Hepatocellular carcinoma prognosis evaluation method and system based on cell pixel density

Abstract

The invention discloses a hepatocellular carcinoma prognosis evaluation method and system based on cell pixel density, and the method comprises the steps: obtaining a full-slide image, carrying out the image preprocessing, and determining a critical value of a cell pixel-level segmentation result; performing cell pixel density calculation according to a critical value of a cell pixel-level segmentation result, and determining CD8 + TILs density of the tumor region; carrying out standardization and average calculation processing on the CD8 + TILs density of the tumor area to obtain a critical value of ATLS-8; and performing prognosis evaluation on the hepatocellular carcinoma of the patient according to the critical value of the ATLS-8. By using the method, the cell density on the whole slide can be accurately quantified, so that the accuracy of cell density measurement is improved, and the accuracy of prognosis evaluation of hepatocellular carcinoma of a patient is improved. The hepatocellular carcinoma prognosis evaluation method and system based on the cell pixel density can be widely applied to the technical field of hepatocellular carcinoma prognosis evaluation.

Description

Technical Field

[0001] The present invention relates to the technical field of hepatocellular carcinoma prognosis assessment, and in particular to a hepatocellular carcinoma prognosis assessment method and system based on cell pixel density. Background Art

[0002] In view of the complexity of tumor heterogeneity, relevant technologies are committed to identifying new prognostic parameters and indicators as comprehensive biomarkers for a wider range of HCC patients. At present, a variety of biomarkers have been discovered and used for the biological characteristics of tumors and prognostic assessment. The tumor microenvironment (TME) is composed of tumor cells, stromal cells, immune cells, and epithelial tissue, and is an important component affecting tumor survival and progression. Tumor heterogeneity plays a vital role in limiting the effectiveness of a wide range of drug and radiotherapy for cancer patients, and is closely related to TME. TILs are widely regarded as immune cells associated with the prognosis of patients with solid tumors. This includes subtypes such as CD3, CD8, and CD45 lymphocytes, among which CD8+T cells have been confirmed to be key members in eliminating tumor cells. The effectiveness and versatility of using the proportion of T lymphocytes in the tumor core and infiltrated edge as the main evaluation index have been verified in various cancer types. Various TIL-centered scoring methods have been widely studied and used for prognostic assessment in a variety of tumor types. However, there is no broad consensus on liver cancer. In previous studies, some pathologists reported using traditional pathological methods to manually score the immune response of tissues and tumor areas to assess prognosis. These methods have limitations and subjectivity, resulting in conflicting results from different studies. Due to the large number of cells in WSI, traditional pathological quantification methods can hardly accurately quantify the cell density on the entire slide. Traditional methods often have biases in the selection of regions of interest (ROI) and rely heavily on the subjective expertise of pathologists, which affects the accuracy of cell density measurement and limits it to qualitative or semi-quantitative evaluation. Summary of the invention

[0003] In order to solve the above technical problems, the purpose of the present invention is to provide a hepatocellular carcinoma prognosis assessment method and system based on cell pixel density, which can accurately quantify the cell density on the entire slide, thereby improving the accuracy of cell density measurement and improving the accuracy of prognosis assessment of hepatocellular carcinoma in patients.

[0004] The first technical solution adopted by the present invention is: a method for evaluating the prognosis of hepatocellular carcinoma based on cell pixel density, comprising the following steps:

[0005] Acquire whole slide images and perform image preprocessing to determine the critical value of cell pixel-level segmentation results;

[0006] According to the critical value of the cell pixel-level segmentation result, the cell pixel density is calculated to determine the CD8+TILs density in the tumor area;

[0007] The density of CD8+TILs in the tumor area was standardized and averaged to obtain the critical value of ATLS-8;

[0008] The prognosis of patients with hepatocellular carcinoma was evaluated according to the critical value of ATLS-8.

[0009] Furthermore, the step of acquiring whole slide imaging and performing image preprocessing to determine the critical value of the cell pixel-level segmentation result specifically includes:

[0010] Perform image preprocessing on the whole slide imaging to obtain a mask image of the tumor area;

[0011] Positive cells were identified and segmented on the whole slide imaging and the tumor area mask image, and the critical value of the cell pixel-level segmentation result was determined.

[0012] Furthermore, the step of performing image preprocessing on the whole slide imaging to obtain a tumor region mask image specifically includes:

[0013] Acquire whole slide imaging and perform staining to obtain H&E stained section images;

[0014] Based on the quality control instrument, non-overlapping plaques were extracted at 10 times magnification of the H&E stained slice image and plaques smaller than the preset threshold were eliminated to obtain the image to be segmented in the tumor area;

[0015] The image to be segmented in the tumor area is segmented using the nnU-Net deep neural network to obtain a mask image of the tumor area.

[0016] Furthermore, the step of identifying and segmenting positive cells in the whole slide imaging and the tumor region mask image and determining a critical value of the cell pixel-level segmentation result specifically includes:

[0017] The whole slide images with different staining of the same tissue are registered according to the H&E stained slice images to obtain the aligned whole slide images;

[0018] Import the aligned whole slide image into the Qupath platform for observation and export the tiled image;

[0019] The tiled image and the tumor area mask image were registered and pixel-segmented using the random forest algorithm to determine the critical value of the cell pixel-level segmentation result.

[0020] Furthermore, the step of calculating the cell pixel density according to the critical value of the cell pixel-level segmentation result to determine the CD8+TILs density in the tumor area specifically includes:

[0021] According to the critical value of the cell pixel-level segmentation result, the area of ​​the tumor region and the area of ​​CD8+ pixels were determined;

[0022] The area ratio of CD8+ cells in the tumor tissue was determined based on the area of ​​the tumor region and the area of ​​CD8+ pixels;

[0023] The area ratio of CD8+ cells in tumor tissue was unit-normalized, and the pixel density was converted into the number of cells per unit area to obtain the CD8+TILs density in the tumor area.

[0024] Furthermore, the step of normalizing and averaging the density of CD8+TILs in the tumor area to obtain the critical value of ATLS-8 specifically includes:

[0025] The density of CD8+TILs in the tumor area was normalized by the normalization factor to obtain the ratio of normalized patch area to unit area;

[0026] The non-normalized cell density was obtained and multiplied with the ratio of the normalized patch area to the unit area to obtain the cell density per unit area of ​​the patch;

[0027] The cell density per unit area of ​​the patch was averaged to obtain the critical value of ATLS-8.

[0028] The second technical solution adopted by the present invention is: a hepatocellular carcinoma prognosis assessment system based on cell pixel density, comprising:

[0029] The first module is used to obtain whole slide imaging and perform image preprocessing to determine the critical value of cell pixel-level segmentation results;

[0030] The second module is used to calculate the cell pixel density based on the critical value of the cell pixel-level segmentation result and determine the CD8+TILs density in the tumor area;

[0031] The third module is used to standardize and average the CD8+TILs density in the tumor area to obtain the critical value of ATLS-8;

[0032] The fourth module is used to evaluate the prognosis of patients with hepatocellular carcinoma based on the critical value of ATLS-8.

[0033] The beneficial effects of the method and system of the present invention are as follows: the present invention determines the CD8+TILs density of the tumor area by determining the critical value of the cell pixel-level segmentation result, standardizes and averages the CD8+TILs density of the tumor area to obtain the critical value of ATLS-8, and finally performs a prognosis assessment of the patient's hepatocellular carcinoma based on the critical value of ATLS-8. The cell density on the entire slide can be accurately quantified, thereby improving the accuracy of cell density measurement and improving the accuracy of the prognosis assessment of the patient's hepatocellular carcinoma. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 It is a flowchart of the steps of a method for evaluating the prognosis of hepatocellular carcinoma based on cell pixel density of the present invention;

[0035] Figure 2 It is a structural block diagram of a hepatocellular carcinoma prognosis evaluation system based on cell pixel density of the present invention;

[0036] Figure 3 It is a schematic diagram of the results of Bland-Altman consistency correlation analysis of 40 plaques selected according to a specific embodiment of the present invention;

[0037] Figure 4 It is a schematic diagram of the Kaplan-Meier survival curve analysis of the discovery cohort provided by a specific embodiment of the present invention;

[0038] Figure 5 It is a schematic diagram of Kaplan-Meier survival curve analysis of validation cohort 1 provided in a specific embodiment of the present invention;

[0039] Figure 6 It is a schematic diagram of the Kaplan-Meier survival curve analysis of the validation cohort 2 provided in a specific embodiment of the present invention. DETAILED DESCRIPTION

[0040] The present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. The step numbers in the following embodiments are only provided for the convenience of explanation and description, and the order between the steps is not limited in any way. The execution order of each step in the embodiment can be adaptively adjusted according to the understanding of those skilled in the art.

[0041] The present invention introduces a new fully automated process that uses an artificial intelligence algorithm to quantify CD8 tumor-infiltrating lymphocytes, or ATLS. In addition, it also uses a CD8+ lymphocyte-based scoring system (called ATLS-8) to assess the prognosis of HCC patients.

[0042] Reference Figure 1 The present invention provides a method for evaluating the prognosis of hepatocellular carcinoma based on cell pixel density, the method comprising the following steps:

[0043] S100, acquiring whole slide imaging and performing image preprocessing to determine a critical value of a cell pixel-level segmentation result;

[0044] S110, performing image preprocessing on the whole slide imaging to obtain a mask image of the tumor region;

[0045] Specifically, whole slide imaging is acquired and stained to obtain H&E stained slice images; based on the quality control instrument, non-overlapping plaques are extracted at a 10-fold magnification of the H&E stained slice images and plaques smaller than a preset threshold are eliminated to obtain images of the tumor area to be segmented; the images of the tumor area to be segmented are segmented using the nnU-Net deep neural network to obtain mask images of the tumor area.

[0046] In this embodiment, in order to remove unnecessary white background, artifacts and other types of noise in WSI, professional pathologists used a quality control instrument to identify usable ROIs in WSI. In order to obtain more accurate tumor segmentation results and higher efficiency, non-overlapping patches (1000x 1000 pixels) were extracted at 10x magnification. Plaques with a tissue ratio less than 0.05 will be discarded. Next, the tumor area is segmented using a pre-trained nnU-Net. Finally, pathologists were invited to assist in refining the tumor mask, which was obtained by aggregating patch-level masks.

[0047] S120, identifying and segmenting positive cells on the whole slide imaging and the tumor region mask image, and determining a critical value of the cell pixel-level segmentation result.

[0048] Specifically, whole slide images with different staining of the same tissue were registered based on the H&E stained section images to obtain aligned whole slide images; the aligned whole slide images were imported into the Qupath platform for observation and tiled images were exported; the tiled images and tumor area mask images were registered and pixel-segmented using the random forest algorithm to determine the critical value of the cell pixel-level segmentation result.

[0049] In this example, a combination of rigid registration and two non-rigid registrations was used to align adjacent tissue WSIs stained with H&E and IHC. This approach ensures that the two slices are as similar as possible in spatial position and morphology. In this step, all WSI images of the same tissue but with different staining are registered based on the H&E staining image as a reference for multiple IHC staining registrations. The aligned images are then imported into the Qupath platform to observe the registration effect. A tiled image of 1024x1024 pixels is generated and exported. Subsequently, the mask is registered with the calibrated IHC slice using a Python program. The tumor and non-tumor areas are segmented into 1024x1024 pixels, and blocks with more than 50% tissue in the mask can be used for further analysis.

[0050] The random forest algorithm model was further used to perform cell segmentation on the tiles. Initially, the pathologists made a small number of precise annotations to observe the segmentation results. Then, based on these observations, the model was trained on a larger scale. The probability generated for each pixel represents the possibility of that pixel being a target cell pixel. The critical value for determining the cell pixel-level segmentation result was finally determined to be P = 0.25.

[0051] S200, calculating the cell pixel density according to the critical value of the cell pixel-level segmentation result to determine the CD8+TILs density in the tumor area;

[0052] S210, determining the area of ​​the tumor region and the area of ​​CD8+ pixels according to the critical value of the cell pixel-level segmentation result;

[0053] S220, determining the area ratio of CD8+ cells in the tumor tissue according to the area of ​​the tumor region and the area of ​​CD8+ pixels;

[0054] In this embodiment, after the cell segmentation is completed, the flat film is divided into tumor areas according to a predetermined strategy. The pixel density of each segment is calculated by determining the area occupied by the pixels marked as CD8+ in the entire segment. This measurement method provides the CD8+ cell pixel density at the WSI level, which represents the area ratio of CD8+ cells in the tumor tissue, and its expression is:

[0055]

[0056] Among them, Pixel CD8 label counting patch It refers to the use of random forest algorithm to segment pixels at the patch level. Therefore, Pixel total label count patchAll pixels within the plaque are included, including CD8+ cells (1) and non-CD8+ cells (0). In this way, the density of CD8+ cells in the tumor area can be calculated at the WSI level. The calculation formula is outlined as follows:

[0057]

[0058] Here, n represents the number of available tumor area patches.

[0059] S230. The area ratio of CD8+ cells in the tumor tissue is normalized, and the pixel density is converted into the number of cells per unit area to obtain the CD8+TILs density in the tumor area.

[0060] In this example, pixel density was converted to number of cells per unit area to standardize with the units used in most studies. This involved measuring the number of cells in each sample based on the cell segmentation marker images. The number of cells / mm was determined by calculating the area fraction and pixel size occupied by each cell. 2 This gives the density of CD8+TILs in the tumor area. The calculation formula is as follows:

[0061]

[0062] In the above formula, Pstch Cell density unnormalised Represents the density of CD8+TILs in the tumor area.

[0063] S300, standardize and average the density of CD8+TILs in the tumor area to obtain the critical value of ATLS-8;

[0064] S310, normalizing the CD8+TILs density in the tumor area by a normalization factor to obtain the ratio of the normalized patch area to the unit area;

[0065] S320, obtaining a non-normalized cell density and multiplying it with the ratio of the normalized patch area to the unit area to obtain a cell density per unit area of ​​the patch;

[0066] S330, performing average calculation on the cell density per unit area of ​​the patch to obtain the critical value of ATLS-8.

[0067] In this example, after calculating the cell density of each plaque, we use the normalization factor to normalize the pixel density of the three centers to reflect the cell size. The calculation formula is as follows

[0068]

[0069] Among them, Fmag represents the actual area of ​​each patch, F nor Represents the ratio of normalized patch area to unit area (unit: mm 2 ). In this experiment, Patch height Height and Patch weight The weight is 1024 pixels. Then, F nor Multiplying by the non-normalized cell density gives the unit area (mm) of each patch. 2 ) The calculation formula is as follows

[0070] Cell density patch-normalised =Cell density patch-unnormalised *F nor

[0071] Finally, the CD8+ cell density of all plaques was averaged to reflect the CD8+ cell density within the tumor area on WSI. This is called ATLS-8. The calculation formula is as follows

[0072]

[0073] Here, n represents the number of available tumor area patches.

[0074] S400. Prognostic evaluation of patients with hepatocellular carcinoma was performed based on the critical value of ATLS-8.

[0075] In summary, the embodiments of the present invention scan H&E stained sections using a whole slide scanner, apply quality control measures, and use deep learning methods to segment images to obtain tumor area masks annotated by professional pathologists. Adjacent CD8 IHC stained sections are registered with H&E images, the registered images are segmented into blocks, and annotated cells are identified and segmented using a machine learning algorithm, the pixel density of positive cells is calculated and converted into cell density, the ATLS-8 critical value is determined according to the maximum selection rank statistics, the patients are divided into a low index group and a high index group, and the Kaplan-Meier (KM) curve is drawn to intuitively display the survival differences between the groups, as shown in FIG. Figure 4 The prognostic value of ATLS-8 was evaluated in two independent external validation groups.

[0076] Finally, the cell segmentation results of the embodiment of the present invention are analyzed through simulation experiments. Figure 3As shown in Figure 2, a total of 40 plaques were randomly selected for cell segmentation. The segmentation results were compared and analyzed with the manual counting results performed by pathologists using traditional pathological counting methods, and the pathologists were blind to the results generated by the algorithm. The comparison results were analyzed using the Bland-Altman method. The agreement between manual counting and automatic counting using our algorithm was good, with an ICC of 0.98 (95% CI, 0.97-0.99; P < 0.001).

[0077] The prognostic value of ATLS-8 analysis was further analyzed. The cutoff value of ATLS-8 was determined by the maximum selection rank statistical method, and the cutoff value of the tumor area was 67.25 cells / mm 2 In the discovery cohort, 90 patients (72.58%) had ATLS-8 values ​​above the cutoff and were classified as low risk, and 34 patients (27.42%) had ATLS-8 values ​​below the cutoff and were classified as high risk. The 5-year overall survival (OS) rate was 65.03% in the low-risk group and 34.05% in the high-risk group. Figure 5 As shown in Figure 1, in validation cohort 1, 74 patients (90.24%) had ATLS-8 values ​​above the cutoff and were classified as low risk, and 8 patients (9.76%) had ATLS-8 values ​​below the cutoff and were classified as high risk. The 5-year OS rate was 68.73% in the low-risk group and 28.57% in the high-risk group. Figure 6 As shown, in validation cohort 2, 27 patients (50.0%) had ATLS-8 values ​​above the cutoff and were classified as low risk, and 27 patients (50.0%) had ATLS-8 values ​​below the cutoff and were classified as high risk. The 5-year OS rate was 81.48% in the low-risk group and 59.26% in the high-risk group.

[0078] The KM curve showed that patients with lower ATLS-8 scores had worse OS compared with those with higher scores (discovery cohort, unadjusted HR, 2.23 (95% CI, 1.27-3.91); P = 0.0042; validation cohort 1, HR 3.38 (95% CI, 1.27-9.02) P = 0.0096; validation cohort 2, HR 2.74 (95% CI, 1.05-7.15); P = 0.031).

[0079] Table 1 Univariate Cox regression analysis results of the three cohorts

[0080]

[0081]

[0082] Table 2 Multivariate Cox regression analysis results of three cohorts

[0083]

[0084]

[0085] Table 1 lists the results of univariate Cox regression analysis of the three cohorts, and Table 2 lists the results of multivariate Cox regression analysis of the three cohorts. Factors that reached statistical significance (P < 0.05) in the univariate analysis of the discovery cohort, including age, HBV, tumor size, differentiation grade, MVI, BCLC stage, and ATLS-8, were included in the multivariate analysis to evaluate the robustness of ATLS-8 in the validation cohort. Data with unknown MVI status were excluded from the univariate and multivariate model analyses. Multivariate analysis showed that after adjusting for other clinicopathological factors, ATLS-8 score was independently associated with OS (discovery cohort, low HR vs. high HR, 2.015 (95% CI: 1.115-3.641); P = 0.0203; validation cohort 1, HR, 4.871 (95% CI: 1.753-13.534); P = 0.0024; validation cohort 2, 3.925 ((95% CI: 1.212-12.714; P = 0.0226). Finally, for the development and validation of the prognostic prediction model, in the discovery cohort, age, HBV, ATLS-8, BCLC stage, and HBV-related factors were significantly associated with OS. , tumor size, differentiation grade, and MVI were considered to be independent factors affecting OS, so we established a complete prognostic prediction model (age + HBV + ATLS-8 + BCLC stage + tumor size + differentiation grade + MVI) based on the above factors (so we call this model the complete model). As shown in Table 3, we further compared the performance of the complete model with the other four models, including the BCLC stage model, the ATLS-8 model, the BCLC stage and ATLS-8 combined model, and the clinicopathological model (age, HBV, BCLC stage, tumor size, differentiation grade, and MVI combined model).

[0086] Table 3 Performance indicators of ATLS-8 model and multi-reference model

[0087]

[0088]

[0089] In all three cohorts, the full model had superior discrimination (assessed by C-index) and calibration (assessed by AIC) compared with the clinicopathological model.

[0090] In the discovery cohort, the C index of the full model was 0.770, compared with 0.757 for the clinicopathological model; the AIC was 381.6, compared with 384.8 for the clinicopathological model. In validation cohort 1, the C index was 0.769 vs 0.727, and the AIC was 285.8 vs 291.3. In validation cohort 2, the C index was 0.712 vs 0.642, and the AIC was 152.7 vs 156.6. The BCLC grading and ATLS-8 model also showed better discrimination and calibration than the BCLC grading and ATLS-8 model.

[0091] In all three cohorts, the BCLC grading and ATLS-8 models also showed better discrimination and calibration than the BCLC grading model. In the discovery cohort, the C index was 0.715 vs 0.655, and the AIC was 389.4 vs 392.8. In the validation cohort 1, the C index was 0.619 vs 0.674, and the AIC was 292.6 vs 296. In the validation cohort 2, the C index was 0.643 vs 0.516, and the AIC was 146.2 vs 160.

[0092] Incorporating the ATLS-8 score into the BCLC staging model can improve the prediction of OS (likelihood ratio P = 0.0013). In addition, incorporating the ATLS-8 score into the clinicopathological model also improves the predictive ability of OS (likelihood ratio P = 0.0011).

[0093] Reference Figure 2 , a hepatocellular carcinoma prognosis assessment system based on cell pixel density, comprising:

[0094] The first module is used to obtain whole slide imaging and perform image preprocessing to determine the critical value of cell pixel-level segmentation results;

[0095] The second module is used to calculate the cell pixel density based on the critical value of the cell pixel-level segmentation result and determine the CD8+TILs density in the tumor area;

[0096] The third module is used to standardize and average the CD8+TILs density in the tumor area to obtain the critical value of ATLS-8;

[0097] The fourth module is used to evaluate the prognosis of patients with hepatocellular carcinoma based on the critical value of ATLS-8.

[0098] The contents of the above method embodiments are all applicable to the present system embodiments. The functions specifically implemented by the present system embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0099] The above is a specific description of the preferred implementation of the present invention, but the invention is not limited to the embodiments. Those skilled in the art may make various equivalent modifications or substitutions without violating the spirit of the present invention. These equivalent modifications or substitutions are all included in the scope defined by the claims of this application.

Claims

1. A method for evaluating the prognosis of hepatocellular carcinoma based on cell pixel density, characterized in that: The following steps are involved: Acquire whole slide images and perform image preprocessing to determine the critical value of cell pixel-level segmentation results; According to the critical value of the cell pixel-level segmentation result, the cell pixel density is calculated to determine the CD8+TILs density in the tumor area; The density of CD8+TILs in the tumor area was standardized and averaged to obtain the critical value of ATLS-8; The prognosis of patients with hepatocellular carcinoma was evaluated according to the critical value of ATLS-8.

2. A method for evaluating the prognosis of hepatocellular carcinoma based on cell pixel density according to claim 1, characterized in that: The step of acquiring whole slide imaging and performing image preprocessing to determine the critical value of the cell pixel-level segmentation result specifically includes: Perform image preprocessing on the whole slide imaging to obtain a mask image of the tumor area; Positive cells were identified and segmented on the whole slide imaging and the tumor area mask image, and the critical value of the cell pixel-level segmentation result was determined.

3. A method for evaluating the prognosis of hepatocellular carcinoma based on cell pixel density according to claim 2, characterized in that: The step of performing image preprocessing on the whole slide imaging to obtain a tumor region mask image specifically includes: Acquire whole slide imaging and perform staining to obtain H&E stained section images; Based on the quality control instrument, non-overlapping plaques were extracted at 10 times magnification of the H&E stained slice image and plaques smaller than the preset threshold were eliminated to obtain the image to be segmented in the tumor area; The image to be segmented in the tumor area is segmented using the nnU-Net deep neural network to obtain a mask image of the tumor area.

4. A method for evaluating the prognosis of hepatocellular carcinoma based on cell pixel density according to claim 3, characterized in that: The step of identifying and segmenting positive cells on the whole slide imaging and the tumor region mask image and determining the critical value of the cell pixel-level segmentation result specifically includes: The whole slide images with different staining of the same tissue are registered according to the H&E stained slice images to obtain the aligned whole slide images; Import the aligned whole slide image into the Qupath platform for observation and export the tiled image; The tiled image and the tumor area mask image were registered and pixel-segmented using the random forest algorithm to determine the critical value of the cell pixel-level segmentation result.

5. A method for evaluating the prognosis of hepatocellular carcinoma based on cell pixel density according to claim 4, characterized in that: The step of calculating the cell pixel density according to the critical value of the cell pixel level segmentation result to determine the CD8+TILs density in the tumor area specifically includes: According to the critical value of the cell pixel-level segmentation result, the area of ​​the tumor region and the area of ​​CD8+ pixels were determined; The area ratio of CD8+ cells in the tumor tissue was determined based on the area of ​​the tumor region and the area of ​​CD8+ pixels; The area ratio of CD8+ cells in tumor tissue was unit-normalized, and the pixel density was converted into the number of cells per unit area to obtain the CD8+TILs density in the tumor area.

6. A method for evaluating the prognosis of hepatocellular carcinoma based on cell pixel density according to claim 5, characterized in that: The step of normalizing and averaging the CD8+TILs density in the tumor area to obtain the critical value of ATLS-8 specifically includes: The CD8+TILs density in the tumor area was normalized by the normalization factor to obtain the ratio of normalized patch area to unit area; The non-normalized cell density was obtained and multiplied with the ratio of the normalized patch area to the unit area to obtain the cell density per unit area of ​​the patch; The cell density per unit area of ​​the patch was averaged to obtain the critical value of ATLS-8.

7. A hepatocellular carcinoma prognosis assessment system based on cell pixel density, characterized in that: Includes the following modules: The first module is used to obtain whole slide imaging and perform image preprocessing to determine the critical value of cell pixel-level segmentation results; The second module is used to calculate the cell pixel density based on the critical value of the cell pixel-level segmentation result and determine the CD8+TILs density in the tumor area; The third module is used to standardize and average the CD8+TILs density in the tumor area to obtain the critical value of ATLS-8; The fourth module is used to evaluate the prognosis of patients with hepatocellular carcinoma based on the critical value of ATLS-8.

Images