Prognosis evaluation method, device and equipment for hepatocellular carcinoma patient and medium

Abstract

The invention relates to the technical field of prognosis analysis, and discloses a prognosis evaluation method, device and equipment for a hepatocellular carcinoma patient and a medium. The method comprises the following steps: acquiring a first target feature vector of at least one hepatocellular carcinoma patient; performing feature screening on the first target feature vectors of the hepatocellular carcinoma patients based on a cross validation method and an akaike information criterion to obtain target features; taking the target feature as a training sample of a preset prognosis model so as to train the prognosis model; and acquiring histopathological section data of a target hepatocellular carcinoma patient, and inputting the histopathological section data into the trained prognosis model to obtain a prognosis evaluation score of the target hepatocellular carcinoma patient. The prognosis performance of the hepatocellular carcinoma patient can be accurately evaluated through the trained prognosis model, and the accuracy of prognosis evaluation is improved.

Description

Background Art

[0002] At the cellular and tissue level, histopathological sections contain rich characteristic information about tumors and their microenvironment. In clinical practice, pathologists observe histopathological sections under a high-power microscope to qualitatively evaluate histopathological patterns and predict cancer behavior to a certain extent. However, manual evaluation methods are time-consuming and easily affected by human subjectivity, resulting in low accuracy in evaluating the prognostic performance of tissue cells. Summary of the invention

[0003] The present application provides a method, device, equipment and medium for evaluating the prognosis of patients with hepatocellular carcinoma, which can accurately evaluate the prognostic performance of patients with hepatocellular carcinoma through a trained prognostic model, thereby improving the accuracy of the prognostic evaluation.

[0004] Other features and advantages of the present application will become apparent from the following detailed description, or may be learned in part by the practice of the present application.

[0005] According to one aspect of an embodiment of the present application, a method for evaluating the prognosis of a patient with hepatocellular carcinoma is provided, the method comprising: Obtaining a first target feature vector of at least one hepatocellular carcinoma patient; Performing feature screening on the first target feature vector of each of the hepatocellular carcinoma patients based on a cross-validation method and Akaike information criterion to obtain a target feature; Using the target feature as a training sample of a preset prognostic model to train the prognostic model; Acquiring tissue pathology section data of a target hepatocellular carcinoma patient, and inputting the tissue pathology section data into the trained prognostic model to obtain a prognostic assessment score of the target hepatocellular carcinoma patient; The first target feature vector of a single hepatocellular carcinoma patient can be obtained by the following steps: Obtaining a digital pathological section of the hepatocellular carcinoma patient, wherein the digital pathological section is a tissue pathological section of the hepatocellular carcinoma patient stained with hematoxylin and eosin; Performing image block cutting on the digital pathological sections to obtain a first preset number of image blocks; Performing cell nucleus segmentation on each of the image blocks respectively to obtain cell nucleus masks corresponding to each of the image blocks respectively, wherein the cell nucleus masks are image data having features of all cell nuclei in the image blocks; Perform feature extraction based on each of the cell nucleus masks to obtain second target feature vectors each having a cell nucleus shape feature, a cell nucleus texture feature, and a cell nucleus interaction feature, and each of the image blocks has one second target feature vector; A second target feature vector of the target image block is obtained to generate a feature matrix corresponding to the target image block and the second target feature vector, and feature concatenation is performed on the feature matrix to obtain the first target feature vector.

[0006] In one embodiment of the present application, based on the above solution, a single second target feature vector can be obtained by the following steps: Extracting shape features of each cell nucleus in the image block according to the cell nucleus mask to obtain a first feature vector having first dimension information, wherein the first feature vector is used to characterize shape features of the cell nucleus in the image block; Extracting texture features of each cell nucleus in the image block according to the cell nucleus mask to obtain a second feature vector having second dimension information, wherein the second feature vector is used to characterize the cell nucleus texture features of the image block; Extracting the interaction features of each cell nucleus in the image block according to the cell nucleus mask to obtain a third feature vector having third dimension information, wherein the third feature vector is used to characterize the cell nucleus interaction features of the image block; The first feature vector, the second feature vector, and the third feature vector are horizontally concatenated to obtain the second target feature vector having the sum of the first dimensional information, the second dimensional information, and the third dimensional information.

[0007] In one embodiment of the present application, based on the above solution, the first feature vector having the first dimension information can be obtained by the following steps: Extracting a first target parameter feature of each of the cell nuclei according to the cell nucleus mask, where the first target parameter feature is a feature vector having first initial dimension information; Performing a first preset statistical operation on each of the first target parameter features to obtain a second target parameter feature of each of the cell nuclei, wherein the second target parameter feature is a feature vector having second initial dimension information; Performing feature transverse splicing of the first target parameter feature and the second target parameter feature to obtain a first feature vector having the first dimensional information; The first dimension information is the product of the first initial dimension information and the second initial dimension information.

[0008] In one embodiment of the present application, based on the above solution, the second feature vector having the second dimension information can be obtained by the following steps: Extracting a third target parameter feature of each of the cell nuclei according to the cell nucleus mask, wherein the third target parameter feature is a feature vector having third initial dimension information; Performing a second preset statistical operation on each of the third target parameter features to obtain a fourth target parameter feature of each of the cell nuclei, wherein the fourth target parameter feature is a feature vector having fourth initial dimension information; Performing feature transverse splicing of the third target parameter feature and the fourth target parameter feature to obtain a second feature vector having the second dimension information; The second dimensional information is the product of the third initial dimensional information and the fourth initial dimensional information.

[0009] In one embodiment of the present application, based on the above solution, the third feature vector having the third dimension information can be obtained by the following steps: Extracting a fifth target parameter feature of each of the cell nuclei according to the cell nucleus mask, wherein the fifth target parameter feature is a feature vector having fifth initial dimension information; Performing a third preset statistical operation on each of the fifth target parameter features to obtain a sixth target parameter feature of each of the cell nuclei, wherein the sixth target parameter feature is a feature vector having sixth initial dimension information; Performing feature horizontal splicing of the fifth target parameter feature and the sixth target parameter feature to obtain a third feature vector having the third dimension information; The third dimensional information is obtained according to the fifth initial dimensional information and the sixth initial dimensional information.

[0010] In one embodiment of the present application, based on the above scheme, the second target feature vector is a feature vector having the sum of the first dimension information, the second dimension information and the third dimension information, and the step of performing feature concatenation on the feature matrix to obtain the first target feature vector includes: Performing a third preset statistical operation on each of the second target feature vectors in the feature matrix to obtain a third target feature vector having fourth dimension information; The second target feature vector is concatenated with the third target feature vector to obtain the first target feature vector.

[0011] In one embodiment of the present application, based on the above scheme, the first target feature vector of each hepatocellular carcinoma patient is subjected to feature screening based on the cross-validation method and the Akaike information criterion to obtain the target feature, including: Performing a preset round of 10-fold cross validation on the first target feature vector of each of the hepatocellular carcinoma patients, and determining the target screening feature obtained by screening in each round using a two-way selection stepwise regression method based on the Akaike information criterion, wherein each round corresponds to an Akaike information criterion value; A target round with the smallest Akaike information criterion value is selected in each round, and the target screening feature screened by the target round is obtained as the target feature.

[0012] According to one aspect of an embodiment of the present application, a device for evaluating the prognosis of a patient with hepatocellular carcinoma is provided, the device comprising: A first acquisition unit, configured to acquire a first target feature vector of at least one hepatocellular carcinoma patient; A screening unit, configured to perform feature screening on the first target feature vector of each of the hepatocellular carcinoma patients based on a cross-validation method and an Akaike information criterion to obtain a target feature; A training unit, used for taking the target feature as a training sample of a preset prognosis model to train the prognosis model; a second acquisition unit, configured to acquire the histopathological section data of the target hepatocellular carcinoma patient, and input the histopathological section data into the trained prognostic model to obtain the prognostic evaluation score of the target hepatocellular carcinoma patient; The first target feature vector of a single hepatocellular carcinoma patient can be obtained by the following steps: Obtaining a digital pathological section of the hepatocellular carcinoma patient, wherein the digital pathological section is a tissue pathological section of the hepatocellular carcinoma patient stained with hematoxylin and eosin; Performing image block cutting on the digital pathological sections to obtain a first preset number of image blocks; Performing cell nucleus segmentation on each of the image blocks respectively to obtain cell nucleus masks corresponding to each of the image blocks respectively, wherein the cell nucleus masks are image data having features of all cell nuclei in the image blocks; Perform feature extraction based on each of the cell nucleus masks to obtain second target feature vectors each having a cell nucleus shape feature, a cell nucleus texture feature, and a cell nucleus interaction feature, and each of the image blocks has one second target feature vector; A second target feature vector of the target image block is obtained to generate a feature matrix corresponding to the target image block and the second target feature vector, and feature concatenation is performed on the feature matrix to obtain the first target feature vector.

[0013] According to one aspect of an embodiment of the present application, a computer-readable storage medium is provided, on which a computer program is stored. The computer program includes executable instructions. When the executable instructions are executed by a processor, the method described in the above embodiment is implemented.

[0014] According to one aspect of an embodiment of the present application, an electronic device is provided, comprising: one or more processors; and a memory for storing executable instructions of the processors, wherein when the executable instructions are executed by the one or more processors, the one or more processors implement the methods described in the above embodiments.

[0015] The present application obtains a first target feature vector of at least one hepatocellular carcinoma patient, wherein the digital pathological sections of the hepatocellular carcinoma patient are cut through image blocks and the nucleus is segmented to obtain a nucleus mask, and the analysis of the nucleus mask can more clearly obtain the nucleus shape features, nucleus texture features and nucleus interaction features. Each image block corresponds to a second target feature vector, and the second target feature vector is a feature vector with multiple dimensions, that is, it can be used to characterize multiple features of the nucleus. Therefore, the first target feature vector obtained by feature splicing of the feature matrix generated by the second target feature vector can be used to characterize the nucleus feature set of a single patient.

[0016] Furthermore, by using the first target feature vector to perform sample training on the preset prognostic model, the prognostic evaluation of the prognostic model can be made more accurate. It is only necessary to input the tissue pathology section data of the target hepatocellular carcinoma patient into the trained prognostic model, and then an accurate prognostic evaluation score will be automatically output, thereby achieving efficient and automated prognostic evaluation, and solving the problems of inefficiency and low accuracy caused by manual evaluation in the prior art.

[0017] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The drawings herein are incorporated into the specification and constitute a part of the specification, showing embodiments consistent with the present application, and together with the specification, are used to explain the principles of the present application. Obviously, the drawings described below are only some embodiments of the present application, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. In the drawings: Figure 1 This is a flow chart of a method for evaluating the prognosis of a patient with hepatocellular carcinoma according to an embodiment of the present application; Figure 2 It is a full logic diagram of the prognosis model shown according to the embodiment of the present application; Figure 3 is a block diagram of a prognosis assessment device for hepatocellular carcinoma patients according to an embodiment of the present application; Figure 4 It is a schematic diagram of the system structure of an electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0019] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in a variety of forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided so that this application will be more comprehensive and complete and fully convey the concept of the example embodiments to those skilled in the art.

[0020] In addition, described feature, structure or characteristic can be combined in one or more embodiments in any suitable manner. In the following description, many specific details are provided to provide a full understanding of the embodiments of the present application. However, those skilled in the art will appreciate that the technical scheme of the present application can be put into practice without one or more of the specific details, or other methods, components, devices, steps, etc. can be adopted. In other cases, known methods, devices, realizations or operations are not shown or described in detail to avoid blurring the various aspects of the application.

[0021] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically independent entities. That is, these functional entities may be implemented in software form, or in one or more hardware modules or integrated circuits, or in different networks and / or processor devices and / or micro-control node devices.

[0022] The flowcharts shown in the accompanying drawings are only exemplary and do not necessarily include all the contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps can be decomposed, and some operations / steps can be combined or partially combined, so the actual execution order may change according to actual conditions.

[0023] It should be noted that the "multiple" mentioned in this article refers to two or more. "And / or" describes the association relationship of the associated objects, indicating that there can be three relationships. For example, A and / or B can mean: A exists alone, A and B exist at the same time, and B exists alone. The character " / " generally indicates that the associated objects before and after are in an "or" relationship.

[0024] The following is a detailed introduction to the background technology corresponding to the embodiments of the present application: At the cellular and tissue level, histopathological sections contain rich characteristic information about the tumor and its microenvironment. In clinical practice, pathologists observe H&E-stained sections under a high-power microscope to qualitatively evaluate histopathological patterns and predict cancer behavior to a certain extent. However, manual evaluation is time-consuming and subjective, and cannot capture sub-visual features. To address the above problems, many prognostic models have been developed to quantify the prognostic status of postoperative hepatocellular carcinoma patients, such as clinical pathological models developed based on clinical and pathological information, deep learning models developed based on deep neural networks, and manual feature models developed based on domain knowledge and mathematical statistical theory. These prognostic models have their own advantages and disadvantages in specific practical applications, as described below: Clinical pathological model: The clinical pathological model refers to a prognostic model constructed based on the patient's clinical and pathological information, such as the patient's age, gender and other information. It mainly quantifies the characteristics of hepatocellular carcinoma at the macroscopic and tissue levels. The combination of more clinical and pathological information can achieve a higher prognostic performance in patients with postoperative hepatocellular carcinoma. However, more effective clinical and pathological information often means higher detection technology (such as molecular, genetic testing, etc.) and material costs, which to a certain extent limits its widespread application in clinical practice. In addition, due to the heterogeneity of tumors and their microenvironment, patients with the same clinical pathological information will have significant differences in postoperative prognosis, which also limits the further development of clinical pathological models in clinical applications.

[0025] Deep learning model: A deep learning model refers to a prognostic model built based on a deep neural network. With the advancement of big data analysis and the increasing number of digital pathology sections, deep learning has become the mainstream method of choice for automated tissue image analysis. Deep convolutional neural networks, deep residual networks and other deep learning models have indeed improved prognostic performance in assisted diagnosis. However, due to the "black box" nature of deep learning, its decision-making process is difficult to explain and understand, which makes it difficult for deep learning models to be widely accepted in clinical practice.

[0026] Manual feature model: Manual feature model refers to a prognostic model developed based on domain knowledge and mathematical statistics theory. For example, the shape features of the region of interest (ROI), texture features, cell density, spatial distribution features and other indicators, under certain input conditions, the model will have clear output results, with high interpretability, which is easy for doctors to understand. In recent years, a large number of studies have emerged on the construction of prognostic models based on the deep features of tumors and their microenvironments extracted by manual feature models, such as the prognostic model for postoperative lung adenocarcinoma patients developed based on the texture features of the tumor area, and the prognostic model for hepatocellular carcinoma patients developed based on the shape and texture features of the tumor and its microenvironment. However, the prognostic efficacy of comprehensive nuclear features in postoperative hepatocellular carcinoma patients is still unknown.

[0027] Therefore, based on the problems existing in the above background technology, this application proposes a preset prognostic model (multi-dimensional fusion cell nucleus feature model) for the prognosis of patients with postoperative hepatocellular carcinoma, which quantifies the patient's prognostic status from three perspectives: cell nucleus shape, texture and interaction, using quantitative features to improve the accuracy of prognostic evaluation.

[0028] The implementation details of the technical solution of the embodiment of the present application are described in detail below: According to one aspect of the present application, a method for evaluating the prognosis of a patient with hepatocellular carcinoma is provided. Figure 1 The flowchart of the prognosis evaluation method for hepatocellular carcinoma patients according to the embodiment of the present application is shown. The prognosis evaluation method for hepatocellular carcinoma patients at least includes steps 110 to 140, which are described in detail as follows: In step 110, a first target feature vector of at least one hepatocellular carcinoma patient is obtained.

[0029] Specifically, the first target feature vector is a feature vector for characterizing a set of cell nucleus features of a patient with hepatocellular carcinoma. The first target feature vector is a feature vector having multiple dimensional information, such as the area ratio of the cell nucleus to the cell nucleus mask, Haralick texture features, cell nucleus interaction features, etc. In order to prevent the problem of overfitting caused by too much dimensional information in the first target feature vector, the first target feature vectors of each of the hepatocellular carcinoma patients are feature screened by cross-validation and Akaike information criterion to obtain target features, reduce overfitting problems, and obtain more representative target features.

[0030] In one embodiment of the present application, the first target feature vector of a single hepatocellular carcinoma patient may be obtained by the following steps: Obtaining a digital pathological section of the hepatocellular carcinoma patient, wherein the digital pathological section is a tissue pathological section of the hepatocellular carcinoma patient stained with hematoxylin and eosin; Performing image block cutting on the digital pathological sections to obtain a first preset number of image blocks; Performing cell nucleus segmentation on each of the image blocks respectively to obtain cell nucleus masks corresponding to each of the image blocks respectively, wherein the cell nucleus masks are image data having features of all cell nuclei in the image blocks; Perform feature extraction based on each of the cell nucleus masks to obtain second target feature vectors each having a cell nucleus shape feature, a cell nucleus texture feature, and a cell nucleus interaction feature, and each of the image blocks has one second target feature vector; A second target feature vector of the target image block is obtained to generate a feature matrix corresponding to the target image block and the second target feature vector, and feature concatenation is performed on the feature matrix to obtain the first target feature vector.

[0031] Specifically, the digital pathological sections of patients with hepatocellular carcinoma are tissue pathological sections of the patients with hepatocellular carcinoma stained with hematoxylin and eosin, that is, image data that can be recognized by computers and other equipment.

[0032] Please refer to Figure 2 As shown, Figure 2 For a complete processing flow diagram, the digital pathological slices are cut into image blocks to obtain the first preset number of image blocks, which can be specifically as follows: the tissue area edge is determined by the Otsu method (OTSU) and the tissue area is segmented, and then a tissue area mask is generated. Based on the tissue area mask, we cut the tissue area in the digital pathological slice into a 1000×1000 pixel non-overlapping Png format image block. To obtain a relatively dense tissue area, we filter out the image blocks with a tissue area ratio of less than 70%. Subsequently, we use a pre-trained segmentation model HoverNet to perform cell nucleus segmentation on these image blocks to generate a cell nucleus mask corresponding to each image block. It should be noted that an image block can contain multiple cell nuclei, so the first preset number mentioned in the embodiment of the present application can be any value, and the digital pathological slices in the present application can be cut into 1000 image blocks with 1000×1000 pixels. In other embodiments, the first preset number can also be set to other values, and the pixel setting of the image block can also be 500×500, and the pixels of the image block and the number of image blocks are not limited here.

[0033] Furthermore, cell nucleus segmentation is performed on each of the image blocks to obtain cell nucleus masks corresponding to each of the image blocks. The cell nucleus masks are image data having the characteristics of all cell nuclei in the image block, that is, each cell nucleus mask is actually an image data with only black and white pixels, where black is the background and white is the cell nucleus.

[0034] Further, feature extraction is performed based on each of the cell nucleus masks to obtain second target feature vectors each having a cell nucleus shape feature, a cell nucleus texture feature, and a cell nucleus interaction feature. Each of the image blocks has a second target feature vector, that is, we extract three different quantitative features on the image block, corresponding to the shape feature, texture feature, and interaction feature of the cell nucleus, respectively, as described below: In one embodiment of the present application, a single second target feature vector may be obtained by the following steps: Extracting shape features of each cell nucleus in the image block according to the cell nucleus mask to obtain a first feature vector having first dimension information, wherein the first feature vector is used to characterize shape features of the cell nucleus in the image block; Extracting texture features of each cell nucleus in the image block according to the cell nucleus mask to obtain a second feature vector having second dimension information, wherein the second feature vector is used to characterize the cell nucleus texture features of the image block; Extracting the interaction features of each cell nucleus in the image block according to the cell nucleus mask to obtain a third feature vector having third dimension information, wherein the third feature vector is used to characterize the cell nucleus interaction features of the image block; The first feature vector, the second feature vector, and the third feature vector are horizontally concatenated to obtain the second target feature vector having the sum of the first dimensional information, the second dimensional information, and the third dimensional information.

[0035] Specifically, the first eigenvector of the first dimension information can be specifically a 100-dimensional cell nucleus shape feature. The cell nucleus shape feature is used to quantify the morphological changes of a single cell nucleus. Specific description parameters include area ratio, distance ratio, distance standard deviation, perimeter area ratio, Fourier description operator of boundary points, etc. Ultimately, 25-dimensional features will be extracted from each cell nucleus, and each image block will obtain a feature matrix of N rows and 25 columns, where N represents the number of cell nuclei in the image block. For each column of this feature matrix, the average, standard deviation, median, and minimum to maximum values ​​are calculated to obtain four eigenvectors. These four eigenvectors are horizontally spliced ​​to obtain a 100-dimensional feature vector to characterize an image block.

[0036] In one embodiment of the present application, the first feature vector having the first dimension information may be obtained by the following steps: Extracting a first target parameter feature of each of the cell nuclei according to the cell nucleus mask, where the first target parameter feature is a feature vector having first initial dimension information; Performing a first preset statistical operation on each of the first target parameter features to obtain a second target parameter feature of each of the cell nuclei, wherein the second target parameter feature is a feature vector having second initial dimension information; Performing feature transverse splicing of the first target parameter feature and the second target parameter feature to obtain a first feature vector having the first dimensional information; The first dimension information is the product of the first initial dimension information and the second initial dimension information.

[0037] Specifically, the first target parameter feature is area ratio, distance ratio, distance standard deviation, perimeter area ratio, Fourier description operator of boundary points, etc., and the 25-dimensional (i.e., the first initial dimensional information) feature vector corresponding to each cell nucleus. The first preset statistical operation is to calculate the average, standard deviation, median, and ratio of the minimum and maximum values ​​of the 25-dimensional feature vectors corresponding to each cell nucleus, that is, the second target parameter feature of four dimensions (i.e., the second initial dimensional information), and then obtain four feature vectors. Then, each 25-dimensional feature vector is spliced ​​with the four statistics to obtain a first feature vector of 25×4=100 dimensions.

[0038] In one embodiment of the present application, the second feature vector having the second dimension information may be obtained by the following steps: Extracting a third target parameter feature of each of the cell nuclei according to the cell nucleus mask, wherein the third target parameter feature is a feature vector having third initial dimension information; Performing a second preset statistical operation on each of the third target parameter features to obtain a fourth target parameter feature of each of the cell nuclei, wherein the fourth target parameter feature is a feature vector having fourth initial dimension information; Performing feature transverse splicing of the third target parameter feature and the fourth target parameter feature to obtain a second feature vector having the second dimension information; The second dimensional information is the product of the third initial dimensional information and the fourth initial dimensional information.

[0039] Specifically, the second eigenvector is used to characterize the texture features of the cell nucleus: the texture features of the cell nucleus are used to quantify the distribution and co-occurrence of pixels in a single cell nucleus. The specific description parameters include the 5-dimensional Haralick texture features, which include information entropy, energy, information metric 1, and information metric 2. According to the grayscale co-occurrence matrix of each cell nucleus, the contrast matrix and the average intensity matrix are obtained, and the average, variance, energy, information entropy and inverse moment of the contrast matrix are calculated to obtain the 5-dimensional features. The average, variance, and information entropy of the average intensity matrix are calculated to obtain the 3-dimensional features. Finally, 5+5+3, that is, 13-dimensional (third initial dimension information) feature vectors will be extracted from each cell nucleus, which is the third target parameter feature. Each image block will get a feature matrix with N rows and 13 columns, where N represents the number of cell nuclei in the image block. The average, standard deviation, median, range, skewness, and kurtosis of each column of this feature matrix, that is, 6 dimensions (the fourth initial dimension information) and 6 statistics, are calculated to obtain 6 eigenvectors, which are the fourth target parameter features. These 6 eigenvectors are horizontally spliced, and finally a 78-dimensional second eigenvector of 6×13 is obtained to characterize an image block.

[0040] In one embodiment of the present application, the third feature vector having the third dimension information may be obtained by the following steps: Extracting a fifth target parameter feature of each of the cell nuclei according to the cell nucleus mask, wherein the fifth target parameter feature is a feature vector having fifth initial dimension information; Performing a third preset statistical operation on each of the fifth target parameter features to obtain a sixth target parameter feature of each of the cell nuclei, wherein the sixth target parameter feature is a feature vector having sixth initial dimension information; Performing feature horizontal splicing of the fifth target parameter feature and the sixth target parameter feature to obtain a third feature vector having the third dimension information; The third dimensional information is obtained according to the fifth initial dimensional information and the sixth initial dimensional information.

[0041] Specifically, the third eigenvector is used to characterize the cell nucleus interaction feature: the cell nucleus interaction feature is used to quantify the structural characteristics of the feature-driven local cell map (FLocK) and the interaction between the feature-driven local cell maps. First, according to the centroid and area of ​​the cell nucleus in each image block, the cell nucleus is clustered by the mean shift clustering algorithm within a bandwidth of 150 to obtain FLocKs of different categories. Based on the clustered FLocK, the specific description parameters are calculated, including the absolute area of ​​the intersection of two FLocKs, the proportion of the number of intersecting FLocKs to the total number of FLocKs, the area ratio between the intersection area and each FLocK, the fifth target parameter features of the Voronoi global map related parameters constructed based on the centroid of FLocK, the Delaunay triangulation related parameters, the minimum spanning tree related parameters, etc. (fifth initial dimension information), and the sixth target parameter features of the average, standard deviation, and range (sixth initial dimension information) of these fifth target parameter features are calculated, and finally a 222-dimensional third eigenvector is obtained to characterize an image block.

[0042] In one embodiment of the present application, the second target feature vector is a feature vector having the sum of the first dimensional information, the second dimensional information and the third dimensional information, and the step of performing feature concatenation on the feature matrix to obtain the first target feature vector includes: Performing a third preset statistical operation on each of the second target feature vectors in the feature matrix to obtain a third target feature vector having fourth dimension information; The second target feature vector is concatenated with the third target feature vector to obtain the first target feature vector.

[0043] Specifically, by horizontally splicing the first eigenvector, the second eigenvector, and the third eigenvector, a second target eigenvector (400 dimensions) having the sum of the first dimensional information (100 dimensions), the second dimensional information (78 dimensions), and the third dimensional information (dimensions) is obtained.

[0044] Furthermore, among the cut image blocks, the top 400 image blocks with the largest number of cell nuclei are selected to represent a digital pathology section / a patient, that is, a 400×400 matrix is ​​formed, in which the first 400 represents the number of target image blocks, and the second 400 represents the dimension of the second target feature vector, that is, the first dimension information. For each column of this feature matrix, the five statistics of mean, median, standard deviation, kurtosis, and skewness are calculated, and then 5 feature vectors are obtained, which are the second feature vectors with the second dimension information (5 dimensions). These 5 second feature vectors are horizontally spliced, that is, 5×400, and finally a 2000-dimensional feature vector is obtained to represent a digital pathology section / a patient.

[0045] In step 120, the first target feature vector of each of the hepatocellular carcinoma patients is screened based on a cross-validation method and an Akaike information criterion to obtain a target feature.

[0046] Specifically, the first target feature vector obtained by the above-mentioned feature splicing has a feature vector of 2000 dimensions and has an overfitting problem. Therefore, in this step 120, the first target feature vector of each hepatocellular carcinoma patient is feature screened by cross-validation method and Akaike information criterion to obtain the target feature, and the most representative target feature is selected as the training sample of the prognosis model, so as to make the training of the prognosis model more efficient and accurate.

[0047] In one embodiment of the present application, the first target feature vector of each hepatocellular carcinoma patient is subjected to feature screening based on the cross-validation method and the Akaike information criterion to obtain the target feature, including: Performing a preset round of 10-fold cross validation on the first target feature vector of each of the hepatocellular carcinoma patients, and determining the target screening feature obtained by screening in each round using a two-way selection stepwise regression method based on the Akaike information criterion, wherein each round corresponds to an Akaike information criterion value; A target round with the smallest Akaike information criterion value is selected in each round, and the target screening feature screened by the target round is obtained as the target feature.

[0048] Specifically, the first target feature vector of each hepatocellular carcinoma patient is used as a discovery cohort, that is, there are multiple first target feature vectors in this discovery cohort, and then a feature screening operation based on the cross-validation method and Akaike information criterion is performed on the discovery cohort.

[0049] The purpose of the embodiment of the present application is to select the target feature with the best prognostic efficacy, that is, the Top feature, and the target feature can be one or more.

[0050] First, the Z-scores standardization operation is performed on each eigenvalue of the first target feature vector; secondly, a LASSO Cox regression model is constructed on the discovery cohort, and 500 rounds (preset rounds) of 10-fold cross validation are run, and the top features can be added and subtracted in the cross validation; after each round of 10-fold cross validation, further feature screening is performed through two-way selection stepwise regression based on the Akaike Information Criterion (AIC), and finally the top feature with the smallest AIC value (Akaike Information Criterion value) in 500 rounds (preset rounds) is selected to construct a multidimensional fusion nuclear feature prognostic model, that is, the trained prognostic model described in the embodiment of the present application.

[0051] In step 130, the target feature is used as a training sample of a preset prognostic model to train the prognostic model.

[0052] Specifically, by using the target features as training samples of a preset prognostic model, the trained prognostic model can more accurately and quickly evaluate the prognostic performance of target hepatocellular carcinoma patients.

[0053] In step 140, the histopathological section data of the target hepatocellular carcinoma patient is obtained, and the histopathological section data is input into the trained prognostic model to obtain the prognostic evaluation score of the target hepatocellular carcinoma patient.

[0054] Specifically, the weighted linear combination of the top features and their corresponding regression coefficients is used as the patient's risk score (S), and the calculation formula is as follows:

[0055] in xj Represents the feature value of the jth Top feature (i.e., one of the target features), j represents the regression coefficient corresponding to the jth Top feature, and N represents the number of Top features (target features). In the discovery cohort, we determined the median of the risk score as the optimal cutoff threshold for dividing patients into high-risk and low-risk. Patients with risk scores less than or equal to this threshold are classified as low-risk, and vice versa. The optimal stage threshold determined from the discovery cohort is then applied to the external validation set to achieve high- and low-risk classification of patients.

[0056] Therefore, the present application can obtain the tissue pathology section data of the target hepatocellular carcinoma patient, and quickly find the characteristic value corresponding to the target feature, input it into the above formula, and obtain the corresponding prognostic evaluation score, and evaluate the risk coefficient corresponding to the score based on the above optimal cutoff threshold.

[0057] In order to verify whether the combination of the prognostic model proposed in this application and clinical pathological variables can significantly enhance the prognostic stratification, we performed the following operations: first, a total of 5 clinical pathological variables including BCLC (Barcelona) staging, tumor grade (Differentiation), MVI, age, and gender were included; secondly, the predictive factors with p < 0.05 among the 5 clinical pathological variables were screened out through univariate analysis; finally, the model developed in the present invention was subjected to Cox multivariate analysis together with the predictive factors with p < 0.05, and the prognostic factors of the model with the smallest final AIC value were determined by using AIC-based two-way stepwise regression to determine the full model formed by the combination of the prognostic model proposed in the embodiment of the present application and clinical pathological variables.

[0058] Figure 3 This is a block diagram of a prognosis assessment device 300 for hepatocellular carcinoma patients according to an embodiment of the present application. According to the prognosis assessment device 300 for hepatocellular carcinoma patients according to an embodiment of the present application, the device 300 includes: a first acquisition unit 301, a screening unit 302, a training unit 303, and a second acquisition unit 304.

[0059] A first acquisition unit 301 is used to acquire a first target feature vector of at least one hepatocellular carcinoma patient; A screening unit 302 is used to perform feature screening on the first target feature vector of each of the hepatocellular carcinoma patients based on a cross-validation method and an Akaike information criterion to obtain a target feature; A training unit 303, configured to use the target feature as a training sample of a preset prognosis model to train the prognosis model; The second acquisition unit 304 is used to acquire the tissue pathology section data of the target hepatocellular carcinoma patient, and input the tissue pathology section data into the trained prognostic model to obtain the prognostic evaluation score of the target hepatocellular carcinoma patient; The first target feature vector of a single hepatocellular carcinoma patient can be obtained by the following steps: Obtaining a digital pathological section of the hepatocellular carcinoma patient, wherein the digital pathological section is a tissue pathological section of the hepatocellular carcinoma patient stained with hematoxylin and eosin; Performing image block cutting on the digital pathological sections to obtain a first preset number of image blocks; Performing cell nucleus segmentation on each of the image blocks respectively to obtain cell nucleus masks corresponding to each of the image blocks respectively, wherein the cell nucleus masks are image data having features of all cell nuclei in the image blocks; Perform feature extraction based on each of the cell nucleus masks to obtain second target feature vectors each having a cell nucleus shape feature, a cell nucleus texture feature, and a cell nucleus interaction feature, and each of the image blocks has one second target feature vector; A second target feature vector of the target image block is obtained to generate a feature matrix corresponding to the target image block and the second target feature vector, and feature concatenation is performed on the feature matrix to obtain the first target feature vector.

[0060] As another aspect, the present application further provides a computer-readable storage medium on which a program product capable of implementing the method provided above in this specification is stored. In some possible implementations, various aspects of the present application may also be implemented in the form of a program product, which includes a program code, and when the program product is run on a terminal device, the program code is used to cause the terminal device to execute the steps according to various exemplary implementations of the present application described in the above "Embodiment Method" section of this specification.

[0061] According to the program product for implementing the above method in the embodiment of the present application, it can adopt a portable compact disk read-only memory (CD-ROM) and include program code, and can be run on a terminal device, such as a personal computer. However, the program product of the present application is not limited thereto. In this document, a readable storage medium can be any tangible medium containing or storing a program, which can be used by or in combination with an instruction execution system, an apparatus or a device.

[0062] The program product may be any combination of one or more readable media. The readable medium may be a readable signal medium or a readable storage medium. The readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination thereof. More specific examples (non-exhaustive list) of readable storage media include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof.

[0063] Computer readable signal media may include data signals propagated in baseband or as part of a carrier wave, in which readable program code is carried. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Readable signal media may also be any readable medium other than a readable storage medium, which may send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device.

[0064] The program code embodied on the readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wired, optical cable, RF, etc., or any suitable combination of the foregoing.

[0065] Program code for performing the operations of the present application may be written in any combination of one or more programming languages, including object-oriented programming languages, such as Java, C++, etc., and conventional procedural programming languages, such as "C" or similar programming languages. The program code may be executed entirely on the user computing device, partially on the user device, as a separate software package, partially on the user computing device and partially on a remote computing device, or entirely on a remote computing device or server. In the case of a remote computing device, the remote computing device may be connected to the user computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computing device (e.g., using an Internet service provider to connect through the Internet).

[0066] As another aspect, the present application also provides an electronic device capable of implementing the above method.

[0067] Those skilled in the art will appreciate that various aspects of the present application may be implemented as a system, method or program product. Therefore, various aspects of the present application may be specifically implemented in the following forms, namely: a complete hardware implementation, a complete software implementation (including firmware, microcode, etc.), or a combination of hardware and software, which may be collectively referred to as "circuit", "module" or "system" herein.

[0068] Refer to the following Figure 4 The electronic device 400 according to this embodiment of the present application is described. Figure 4 The electronic device 400 shown is merely an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0069] like Figure 4As shown, the electronic device 400 is in the form of a general computing device. The components of the electronic device 400 may include but are not limited to: at least one processing unit 410, at least one storage unit 420, and a bus 430 connecting different system components (including the storage unit 420 and the processing unit 410).

[0070] The storage unit stores program codes, which can be executed by the processing unit 410, so that the processing unit 410 executes the steps described in the above “Example Method” section of this specification according to various exemplary implementations of the present application.

[0071] The storage unit 420 may include a readable medium in the form of a volatile storage unit, such as a random access memory unit (RAM) 421 and / or a cache memory unit 422 , and may further include a read-only memory unit (ROM) 423 .

[0072] The storage unit 420 may also include a program / utility 424 having a set (at least one) of program modules 425, such program modules 425 including but not limited to: an operating system, one or more application programs, other program modules, and program data, each of which or some combination may include an implementation of a network environment.

[0073] Bus 430 may represent one or more of several types of bus structures, including a memory unit bus or memory unit controller node, a peripheral bus, an accelerated graphics port, a processing unit, or a local bus using any of a variety of bus architectures.

[0074] The electronic device 400 may also communicate with one or more external devices 1200 (e.g., keyboards, pointing devices, Bluetooth devices, etc.), may also communicate with one or more devices that enable a user to interact with the electronic device 400, and / or communicate with any device that enables the electronic device 400 to communicate with one or more other computing devices (e.g., routers, modems, etc.). This communication may be performed through an input / output (I / O) interface 450. In addition, the electronic device 400 may also communicate with one or more networks (e.g., local area networks (LANs), wide area networks (WANs), and / or public networks, such as the Internet) through a network adapter 460. As shown, the network adapter 460 communicates with other modules of the electronic device 400 through a bus 430. It should be understood that, although not shown in the figure, other hardware and / or software modules may be used in conjunction with the electronic device 400, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems, etc.

[0075] Through the description of the above implementation methods, it is easy for those skilled in the art to understand that the example implementation methods described here can be implemented by software, or by combining software with necessary hardware. Therefore, the technical solution according to the implementation method of the present application can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.) or on a network, including instructions to enable a computing device (which can be a personal computer, a server, a terminal device, or a network device, etc.) to execute the method according to the implementation method of the present application.

[0076] In addition, the above-mentioned figures are only schematic illustrations of the processes included in the method according to the exemplary embodiments of the present application, and are not intended to be limiting. It is easy to understand that the processes shown in the above-mentioned figures do not indicate or limit the time sequence of these processes. In addition, it is also easy to understand that these processes can be performed synchronously or asynchronously, for example, in multiple modules.

[0077] It should be understood that the present application is not limited to the precise structures that have been described above and shown in the drawings, and that various modifications and changes may be performed without departing from the scope thereof. The scope of the present application is limited only by the appended claims.

Claims

1. A method for evaluating the prognosis of a patient with hepatocellular carcinoma, characterized in that: The method comprises: Obtaining a first target feature vector of at least one hepatocellular carcinoma patient; Performing feature screening on the first target feature vector of each of the hepatocellular carcinoma patients based on a cross-validation method and Akaike information criterion to obtain a target feature; Using the target feature as a training sample of a preset prognostic model to train the prognostic model; Acquiring tissue pathology section data of a target hepatocellular carcinoma patient, and inputting the tissue pathology section data into the trained prognostic model to obtain a prognostic assessment score of the target hepatocellular carcinoma patient; The first target feature vector of a single hepatocellular carcinoma patient can be obtained by the following steps: Obtaining a digital pathological section of the hepatocellular carcinoma patient, wherein the digital pathological section is a tissue pathological section of the hepatocellular carcinoma patient stained with hematoxylin and eosin; Performing image block cutting on the digital pathological sections to obtain a first preset number of image blocks; Performing cell nucleus segmentation on each of the image blocks respectively to obtain cell nucleus masks corresponding to each of the image blocks respectively, wherein the cell nucleus masks are image data having features of all cell nuclei in the image blocks; Perform feature extraction based on each of the cell nucleus masks to obtain second target feature vectors each having a cell nucleus shape feature, a cell nucleus texture feature, and a cell nucleus interaction feature, and each of the image blocks has one second target feature vector; A second target feature vector of the target image block is obtained to generate a feature matrix corresponding to the target image block and the second target feature vector, and feature concatenation is performed on the feature matrix to obtain the first target feature vector.

2. The method for evaluating the prognosis of a patient with hepatocellular carcinoma according to claim 1, characterized in that: A single second target feature vector can be obtained by the following steps: Extracting shape features of each cell nucleus in the image block according to the cell nucleus mask to obtain a first feature vector having first dimension information, wherein the first feature vector is used to characterize shape features of the cell nucleus in the image block; Extracting texture features of each cell nucleus in the image block according to the cell nucleus mask to obtain a second feature vector having second dimension information, wherein the second feature vector is used to characterize the cell nucleus texture features of the image block; Extracting the interaction features of each cell nucleus in the image block according to the cell nucleus mask to obtain a third feature vector having third dimension information, wherein the third feature vector is used to characterize the cell nucleus interaction features of the image block; The first feature vector, the second feature vector, and the third feature vector are horizontally concatenated to obtain the second target feature vector having the sum of the first dimensional information, the second dimensional information, and the third dimensional information.

3. The method for evaluating the prognosis of a patient with hepatocellular carcinoma according to claim 2, characterized in that: The first feature vector having the first dimension information can be obtained by the following steps: Extracting a first target parameter feature of each of the cell nuclei according to the cell nucleus mask, where the first target parameter feature is a feature vector having first initial dimension information; Performing a first preset statistical operation on each of the first target parameter features to obtain a second target parameter feature of each of the cell nuclei, wherein the second target parameter feature is a feature vector having second initial dimension information; Performing feature transverse splicing of the first target parameter feature and the second target parameter feature to obtain a first feature vector having the first dimensional information; The first dimension information is the product of the first initial dimension information and the second initial dimension information.

4. The method for evaluating the prognosis of a patient with hepatocellular carcinoma according to claim 3, characterized in that: The second feature vector having the second dimension information can be obtained by the following steps: Extracting a third target parameter feature of each of the cell nuclei according to the cell nucleus mask, wherein the third target parameter feature is a feature vector having third initial dimension information; Performing a second preset statistical operation on each of the third target parameter features to obtain a fourth target parameter feature of each of the cell nuclei, wherein the fourth target parameter feature is a feature vector having fourth initial dimension information; Performing feature transverse splicing of the third target parameter feature and the fourth target parameter feature to obtain a second feature vector having the second dimension information; The second dimensional information is the product of the third initial dimensional information and the fourth initial dimensional information.

5. The method for evaluating the prognosis of a patient with hepatocellular carcinoma according to claim 4, characterized in that: The third eigenvector having the third dimension information can be obtained by the following steps: Extracting a fifth target parameter feature of each of the cell nuclei according to the cell nucleus mask, wherein the fifth target parameter feature is a feature vector having fifth initial dimension information; Performing a third preset statistical operation on each of the fifth target parameter features to obtain a sixth target parameter feature of each of the cell nuclei, wherein the sixth target parameter feature is a feature vector having sixth initial dimension information; Performing feature horizontal splicing of the fifth target parameter feature and the sixth target parameter feature to obtain a third feature vector having the third dimension information; The third dimensional information is obtained according to the fifth initial dimensional information and the sixth initial dimensional information.

6. The method for evaluating the prognosis of a patient with hepatocellular carcinoma according to claim 5, characterized in that: The second target feature vector is a feature vector having the sum of the first dimension information, the second dimension information and the third dimension information, and the step of performing feature concatenation on the feature matrix to obtain the first target feature vector includes: Performing a third preset statistical operation on each of the second target feature vectors in the feature matrix to obtain a third target feature vector having fourth dimension information; The second target feature vector is concatenated with the third target feature vector to obtain the first target feature vector.

7. The method for evaluating the prognosis of a patient with hepatocellular carcinoma according to claim 6, characterized in that: The method of performing feature screening on the first target feature vector of each of the hepatocellular carcinoma patients based on the cross-validation method and the Akaike information criterion to obtain the target feature includes: Performing a preset round of 10-fold cross validation on the first target feature vector of each of the hepatocellular carcinoma patients, and determining the target screening feature obtained by screening in each round using a two-way selection stepwise regression method based on the Akaike information criterion, wherein each round corresponds to an Akaike information criterion value; A target round with the smallest Akaike information criterion value is selected in each round, and the target screening feature screened by the target round is obtained as the target feature.

8. A prognosis assessment device for patients with hepatocellular carcinoma, characterized in that: The device comprises: A first acquisition unit, configured to acquire a first target feature vector of at least one hepatocellular carcinoma patient; A screening unit, configured to perform feature screening on the first target feature vector of each of the hepatocellular carcinoma patients based on a cross-validation method and an Akaike information criterion to obtain a target feature; A training unit, used for taking the target feature as a training sample of a preset prognosis model to train the prognosis model; a second acquisition unit, configured to acquire the histopathological section data of the target hepatocellular carcinoma patient, and input the histopathological section data into the trained prognostic model to obtain the prognostic evaluation score of the target hepatocellular carcinoma patient; The first target feature vector of a single hepatocellular carcinoma patient can be obtained by the following steps: Obtaining a digital pathological section of the hepatocellular carcinoma patient, wherein the digital pathological section is a tissue pathological section of the hepatocellular carcinoma patient stained with hematoxylin and eosin; Performing image block cutting on the digital pathological sections to obtain a first preset number of image blocks; Performing cell nucleus segmentation on each of the image blocks respectively to obtain cell nucleus masks corresponding to each of the image blocks respectively, wherein the cell nucleus masks are image data having features of all cell nuclei in the image blocks; Perform feature extraction based on each of the cell nucleus masks to obtain second target feature vectors each having a cell nucleus shape feature, a cell nucleus texture feature, and a cell nucleus interaction feature, and each of the image blocks has one second target feature vector; A second target feature vector of the target image block is obtained to generate a feature matrix corresponding to the target image block and the second target feature vector, and feature concatenation is performed on the feature matrix to obtain the first target feature vector.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores at least one program code, and the at least one program code is loaded and executed by a processor to implement the operations performed by the method according to any one of claims 1 to 7.

10. An electronic device, characterized in that: The electronic device includes one or more processors and one or more memories, wherein the one or more memories store at least one program code, and the at least one program code is loaded and executed by the one or more processors to implement the operations performed by the method according to any one of claims 1 to 7.

Images