment provides an auxiliary diagnosis system for primary liver cancer based on imaging data, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the method for auxiliary diagnosis of primary liver cancer based on imaging data in steps S001 to S004.

[0052] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the principles of the present invention shall be included in the protection scope of the present invention.

Claims

1. An auxiliary diagnosis method for primary liver cancer based on imaging data, characterized in that, The method includes the following steps: Obtain an initial MRI image, an arterial phase MRI image, a portal venous phase MRI image, an equilibrium phase MRI image, and a delayed phase MRI image, and obtain the human body region in each MRI image; Obtain the blur degree of each pixel point in the human body region according to the pixel points within the local range of each pixel point in the human body region; obtain the possibility that adjacent pixel points in the human body region are located in the same tissue region according to the gray values and blur degrees of adjacent pixel points in the human body region; divide the human body region into several pixel point cluster classes according to the possibility that adjacent pixel points in the human body region are located in the same tissue region; Obtain the first abnormal degree, the second abnormal degree, the third abnormal degree, and the fourth abnormal degree of the overall pixel point cluster class in the arterial phase MRI image according to the pixel point cluster classes in different MRI images; obtain the first liver cancer factor and the second liver cancer factor of the pixel point cluster class in the arterial phase MRI image according to the first abnormal degree, the second abnormal degree, the third abnormal degree, and the fourth abnormal degree of the overall pixel point cluster class in the arterial phase MRI image; Obtain the possibility that the corresponding region of the pixel point cluster class in the arterial phase MRI image is a liver cancer region according to the first liver cancer factor and the second liver cancer factor of the pixel point cluster class in the arterial phase MRI image, and determine whether the corresponding region of the pixel point cluster class in the arterial phase MRI image is a liver cancer region.

2. The primary liver cancer auxiliary diagnosis method based on imaging data according to claim 1, wherein The specific method for obtaining the blur degree of each pixel point in the human body region according to the pixel points within the local range of each pixel point in the human body region includes: For any pixel point in any human body region, a -sized local window is constructed with the pixel point in the human body region as the center to obtain the local window of the pixel point in the human body region; for the -th pixel point in the local window of the pixel point in the human body region, if the gray values of the pixel points within the eight-neighborhood of the -th pixel point are all less than the gray value of each pixel point, then the -th pixel point is recorded as a maximum point; For the th maximum point in the local window of the pixel point in the human body region, connect the th maximum point in the local window of the pixel point in the human body region with each other maximum point to obtain several line segments of the th maximum point, and use the pixel points on the line segment located at the th maximum point as the reference points of the th maximum point; according to the distances between all the maximum points in the local window of the pixel point in the human body region and the differences in gray level between the reference points of all the maximum points, obtain the degree of blurriness of the pixel point in the human body region, and its specific calculation formula is: ； In the formula, represents the blurring degree of the pixel point in the human body area; represents the number of maximum points in the local window of the pixel point in the human body area; represents the th and th maximum points in the local window of the pixel point in the human body area; represents the number of control points of the th maximum point in the local window of the pixel point in the human body area; represents the gray value of the th maximum point in the local window of the pixel point in the human body area; represents the gray value of the th th control point of the represents the absolute value function; represents the linear normalization function.

3. The primary liver cancer auxiliary diagnosis method based on imaging data according to claim 1, characterized in that The specific method for obtaining the possibility that adjacent pixel points in the human body region are located in the same tissue region according to the gray values and blur degrees of adjacent pixel points in the human body region includes: For any two adjacent pixel points in any human body region, obtain the similarity degree of the two adjacent pixel points in the human body region according to the gray values and blur degrees of the two adjacent pixel points in the human body region, and its specific calculation formula is: ； In the formula, represents the probability that the two adjacent pixel points in the human body area are located in the same tissue area; represents the difference in the degree of blurriness between the two adjacent pixel points in the human body area; represents the difference in the gray value between the two adjacent pixel points in the human body area; represents the average value of the degree of blurriness of the two adjacent pixel points in the human body area; represents the absolute value function; represents the exponential function with the natural constant as the base.

4. The primary liver cancer auxiliary diagnosis method based on imaging data according to claim 1, wherein The specific method for obtaining the first abnormal degree of the overall pixel point cluster class in the arterial phase MRI image includes: For the human body region in any MRI image, take the average gray value of all pixel points in the human body region in the MRI image as the overall gray value of the MRI image; For the th pixel point in any pixel point cluster in the arterial phase MRI image, the pixel points with the same coordinates as the th pixel point in all MRI images are used as the corresponding pixel points of the th pixel point; the difference in gray value between the pixel point corresponding to the th pixel point in the initial MRI image and the th pixel point is used as the first abnormal feature of the th pixel point; Obtain the first abnormal features of all pixel points in the pixel point cluster class in the arterial phase MRI image, and take the average value of the first abnormal features of all pixel points in the pixel point cluster class in the arterial phase MRI image as the first abnormal feature of the overall pixel point cluster class in the arterial phase MRI image; Take the difference obtained by subtracting the difference in the overall gray values between the initial MRI image and the arterial phase MRI image from the first abnormal feature of the overall pixel point cluster class in the arterial phase MRI image as the first abnormal degree of the pixel point cluster class in the arterial phase MRI image.

5. The primary liver cancer auxiliary diagnosis method based on imaging data according to claim 4, wherein, The specific method for obtaining the second abnormal degree is: Take the difference in gray value between the pixel corresponding to the th pixel point in the portal venous phase MRI image and the th pixel point as the second abnormal feature of the th pixel point; Obtain the second abnormal features of all the pixel points in the pixel point cluster class in the arterial phase MRI image, and take the mean value of the second abnormal features of all the pixel points in the pixel point cluster class in the arterial phase MRI image as the second abnormal feature of the pixel point cluster class as a whole in the arterial phase MRI image; Subtract the difference between the second abnormal feature of the pixel point cluster class as a whole in the arterial phase MRI image and the difference in overall gray level between the arterial phase MRI image and the portal venous phase MRI image, and take it as the second abnormal degree of the pixel point cluster class in the arterial phase MRI image.

6. The primary liver cancer auxiliary diagnosis method based on imaging data according to claim 4, wherein The specific method for obtaining the third abnormal degree is as follows: Take the difference in pixel values between the pixel corresponding to the th pixel in the balanced-phase MRI image and the th pixel as the third abnormal feature of the th pixel; Obtain the third abnormal features of all the pixel points in the pixel point cluster class in the arterial phase MRI image, and take the mean value of the third abnormal features of all the pixel points in the pixel point cluster class in the arterial phase MRI image as the third abnormal feature of the pixel point cluster class as a whole in the arterial phase MRI image; Subtract the difference between the third abnormal feature of the pixel point cluster class as a whole in the arterial phase MRI image and the difference in overall gray level between the arterial phase MRI image and the equilibrium phase MRI image, and take it as the third abnormal degree of the pixel point cluster class in the arterial phase MRI image.

7. The method for auxiliary diagnosis of primary liver cancer based on imaging data according to claim 4, characterized in that, The specific method for obtaining the fourth abnormal degree is as follows: Take the difference in gray value between the pixel corresponding to the th pixel in the equilibrium-phase MRI image and the pixel corresponding to the th pixel in the delayed-phase MRI image as the fourth abnormal feature of the th pixel; Obtain the fourth abnormal features of all the pixel points in the pixel point cluster class in the arterial phase MRI image, and take the mean value of the fourth abnormal features of all the pixel points in the pixel point cluster class in the arterial phase MRI image as the fourth abnormal feature of the pixel point cluster class as a whole in the arterial phase MRI image; Subtract the difference between the fourth abnormal feature of the pixel point cluster class as a whole in the arterial phase MRI image and the difference in overall gray level between the equilibrium phase MRI image and the delayed phase MRI image, and take it as the fourth abnormal degree of the pixel point cluster class in the arterial phase MRI image.

8. The primary liver cancer auxiliary diagnosis method based on imaging data according to claim 1, wherein The specific method for obtaining the first liver cancer factor and the second liver cancer factor of the pixel point cluster class in the arterial phase MRI image includes: For any pixel point cluster class in the arterial phase MRI image, take the absolute value of the product of the first abnormality, the second abnormal degree, and the third abnormal degree of the pixel point cluster class as the first liver cancer factor of the pixel point cluster class in the arterial phase MRI image; For any pixel point cluster class in the arterial phase MRI image, take the pixel point cluster class adjacent to the pixel point cluster class in the arterial phase MRI image as the neighborhood cluster class of the pixel point cluster class; according to the differences in the first abnormal degree, the second abnormal degree, and the third abnormal degree between the neighborhood cluster class of the pixel point cluster class and the pixel point cluster class, obtain the second liver cancer factor of the pixel point cluster class in the arterial phase MRI image, and its specific calculation formula is: ； In the formula, represents the second liver cancer factor of the pixel cluster class in the arterial phase MRI image; represents the number of neighborhood cluster classes of the pixel cluster class in the arterial phase MRI image; represents the difference between the neighborhood cluster class of the pixel cluster class and the pixel cluster class in the first degree of abnormality; represents the difference between the neighborhood cluster class of the pixel cluster class and the pixel cluster class in the second degree of abnormality; represents the difference between the neighborhood cluster class of the pixel cluster class and the pixel cluster class in the third degree of abnormality; represents the difference between the neighborhood cluster class of the pixel cluster class and the pixel cluster class in the fourth degree of abnormality; represents the linear normalization function.

9. The primary liver cancer auxiliary diagnosis method based on imaging data according to claim 1, wherein The specific method for obtaining the possibility that the area corresponding to the pixel point cluster class in the arterial phase MRI image is a liver cancer area based on the first liver cancer factor and the second liver cancer factor of the pixel point cluster class in the arterial phase MRI image, and determining whether the area corresponding to the pixel point cluster class in the arterial phase MRI image is a liver cancer area includes: For any pixel cluster class in the arterial-phase MRI image, normalize the product of the first liver cancer factor and the second liver cancer factor of the pixel cluster class in the arterial-phase MRI image, and use the normalization result as the possibility that the corresponding region of the pixel cluster class in the arterial-phase MRI image is a liver cancer region; Preset a probability threshold , for any pixel cluster class in the arterial-phase MRI image, if the probability that the corresponding region of the pixel cluster class in the arterial-phase MRI image is a liver cancer region is greater than or equal to , then the corresponding region of the pixel cluster class in the arterial-phase MRI image is a liver cancer region.

10. An auxiliary diagnosis system for primary liver cancer based on imaging data, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, When the computer program is executed by a processor, it implements the steps of the primary liver cancer auxiliary diagnosis method based on imaging data according to any one of claims 1-9.

Images