49 results about "Tumor Pathology" patented technology

The invention discloses a multi-modal tumor identification method based on a frequency domain attention mechanism. The method comprises the following steps: S1, acquiring multi-frequency bioelectrical impedance signals of tumor tissues and normal tissues and corresponding medical tumor pathological images; s2, extracting a frequency-time feature of the electrical impedance signal and a spatial feature of the pathological image through a multi-modal feature coding module, and generating a multi-modal fusion feature; s3, converting the multi-modal fusion features into frequency domain representation, performing weighted enhancement on the multi-modal fusion features through a frequency domain attention module, and outputting the features; and S4, fusing the features output by the frequency domain attention module and the superficial layer features in the coding stage through jump connection to obtain enhanced features, inputting the enhanced features into a decoding network, and outputting a tumor classification result. According to the method, the bioelectrical impedance signal and the multi-modal information of the pathological image are fused, and a frequency domain attention mechanism is introduced to effectively improve the expression ability of key frequency characteristics.

The invention discloses a tumor pathological image segmentation method and system based on a state space model, the system comprises a state space feature encoder, a multi-scale feature decoder, a side output module and a splicing fusion module, the state space feature encoder comprises four stages of QSSBlock feature encoders, the multi-scale feature decoder comprises a multi-scale feature fusion module and a multi-scale convolution attention module. The multi-scale feature fusion module comprises an upper convolution module and a grouping attention module. According to the method, the global features in the tumor pathological image can be effectively captured, the long-range correlation between the instances can be established, the problem that the instances are mutually independent is solved, segmentation mapping is greatly simplified, and the weak supervision method is more interpretable. By introducing a deep supervision multi-stage side output mode, the hierarchical information of the multi-scale feature map is effectively utilized, high-throughput tumor pathological image segmentation can be quickly and accurately realized, and the method has wide application prospects and important practical values in the field of medical image processing.

The invention discloses a stomach cancer postoperative survival prediction method and system based on machine learning, and belongs to the technical field of medical worker crossing and medical worker combination. According to the technical scheme, the method comprises the following steps: acquiring clinical data of a gastric cancer patient, wherein the clinical data comprises demographic characteristics, tumor pathology characteristics, operation related parameters and laboratory detection indexes; filling missing values in the clinical data by using an iterative random forest missing value filling method based on mutual information weighting; on the basis of the filled data, a feature subset with the most information content for postoperative three-year survival prediction is screened out through a dual feature selection strategy; training a machine learning model by using the feature subset so as to predict the survival risk of the gastric cancer patient in three years after operation; and outputting a prediction result. The method has the beneficial effects that a plurality of key challenges from data preprocessing, feature engineering and model construction to interpretability and clinical application are systematically solved, and an accurate, reliable, transparent and practical gastric cancer postoperative survival prediction solution is finally formed.

The invention is suitable for the field of medical image analysis, and provides a tumor pathological image segmentation method based on a U-Net neural network, and the method comprises the steps: extracting the gray, texture and shape multi-dimensional features of tumor cells through a U-Net encoder, constructing a feature cluster, building an invasion direction probability model based on the cluster, analyzing the morphological features of a necrotic region, and speculating the growth speed. And dynamically adjusting the segmentation sensitivity threshold of the decoder, and finally outputting a three-channel segmentation map containing a tumor core region, a false envelope invasion region and a capillary invasion region. According to the scheme, through multi-dimensional feature fusion and biological behavior modeling, precise characterization of tumor heterogeneity is achieved, the adaptability of a segmentation model to different invasion active areas is improved through a dynamic sensitivity regulation mechanism, a segmentation result with anatomical positioning and biological evaluation values is provided for clinic, and the segmentation accuracy is improved. And diagnosis and treatment decision of tumors are effectively assisted. The method is easy and convenient to operate and high in automation degree and has remarkable clinical application value.

The invention provides a method for inducing HL-60 cells to be differentiated into neutrophil-like cells, and belongs to the technical field of cell culture. According to the method, the induced differentiation efficiency of the HL-60 cells to the neutrophil-like cells (D-HL-60) and the cell quality are remarkably improved, high-proportion functional D-HL-60 cells can be obtained, a good motility rate can be maintained, and a stable and reliable cell model is provided for researching a neutrophil differentiation mechanism, screening related drugs and analyzing immune functions. Particularly, the functional D-HL-60 cells obtained on the basis of the method can be used as an ideal model, deep analysis of neutrophil trapping nets (NETs) in inflammation and tumor pathological mechanisms is assisted, and an experimental basis is provided for development of therapeutic strategies of targeted NETs.

The present invention relates to the field of pathology analysis technology, and more specifically to a tumor pathology analysis method based on multimodal artificial intelligence, comprising the following steps: acquiring multi-center, multi-type whole-slice tissue images, cutting the whole-slice images into pathology images, and dividing the standardized pre-processed pathology images into training and test sets; constructing a contrastive language and image model for capturing key features in pathology images through a two-stage pre-training process; and fine-tuning the model using the pre-trained weights of the contrastive language and image model as initial weights and adding task-specific fully connected layers to obtain a pathology analysis model for implementing different classification and regression tasks in tumor pathology analysis. By integrating multimodal information from images and text, the present invention can more comprehensively assess tumor type, stage, and prognosis, thereby fully leveraging the advantages of multi-source data and effectively addressing the current bottlenecks in tumor pathology analysis technology.

The invention relates to the technical field of image recognition, and particularly discloses an auxiliary blood tumor pathological diagnosis system and method based on artificial intelligence, and the system extracts a pathological image group of a patient through a pathological diagnosis auxiliary platform, analyzes the data of each pathological image, and judges the effective feature value of each pathological image; the effective feature values of the pathological images are compared with a predefined effective feature threshold value to obtain a comparison result, and the pathological diagnosis auxiliary platform judges whether the effective features of the pathological images are enhanced or not based on the comparison result; and extracting cell characteristic data in each pathological image according to each pathological image, judging the characteristic complexity of each pathological image, and performing difference analysis on the pathological image group to complete auxiliary blood tumor pathological diagnosis.

The present invention discloses an intelligent virtual staining method and device for Epstein-Barr virus-related tumor pathology slice images, and relates to the field of image intelligent processing technology. The method comprises: obtaining H&E images of in vitro tissue slices of Epstein-Barr virus-positive cells, performing EBER in situ hybridization on the slices after elution and staining, and then scanning them into slice images, performing image registration on the above two images, and obtaining training images after marking Epstein-Barr virus-positive cells; then obtaining images of in vitro tissue slices of Epstein-Barr virus-negative cells as comparison images, and training a virtual staining model that can intelligently distinguish and virtually stain Epstein-Barr virus-positive cells in H&E slice images based on the training images and the comparison images. The present invention can reduce the complexity of the operation process of artificial intelligence processing slice image technology, improve the intelligence level of artificial intelligence processing slice image technology, and replace the biological marker method with intelligent virtual staining, which has great scientific significance, social value and economic benefits.

The utility model relates to the technical field of pathological specimen manufacturing, and discloses a bone tumor pathological specimen sampling adjustable fixator which comprises a base, a fixing plate is fixedly connected to the base, at least one evenly-arranged slide making groove is formed in the fixing plate, a stroke through groove is formed in the position, above the slide making groove, in the fixing plate, and the stroke through groove is communicated with the slide making groove. A long moving rod is slidably arranged in the stroke through groove, holding plates matched with the sheet making grooves are fixedly connected to the long moving rod, electric push rods are arranged in the holding plates, pressing plates matched with the upper glass sheet are fixedly connected to the lower ends of the electric push rods, and limiting moving plates are fixedly connected to the two ends of the long moving rod; and fixed limiting plates are fixedly connected to the two sides of the base correspondingly, plate body stroke grooves matched with the limiting moving plates are formed in the fixed limiting plates correspondingly, and sliding rods are fixedly connected to the fixed limiting plates correspondingly. Compared with the prior art, the upper glass fixing device has the advantage that upper glass of different sizes can be fixed.

The invention discloses a survival prediction method based on a tumor pathological image, and the method comprises the steps: obtaining a key region of the tumor pathological image through deep learning and an attention mechanism, and obtaining a cell nucleus boundary feature, a cell nucleus texture feature and / or a cell nucleus image structure feature based on the key region of the tumor pathological image; and survival prediction is carried out through the trained machine learning model. According to the survival prediction method, the accuracy of survival prediction of tumor patients is improved, and meanwhile, the defect that a traditional deep learning method is poor in interpretability is overcome.

The present patent application relates to nanoparticles for the delivery and targeting of a non-tumour-specific antigen in cancer cells, comprising a matrix support based on a biocompatible material, the non-tumour-specific antigen, and an adjuvant, for use in recalling, in cancer patients who have a specific immunity for the non-tumour- specific antigen pre-existing to the tumour pathology, the immune response specific to the non-tumour-specific antigen against the cancer cells. A further object of the application is anti-tumour pharmaceutical formulations comprising nanoparticles and kits comprising the aforementioned anti-tumour pharmaceutical formulations in combination with traditional anti-tumour vaccines.

The invention relates to negative functional modulators of at least one variant of the non-erythrogenic erythropoietin (V-EPO) and pharmaceutical compositions or kits containing them. Such functional negative modulators of V-EPO may be a mono- or multi-specific antibody anti-EV3, anti-EV4, anti-EV1-4, anti-EV1-5, anti-EV1-1, or EV2-1, or anti-Epo receptor (EpoR) anti-EPHB4, anti-CSF2RB, an antisense oligonucleotide, DNA decoy, RNA decoy, a ribozyme, an antagomiR, a shRNA, LNA or siRNA.Several uses of these functional modulators are described, which have been advantageously employed as a medicament and for the treatment of an oncological pathology, a proliferative pathology, chronic inflammatory diseases on an autoimmune and non-autoimmune basis, of neurodegenerative diseases, and in the treatment of patients undergoing organ or tissue transplantation.The invention also describes variants of EPO for use in the diagnosis and in the treatment of an oncological pathology, a proliferative pathology, chronic inflammatory diseases on an autoimmune and non-autoimmune basis, of neurodegenerative diseases, and in the treatment of patients undergoing an organ or tissue transplantation and as a diagnostic agent.According to another aspect, a monoclonal antibody to at least one of the variants of the erythropoietin is described.According to yet another aspect, the use of at least one alternative splicing variant of non-erythrogenic EPO and the measurement thereof at tissue and / or systemic level is described, as well as the study of the methylation status of the promoters of the genes involved in the EPO signalling pathway (by way of example EPO, EPOR, EPHB4, CSF2RB), as diagnostic, prognostic and predictive markers of an oncological pathology, a proliferative pathology, neurodegenerative or inflammatory pathology.

The invention belongs to the technical field of tumor medicine, and discloses a tumor pathological specimen storage device which comprises a bottom plate, and a rotating device, a temperature control sealing device and a linkage device are arranged above the bottom plate. Through organic combination of a plurality of devices, a storage system which is compact in structure and comprehensive in function is formed. The temperature control pipe accurately controls the temperature, especially independent temperature adjustment can be carried out according to the characteristics of different tumor pathological specimens, a personalized and stable storage environment is provided for the specimens, the quality of different types of specimens is effectively guaranteed, diversified storage requirements are met, the device is more practical and scientific, and actual application and operation are facilitated.

The invention relates to the technical field of tumor pathological specimen fixing and preservation, in particular to a tumor pathological specimen fixer which comprises a fixing box, a storage drawer is inserted into the fixing box, and a locking mechanism for conveniently limiting the drawing position of the storage drawer is arranged in the fixing box. An identifying and fixing mechanism which is convenient for fixing and identifying specimens is arranged in the storage drawer, and a taking and placing mechanism which is convenient for quickly taking and placing the specimens in a non-contact manner and is convenient for taking and placing the specimens is arranged above the storage drawer. According to the tumor pathological specimen fixing device, a tumor pathological slide specimen can be conveniently clamped, stored and fixed through a movable cavity formed in a storage drawer and a lifting rod, a fixing sleeve and a buffering cushion strip arranged in a lifting groove, and a sliding seat and a second magnet are slidably connected to a sliding rail fixedly connected to the bottom of an inner cavity of the movable cavity; and the second magnet slides below the first magnet, so that the first magnet and the second magnet are conveniently utilized to jack up the fixed sleeve and the taking and placing mechanism for observation and taking.

The invention provides a tumor pathological image typing and grading method and device based on deep learning and a residual network technology, and relates to the technical field of image processing, and the method comprises the steps: collecting and preprocessing a tumor pathological image, segmenting a tissue region, carrying out the color normalization, and dividing the tumor pathological image into image blocks; constructing a deep convolutional feature extraction network based on a residual network, and extracting shallow and deep feature maps; performing up-sampling on the deep feature map and performing weighted fusion on the deep feature map and the shallow feature map, and inputting the deep feature map and the shallow feature map into parallel convolution branches to obtain typing and grading feature maps; calculating a feature deviation degree based on the hierarchical feature map to generate a space attention map, and determining a region focusing degree; enhancing the typing feature map by using the attention weight; respectively carrying out global pooling and classification on the enhanced feature map and the graded feature map, and outputting a typing and grading result; screening high-focus points according to an area focusing degree set threshold value, and generating a key area thermodynamic diagram; according to the invention, typing and grading cooperative processing is realized, and the diagnosis efficiency and accuracy are improved.

The present invention discloses a bone tumor image enhancement method and system, comprising: step 1: acquiring a bone tumor pathology image, and performing fuzzy processing on the bone tumor pathology image, using the fuzzy processed bone tumor pathology image as an input data set, and using the original bone tumor pathology image as an output data set; step 2: constructing an image super-resolution enhancement model, and training the image super-resolution enhancement model using the input data set and the output data set until the loss function value of the image super-resolution enhancement model tends to converge; step 3: acquiring the bone tumor pathology image to be enhanced in real time, inputting the trained image super-resolution enhancement model, and generating a bone tumor pathology image with completed image super-resolution enhancement. The present invention solves the problem of image distortion of subtle pathological tissue parts in pathology images caused by conventional image super-resolution enhancement methods.

The present invention discloses a tumor pathology image segmentation method and system based on a graph convolutional contrastive learning network, which relates to the field of image segmentation technology. The method includes: obtaining a tumor pathology image and a segmentation label corresponding to the tumor pathology image, generating a tumor pathology image training set based on the tumor pathology image and the segmentation label corresponding to the tumor pathology image; preprocessing the tumor pathology image training set, inputting the preprocessed tumor pathology image training set into a pre-established offset graph convolutional contrastive learning network model for training, and obtaining a trained offset graph convolutional contrastive learning network model; receiving a tumor pathology image to be segmented, inputting the tumor pathology image to be segmented into the trained offset graph convolutional contrastive learning network model, and outputting a tumor pathology image segmentation result to achieve accurate image segmentation.

The invention relates to the technical field of medical instruments, and discloses a bone tumor pathological tissue sampling device integrated with a cutting function, which comprises a left side shell and a right side shell which are connected through a screw, and also comprises a cutting assembly and a tumor sampling assembly, the front ends of the left side shell and the right side shell are connected with end covers through screws, the end faces of the end covers are fixedly connected with sleeves, and the outer walls of the sleeves are fixedly connected with lug blocks and side blocks. Sampling and cutting functions are ingeniously combined, after cutting is completed, a buckle plate is pulled to drive a string, a micro winch and other parts, so that a biting block rotates around a vertical shaft to be close to a cutting blade, the biting block and the cutting blade are matched to bite tumor pathological tissue, and sawteeth on the side edge of the biting block facilitate easy sampling; by means of the integrated design, the sampling process of bone tumor pathological tissue is effectively simplified, the tedious step of switching multiple tools in traditional operation is avoided, and the sampling efficiency is greatly improved.

The invention discloses an intelligent closestool control system which comprises a negative pressure suction module, a deodorization module, a tumor disease detection module, a urination function health state evaluation module and an interaction module. An intelligent closestool control method comprises the following steps that S1, the working mode of a negative pressure pump is adjusted; s2, starting a deodorization function of the intelligent closestool and a ventilation system of the toilet; s3, collecting part of the excreta diluent as a sample liquid, obtaining and analyzing genome sequence information of cells in the sample liquid, and analyzing and predicting the correlation of the genome sequence information and tumor pathology by using bioinformatics; s4, judging whether the user has urination dysfunction or not through continuous monitoring and analysis of the urine flow rate data and the urine volume data; and S5, obtaining a tumor disease analysis result and a urination function health state analysis result, and sending the analysis results to the user side APP and the intelligent closestool health detection display screen. The invention further discloses a closestool intelligent control method corresponding to the system and a closestool applying the intelligent control system.

The present invention belongs to the field of medical device technology and discloses a tumor pathology section dehydration device, comprising a workbench, a mounting block mounted on the workbench via a locking assembly; a dehydration cylinder detachably connected to the mounting block, a dehydration box for containing tissue samples disposed in the dehydration cylinder; an emptying assembly comprising an emptying pipe disposed in an assembly slot, the top end of the emptying pipe being connected to the dehydration cylinder, and the bottom end of the emptying pipe being respectively connected to a waste liquid cylinder and an exhaust gas cylinder disposed in the workbench; a centrifugal assembly disposed on the workbench, the output end of the centrifugal assembly being transmission-connected to the dehydration box to drive the dehydration box for centrifugal dehydration; and a liquid inlet assembly disposed on the workbench and connected to the dehydration box. The present invention has a compact structure, a high degree of automation, and organically combines physical dehydration and chemical dehydration, thereby overcoming the shortcomings of physical dehydration and chemical dehydration, improving the dehydration efficiency and quality of tissue samples, reducing dehydration costs, and accelerating the efficiency and accuracy of determining tumor properties.

The invention discloses a brain tumor specimen space positioning and section manufacturing device, and belongs to the technical field of brain tumor pathological sections, and the brain tumor specimen space positioning and section manufacturing device comprises a bottom plate, a fixing frame, a directional puncture device, tumor forceps and a proofreading device. The system can automatically trigger and record the space coordinates at two key action points for clamping and releasing a tumor, accurately calculates the position and angle change of the tumor from the inside to the outside by calculating a space transformation matrix, solves the fundamental problem of in-vitro specimen orientation loss, and improves the accuracy of the in-vitro specimen. A two-stage posture adjusting system composed of a rotating assembly and an angle fine adjustment assembly is used for driving a puncture needle to conduct puncture and dyeing marking at the calculated accurate angle, the marking direction corresponds to the original direction of a tumor in a body, and a reliable space reference is provided for pathological analysis; the calibration device is arranged to perform multi-attitude high-precision calibration on the three-dimensional directional monitoring assembly of the tumor forceps.

The invention discloses an iodine contrast agent adverse reaction risk prediction method and system for tumor patients, and relates to the technical field of medical data processing and risk prediction. According to the technical key points, the method comprises the following steps of: constructing a tumor patient exclusive multi-dimensional time sequence data set, preprocessing data, constructing a tumor patient exclusive risk factor system, constructing a pharmacological perception drug embedding layer, constructing a multi-scale self-adaptive time sequence feature extraction core module, and coupling and modulating a tumor pathological state and a time sequence risk. Performing multi-feature fusion and risk prediction model training verification, risk prediction and interpretability clinical decision support and clinical iterative optimization of the model; by constructing a multi-dimensional data set, time sequence modes of different anti-tumor drugs in different short, medium and long risk periods are dynamically captured, and the specific pathological state of a tumor patient is used as a modulation factor to non-linearly amplify or weaken the risk intensity of a corresponding time window, so that the accuracy and clinical credibility of risk prediction are improved.

The invention relates to the technical field of tumor immunity, and discloses a sample treatment method for detecting a circulating mononuclear cell immune function and a tumor microenvironment macrophage immune function, which comprises the following steps: treating a peripheral blood sample and a tumor pathological tissue sample; according to the treatment method, the mononuclear cell purity is improved through CD14 magnetic bead sorting, the macrophage recovery rate is improved through mixed enzymatic hydrolysate, and the core improvement that mild reagents (a hypotonic hemolysis buffer solution and a BSA-containing stationary solution) protect the cell activity and the antigen integrity is achieved, so that the treatment method has the important significance in the aspects of the mononuclear cell purity, the macrophage recovery rate, the cell activity and the antigen fluorescence intensity. Compared with the prior art, the method provided by the invention has the advantages that the method provided by the invention has the advantages that the method provided by the invention is obviously superior to a conventional treatment method, the defects of low monocyte enrichment efficiency, poor macrophage recovery rate and insufficient cell activity and antigen integrity of the conventional method are overcome, and a higher-quality sample basis is provided for subsequent circulating monocyte subtype typing (Mo1-Mo4) and tumor microenvironment macrophage immune function detection.

The invention discloses application of a reagent for detecting RELL2 gene expression quantity in preparation of a glioblastoma prognosis preparation, and relates to the field of tumor molecular biology. Researches find that by detecting the transcription expression level of the RELL2 gene in a glioblastoma patient, the prognosis condition of the glioblastoma can be detected in a high-sensitivity and high-specificity mode, prognosis layering is independent of tumor pathological staging, the RELL2 gene can be applied to primary and recurrent glioblastoma patients, and the RELL2 gene can be used for detecting the prognosis condition of the glioblastoma in a high-sensitivity and high-specificity mode. A powerful molecular biology basis is provided for prognosis detection of glioblastoma, important guiding significance is provided for precise medical treatment of patients, and the kit has extremely high clinical application value.

The invention provides a tumor pathological image detection method. The tumor pathological image detection method comprises the following steps: sectioning an in-vitro tissue separated from a tumor focus along a preset plane to form a sectioning plane; the coloring agent is used for coloring the section cutting plane to form a to-be-detected object; translating, scanning and imaging along the profile contour of the to-be-detected object by using an imaging device so as to continuously obtain a plurality of local pathological images; and splicing the local pathological images according to the imaging sequence to form an annular image at least covering the profile contour, judging whether the annular image contains tumor cells or not through a machine learning algorithm, and judging that the in-vitro tissue contains a complete tumor when the annular image does not contain the tumor cells. According to the invention, the calculation cost for judging whether the annular image contains the tumor cells or not by a machine learning algorithm is remarkably reduced, and the detection time is remarkably shortened.

The present application relates to the technical field of necrosis region segmentation, and particularly relates to a kind of automatic segmentation method of digestive tract tumor pathological image necrosis region.The present application obtains the necrosis probability of each image block according to the pixel value distribution of different pixel points in each image block and the morphological characteristics of different regions;for any image block, according to the pixel value distribution of different regions, the necrosis probability and the preset pixel fixed step, obtain the new clustering center corresponding to the cytoplasm region in each image block, and the iteration clustering of all pixel points is carried out until the preset iteration number is reached, to obtain the optimized clustering center of cytoplasm region;according to the optimized clustering center corresponding to the cytoplasm region in each image block, clustering is carried out, the new necrosis probability of each image block is analyzed, and the necrosis block is screened out for segmentation.The present application improves the accuracy of segmentation by adaptively adjusting the clustering center in the clustering algorithm.

The invention provides a tumor prediction method based on a hierarchical visual language and electronic equipment, the method is applied to the field of artificial intelligence, and the method comprises the following steps: obtaining a three-dimensional medical image of a target patient and a tumor region of interest, key sub-region features of the region are extracted to obtain a local semantic feature set, tumor and adjacent tissue features are extracted to obtain a global context feature set, and the local semantic feature set and the global context feature set form a visual feature set; the method comprises the following steps: constructing clinical prior query of a target cancer species and an image mode, generating prognosis related language prompts through a large language model, encoding, solving an optimal matching plan matrix of a semantic matching optimal problem, and screening an elite feature subset. And enhancing a global feature set through gating fusion, applying a contrast learning consistency constraint, fusing features to form a representation vector, inputting the representation vector into a prediction layer, and obtaining a survival risk score. According to the method, priori knowledge and tumor pathological features can be fused, so that an accurate and reliable survival risk score is finally output, and the method is more reliable to adapt to clinical actual evaluation requirements.

The invention discloses a tumor chemotherapy effect prediction model construction method and system based on artificial intelligence. The method comprises the following steps: multi-dimensional data acquisition: acquiring tumor biological data, individual clinical heterogeneity data and patient report outcome data of a target cancer type patient; performing feature engineering processing: performing quantitative processing on the individual clinical heterogeneity data to generate quantitative indexes, and performing feature extraction on the tumor biological data to form a tumor microenvironment feature set; basic model training: taking multi-dimensional data of large-sample general cancer species as a training set, adopting a multi-branch feature fusion network to construct a general chemotherapy effect prediction basic model, and taking chemotherapy effect prediction and life quality influence prediction as double optimization targets for training, according to the invention, by collecting tumor biological data, individual clinical heterogeneity data and patient report outcome data, a multi-dimensional data system is constructed, and tumor pathological characteristics and microenvironment states can be comprehensively covered.

The application discloses a kind of multi-view child tumor pathological image classification deep network incremental learning method, this method is first to the child tumor pathological image feature modeling, then under the condition of constantly obtaining new data, training obtains incremental learning model, to improve its accuracy and obtain robust classification ability.The innovation of the present application lies in the characteristics of medical pathological image are designed, a multi-level knowledge distillation regularization method is proposed to alleviate the catastrophic forgetting problem of old model in incremental learning when adapting to new data, increase the robustness and reliability of algorithm.The method of the present application can be applied to the field of medical image processing and other tasks requiring incremental learning.

The invention relates to an adrenal cortical carcinoma prognosis evaluation method based on a peripheral blood inflammation marker, which comprises the following steps: acquiring a peripheral blood detection result and tumor pathological characteristics of a patient before treatment, and calculating a derived neutrophil-lymphocyte ratio and a platelet-leukocyte ratio based on the peripheral blood detection result; inputting the derived neutrophil-lymphocyte ratio, the platelet-leukocyte ratio, the age of the patient and the tumor pathological characteristics into a prognostic scoring model to obtain a risk score; wherein the prognosis scoring model is a regression model which is screened by adopting Cox regression analysis and is established with variables related to the postoperative disease-free lifetime. According to the invention, the accuracy of prognosis evaluation of the adrenal cortex cancer can be improved.