Pathological image-based oral mucosa canceration prediction method, equipment and program product

Abstract

The invention relates to the field of intelligent medical treatment, in particular to an oral mucosa canceration prediction method and device based on a pathological image and a program product. Comprising the following steps: S1, acquiring a pathological image of oral mucosa tissue of a to-be-detected person; s2, phenotypes of microvessels are obtained based on the pathological image, the phenotypes comprise an elliptic type and a non-elliptic type, when the phenotype is the non-elliptic type, ANGPTL2 gene detection is carried out to obtain a gene expression quantity, and when the gene expression quantity of ANGPTL2 is larger than a first preset threshold value, it is judged that a prediction result of oral mucosa canceration occurs, and if the prediction result of oral mucosa canceration occurs, the prediction result of oral mucosa canceration does not occur. The method can be used for oral mucosa canceration prediction and progress prediction, and has a very good clinical value.

Description

Technical Field

[0001] The present application relates to the field of intelligent medicine, and specifically relates to a method, device, program product, and computer-readable storage medium for predicting oral mucosal canceration based on pathological images. Background Art

[0002] The oral mucosal canceration process usually goes through three stages: normal oral mucosa, potentially malignant oral lesions, and oral cancer. The growth and metastasis of tumors are closely related to the angiogenesis of tumor microvessels. That is, the formation of microvessels is an important sign of tumor occurrence and development, which can promote genetic changes in normal cells and accelerate the transformation from the potentially malignant lesion stage to the cancer stage. Research shows that there are significant phenomena of neovascularization in oral mucosal diseases such as lichen planus and oral submucous fibrosis, which manifest as potentially malignant disorders. This suggests that the formation, characteristics, and differentiation of microvessels may run through the occurrence and development process of oral mucosal canceration and play a crucial role in regulating this process. Therefore, there is an urgent hope to establish an objective pathological diagnosis standard for predicting the oral mucosal canceration process, aiming to objectively and accurately predict oral mucosal canceration and each stage of the canceration process based on this standard, so as to guide clinical targeted vascular treatment methods to improve the prognosis and survival of patients. Summary of the Invention

[0003] In view of the above problems, the present invention provides a method for predicting oral mucosal canceration based on pathological images, specifically including: S1. Obtain the pathological image of the oral mucosal tissue of the person to be tested; S2. Obtain the phenotype of the microvessels based on the pathological image, and the phenotype includes elliptical and non-elliptical. When the phenotype is non-elliptical, perform ANGPTL2 gene detection to obtain the gene expression level. When the gene expression level of ANGPTL2 is greater than the first preset threshold, it is determined that the prediction result of oral mucosal canceration has occurred.

[0004] The first preset threshold is the gene expression level of ANGPTL2 in the normal state; Optionally, the ANGPTL2 gene is replaced by the CK15 gene. When the gene expression level of CK15 is greater than the second preset threshold, it is determined that the prediction result of oral mucosal canceration has occurred; Optionally, the second preset threshold is the gene expression level of CK15 in the normal state; Optionally, the ANGPTL2 gene is replaced by the EEF1E1 gene. When the gene expression level of EEF1E1 is greater than the third preset threshold, it is determined that the prediction result of oral mucosal canceration has occurred; Optionally, the third preset threshold is the gene expression level of EEF1E1 in the normal state.

[0005] Optionally, the gene detection further includes one or more of the following: CK15, EEF1E1.

[0006] The method further includes judging the progression of oral mucosal carcinogenesis, and the progression of oral mucosal carcinogenesis includes well-differentiated squamous cell carcinoma stage, moderately-differentiated squamous cell carcinoma stage, and poorly-differentiated squamous cell carcinoma stage; detecting the gene expression levels of any one or more of ANGPTL2, CK15, and EEF1E1 to obtain gene expression levels, and inputting them into the first judgment model to obtain the progression of oral mucosal carcinogenesis; Optionally, the first judgment model preset the thresholds of any one or more of ANGPTL2, CK15, and EEF1E1 in the normal stage, well / moderately / poorly-differentiated squamous cell carcinoma stage, and obtain the judgment result by comparing the gene expression level with the preset threshold; Optionally, when the gene expression level of ANGPTL2 is greater than the threshold in the normal stage, it is determined as the well-differentiated squamous cell carcinoma stage; when the gene expression level of ANGPTL2 is less than the threshold in the well-differentiated squamous cell carcinoma stage, it is determined as the moderately-differentiated squamous cell carcinoma stage or the poorly-differentiated squamous cell carcinoma stage; Optionally, when the gene expression level of CK15 is greater than the threshold in the normal stage, it is determined as the well-differentiated squamous cell carcinoma stage; when the gene expression level of CK15 is greater than the threshold in the well-differentiated squamous cell carcinoma stage, it is determined as the moderately-differentiated squamous cell carcinoma stage or the poorly-differentiated squamous cell carcinoma stage; Optionally, when the gene expression level of EEF1E1 is greater than the threshold in the normal stage, it is determined as the well-differentiated squamous cell carcinoma stage or the moderately-differentiated squamous cell carcinoma stage; when the gene expression level of EEF1E1 is less than the threshold in the moderately-differentiated squamous cell carcinoma stage, it is determined as the poorly-differentiated squamous cell carcinoma stage; Optionally, when the gene expression levels of ANGPTL2, CK15, and EEF1E1 are greater than the threshold in the normal stage, it is determined as the well-differentiated squamous cell carcinoma stage; when the gene expression level of ANGPTL2 is less than the threshold in the well-differentiated squamous cell carcinoma, the gene expression level of CK15 is greater than the threshold in the well-differentiated squamous cell carcinoma, and the gene expression level of EEF1E1 is greater than the threshold in the well-differentiated squamous cell carcinoma, it is determined as the moderately-differentiated squamous cell carcinoma stage; when the gene expression level of ANGPTL2 is less than the threshold in the moderately-differentiated squamous cell carcinoma, the gene expression level of CK15 is greater than the threshold in the moderately-differentiated squamous cell carcinoma, and the gene expression level of EEF1E1 is less than the threshold in the moderately-differentiated squamous cell carcinoma, it is determined as the poorly-differentiated squamous cell carcinoma stage.

[0007] The method further includes TSCs detection. When the phenotype is non-elliptical, TSCs detection and ANGPTL2 gene detection are performed. When TSCs are detected and the gene expression level of ANGPTL2 is less than the first preset threshold, it is determined as the prediction result of oral mucosal carcinogenesis; Optionally, the ANGPTL2 is replaced by the CK15 gene. When TSCs are detected and the gene expression level of CK15 is greater than the second preset threshold, it is determined that the prediction result is oral mucosal carcinogenesis. Optionally, the ANGPTL2 is replaced by the EEF1E1 gene. When TSCs are detected and the gene expression level of EEF1E1 is greater than the third preset threshold, it is determined that the prediction result is oral mucosal carcinogenesis.

[0008] The method further includes the judgment of the oral mucosal carcinogenesis process. The process of oral mucosal carcinogenesis includes the precancerous lesion stage, well-differentiated squamous cell carcinoma stage, moderately differentiated squamous cell carcinoma stage, and poorly differentiated squamous cell carcinoma stage. The gene expression levels of any one or several genes among ANGPTL2, CK15, and EEF1E1 are detected to obtain the gene expression levels, and are input into the second judgment model to obtain the oral mucosal carcinogenesis process. Optionally, the second judgment model preset the thresholds of any one or several genes among ANGPTL2, CK15, and EEF1E1 in the normal stage, precancerous lesion stage, and well / moderately / poorly differentiated squamous cell carcinoma stage, and the judgment result is obtained by comparing the gene expression level with the preset threshold. Optionally, when the gene expression level of ANGPTL2 is greater than the threshold in the normal stage, it is determined to be the precancerous lesion stage; when the gene expression level of ANGPTL2 is less than the threshold in the precancerous lesion stage, it is determined to be the well / moderately / poorly differentiated squamous cell carcinoma stage. Optionally, when the gene expression level of EEF1E1 is greater than the value in the normal stage, it is determined to be the well-differentiated squamous cell carcinoma stage; when the gene expression level of CK15 is less than the threshold in the well-differentiated squamous cell carcinoma stage, it is determined to be the moderately differentiated squamous cell carcinoma stage or the poorly differentiated squamous cell carcinoma stage.

[0009] Optionally, the non-elliptical type includes the fusion type, budding type, and finger-like advancing type. Optionally, the process judgment of the well / moderately / poorly differentiated squamous cell carcinoma stage further includes phenotypic characteristics. When the gene expression level of ANGPTL2 is less than the threshold in the precancerous lesion stage and the phenotypic characteristic is the fusion type, it is determined to be the well-differentiated squamous cell carcinoma stage; when the gene expression level of ANGPTL2 is less than the threshold in the precancerous lesion stage and the phenotypic characteristic is the budding type, it is determined to be the moderately differentiated squamous cell carcinoma stage; when the gene expression level of ANGPTL2 is less than the threshold in the precancerous lesion stage and the phenotypic characteristic is the finger-like advancing type, it is determined to be the poorly differentiated squamous cell carcinoma stage.

[0010] The S2 is replaced with: obtaining the phenotypes of microvessels based on the pathological image, where the phenotypes include oval and non-oval. When the phenotype is non-oval, TSCs detection is performed. When TSCs are not detected, ANGPTL2 gene detection is performed to obtain the gene expression level. When the gene expression level of ANGPTL2 is greater than the first preset threshold, it is determined as the prediction result of oral mucosal carcinogenesis; when TSCs are detected, ANGPTL2 gene detection is performed to obtain the gene expression level. When the gene expression level of ANGPTL2 is less than the first preset threshold, it is determined as the prediction result of oral mucosal carcinogenesis.

[0011] The pathological image is classified by a microvessel phenotype classification model to obtain the phenotypes of microvessels; Optionally, the microvessel phenotype classification model is obtained by performing feature comparison training on the microvessel features in the microvessel pathological image and the microvessel immunohistochemical image; Optionally, the training process of the microvessel phenotype classification model is as follows: Obtain a microvessel pathological staining image dataset and phenotype labels, and a microvessel immunohistochemical image dataset; Segment the microvessels in the microvessel pathological staining image dataset and the microvessel immunohistochemical image dataset to obtain microvessels of each phenotype in pathological staining and microvessels of each phenotype in immunohistochemistry; Input the microvessels of each phenotype in pathological staining and the phenotype labels into the microvessel classification model to be trained to obtain the phenotype types predicted by the pathological staining microvessels. Compare the predicted pathological staining microvessels with the immunohistochemical microvessels of the corresponding phenotype types to obtain a comparison result. Based on the comparison result, perform model iteration optimization until the loss function remains unchanged to obtain the microvessel classification model.

[0012] The purpose of the present invention is to provide a computer program product, which includes a computer program or instruction, and the computer program or instruction is executed by a processor to implement the above-mentioned method for predicting oral mucosal carcinogenesis based on pathological images.

[0013] The purpose of the present invention is to provide a computer device, which includes a memory, a processor, and a computer program or instruction stored on the memory, and the computer program or instruction is executed by the processor to implement the above-mentioned method for predicting oral mucosal carcinogenesis based on pathological images.

[0014] The purpose of the present invention is to provide a computer-readable storage medium, on which a computer program or instruction is stored, and the computer program or instruction is executed by a processor to implement the above-mentioned method for predicting oral mucosal carcinogenesis based on pathological images.

[0015] Advantages of the present invention: 1. The present invention proposes a method for predicting oral mucosal carcinogenesis based on pathological images, which can classify and judge normal phenotypes (elliptical) and abnormal phenotypes (non-elliptical) through the microvascular phenotypes of pathological images. When an abnormal phenotype appears, the detection of the ANGPTL2 gene expression level is carried out to obtain the prediction result of oral mucosal carcinogenesis. This method combines microvascular phenotypes and gene expression levels to achieve the prediction result of oral mucosal carcinogenesis, which is helpful for the diagnosis and prediction of oral mucosal carcinogenesis, and then for early intervention.

[0016] 2. The present invention predicts the process of oral mucosal carcinogenesis through gene expression level / microvascular phenotype and gene expression level, obtains the exact carcinogenesis process of the tested person, which is helpful for doctors to treat the symptoms, improve the prognosis and survival rate of patients, avoid unnecessary tests in the diagnosis process, and reduce the surgical cost.

[0017] 3. The present invention completes the prediction of oral mucosal carcinogenesis through two processing procedures by detecting tumor stem cells TSCs, including judging the carcinogenesis result by the change of ANGPTL2 gene expression level when there are TSCs and when there are no TSCs, which conforms to the actual clinical situation and provides a solid foundation for clinical application. In addition, it also includes predicting carcinogenesis by the expression levels of CK15 and EEF1E1 genes, providing multiple prediction indicators, which is helpful for flexible clinical application. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present invention. For those skilled in the art, without creative efforts, other drawings can be obtained according to these drawings.

[0019] Figure 1 It is a schematic flowchart of the method for predicting oral mucosal carcinogenesis based on pathological images provided by the embodiment of the present invention; Figure 2 It is a schematic diagram of the system for predicting oral mucosal carcinogenesis based on pathological images provided by the embodiment of the present invention; Figure 3 It is a schematic diagram of the device for predicting oral mucosal carcinogenesis based on pathological images provided by the embodiment of the present invention; Figure 4 It is the standard of the animal disease model for the process of oral mucosal carcinogenesis provided by the embodiment of the present invention; Figure 5 It is the pathological characterization of the animal disease model for the process of oral mucosal carcinogenesis in the control group provided by the embodiment of the present invention; Figure 6 It is the pathological characterization of the animal disease model for the process of oral mucosal carcinogenesis in the experimental group provided by the embodiment of the present invention; Figure 7The microvascular morphology provided by the embodiments of the present invention: (a1) cgn: elliptical; (b1) (b2) cgpl: fused type, budding type; (c1) (c2) cghdscc: fused type, budding type; (d1) (d2) (d3) cgmdscc: fused type, budding type, finger-like propulsion type; (e1) cgpdscc: finger-like propulsion type; Figure 8 The expressions of ANGPTL2, CK15, and EEF1E1 provided by the embodiments of the present invention in the normal group; Figure 9 The expressions of ANGPTL2, CK15, and EEF1E1 provided by the embodiments of the present invention in the precancerous lesion tissue group; Figure 10 The expressions of ANGPTL2, CK15, and EEF1E1 provided by the embodiments of the present invention in the well-differentiated squamous cell carcinoma group; Figure 11 The expressions of ANGPTL2, CK15, and EEF1E1 provided by the embodiments of the present invention in the moderately differentiated squamous cell carcinoma group; Figure 12 The expressions of ANGPTL2, CK15, and EEF1E1 provided by the embodiments of the present invention in the poorly differentiated squamous cell carcinoma group. Detailed implementation manners

[0020] In order to enable those skilled in the art to better understand the solution of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.

[0021] In some processes described in the specification, claims, and the above-mentioned drawings of the present invention, a plurality of operations appear in a specific order. However, it should be clearly understood that these operations may not be executed in the order in which they appear in this document or may be executed in parallel. The serial numbers of the operations, such as S101, S102, etc., are only used to distinguish different operations, and the serial numbers themselves do not represent any execution order. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions such as "first" and "second" in this document are used to distinguish different messages, devices, modules, etc., and do not represent a sequence, nor do they limit that "first" and "second" are of different types.

[0022] Figure 1 The schematic diagram of the oral mucosa canceration prediction method based on pathological images provided by the embodiments of the present invention specifically includes: S1: Obtain the pathological image of the oral mucosa tissue of the person to be tested; In one embodiment, the pathological image is classified by a microvascular phenotype classification model to obtain the phenotype of the microvasculature; Optionally, the microvascular phenotype classification model is obtained by training through feature comparison of microvascular features in microvascular pathological images and microvascular immunohistochemical images.

[0023] In one embodiment, the training process of the microvascular phenotype classification model is as follows: Obtain a microvascular pathological staining image dataset and phenotype labels, and a microvascular immunohistochemical image dataset; Segment the microvessels in the microvascular pathological staining image dataset and the microvascular immunohistochemical image dataset to obtain microvessels of each phenotype in pathological staining and microvessels of each phenotype in immunohistochemistry; Input the microvessels of each phenotype in pathological staining and the phenotype labels into the microvascular classification model to be trained to obtain the predicted phenotype types of the microvessels in pathological staining. Compare the predicted microvessels in pathological staining with the microvessels in immunohistochemistry of the corresponding phenotype types to obtain a comparison result. Iteratively optimize the model based on the comparison result until the loss function remains unchanged to obtain the microvascular classification model.

[0024] S2: Obtain the phenotype of the microvessels based on the pathological image. The phenotypes include elliptical and non-elliptical. When the phenotype is non-elliptical, perform ANGPTL2 gene detection to obtain the gene expression level. When the gene expression level of ANGPTL2 is greater than the first preset threshold, it is determined as the predicted result of oral mucosal carcinogenesis.

[0025] In one embodiment, the first preset threshold is the gene expression level of ANGPTL2 in the normal state.

[0026] In one embodiment, the ANGPTL2 gene is replaced by the CK15 gene. When the gene expression level of CK15 is greater than the second preset threshold, it is determined as the predicted result of oral mucosal carcinogenesis.

[0027] Optionally, the second preset threshold is the gene expression level of CK15 in the normal state.

[0028] In one embodiment, the ANGPTL2 gene is replaced by the EEF1E1 gene. When the gene expression level of EEF1E1 is greater than the third preset threshold, it is determined as the predicted result of oral mucosal carcinogenesis; Optionally, the third preset threshold is the gene expression level of EEF1E1 in the normal state.

[0029] In one embodiment, the gene detection further includes one or more of the following: CK15, EEF1E1. Predict whether oral mucosal carcinogenesis has occurred through two or three genes.

[0030] In one embodiment, the method further includes the determination of the progression of oral mucosal carcinogenesis, and the progression of oral mucosal carcinogenesis includes the well-differentiated squamous cell carcinoma stage, the moderately-differentiated squamous cell carcinoma stage, and the poorly-differentiated squamous cell carcinoma stage; detecting the gene expression levels of any one or more of ANGPTL2, CK15, and EEF1E1 to obtain the gene expression levels, and inputting them into the first judgment model to obtain the progression of oral mucosal carcinogenesis; Optionally, the first judgment model preset the thresholds of any one or more of ANGPTL2, CK15, and EEF1E1 in the normal stage, the well / moderately / poorly-differentiated squamous cell carcinoma stage, and obtain the judgment result by comparing the gene expression level with the preset threshold.

[0031] In one embodiment, when the gene expression level of ANGPTL2 is greater than the threshold in the normal stage, it is determined to be in the well-differentiated squamous cell carcinoma stage; when the gene expression level of ANGPTL2 is less than the threshold in the well-differentiated squamous cell carcinoma stage, it is determined to be in the moderately-differentiated squamous cell carcinoma stage or the poorly-differentiated squamous cell carcinoma stage.

[0032] In one embodiment, when the gene expression level of CK15 is greater than the threshold in the normal stage, it is determined to be in the well-differentiated squamous cell carcinoma stage; when the gene expression level of CK15 is greater than the threshold in the well-differentiated squamous cell carcinoma stage, it is determined to be in the moderately-differentiated squamous cell carcinoma stage or the poorly-differentiated squamous cell carcinoma stage.

[0033] In one embodiment, when the gene expression level of EEF1E1 is greater than the threshold in the normal stage, it is determined to be in the well-differentiated squamous cell carcinoma stage or the moderately-differentiated squamous cell carcinoma stage; when the gene expression level of EEF1E1 is less than the threshold in the moderately-differentiated squamous cell carcinoma stage, it is determined to be in the poorly-differentiated squamous cell carcinoma stage.

[0034] In one embodiment, when the gene expression levels of ANGPTL2, CK15, and EEF1E1 are greater than the threshold in the normal stage, it is determined to be in the well-differentiated squamous cell carcinoma stage; when the gene expression level of ANGPTL2 is less than the threshold in the well-differentiated squamous cell carcinoma, the gene expression level of CK15 is greater than the threshold in the well-differentiated squamous cell carcinoma, and the gene expression level of EEF1E1 is greater than the threshold in the well-differentiated squamous cell carcinoma, it is determined to be in the moderately-differentiated squamous cell carcinoma stage; when the gene expression level of ANGPTL2 is less than the threshold in the moderately-differentiated squamous cell carcinoma, the gene expression level of CK15 is greater than the threshold in the moderately-differentiated squamous cell carcinoma, and the gene expression level of EEF1E1 is less than the threshold in the moderately-differentiated squamous cell carcinoma, it is determined to be in the poorly-differentiated squamous cell carcinoma stage.

[0035] In one embodiment, the method further includes the detection of TSCs. When the phenotype is non-elliptical, the detection of TSCs and the detection of the ANGPTL2 gene are performed. When TSCs are detected and the gene expression level of ANGPTL2 is less than the first preset threshold, it is determined as the predicted result of oral mucosal carcinogenesis.

[0036] In one embodiment, the ANGPTL2 is replaced by the CK15 gene. When TSCs are detected and the gene expression level of CK15 is greater than a second preset threshold, a prediction result of oral mucosal carcinogenesis is determined.

[0037] In one embodiment, the ANGPTL2 is replaced by the EEF1E1 gene. When TSCs are detected and the gene expression level of EEF1E1 is greater than a third preset threshold, a prediction result of oral mucosal carcinogenesis is determined. In another embodiment, oral mucosal carcinogenesis is predicted in the presence of TSCs by any two or three of ANGPTL2, EEF1E1, and CK15.

[0038] In one embodiment, the method further includes determining the process of oral mucosal carcinogenesis. The process of oral mucosal carcinogenesis includes the precancerous lesion stage, well-differentiated squamous cell carcinoma stage, moderately differentiated squamous cell carcinoma stage, and poorly differentiated squamous cell carcinoma stage. The gene expression level of any one or several genes among ANGPTL2, CK15, and EEF1E1 is detected to obtain the gene expression level, which is then input into a second judgment model to obtain the process of oral mucosal carcinogenesis. Optionally, the second judgment model preset thresholds of any one or several genes among ANGPTL2, CK15, and EEF1E1 in the normal stage, precancerous lesion stage, and well / moderately / poorly differentiated squamous cell carcinoma stage. The judgment result is obtained by comparing the gene expression level with the preset threshold.

[0039] In one embodiment, when the gene expression level of ANGPTL2 is greater than the threshold in the normal stage, it is determined as the precancerous lesion stage. When the gene expression level of ANGPTL2 is less than the threshold in the precancerous lesion stage, it is determined as the well / moderately / poorly differentiated squamous cell carcinoma stage.

[0040] In one embodiment, when the gene expression level of EEF1E1 is greater than the value in the normal stage, it is determined as the well-differentiated squamous cell carcinoma stage. When the gene expression level of CK15 is less than the threshold in the well-differentiated squamous cell carcinoma stage, it is determined as the moderately differentiated squamous cell carcinoma stage or the poorly differentiated squamous cell carcinoma stage.

[0041] In one embodiment, the non-elliptical type includes the fusion type, budding type, and finger-like advancing type. Fusion type: The cross-section of the circular microvascular lumen is partially fused. Budding type: The microvascular lumen buds to form vascular branches. Finger-like advancing type: The longitudinal sections of the microvascular lumens are arranged in parallel and aggregated.

[0042] In one embodiment, the process determination of the high / medium / low differentiated squamous cell carcinoma stage further includes phenotypic characteristics. When the gene expression level of ANGPTL2 is less than the threshold value of the precancerous lesion stage and the phenotypic characteristic is the fusion type, it is determined to be the high differentiated squamous cell carcinoma stage; when the gene expression level of ANGPTL2 is less than the threshold value of the precancerous lesion stage and the phenotypic characteristic is the budding type, it is determined to be the medium differentiated squamous cell carcinoma stage; when the gene expression level of ANGPTL2 is less than the threshold value of the precancerous lesion stage and the phenotypic characteristic is the finger-like advancing type, it is determined to be the low differentiated squamous cell carcinoma stage.

[0043] In one embodiment, S2 is replaced with: obtaining the phenotype of the microvessels based on the pathological image, the phenotype including elliptical and non-elliptical. When the phenotype is non-elliptical, TSCs detection is performed. When TSCs are not detected, ANGPTL2 gene detection is performed to obtain the gene expression level. When the gene expression level of ANGPTL2 is greater than the first preset threshold value, it is determined to be the prediction result of oral mucosal carcinogenesis; when TSCs are detected, ANGPTL2 gene detection is performed to obtain the gene expression level. When the gene expression level of ANGPTL2 is less than the first preset threshold value, it is determined to be the prediction result of oral mucosal carcinogenesis.

[0044] In a specific embodiment, an animal disease model of oral mucosal carcinogenesis is established, including a normal group, a precancerous lesion tissue group, and high, medium, and low differentiated squamous cell carcinoma groups, and this process is defined as a control group; ② LCM is used to target and obtain TSCs in human oral squamous carcinoma cells (HSC4), and monoclonal expansion and culture are performed. The expression of stem cell markers CD44, CD133, CD90, ALDH1, and SOX2 in TSCs is detected by immunofluorescence staining, and the TSCs microsphere formation experiment is used to determine the TSCs cell sphere formation efficiency; ③ TSCs (EdU Apollo 567 labeled) interfere with the oral mucosal carcinogenesis animal disease model is established, and this process is defined as an experimental group, that is, EdU Apollo 567 is used to mark TSCs in vitro, and the rat tail vein is injected to interfere with the oral mucosal carcinogenesis animal disease model, and TSCs are subsequently traced in vivo Distribution and localization in each lesion model stage and each organ tissue; ④ Microvascular typing analysis was performed in each stage of the oral mucosal carcinogenesis process in the control group and the experimental group, the composition ratio of microvascular morphological heterogeneity in the tissues at each stage was statistically analyzed, and the differences in the classical characterization of microvascular morphology in the oral mucosal carcinogenesis process between the control group and the experimental group were compared and analyzed; ⑤ LCM was used to target the capture of the classically characterized microvascular morphological structure in each stage of the oral mucosal carcinogenesis process in the control group and the experimental group, TMT was used to analyze and mine the potential phenotypic families of TSCs interfering with the heterogeneity characteristics of microvascular morphology in each stage of the oral mucosal carcinogenesis process, and then immunofluorescence staining and western blot were used to verify the expression of TSCs on the differential phenotypes of MAPH characteristics in each stage of the oral mucosal carcinogenesis process, and the recruitment and integration of VM in the oral mucosal carcinogenesis process by TSCs were explored.

[0045] In a specific embodiment, the results of an animal disease model in which TSCs interfere with the process of oral mucosal carcinogenesis showed that TSCs labeled with EdUApollo 567 entered the vascular system via tail vein injection, circulated through the blood and survived, and resided around the microvessels of tissues and organs at various stages of disease, and at the same time invaded the walls of blood vessels, causing changes in microvascular morphology.

[0046] By comparing the classical microvascular morphological characteristics in the oral mucosa carcinogenesis process between the control group and the experimental group, the statistical results of the microvascular morphology in the carcinogenesis process of the control group showed that in the normal group, the oval microvascular morphology was the main type; in the precancerous lesion tissue group, the fused and budding microvascular morphologies were the main types; in the well-differentiated squamous cell carcinoma group, the fused and budding microvascular morphologies were the main types; in the moderately differentiated squamous cell carcinoma group, the fused, budding, and finger-like advancing microvascular morphologies were the main types; in the poorly differentiated squamous cell carcinoma group, the finger-like advancing microvascular morphology was the main type. The statistical results of the microvascular morphology in the carcinogenesis process of the experimental group showed that in the normal group, the oval microvascular morphology was the main type; in the precancerous lesion tissue group, the fused microvascular morphology was the main type; in the well-differentiated squamous cell carcinoma group, the fused microvascular morphology was the main type; in the moderately differentiated squamous cell carcinoma group, the budding microvascular morphology was the main type; in the poorly differentiated squamous cell carcinoma group, the finger-like advancing microvascular morphology was the main type.

[0047] The LCM targeted and captured the characteristic microvascular structures at each stage of the oral mucosal carcinogenesis process in the control group and the experimental group, and was associated with TMT analysis. A total of 2,215 differentially expressed proteins with different expression levels were identified. Among them, there were 1,025 up-regulated proteins and 1,190 down-regulated proteins in the experimental group of the carcinogenesis process compared with the control group of the carcinogenesis process; there were 5 up-regulated proteins with differential expression and 1 down-regulated protein with differential expression at each stage between the control group and the experimental group of the carcinogenesis process; when comparing each stage of the control group of the carcinogenesis process with each other, there were 4 up-regulated proteins and 17 down-regulated proteins; when comparing each stage of the experimental group of the carcinogenesis process with each other, there were 2 up-regulated proteins and 2 down-regulated proteins. Through statistical analysis, 1 up-regulated protein and 4 down-regulated proteins were screened out by t-test or chi-square test, and differentially expressed proteins (DEPs) with FC>0.8 were selected for subsequent experimental verification, namely angiopoietin-like protein 2 (ANGPTL2), cytokeratin 15 (CK15), and eukaryotic translation elongation factor 1 epsilon 1 (EEF1E1). The results of Western blot and immunohistochemistry showed that: compared with the control group of the carcinogenesis process, the expression level of CK15 showed an upward trend in the experimental group of the carcinogenesis process, while the expression levels of ANGPTL2 and EEF1E1 showed a downward trend. In each stage of oral mucosal carcinogenesis in the control group, the expression level of CK15 gradually increased, the expression level of ANGPTL2 gradually increased from normal to well-differentiated squamous cell carcinoma stage, and gradually decreased from well-differentiated squamous cell carcinoma to poorly-differentiated squamous cell carcinoma stage, with the highest expression level in the well-differentiated squamous cell carcinoma stage. The expression level of EEF1E1 gradually increased from normal to moderately-differentiated squamous cell carcinoma stage, and gradually decreased from moderately-differentiated squamous cell carcinoma to poorly-differentiated squamous cell carcinoma stage, with the highest expression level in the moderately-differentiated squamous cell carcinoma stage. In each stage of oral mucosal carcinogenesis in the experimental group, the expression level of CK15 gradually increased, the expression level of ANGPTL2 gradually increased from normal to precancerous lesion stage, and gradually decreased from precancerous lesion to poorly-differentiated squamous cell carcinoma stage, with the highest expression level in the precancerous lesion stage. The expression level of EEF1E1 gradually increased from normal to well-differentiated squamous cell carcinoma stage, and gradually decreased from well-differentiated squamous cell carcinoma to poorly-differentiated squamous cell carcinoma stage, with the highest expression level in the well-differentiated squamous cell carcinoma stage.

[0048] In a specific embodiment, nude mice were purchased from Beijing Huafukang Bioscience Co., Ltd. They were 4 weeks old, male, with a body weight of about 20 g, and there were 100 in total. They were randomly divided into an experimental group and a control group. Each group was further divided into 5 subgroups, resulting in a total of 10 subgroups with 10 mice in each subgroup. The grouping of the control group and the experimental group was as followsFigure 4 As shown in Figure 4 , primary culture of HGECs: In a sterile laminar flow hood, healthy and non - inflamed human gingival epithelial tissues were washed with normal saline and double antibiotics, and the surface epithelial tissues of the gingiva were separated. The epithelial tissues were cut into small pieces of about 1 mm³ and spread flat on the bottom of the culture flask, with a 5 - mm interval between each small piece of tissue. An appropriate amount of high - glucose culture medium containing 15% fetal bovine serum was added to the bottom of the culture flask. The culture flask was incubated in an incubator at 37°C and 5% CO₂ for 2 h, then gently flipped and placed flat to allow the culture medium to slowly cover the tissue pieces at the bottom of the culture flask for primary culture. The culture medium was changed once every 3 - 5 d. When the cells around the tissue pieces covered about 85% of the bottom of the flask, sub - culture and amplification were carried out at a ratio of 1:2.

[0049] Preparation of HGECs and HSC4 cell suspensions: HGECs and HSC4 in good condition and in the logarithmic growth phase were collected. After digestion and centrifugation, single - cell suspensions were prepared. After cell counting and viability determination, the cell concentration was adjusted (about 1×10⁷ cells / mL) and stored on ice for later use.

[0050] Subcutaneous injection of HGECs and HSC4 cell suspensions: Under sterile conditions, HGECs cell suspension was subcutaneously injected into the right upper limb of cgn nude mice, 0.2 mL per mouse, and a total of 10 mice were injected; HGECs cell suspension was subcutaneously injected into the right upper limb of cgpl nude mice, 0.2 mL per mouse, and a total of 10 mice were injected. After 2 w, when a stable epitheloid structure was formed, HSC4 cell suspension was injected into the epitheloid structure, 0.2 mL per mouse, and a total of 10 mice were injected; HSC4 cell suspension was subcutaneously injected into the right upper limb of cghdscc, cgmdscc, and cgpdscc nude mice, 0.2 mL per mouse, and a total of 30 mice were injected. After injection, the status of the nude mice and the conditions of the injection sites were closely observed. When a mass appeared at the injection site, the body weight of the nude mice and the size of the mass were regularly recorded. After cgn was inoculated with HGECs for 2 w, and cgpl, cghdscc, cgmdscc, and cgpdscc nude mice were subcutaneously injected with HSC4 for 2 w, 4 w, 6 w, and 8 w respectively, euthanasia was performed (using a 5% sodium pentobarbital solution prepared with normal saline, intraperitoneally injected at 0.5 mL / 100 g). Under sterile conditions, the tissues of the oral mucosa canceration process were dissected from the right upper limb of the control group nude mice. Some tissues were fixed with 10% neutral formalin for HE staining; some fresh tissues were frozen - embedded and sectioned onto metal slides, and the microvascular structures with heterogeneous dominant morphology in the tumor tissues were targeted and captured by LCM for TMT detection; some tissues were stored in a - 80°C refrigerator for immunofluorescence staining and western blot.

[0051] In a specific embodiment, immunofluorescence staining of TSCs specific surface markers: The primary antibodies used were anti-CD44 antibody, anti-CD90 antibody, anti-CD133 antibody, anti-ALDH1 antibody and anti-SOX2 antibody, and the fluorescent secondary antibody was Rhodamine (TRITC) Goat Anti-Rabbit IgG (H+L) for immunofluorescence chemical staining. To detect the expression and distribution of stem cell surface markers CD44, CD90, CD133, ALDH1, and SOX2 in TSCs. (1) Take TSCs in good growth state, digest them with 0.25% trypsin, centrifuge, add culture medium to obtain cell suspension, and evenly inoculate it into the treated sterile cover slips, and culture it in an incubator at 37°C and 5% CO2. When the cells grow and fuse to 70% - 80%, take out the cover slips, wash them 3 times with PBS, 5 minutes each time; (2) At room temperature, fix the cell slides with 4% paraformaldehyde for 30 minutes, discard the fixing solution in the slides, wash them 3 times with PBS, 5 minutes each time; (3) Treat with 3% H2O2 for 10 minutes to block endogenous peroxidase, wash 3 times with PBS, 5 minutes each time; (4) Incubate with primary antibodies (TSCs specific surface markers CD44, CD90, CD133, ALDH1 and SOX2, each marker is diluted with PBS solution in a certain proportion and used), overnight in a 4°C refrigerator, wash 3 times with PBS, 5 minutes each time; (5) Incubate with secondary antibody (dilute the secondary antibody goat anti-rabbit Rhodamine with PBS solution at a ratio of 1:100), incubate in the dark at 37°C for 1 h, wash 3 times with PBS, 5 minutes each time; (6) Dilute the DAPI solution to the working concentration (1:1000), at room temperature, incubate in the dark for 30 mim, wash 3 times with PBS, 5 minutes each time; (7) Drop 5% glycerol for mounting, collect images and analyze them under an inverted fluorescence microscope.

[0052] Construction of an animal disease model of TSCs interfering with the oral mucosal carcinogenesis process: 1. HGECs and HSC4 in good condition and in the logarithmic growth phase were collected, made into single-cell suspensions after digestion and centrifugation, and after cell counting and viability determination, the cell concentration was adjusted (about 1×107 cells / mL) and placed on ice for later use. 2. Under sterile conditions, HGECs cell suspension was subcutaneously injected into the right upper limb of egn nude mice, 0.2 mL per mouse, and a total of 10 mice were injected; HGECs cell suspension was subcutaneously injected into the right upper limb of egpl nude mice, 0.2 mL per mouse, and a total of 10 mice were injected. After 2 weeks, a stable epithelial-like structure was formed, and HSC4 cell suspension was injected into the epithelial-like structure, 0.2 mL per mouse, and a total of 10 mice were injected; HSC4 cell suspension was subcutaneously injected into the right upper limb of eghdscc, egmdscc, and egpdscc nude mice, 0.2 mL per mouse, and a total of 30 mice were injected. After injection, the status of the nude mice and the conditions of the injection sites were closely observed, and the body weights of the nude mice were regularly recorded when lumps appeared at the injection sites. 3. One week after injecting HGECs into egn nude mice, and 1 week, 3 weeks, 5 weeks, and 7 weeks after inoculating HSC4 into egpl, eghdscc, egmdscc, and egpdscc nude mice respectively, the suspension of TSCs labeled with EdU Apollo 567 (the cell count was 1×107 cells / mL) was injected into the tail vein of the nude mice, 0.1 mL per mouse, and a total of 50 mice were injected. 4. One week after receiving the tail vein injection of the TSCs suspension, egn, egpl, eghdscc, egmdscc, and egpdscc nude mice were euthanized (a 5% sodium pentobarbital solution prepared with normal saline, intraperitoneally injected at 0.5 mL / 100 g). Under sterile conditions, the tissues of the right upper limb of the nude mice with oral mucosal carcinogenesis process in the experimental group were dissected, and samples were taken from the heart, liver, spleen, lungs, and kidneys labeled with EdU Apollo 567, and marks were made respectively. Some tissues of the oral mucosal carcinogenesis process and various organ tissues were fixed with 10% neutral formalin and stained with HE and EdU Apollo 567 in vivo tracer staining; some fresh tissues were frozen-embedded and sectioned onto metal slides, and the microvascular structure with heterogeneity-dominated morphology in the tumor tissue was targeted and captured by LCM for TMT detection; some tissues were stored in a -80°C refrigerator for immunofluorescence staining and western blot.

[0053] In a specific embodiment, morphological identification of the microvascular structure of the oral mucosa carcinogenesis process in the control group and the experimental group: HE staining of the oral mucosa carcinogenesis process tissue: (1) Dewaxing of paraffin sections: Dewax with xylene I and xylene II for 12 minutes each, and treat with ethanol gradients (100%, 95%, 85%, 75%) for 2 minutes each; (2) Treat with hematoxylin for 5 minutes and rinse with water for 2 s; (3) Differentiate with hydrochloric acid alcohol for 2 s and rinse with water for 2 s; (3) Rinse with 1% ammonia water for 1 minute and rinse with water for 2 s; (4) Stain with eosin stain for 30 s and rinse with water for 2 s; (5) Dehydrate with ascending gradient ethanol (75%, 85%, 95%) for 2 s each, and dehydrate with absolute ethanol for 10 minutes; (6) Clear with xylene for at least 5 minutes; (7) Mount with neutral gum, observe the staining results under an upright microscope, collect images and analyze them.

[0054] Analysis of the morphological heterogeneity of the microvascular architecture of the oral mucosa carcinogenesis process in the control group and the experimental group: After identifying the HE-stained oral mucosa carcinogenesis process tissue in the control group, samples were included in each group, with a sample size of n≥3 for each group. Observe the morphological characteristics of microvessels in each group of samples under an upright microscope, and determine the microvascular morphology that dominates the morphological heterogeneity of the microvascular architecture during the oral mucosa carcinogenesis process. Statistically analyze the microvascular morphology composition ratio of n≥3 sections for each group of samples, and perform one-way ANOVA on the proportion of microvascular morphology at each stage of the oral mucosa carcinogenesis process in the control group.

[0055] Immunofluorescence chemical staining: Immunofluorescence staining was performed on the tissues at each stage of the oral mucosa carcinogenesis process dissected from the experimental group and the normal group. (1) The specific steps for preparing and dewaxing paraffin sections are the same as in section 2.7 of the second part; (2) Treat with 3% H2O2 for 10 minutes to block endogenous peroxidase, wash with PBS 3 times, 5 minutes each time; (3) Incubate with primary antibodies (rabbit anti-human ANGPTL2, CK15, and EEF1E1 antibodies are diluted with PBS solution in a certain proportion and used), overnight in a 4°C refrigerator, wash with PBS 3 times, 5 minutes each time; (4) Incubate with secondary antibodies (dilute secondary antibody goat anti-rabbit Rhodamine with PBS solution at a ratio of 1:100), incubate at 37°C in the dark for 1 h, wash with PBS 3 times, 5 minutes each time; (5) Dilute DAPI solution to the working concentration (1:1000), incubate at room temperature in the dark for 30 mim, wash with PBS 3 times, 5 minutes each time; (6) Drop 5% glycerol for mounting, collect images and analyze them under an inverted fluorescence microscope.

[0056] In one embodiment, the pathological characteristics of the animal model of oral mucosa carcinogenesis in the control group are as Figure 5 shown, and the experimental group is as Figure 6as shown. The microvascular morphology of different phenotypes in the control group is as Figure 7 shown.

[0057] In a specific embodiment, Western blot was used to analyze the expression of each protein in cgn, cgpl, cghdscc, cgmdscc, cgpdscc, egn, egpl, eghdscc, egmdscc, and egpdscc. The expressions of ANGPTL2, CK15, and EEF1E1 were detected in all 10 groups, as Figure 8 , Figure 9 , Figure 10 , Figure 11 , Figure 12 shown. Normalization was performed using the internal reference GADPH, and the expression of each protein in each group was analyzed by the difference in gray values. Compared with the control group, the expression levels of CK15 in the experimental group showed an upward trend, while the expression levels of ANGPTL2 and EEF1E1 showed a downward trend. In each stage of oral mucosal carcinogenesis in the control group, the expression level of CK15 gradually increased. The expression level of ANGPTL2 gradually increased from the normal stage to the well-differentiated squamous cell carcinoma stage, and gradually decreased from the well-differentiated squamous cell carcinoma stage to the poorly-differentiated squamous cell carcinoma stage, with the highest expression level in the well-differentiated squamous cell carcinoma stage. The expression level of EEF1E1 gradually increased from the normal stage to the moderately-differentiated squamous cell carcinoma stage, and gradually decreased from the moderately-differentiated squamous cell carcinoma stage to the poorly-differentiated squamous cell carcinoma stage, with the highest expression level in the moderately-differentiated squamous cell carcinoma stage. In each stage of oral mucosal carcinogenesis in the experimental group, the expression level of CK15 gradually increased. The expression level of ANGPTL2 gradually increased from the normal stage to the precancerous lesion stage, and gradually decreased from the precancerous lesion stage to the poorly-differentiated squamous cell carcinoma stage, with the highest expression level in the precancerous lesion stage. The expression level of EEF1E1 gradually increased from the normal stage to the well-differentiated squamous cell carcinoma stage, and gradually decreased from the well-differentiated squamous cell carcinoma stage to the poorly-differentiated squamous cell carcinoma stage, with the highest expression level in the well-differentiated squamous cell carcinoma stage.

[0058] In a specific embodiment, the morphological heterogeneity of microvessels in the oral mucosa carcinogenesis process presents different characteristic features at each stage. Among them, the elliptical microvessel morphology is the main type in the normal stage, the fused type and budding type microvessel morphologies are the main types in the carcinomatous lesion stage, the fused type and budding type microvessel morphologies are the main types in the well-differentiated squamous cell carcinoma stage, the fused type, budding type and finger-like advancing type microvessel morphologies are the main types in the moderately differentiated squamous cell carcinoma stage, and the finger-like advancing type microvessel morphology is the main type in the poorly differentiated squamous cell carcinoma stage; 2 TSCs interfere with the oral mucosa carcinogenesis process and induce characteristic changes in microvessel morphological heterogeneity, promoting tumor growth. Among them, the elliptical microvessel morphology is the main type in the normal stage, the fused type microvessel morphology is the main type in the precancerous lesion stage, the fused type microvessel morphology is the main type in the well-differentiated squamous cell carcinoma stage, the budding type microvessel morphology is the main type in the moderately differentiated squamous cell carcinoma stage, and the finger-like advancing type microvessel morphology is the main type in the poorly differentiated squamous cell carcinoma stage; 3 In the oral mucosa carcinogenesis process, the expression level of ANGPTL2 gradually increases from the normal stage to the well-differentiated squamous cell carcinoma stage and gradually decreases from the well-differentiated squamous cell carcinoma stage to the poorly differentiated squamous cell carcinoma stage. After TSCs interfere with the oral mucosa carcinogenesis process, the expression level of ANGPTL2 gradually increases from the normal stage to the precancerous lesion stage and gradually decreases from the precancerous lesion stage to the poorly differentiated squamous cell carcinoma stage; 4 In the oral mucosa carcinogenesis process, the expression level of CK15 gradually increases. After TSCs interfere with the oral mucosa carcinogenesis process, the expression level of CK15 still gradually increases; 5 In the oral mucosa carcinogenesis process, the expression level of EEF1E1 gradually increases from the normal stage to the moderately differentiated squamous cell carcinoma stage and gradually decreases from the moderately differentiated squamous cell carcinoma stage to the poorly differentiated squamous cell carcinoma stage. After TSCs interfere with the oral mucosa carcinogenesis process, the expression level of EEF1E1 gradually increases from the normal stage to the well-differentiated squamous cell carcinoma stage and gradually decreases from the well-differentiated squamous cell carcinoma stage to the poorly differentiated squamous cell carcinoma stage; 6 Compared with each stage of the oral mucosa carcinogenesis process, the expression level of CK15 shows an upward trend in each stage of TSCs interfering with oral mucosa carcinogenesis, while the expression levels of ANGPTL2 and EEF1E1 show downward trends; 7 Based on the characteristics of microvessel architectural morphological heterogeneity in the oral mucosa carcinogenesis process, targeting microvessels by combining ANGPTL2, CK15 and EEF1E1 may be the best strategy for anti-angiogenic therapy in the oral mucosa carcinogenesis process.

[0059] The disclosed embodiments of the present invention also provide a computer program product or system, including a computer program, which implements the steps of the above-mentioned oral mucosa carcinogenesis prediction method based on pathological images when executed by a processor.

[0060] Figure 2 The schematic diagram of the oral mucosa carcinogenesis prediction system provided by the embodiments of the present invention specifically includes: Acquisition unit: Acquire the pathological image of the oral mucosa tissue of the person to be tested; Prediction unit: obtaining the phenotype of the microvessels based on the pathological image, where the phenotype includes elliptical and non-elliptical. When the phenotype is non-elliptical, perform ANGPTL2 gene detection to obtain the gene expression level. When the gene expression level of ANGPTL2 is greater than the first preset threshold, determine the prediction result of oral mucosal carcinogenesis.

[0061] Figure 3 The schematic diagram of the oral mucosal carcinogenesis prediction device based on the pathological image provided by the embodiment of the present invention specifically includes: A memory and a processor; the memory is used to store program instructions; the processor is used to call the program instructions, and when the program instructions are executed, any of the above-mentioned oral mucosal carcinogenesis prediction methods based on the pathological image.

[0062] The disclosed embodiment of the present invention also provides a computer-readable storage medium, where the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, any of the above-mentioned oral mucosal carcinogenesis prediction methods based on the pathological image.

[0063] The verification results of this verification embodiment show that assigning fixed weights to indications can improve the performance of this method compared to the default settings. Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the systems, devices, and units described above can refer to the corresponding processes in the foregoing method embodiments, and will not be elaborated herein. In several embodiments provided by this application, it should be understood that the disclosed systems, devices, and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the units is only a logical function division, and there may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed couplings or direct couplings or communication connections to each other can be through some interfaces, and the indirect couplings or communication connections of the devices or units can be in electrical, mechanical, or other forms. The units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they can be located in one place or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment. In addition, in each embodiment of the present invention, the functional units can be integrated into one processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit. The above-mentioned integrated units can be implemented in the form of hardware or in the form of software functional units. Those of ordinary skill in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing relevant hardware through a program, and this program can be stored in a computer-readable storage medium. The storage medium can include: read-only memory (ROM, Read Only Memory), random access memory (RAM, Random Access Memory), magnetic disk, or optical disc, etc.

[0064] Those of ordinary skill in the art can understand that all or part of the steps in implementing the methods of the above embodiments can be completed by instructing relevant hardware through a program, and the program can be stored in a computer-readable storage medium. The above-mentioned medium storage can be read-only memory, magnetic disk, or optical disc, etc.

[0065] The above has introduced in detail a computer device provided by the present invention. For those of ordinary skill in the art, according to the idea of the embodiments of the present invention, there will be changes in the specific implementation manners and application scopes. In summary, the content of this specification should not be construed as a limitation to the present invention.

Claims

1. A method for predicting oral mucosal canceration based on pathological images, characterized in that: include: S1. Obtaining a pathological image of the oral mucosal tissue of the subject; S2. obtaining a phenotype of microvessels based on the pathological image, wherein the phenotype includes an elliptical type and a non-elliptical type. When the phenotype is non-elliptical, ANGPTL2 gene detection is performed to obtain the gene expression level, and when the ANGPTL2 gene expression level is greater than a first preset threshold, it is determined as a prediction result of oral mucosal canceration.

2. The method for predicting oral mucosal canceration based on pathological images according to claim 1, characterized in that: The first preset threshold is the gene expression level of ANGPTL2 under normal conditions; Optionally, the ANGPTL2 gene is replaced by the CK15 gene, and when the gene expression level of the CK15 is greater than a second preset threshold, it is determined as a prediction result of oral mucosal canceration; Optionally, the second preset threshold is the gene expression level of CK15 under normal conditions; Optionally, the ANGPTL2 gene is replaced by the EEF1E1 gene, and when the gene expression level of the EEF1E1 is greater than a third preset threshold, it is determined as a prediction result of oral mucosal canceration; Optionally, the third preset threshold is the gene expression level of EEF1E1 under normal conditions; Optionally, the gene detection also includes one or more of the following: CK15, EEF1E1.

3. The method for predicting oral mucosal canceration based on pathological images according to claim 2, characterized in that: The method further includes determining the progression of oral mucosal carcinoma, wherein the progression of oral mucosal carcinoma includes a well-differentiated squamous cell carcinoma stage, a moderately differentiated squamous cell carcinoma stage, and a poorly differentiated squamous cell carcinoma stage; detecting the gene expression of any one or more genes among ANGPTL2, CK15, and EEF1E1 to obtain the gene expression, and inputting the gene expression into the first determination model to obtain the progression of oral mucosal carcinoma; Optionally, the first judgment model pre-sets the thresholds of any one or several genes of ANGPTL2, CK15, and EEF1E1 in the normal stage and the high / medium / poorly differentiated squamous cell carcinoma stage, and obtains the judgment result by comparing the gene expression amount with the preset threshold; Optionally, when the gene expression level of ANGPTL2 is greater than the threshold of the normal stage, it is determined to be at the well-differentiated squamous cell carcinoma stage; when the gene expression level of ANGPTL2 is less than the threshold of the well-differentiated squamous cell carcinoma stage, it is determined to be at the moderately differentiated squamous cell carcinoma stage or the poorly differentiated squamous cell carcinoma stage; Optionally, when the gene expression level of CK15 is greater than the threshold of the normal stage, it is determined to be at the well-differentiated squamous cell carcinoma stage; when the gene expression level of CK15 is greater than the threshold of the well-differentiated squamous cell carcinoma stage, it is determined to be at the moderately differentiated squamous cell carcinoma stage or the poorly differentiated squamous cell carcinoma stage; Optionally, when the gene expression level of EEF1E1 is greater than the threshold of the normal stage, it is determined to be at the well-differentiated squamous cell carcinoma stage or the moderately differentiated squamous cell carcinoma stage; when the gene expression level of EEF1E1 is less than the threshold of the moderately differentiated squamous cell carcinoma stage, it is determined to be at the poorly differentiated squamous cell carcinoma stage; Optionally, when the gene expression levels of ANGPTL2, CK15, and EEF1E1 are greater than the threshold of the normal stage, it is judged to be the well-differentiated squamous cell carcinoma stage; when the gene expression level of ANGPTL2 is less than the threshold of well-differentiated squamous cell carcinoma, the gene expression level of CK15 is greater than the threshold of well-differentiated squamous cell carcinoma, and the gene expression level of EEF1E1 is greater than the threshold of well-differentiated squamous cell carcinoma, it is judged to be the moderately differentiated squamous cell carcinoma stage; when the gene expression level of ANGPTL2 is less than the threshold of moderately differentiated squamous cell carcinoma, the gene expression level of CK15 is greater than the threshold of moderately differentiated squamous cell carcinoma, and the gene expression level of EEF1E1 is less than the threshold of moderately differentiated squamous cell carcinoma, it is judged to be the poorly differentiated squamous cell carcinoma stage.

4. The method for predicting oral mucosal canceration based on pathological images according to claim 1, characterized in that: The method further comprises TSCs detection, when the phenotype is non-elliptical, TSCs detection and ANGPTL2 gene detection are performed, when TSCs are detected and the gene expression level of ANGPTL2 is less than a first preset threshold, it is determined as a prediction result of oral mucosal carcinogenesis; Optionally, the ANGPTL2 is replaced by the CK15 gene, and when TSCs are detected and the gene expression level of CK15 is greater than a second preset threshold, it is determined as a prediction result of oral mucosal canceration; Optionally, the ANGPTL2 is replaced by the EEF1E1 gene, and when TSCs are detected and the gene expression level of EEF1E1 is greater than a third preset threshold, it is determined as a prediction result of oral mucosal cancer.

5. The method for predicting oral mucosal canceration based on pathological images according to claim 4, characterized in that: The method further includes judging the progression of oral mucosal canceration, wherein the progression of oral mucosal canceration includes the precancerous lesion stage, the well-differentiated squamous cell carcinoma stage, the moderately differentiated squamous cell carcinoma stage, and the poorly differentiated squamous cell carcinoma stage; detecting the gene expression of any one or more genes among ANGPTL2, CK15, and EEF1E1 to obtain the gene expression amount, and inputting the gene expression amount into the second judgment model to obtain the progression of oral mucosal canceration; Optionally, the second judgment model pre-sets thresholds of any one or more genes of ANGPTL2, CK15, and EEF1E1 in the normal stage, precancerous lesion stage, and high / medium / poorly differentiated squamous cell carcinoma stage, and obtains the judgment result by comparing the gene expression amount with the preset threshold value; Optionally, when the gene expression level of ANGPTL2 is greater than the threshold of the normal stage, it is determined to be a precancerous lesion stage; when the gene expression level of ANGPTL2 is less than the threshold of the precancerous lesion stage, it is determined to be a well / moderately / poorly differentiated squamous cell carcinoma stage; Optionally, when the gene expression level of EEF1E1 is greater than the value of the normal stage, it is determined to be a well-differentiated squamous cell carcinoma stage; when the gene expression level of CK15 is less than the threshold value of the well-differentiated squamous cell carcinoma stage, it is determined to be a moderately differentiated squamous cell carcinoma stage or a poorly differentiated squamous cell carcinoma stage; Optionally, the non-elliptical type includes a fusion type, a budding type, and a finger-like propulsion type; Optionally, the progression judgment of the highly / moderately / poorly differentiated squamous cell carcinoma stage also includes phenotypic characteristics. When the gene expression level of ANGPTL2 is less than the threshold of the precancerous lesion stage and the phenotypic characteristic is a fusion type, it is determined to be a highly differentiated squamous cell carcinoma stage; when the gene expression level of ANGPTL2 is less than the threshold of the precancerous lesion stage and the phenotypic characteristic is a budding type, it is determined to be a moderately differentiated squamous cell carcinoma stage; when the gene expression level of ANGPTL2 is less than the threshold of the precancerous lesion stage and the phenotypic characteristic is a finger-like propulsion type, it is determined to be a poorly differentiated squamous cell carcinoma stage.

6. The method for predicting oral mucosal canceration based on pathological images according to claim 1, characterized in that: The S2 is replaced by: obtaining a phenotype of microvessels based on the pathological image, the phenotype including an elliptical type and a non-elliptical type, performing TSCs detection when the phenotype is non-elliptical, and performing ANGPTL2 gene detection to obtain a gene expression level when TSCs are not detected, and determining a prediction result of oral mucosal canceration when the ANGPTL2 gene expression level is greater than a first preset threshold value; and performing ANGPTL2 gene detection to obtain a gene expression level when TSCs are detected, and determining a prediction result of oral mucosal canceration when the ANGPTL2 gene expression level is less than the first preset threshold value.

7. The method for predicting oral mucosal canceration based on pathological images according to claim 1, characterized in that: The pathological image is classified by a microvascular phenotype classification model to obtain a microvascular phenotype; Optionally, the microvascular phenotype classification model is obtained by training by comparing features of microvascular features in microvascular pathology images and microvascular immunohistochemistry images; Optionally, the training process of the microvascular phenotype classification model is: Obtain microvascular pathological staining image datasets and phenotype labels, and microvascular immunohistochemistry image datasets; Segmenting the microvessels in the microvascular pathological staining image dataset and the microvascular immunohistochemical image dataset to obtain microvessels of various phenotypes of pathological staining and immunohistochemical microvessels of various phenotypes; Each phenotype of pathologically stained microvessels and phenotype labels are input into the microvessel classification model to be trained to obtain the predicted phenotypic type of pathologically stained microvessels, and the predicted pathologically stained microvessels are feature-compared with the immunohistochemical microvessels of the corresponding phenotypic type to obtain a comparison result. The model is iteratively optimized based on the comparison result until the loss function remains unchanged to obtain the microvessel classification model.

8. A computer program product comprising a computer program or instructions, characterized in that: The computer program or instructions are executed by a processor to implement the method for predicting oral mucosal canceration based on pathological images as described in any one of claims 1 to 7.

9. A computer device comprising a memory, a processor and a computer program or instruction stored in the memory, characterized in that: The computer program or instructions are executed by a processor to implement the method for predicting oral mucosal canceration based on pathological images as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program or instruction stored thereon, characterized in that: The computer program or instructions are executed by a processor to implement the method for predicting oral mucosal canceration based on pathological images as described in any one of claims 1 to 7.

Images