Monitoring system and method for preoperative immunotherapy microenvironment remodeling

Abstract

The invention discloses a monitoring system for preoperative immunotherapy microenvironment remodeling, which comprises a clinical data acquisition module, a magnetic resonance image acquisition module, a first prediction module and a monitoring module, and in addition, the invention also provides a monitoring method for preoperative immunotherapy microenvironment remodeling. According to the invention, the parameter change of the microenvironment cells in the tumor region of interest of the to-be-predicted patient is combined with the clinical data information, and the absolute value of the TIL infiltration density and the change of the TIL infiltration density in the treatment process of the to-be-predicted patient are predicted on the basis of the pre-trained prediction model; and the immunotherapy microenvironment remodeling of the to-be-predicted patient is indirectly monitored.

Description

Technical Field

[0001] The present application relates to the field of bioinformatics, and in particular to a monitoring system and method for preoperative immunotherapy microenvironment remodeling. Background Art

[0002] Immune checkpoint inhibitors (ICI) therapy is a new clinical tumor treatment method. During treatment, the immune microenvironment remodeling process affects the efficacy, and the efficacy can be predicted in real time by monitoring this process.

[0003] In the relevant technologies, the most important and most recognized microenvironment remodeling factor that affects the effect of immunotherapy is the local aggregation of tumor infiltration lymphocytes (TIL) in the tumor. Since the average diameter of human TIL cells is significantly smaller than that of most cancer cells in solid tumors, as immunotherapy takes effect, the increase in TIL infiltration is accompanied by an increase in tumor cell apoptosis and necrosis, and the average cell diameter in the local tumor microenvironment will be significantly reduced. In other words, the higher the TIL density, the better the efficacy. However, this process cannot be monitored in the entity at present, that is, neither conventional imaging nor fluorescently labeled molecular imaging can be used to monitor the TIL infiltration density of patients. Summary of the invention

[0004] In view of this, the present application provides a monitoring system and method for preoperative immunotherapy microenvironment remodeling, which is used to solve the problem in the prior art that neither conventional imaging nor fluorescently labeled molecular imaging can be applied to the monitoring of TIL infiltration density in patients.

[0005] According to the first group of aspects of the present application, a monitoring system for preoperative immunotherapy microenvironment remodeling is provided, comprising: a clinical data acquisition module, a magnetic resonance image acquisition module, a first prediction module and a monitoring module, wherein:

[0006] The clinical data acquisition module is used to acquire clinical data information of the patient to be predicted;

[0007] The magnetic resonance image acquisition module is used to acquire two sets of magnetic resonance diffusion weighted images marked with tumor regions of interest, wherein the two sets of magnetic resonance diffusion weighted images include pulse gradient spin echo sequence images and oscillating gradient spin echo sequence images obtained by magnetic resonance diffusion weighted time diffusion spectrum imaging function;

[0008] The first prediction module is used to input the clinical data information obtained by the clinical data acquisition module and the two groups of magnetic resonance diffusion weighted images obtained by the magnetic resonance image acquisition module into a pre-trained prediction model, so as to predict the absolute value of TIL infiltration density and the change of TIL infiltration density of the patient to be predicted during the treatment process according to the microenvironment cell parameters in the tumor region of interest in the magnetic resonance diffusion weighted image and the clinical data information through the prediction model, wherein the microenvironment cell parameters are obtained by fitting the image features in the tumor region of interest in the two groups of magnetic resonance diffusion weighted images;

[0009] The monitoring module is used to monitor the preoperative immunotherapy microenvironment remodeling of the patient to be predicted based on the absolute value of the TIL infiltration density and the change of the TIL infiltration density predicted by the first prediction module.

[0010] Furthermore, the magnetic resonance image acquisition module includes an image acquisition unit, a region of interest delineation unit and an image matching unit, wherein:

[0011] The image acquisition unit is used to acquire two sets of magnetic resonance diffusion weighted images of the patient to be predicted through a preset storage path;

[0012] The region of interest delineation unit is used to delineate and annotate the tumor contour range frame by frame on one of the two groups of magnetic resonance diffusion weighted images obtained by the image acquisition unit by an operator delineation method to obtain a tumor region of interest, so that one of the two groups of magnetic resonance diffusion weighted images is annotated with the tumor region of interest;

[0013] The image matching unit is used to match another set of magnetic resonance diffusion weighted images according to a set of magnetic resonance diffusion weighted images annotated with tumor regions of interest obtained by the region of interest delineation unit, so that the two sets of magnetic resonance sequence diffusion weighted images are annotated with the same tumor regions of interest.

[0014] Furthermore, the system also includes: a training sample acquisition module, a model training module and a model construction module, wherein:

[0015] The training sample acquisition module is used to acquire training sample data of the machine learning model before inputting the clinical data information obtained by the clinical data acquisition module and the two sets of magnetic resonance diffusion weighted images obtained by the magnetic resonance image acquisition module into the pre-trained prediction model, wherein the training sample data includes training input data and training output data;

[0016] The model training module is used to input the training sample data obtained by the training sample acquisition module into the machine learning model, so as to train the mapping relationship between the training input data and the training output data through the machine learning model;

[0017] The model building module is used to build a prediction model according to the mapping relationship obtained through training by the model training module when the training meets the iteration stopping condition.

[0018] Furthermore, the training sample acquisition module includes: an input data construction unit and an output data construction unit, wherein:

[0019] The input data construction unit is used to construct training input data according to clinical information of different patient samples and two groups of magnetic resonance diffusion weighted images annotated with corresponding tumor regions of interest;

[0020] The output data construction unit is used to construct training output data according to the absolute value of TIL infiltration density and the change of TIL infiltration density of different patient samples at different time points during the treatment process.

[0021] Furthermore, the system also includes: a CD8 expression score input module and a median value calculation module, wherein:

[0022] The CD8 expression score input module is used to input the CD8 expression score representing the CD8 expression level at different sites of the tumor obtained by immunohistochemical staining of multiple biopsy tissue sections of the patient sample at the time point to be monitored during immunotherapy in the training input data, so as to obtain the absolute value and change of the CD8 molecule expression score of the tumor microenvironment of the patient sample during the immunotherapy process;

[0023] The median calculation module is used to calculate the median value of multiple groups of CD8 molecule expression score absolute values ​​and change values ​​relative to the initial expression score according to the absolute value and change of the tumor microenvironment CD8 molecule expression score of the patient sample obtained by the CD8 expression score input module during immunotherapy. The median value reflects the actual TIL infiltration density absolute value and TIL infiltration density change of the patient sample at a preset time point during the treatment process.

[0024] Furthermore, the model training module includes a training parameter acquisition unit and a model training unit, wherein:

[0025] The training parameter acquisition unit is used to obtain microenvironment cell parameters and clinical factors that significantly affect the absolute value of TIL infiltration density and the change of TIL infiltration density in the immunotherapy microenvironment based on the training input data;

[0026] The model training unit is used to input the microenvironment parameters and clinical factors obtained by the training parameter acquisition module into the machine learning model for training. During the training process, the absolute value of TIL infiltration density and the change of TIL infiltration density of the corresponding patient samples at different time points during immunotherapy are used as training output data to train the mapping relationship between the training input data and the training output data.

[0027] Furthermore, the training parameter acquisition unit fits the image features within the tumor region of interest in the two groups of magnetic resonance diffusion weighted images on the basis of the training input data to obtain the microenvironment cell parameters within the tumor region of interest; performs univariate regression analysis on the clinical data information of the different patient samples to determine the clinical factors that affect the absolute value of TIL infiltration density and the change of TIL infiltration density in the immunotherapy microenvironment; performs multivariate logistic regression analysis on the microenvironment cell parameters within the tumor region of interest and the clinical factors that affect the absolute value of TIL infiltration density and the change of TIL infiltration density in the immunotherapy microenvironment to obtain the microenvironment cell parameters and clinical factors that significantly affect the absolute value of TIL infiltration density and the change of TIL infiltration density in the immunotherapy microenvironment.

[0028] Furthermore, the training parameter acquisition unit uses a preset fitting algorithm to fit the image features in the tumor region of interest in two sets of magnetic resonance diffusion weighted images for a preset number of times based on the training input data; after randomly initializing and repeating the fitting for a preset number of times, the microenvironment cell parameters in the tumor region of interest are obtained, and the microenvironment cell parameters include the average cell diameter, intracellular volume fraction, extracellular diffusion rate and cell density in the tumor region of interest.

[0029] Furthermore, the system also includes: a verification parameter acquisition module, a comparison module and an evaluation module, wherein:

[0030] The verification parameter acquisition module is used to obtain the absolute value of TIL infiltration density and the change of TIL infiltration density evaluated in real time during the treatment of the patient to be verified;

[0031] The comparison module is used to compare the absolute value of TIL infiltration density and the change of TIL infiltration density obtained by the verification parameter acquisition module during the treatment process as the actual measurement result with the absolute value of TIL infiltration density and the change of TIL infiltration density predicted by the prediction model to obtain a receiver operating characteristic curve predicted by the model;

[0032] The evaluation module is used to evaluate the prediction accuracy of the pre-trained prediction model according to the receiver operating characteristic curve predicted by the model.

[0033] According to a second aspect of the present application, a method for monitoring preoperative immunotherapy microenvironment remodeling is provided, comprising:

[0034] S1. A clinical data acquisition module acquires clinical data information of a patient to be predicted, and transmits the clinical data information to a first prediction module;

[0035] S2, the magnetic resonance image acquisition module acquires two sets of magnetic resonance diffusion weighted images marked with tumor regions of interest, and transmits the two sets of magnetic resonance diffusion weighted images to the first prediction module, wherein the two sets of magnetic resonance diffusion weighted images include pulse gradient spin echo sequence images and oscillating gradient spin echo sequence images obtained by magnetic resonance diffusion weighted time diffusion spectrum imaging function; specifically including:

[0036] S2.1, the image acquisition unit acquires two sets of magnetic resonance diffusion weighted images of the patient to be predicted through a preset storage path;

[0037] S2.2, the region of interest delineation unit delineates and annotates the tumor contour range frame by frame on one of the two groups of magnetic resonance diffusion weighted images obtained by the image acquisition unit to obtain a tumor region of interest, so that one of the two groups of magnetic resonance diffusion weighted images is annotated with the tumor region of interest;

[0038] S2.3, the image matching unit matches another set of magnetic resonance diffusion weighted images according to the set of magnetic resonance diffusion weighted images annotated with the tumor region of interest obtained by the region of interest delineation unit, so that the two sets of magnetic resonance sequence diffusion weighted images are annotated with the same tumor region of interest;

[0039] S3, the first prediction module inputs the clinical data information obtained by the clinical data acquisition module and the two groups of magnetic resonance diffusion weighted images obtained by the magnetic resonance image acquisition module into a pre-trained prediction model, so as to predict the absolute value of TIL infiltration density and the change of TIL infiltration density of the patient to be predicted during the treatment process according to the microenvironment cell parameters in the tumor region of interest in the magnetic resonance diffusion weighted image and the clinical data information through the prediction model, wherein the microenvironment cell parameters are obtained by fitting the image features in the tumor region of interest in the two groups of magnetic resonance diffusion weighted images;

[0040] S4. The monitoring module monitors the remodeling of the preoperative immunotherapy microenvironment of the patient to be predicted based on the absolute value of the TIL infiltration density and the change of the TIL infiltration density predicted by the first prediction module.

[0041] After step S2, the method further comprises:

[0042] S5, a feature fitting module performs feature fitting on the two groups of magnetic resonance diffusion weighted images obtained by the magnetic resonance image acquisition module to obtain microenvironment cell parameters in the tumor region of interest;

[0043] S6. The analysis module screens out clinical factors that significantly affect the absolute value of TIL infiltration density and the change of TIL infiltration density in the immunotherapy microenvironment according to the clinical data information obtained by the clinical data acquisition module, and combines the clinical factors that significantly affect the absolute value of TIL infiltration density and the change of TIL infiltration density in the immunotherapy microenvironment with the microenvironment cell parameters to perform multivariate logistic regression analysis to obtain microenvironment cell parameters and clinical factors that significantly affect the absolute value of TIL infiltration density and the change of TIL infiltration density in the immunotherapy microenvironment;

[0044] Correspondingly, S3, the first prediction module inputs the microenvironment cell parameters and clinical factors that significantly affect the absolute value of TIL infiltration density and the change of TIL infiltration density in the immunotherapy microenvironment obtained by the analysis module into the pre-trained prediction model, so as to predict the absolute value of TIL infiltration density and the change of TIL infiltration density of the patient to be predicted during the treatment process through the prediction model;

[0045] Before step S3, the method further includes:

[0046] S7, the training sample acquisition module acquires the training sample data of the machine learning model, wherein the training sample data includes training input data and training output data; specifically includes:

[0047] S7.1. Construct training input data based on clinical information of different patient samples and two sets of magnetic resonance diffusion-weighted images annotated with corresponding tumor regions of interest;

[0048] S7.2. Construct training output data based on the absolute value of TIL infiltration density and the change of TIL infiltration density of different patient samples at different time points during treatment;

[0049] S8, the model training module inputs the training sample data obtained by the training sample acquisition module into the machine learning model, so as to train the mapping relationship between the training input data and the training output data through the machine learning model; specifically including:

[0050] S8.1. Based on the training input data, obtain microenvironment cell parameters and clinical factors that significantly affect the absolute value of TIL infiltration density and the change of TIL infiltration density in the immunotherapy microenvironment;

[0051] S8.2, inputting the microenvironment parameters and significant clinical factors into the machine learning model for training, using the absolute value of TIL infiltration density and the change of TIL infiltration density of the corresponding patient samples at different time points during the treatment process as training output data, and training the mapping relationship between the training input data and the training output data;

[0052] S9. When the training of the model building module meets the iteration stopping condition, a prediction model is built according to the mapping relationship obtained by the training of the model training module.

[0053] Beneficial effects of the present invention: Compared with the current methods of using conventional imaging or fluorescently labeled molecular imaging to achieve preoperative immunotherapy microenvironment remodeling, the present application combines two groups of magnetic resonance diffusion-weighted images marked with tumor regions of interest to detect the unique effects of various microstructures in the cell body on the diffusion of water molecules, so as to accurately reflect the structural characteristics of microenvironment cells. On the basis of a pre-trained prediction model, the absolute value of TIL infiltration density and the change of TIL infiltration density during the treatment of the patient to be predicted are predicted by combining the changes in microenvironment cell parameters in the tumor region of interest with clinical data information, thereby applying the absolute value of TIL infiltration density and the change of TIL infiltration density to preoperative immunotherapy microenvironment remodeling, so as to monitor the immunotherapy microenvironment remodeling of the patient to be predicted.

[0054] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0056] Figure 1 It is a structural block diagram of a monitoring system for preoperative immunotherapy microenvironment remodeling provided in an embodiment of the present application;

[0057] Figure 2 is a line chart of relevant parameters in the model prediction process provided by the embodiment of the present application;

[0058] Figure 3 is the receiver operating characteristic curve predicted by the model provided in the embodiments of the present application;

[0059] Figure 4 This is a structural block diagram of another monitoring system for preoperative immunotherapy microenvironment remodeling provided in an embodiment of the present application. DETAILED DESCRIPTION

[0060] Now the content of the present invention will be discussed with reference to several exemplary embodiments. It should be understood that these embodiments are discussed only to enable those skilled in the art to better understand and thus implement the content of the present invention, rather than implying any limitation on the scope of the present invention.

[0061] As used herein, the term "including" and variations thereof are to be interpreted as open-ended terms meaning "including but not limited to." The term "based on" is to be interpreted as "based, at least in part, on." The terms "a set of embodiments" and "an embodiment" are to be interpreted as "at least one set of embodiments." The term "another set of embodiments" is to be interpreted as "at least one set of other embodiments."

[0062] In the related technologies, the most important and most recognized microenvironment remodeling factor that affects the effect of immunotherapy is the local accumulation of tumor infiltration lymphocytes (TIL) in the tumor. Since the average diameter of human TIL cells is significantly smaller than that of most cancer cells in solid tumors, as immunotherapy takes effect, the increase in TIL infiltration is accompanied by an increase in tumor cell apoptosis and necrosis, and the average cell diameter in the local tumor microenvironment will be significantly reduced. In other words, the higher the TIL density, the better the efficacy. However, currently, conventional imaging or fluorescently labeled molecular imaging cannot be applied to the monitoring of TIL infiltration density in patients, making it difficult to achieve real-time efficacy evaluation in vivo.

[0063] In order to solve this problem, this embodiment provides a monitoring system for preoperative immunotherapy microenvironment remodeling, such as Figure 1 As shown, it includes: a clinical data acquisition module 11, a magnetic resonance image acquisition module 12, a first prediction module 13 and a monitoring module 14, wherein:

[0064] The clinical data acquisition module 11 is used to acquire clinical data information of the patient to be predicted;

[0065] The magnetic resonance image acquisition module 12 is used to acquire two sets of magnetic resonance diffusion weighted images marked with tumor regions of interest, wherein the two sets of magnetic resonance diffusion weighted images include pulse gradient spin echo sequence images and oscillating gradient spin echo sequence images obtained by magnetic resonance diffusion weighted time diffusion spectrum imaging function;

[0066] The first prediction module 13 is used to input the clinical data information obtained by the clinical data acquisition module and the two groups of magnetic resonance diffusion weighted images obtained by the magnetic resonance image acquisition module into a pre-trained prediction model, so as to predict the absolute value of TIL infiltration density and the change of TIL infiltration density of the patient to be predicted during the treatment process according to the microenvironment cell parameters in the tumor region of interest in the magnetic resonance diffusion weighted image and the clinical data information through the prediction model, wherein the microenvironment cell parameters are obtained by fitting the image features in the tumor region of interest in the two groups of magnetic resonance diffusion weighted images;

[0067] The monitoring module 14 is used to monitor the preoperative immunotherapy microenvironment remodeling of the patient to be predicted based on the absolute value of the TIL infiltration density and the change of the TIL infiltration density predicted by the first prediction module.

[0068] In the above-mentioned clinical data acquisition module, the clinical data information includes but is not limited to the patient's gender, age, height, weight, blood routine and serological tumor marker results, preoperative clinical staging, degree of differentiation of living tumor and the adopted immunotherapy plan. For example, the preoperative clinical staging here can be the preoperative TNM clinical staging of rectal cancer, and correspondingly, the degree of differentiation of living tumor can be the degree of differentiation of living tumor by colonoscopy.

[0069] It is understandable that by real-time monitoring of changes in relevant parameters such as the average cell diameter in the predicted tumor region of interest during preoperative immunotherapy, the TIL infiltration and cell apoptosis and necrosis process in the microenvironment of the predicted tumor region of interest during treatment can be indirectly monitored, thereby realizing the monitoring of preoperative immunotherapy microenvironment remodeling.

[0070] Furthermore, the magnetic resonance image acquisition module includes an image acquisition unit, a region of interest delineation unit and an image matching unit, wherein:

[0071] The image acquisition unit is used to acquire two sets of magnetic resonance diffusion weighted images of the patient to be predicted through a preset storage path;

[0072] The region of interest delineation unit is used to delineate and annotate the tumor contour range frame by frame on one of the two groups of magnetic resonance diffusion weighted images obtained by the image acquisition unit by an operator delineation method to obtain a tumor region of interest, so that one of the two groups of magnetic resonance diffusion weighted images is annotated with the tumor region of interest;

[0073] The image matching unit is used to match another set of magnetic resonance diffusion weighted images according to a set of magnetic resonance diffusion weighted images annotated with tumor regions of interest obtained by the region of interest delineation unit, so that the two sets of magnetic resonance sequence diffusion weighted images are annotated with the same tumor regions of interest.

[0074] In the above-mentioned magnetic resonance image acquisition module, the two sets of magnetic resonance diffusion weighted images include pulsed gradient spin echo sequence images and oscillating gradient spin echo sequence images obtained by the magnetic resonance diffusion weighted time diffusion spectrum imaging function. Both sets of magnetic resonance diffusion weighted images are common medical format images, i.e., nii format images, where the tumor region of interest can be delineated in the two sets of magnetic resonance diffusion weighted images by professional medical personnel, and the tumor regions of interest in the two magnetic resonance diffusion weighted images are consistent.

[0075] Exemplarily, the pulse gradient spin echo sequence images and oscillating gradient spin echo sequence images of the patient's MRI pre-stored in the electronic computer system space can be retrieved through the path provided by the operator. The sequence images are all nii format images. Then, a professional doctor outlines the tumor region of interest of the pulse gradient spin echo sequence image, and based on the tumor region of interest in the pulse gradient spin echo sequence image, the oscillating gradient spin echo sequence image is outlined by matching, so that the tumor regions of interest in the two magnetic resonance diffusion weighted images are consistent. It is further explained that the region of interest is a part of the image, which is generated by selecting on the image or using methods such as setting a threshold or converting from other files (such as vectors). The region of interest can be irregular shapes such as points, lines, and surfaces, and is usually used as samples, masks, cropping areas, and other operations for image classification.

[0076] Currently, there is no non-invasive method for detecting cell volume in the human body. Diffusion-weighted imaging (DWI) is a commonly used clinical magnetic resonance diffusion weighting that indirectly reflects the cell density in tissues. Its principle is based on pulsed gradient spin echo (PGSE) to detect the diffusion of water molecules in tissues. However, due to the long gradient diffusion time of conventional DWI sequences (about 30 to 100 ms) and the single time dimension, the obtained apparent diffusion coefficient represents the average diffusion characteristics of all microstructures in each voxel, which reduces the specificity of detecting cell size.

[0077] In order to solve the above problems, temporal diffusion spectroscopy (TDS) imaging combines two types of scanning sequences with different diffusion times, PGSE and oscillating gradient spin echo (OGSE). After discrete point acquisition, they are arranged in order according to time, and the information of tissue cells of different sizes is fitted using computational models such as Imaging Microstructural Paramters Using Limited Spectrally Edited Diffusion (IMPULSED). Since OGSE uses an oscillating gradient to achieve a shorter diffusion time (about 5-10ms), its quantifiable cell diameter is 10-20um, which is close to the size of general tumor cells. In addition, the combination of PGSE and OGSE can detect the independent effects of various pericellular structures in cells on the diffusion of water molecules, and its quantitative parameters can more accurately reflect its microstructural characteristics such as cell size, and construct a habitat map of cell distribution. Therefore, the application of TDS can evaluate the prognosis of TILs and tumor cells in real time according to the average cell size and changes in the tumor area of ​​interest without cell labeling, and effectively monitor the microenvironment remodeling during immunotherapy.

[0078] Furthermore, the system also includes: a training sample acquisition module, a model training module and a model construction module, wherein:

[0079] The training sample acquisition module is used to acquire training sample data of the machine learning model before inputting the clinical data information obtained by the clinical data acquisition module and the two sets of magnetic resonance diffusion weighted images obtained by the magnetic resonance image acquisition module into the pre-trained prediction model, wherein the training sample data includes training input data and training output data;

[0080] The model training module is used to input the training sample data obtained by the training sample acquisition module into the machine learning model, so as to train the mapping relationship between the training input data and the training output data through the machine learning model;

[0081] The model building module is used to build a prediction model according to the mapping relationship obtained through training by the model training module when the training meets the iteration stopping condition.

[0082] Furthermore, the training sample acquisition module includes: an input data construction unit and an output data construction unit, wherein:

[0083] The input data construction unit is used to construct training input data according to clinical information of different patient samples and two groups of magnetic resonance diffusion weighted images annotated with corresponding tumor regions of interest;

[0084] The output data construction unit is used to construct training output data according to the absolute value of TIL infiltration density and the change of TIL infiltration density of different patient samples at different time points during the treatment process.

[0085] The system further comprises: a CD8 expression score input module and a median value calculation module, wherein:

[0086] The CD8 expression score input module is used to input the CD8 expression score representing the CD8 expression level at different sites of the tumor obtained by immunohistochemical staining of multiple biopsy tissue sections of the patient sample at the time point to be monitored during immunotherapy in the training input data, so as to obtain the absolute value and change of the CD8 molecule expression score of the tumor microenvironment of the patient sample during the immunotherapy process;

[0087] The median calculation module is used to calculate the median value of multiple groups of CD8 molecule expression score absolute values ​​and change values ​​relative to the initial expression score according to the absolute value and change of the tumor microenvironment CD8 molecule expression score of the patient sample obtained by the CD8 expression score input module during immunotherapy. The median value reflects the actual TIL infiltration density absolute value and TIL infiltration density change of the patient sample at a preset time point during the treatment process.

[0088] It is understandable that before building a prediction model, it is necessary to add a CD8 expression score to the training input data. Specifically, the CD8 expression score can be obtained by the expression level and change of CD8 molecules in the biopsy specimens of different patient samples and the surgical specimens after immunotherapy. CD8 is a TIL molecular marker, and its expression level and the molecular expression level and change of immunohistochemical chromosome markers can reach the infiltration density and change of TIL molecules in the microenvironment. Specifically, the image processing software can be used to calculate the expression level of CD8 immunohistochemically stained sections, and the formula H-score = [(1×% cells intensity = 1) + (2×% cells intensity = 2) + (3×% cells intensity = 3)] is used. The image processing software automatically calculates the H-score of CD8 expression, and the H-score of CD8 expression is a continuous variable.

[0089] It should be noted that in the model training stage, the training input data includes the changes in the expression of CD8 molecules in the tumor microenvironment of the patient samples during the immunotherapy process, so that the mapping relationship between the training input data and the training output data is trained through the machine learning model based on the changes in the expression of CD8 molecules in the tumor microenvironment of the patient samples during the immunotherapy process. In the model prediction stage, the actual input data does not include the changes in the expression of CD8 molecules in the tumor microenvironment of the patient to be predicted during the immunotherapy process, so that the absolute value of the TIL infiltration density and the change of the TIL infiltration density at different time periods of the patient to be predicted during the immunotherapy process are predicted through the prediction model.

[0090] Furthermore, the model training module includes a training parameter acquisition unit and a model training unit, wherein:

[0091] The training parameter acquisition unit is used to obtain microenvironment cell parameters and clinical factors that significantly affect the absolute value of TIL infiltration density and the change of TIL infiltration density in the immunotherapy microenvironment based on the training input data;

[0092] The model training unit is used to input the microenvironment parameters and clinical factors obtained by the training parameter acquisition module into the machine learning model for training. During the training process, the absolute value of TIL infiltration density and the change of TIL infiltration density of the corresponding patient samples at different time points during immunotherapy are used as training output data to train the mapping relationship between the training input data and the training output data.

[0093] In the above-mentioned model training unit, the machine learning model may use but is not limited to any one of the machine learning algorithms such as decision tree, random forest, support vector machine, naive Bayes, etc., so as to train the mapping relationship between the training input data and the training output data by selecting the training process of the machine learning algorithm on the basis of the known training input data and the training output data.

[0094] Furthermore, the training parameter acquisition unit fits the image features within the tumor region of interest in the two groups of magnetic resonance diffusion weighted images on the basis of the training input data to obtain the microenvironment cell parameters within the tumor region of interest; performs univariate regression analysis on the clinical data information of the different patient samples to determine the clinical factors that affect the absolute value of TIL infiltration density and the change of TIL infiltration density in the immunotherapy microenvironment; performs multivariate logistic regression analysis on the microenvironment cell parameters within the tumor region of interest and the clinical factors that affect the absolute value of TIL infiltration density and the change of TIL infiltration density in the immunotherapy microenvironment to obtain the microenvironment cell parameters and clinical factors that significantly affect the absolute value of TIL infiltration density and the change of TIL infiltration density in the immunotherapy microenvironment.

[0095] Furthermore, the training parameter acquisition unit uses a preset fitting algorithm to fit the image features in the tumor region of interest in two sets of magnetic resonance diffusion weighted images for a preset number of times based on the training input data; after randomly initializing and repeating the fitting for a preset number of times, the microenvironment cell parameters in the tumor region of interest are obtained, and the microenvironment cell parameters include the average cell diameter, intracellular volume fraction, extracellular diffusion rate and cell density in the tumor region of interest.

[0096] In the above training parameter acquisition unit, the microenvironment cell parameters are obtained by fitting the image features in the region of interest representing the tumor in the two sets of magnetic resonance diffusion weighted images. Specifically, the image features in the tumor region of interest in the two magnetic resonance diffusion weighted images can be fitted using least squares curve fitting, and randomly initialized and repeated 100 times to avoid local minimum values. The microenvironment cell parameters at least include the average cell diameter (d), intracellular volume fraction (Vin), extracellular diffusion rate (Dex) and cell density (Cellularity) in the tumor region of interest, wherein the intracellular diffusion rate is fixed at 1.0 μm2 / ms, and the cell density (Cellularity) = Vin / d·100. For ease of observation, the above microenvironment cell parameters (d, Vin, Dex, Cellularity) can be displayed through a habitat map. The above image feature fitting and parameter calculation and the generation of the habitat map can all be generated by data analysis software.

[0097] Furthermore, the system also includes: a verification parameter acquisition module, a comparison module and an evaluation module, wherein:

[0098] The verification parameter acquisition module is used to obtain the absolute value of TIL infiltration density and the change of TIL infiltration density evaluated in real time during the treatment of the patient to be verified;

[0099] The comparison module is used to compare the absolute value of TIL infiltration density and the change of TIL infiltration density obtained by the verification parameter acquisition module during the treatment process as the actual measurement result with the absolute value of TIL infiltration density and the change of TIL infiltration density predicted by the prediction model to obtain a receiver operating characteristic curve predicted by the model;

[0100] The evaluation module is used to evaluate the prediction accuracy of the pre-trained prediction model according to the receiver operating characteristic curve predicted by the model.

[0101] In order to evaluate the effectiveness of the trained prediction model, the receiver operating characteristic curve predicted by the model can be used to compare the absolute value of TIL infiltration density and the change of TIL infiltration density obtained by actual immunohistochemical staining evaluation during treatment with the predicted absolute value of TIL infiltration density and the change of TIL infiltration density to verify the accuracy and stability of the trained prediction model.

[0102] In the above comparison module, the measured results are equivalent to the absolute value of TIL infiltration density and the change of TIL infiltration density in the set period of time obtained by the actual evaluation of the patient to be verified during the treatment process. Correspondingly, the training sample data of the patient to be verified is input into the trained prediction model, and the absolute value of TIL infiltration density and the change of TIL infiltration density in the set period of time of the patient to be verified during the treatment process can be predicted. Here, the system can display the relevant parameters in the model prediction process based on the line chart operation, as shown in the following example. Figure 2 As shown, Figure 2 The line chart shows the microenvironment cell parameters after immunotherapy, the absolute value of TIL infiltration density in the microenvironment, and the change of TIL infiltration density and other related parameters.

[0103] Exemplarily, the training sample data are sorted according to treatment time and divided into a training group (105 cases) and a validation group (45 cases) in a ratio of 7:3. Two groups of nii-formatted magnetic resonance diffusion-weighted images of patients in the training group are stored in a predetermined folder, and complete information is input into the data acquisition module. The prediction is performed through the above-mentioned first prediction module. Since the training sample data of the validation group is input into the prediction model using the training input data to obtain the prediction result data, in order to facilitate comparison, the training output data can be directly used as the actual result to compare the prediction result data, and the receiver operating characteristic curve predicted by the model can be obtained to verify the accuracy and stability of the prediction model through the receiver operating characteristic curve predicted by the model. The receiver operating characteristic curve of the specific prediction model is as follows: Figure 3 As shown in the figure, the receiver operating characteristic curve is used to verify the deviation between the actual results and the predicted results of the training set cases and the validation set cases. Figure 3 The figure shows the receiver operating characteristic curves of the training set cases and the validation set cases after verification, including the predicted absolute value of TIL infiltration density, which reaches 0.920 and 0.882 for the training set cases and the validation set cases respectively, and the predicted change value of TIL infiltration density, which reaches 0.924 and 0.894 for the training set cases and the validation set cases respectively.

[0104] Further, as Figure 1 The specific implementation of the system, the embodiment of the present application provides another monitoring system for preoperative immunotherapy microenvironment remodeling, such as Figure 4As shown, the system includes: an image acquisition module 21, an image processing module 22, a data acquisition module 23, an analysis module 24, a second prediction module 25 and a display module 26;

[0105] The image acquisition module 21 is used to obtain two sets of magnetic resonance diffusion weighted images marked with tumor regions of interest, and transmit the two sets of magnetic resonance diffusion weighted images to the image processing module, the two sets of magnetic resonance diffusion weighted images include pulse gradient spin echo sequence images and oscillating gradient spin echo sequence images obtained by magnetic resonance diffusion weighted time diffusion spectrum imaging function;

[0106] The image processing module 22 is used to perform feature fitting on the two groups of magnetic resonance diffusion weighted images to obtain microenvironment cell parameters in the tumor region of interest;

[0107] The data collection module 23 is used to obtain clinical data information of the patient to be predicted;

[0108] The analysis module 24 is used to screen out clinical factors that significantly affect the absolute value of TIL infiltration density and the change of TIL infiltration density in the immunotherapy microenvironment based on the clinical data information obtained by the clinical data acquisition module, and combine the clinical factors that significantly affect the absolute value of TIL infiltration density and the change of TIL infiltration density in the immunotherapy microenvironment with the microenvironment cell parameters to perform multivariate logistic regression analysis to obtain the microenvironment cell parameters and clinical factors that significantly affect the absolute value of TIL infiltration density and the change of TIL infiltration density in the immunotherapy microenvironment;

[0109] The second prediction module 25 is used to input the microenvironment cell parameters and clinical factors that significantly affect the absolute value of TIL infiltration density and the change of TIL infiltration density in the immunotherapy microenvironment obtained by the analysis module into the pre-trained prediction model, so as to predict the absolute value of TIL infiltration density and the change of TIL infiltration density of the patient to be predicted during the treatment process through the prediction model;

[0110] The display module 26 is used to display the predicted absolute value of TIL infiltration density and the change of TIL infiltration density in a nomogram manner, so as to monitor the remodeling of the preoperative immunotherapy microenvironment according to the displayed absolute value of TIL infiltration density and the change of TIL infiltration density.

[0111] It can be understood that after the above-mentioned image acquisition module is combined with the image processing module, two data acquisition branches are formed with the data acquisition module. These two data acquisition branches are executed in parallel in actual application, and there is no order of data acquisition. Then the data acquired by the two data acquisition branches are transmitted to the analysis module, and the analysis module combines and analyzes the acquired data to obtain clinical factors that significantly affect the absolute value of TIL infiltration density and the change of TIL infiltration density in the immunotherapy microenvironment. Then, the second prediction module predicts the absolute value of TIL infiltration density and the change of TIL infiltration density of the patient to be predicted during the treatment process according to the data obtained by the analysis module. Finally, the predicted data is output to the display module, and the predicted absolute value of TIL infiltration density and the change of TIL infiltration density are displayed in a nomogram, so as to monitor the remodeling of the preoperative immunotherapy microenvironment according to the displayed absolute value of TIL infiltration density and the change of TIL infiltration density.

[0112] In addition, the present invention also provides a method for monitoring the remodeling of the microenvironment during preoperative immunotherapy, comprising the following steps:

[0113] S1. A clinical data acquisition module acquires clinical data information of a patient to be predicted, and transmits the clinical data information to a first prediction module;

[0114] S2, the magnetic resonance image acquisition module acquires two sets of magnetic resonance diffusion weighted images marked with tumor regions of interest, and transmits the two sets of magnetic resonance diffusion weighted images to the first prediction module, wherein the two sets of magnetic resonance diffusion weighted images include pulse gradient spin echo sequence images and oscillating gradient spin echo sequence images obtained by magnetic resonance diffusion weighted time diffusion spectrum imaging function; specifically including:

[0115] S2.1, the image acquisition unit acquires two sets of magnetic resonance diffusion weighted images of the patient to be predicted through a preset storage path;

[0116] S2.2, the region of interest delineation unit delineates and annotates the tumor contour range frame by frame on one of the two groups of magnetic resonance diffusion weighted images obtained by the image acquisition unit to obtain a tumor region of interest, so that one of the two groups of magnetic resonance diffusion weighted images is annotated with the tumor region of interest;

[0117] S2.3, the image matching unit matches another set of magnetic resonance diffusion weighted images according to the set of magnetic resonance diffusion weighted images annotated with the tumor region of interest obtained by the region of interest delineation unit, so that the two sets of magnetic resonance sequence diffusion weighted images are annotated with the same tumor region of interest;

[0118] S3, the first prediction module inputs the clinical data information obtained by the clinical data acquisition module and the two groups of magnetic resonance diffusion weighted images obtained by the magnetic resonance image acquisition module into a pre-trained prediction model, so as to predict the absolute value of TIL infiltration density and the change of TIL infiltration density of the patient to be predicted during the treatment process according to the microenvironment cell parameters in the tumor region of interest in the magnetic resonance diffusion weighted image and the clinical data information through the prediction model, wherein the microenvironment cell parameters are obtained by fitting the image features in the tumor region of interest in the two groups of magnetic resonance diffusion weighted images;

[0119] S4. The monitoring module monitors the remodeling of the preoperative immunotherapy microenvironment of the patient to be predicted based on the absolute value of the TIL infiltration density and the change of the TIL infiltration density predicted by the first prediction module.

[0120] After step S2, the method further comprises:

[0121] S5, a feature fitting module performs feature fitting on the two groups of magnetic resonance diffusion weighted images obtained by the magnetic resonance image acquisition module to obtain microenvironment cell parameters in the tumor region of interest;

[0122] S6. The analysis module screens out clinical factors that significantly affect the absolute value of TIL infiltration density and the change of TIL infiltration density in the immunotherapy microenvironment according to the clinical data information obtained by the clinical data acquisition module, and combines the clinical factors that significantly affect the absolute value of TIL infiltration density and the change of TIL infiltration density in the immunotherapy microenvironment with the microenvironment cell parameters to perform multivariate logistic regression analysis to obtain microenvironment cell parameters and clinical factors that significantly affect the absolute value of TIL infiltration density and the change of TIL infiltration density in the immunotherapy microenvironment;

[0123] Correspondingly, S3, the first prediction module inputs the microenvironment cell parameters and clinical factors that significantly affect the absolute value of TIL infiltration density and the change of TIL infiltration density in the immunotherapy microenvironment obtained by the analysis module into the pre-trained prediction model, so as to predict the absolute value of TIL infiltration density and the change of TIL infiltration density of the patient to be predicted during the treatment process through the prediction model;

[0124] Before step S3, the method further includes:

[0125] S7, the training sample acquisition module acquires the training sample data of the machine learning model, wherein the training sample data includes training input data and training output data; specifically includes:

[0126] S7.1. Construct training input data based on clinical information of different patient samples and two sets of magnetic resonance diffusion-weighted images annotated with corresponding tumor regions of interest;

[0127] S7.2. Construct training output data based on the absolute value of TIL infiltration density and the change of TIL infiltration density of different patient samples at different time points during treatment;

[0128] S8, the model training module inputs the training sample data obtained by the training sample acquisition module into the machine learning model, so as to train the mapping relationship between the training input data and the training output data through the machine learning model; specifically including:

[0129] S8.1. Based on the training input data, obtain microenvironment cell parameters and clinical factors that significantly affect the absolute value of TIL infiltration density and the change of TIL infiltration density in the immunotherapy microenvironment;

[0130] S8.2, inputting the microenvironment parameters and significant clinical factors into the machine learning model for training, using the absolute value of TIL infiltration density and the change of TIL infiltration density of the corresponding patient samples at different time points during the treatment process as training output data, and training the mapping relationship between the training input data and the training output data;

[0131] S9. When the training of the model building module meets the iteration stopping condition, a prediction model is built according to the mapping relationship obtained by the training of the model training module.

[0132] The monitoring method for preoperative immunotherapy microenvironment remodeling provided in the embodiments of the present application is compared with the current method of using conventional imaging or fluorescent-labeled molecular imaging to achieve preoperative immunotherapy microenvironment remodeling. The present application combines two groups of magnetic resonance diffusion-weighted images marked with tumor regions of interest to detect the unique effects of various microstructures in the cell body on the diffusion of water molecules, so as to accurately reflect the structural characteristics of microenvironment cells. On the basis of a pre-trained prediction model, the changes in microenvironment cell parameters in the tumor region of interest and clinical data information are combined to predict the absolute value of TIL infiltration density and the change in TIL infiltration density during the treatment of the patient to be predicted, thereby applying the absolute value of TIL infiltration density and the change in TIL infiltration density to preoperative immunotherapy microenvironment remodeling, so as to achieve monitoring of the immunotherapy microenvironment remodeling of the patient to be predicted.

[0133] Based on this understanding, the technical solution of the present application can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.), and includes a number of instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each implementation scenario of the present application.

[0134] Based on the above method, and Figure 1 and Figure 4 The system embodiment shown, in order to achieve the above-mentioned purpose, the embodiment of the present application also provides a physical device for monitoring the remodeling of the preoperative immunotherapy microenvironment, which can be specifically a computer, a smart phone, a tablet computer, a smart watch, a server, or a network device, etc. The physical device includes a storage medium and a processor; the storage medium is used to store a computer program; the processor is used to execute the computer program to implement the above-mentioned method for monitoring the remodeling of the preoperative immunotherapy microenvironment.

[0135] Optionally, the physical device may also include a user interface, a network interface, a camera, a radio frequency (RF) circuit, a sensor, an audio circuit, a WI-FI module, etc. The user interface may include a display, an input unit such as a keyboard, etc., and the optional user interface may also include a USB interface, a card reader interface, etc. The network interface may optionally include a standard wired interface, a wireless interface (such as a WI-FI interface), etc.

[0136] Those skilled in the art will understand that the structure of the physical device for monitoring the remodeling of the preoperative immunotherapy microenvironment provided in this embodiment does not constitute a limitation on the physical device, and may include more or fewer components, or a combination of certain components, or different arrangements of components.

[0137] The storage medium may also include an operating system and a network communication module. The operating system is a program that manages the hardware and software resources of the physical device for monitoring the remodeling of the preoperative immunotherapy microenvironment, and supports the operation of information processing programs and other software and / or programs. The network communication module is used to achieve communication between components within the storage medium, as well as communication with other hardware and software in the information processing physical device.

[0138] Through the description of the above implementation methods, those skilled in the art can clearly understand that the present application can be implemented by means of software plus the necessary general hardware platform, or by hardware. By applying the technical solution of the present application, compared with the current existing methods, the present application, based on the pre-trained prediction model, combines the changes in microenvironment cell parameters in the tumor region of interest with clinical data information to predict the absolute value of TIL infiltration density and the change in TIL infiltration density during the treatment of the patient to be predicted, thereby applying the absolute value of TIL infiltration density and the change in TIL infiltration density to the preoperative immunotherapy microenvironment remodeling, so as to monitor the immunotherapy microenvironment remodeling of the patient to be predicted.

[0139] Those skilled in the art will appreciate that the accompanying drawings are only schematic diagrams of a preferred implementation scenario, and the modules or processes in the accompanying drawings are not necessarily necessary for implementing the present application. Those skilled in the art will appreciate that the modules in the devices in the implementation scenario can be distributed in the devices of the implementation scenario according to the description of the implementation scenario, or can be changed accordingly and located in one or more devices different from the present implementation scenario. The modules of the above-mentioned implementation scenario can be combined into one module, or can be further split into multiple submodules.

[0140] The above serial numbers of this application are only for description and do not represent the advantages and disadvantages of the implementation scenarios. The above disclosure is only a few specific implementation scenarios of this application, but this application is not limited to them, and any changes that can be thought of by technicians in this field should fall within the scope of protection of this application.

Claims

1. A monitoring system for preoperative immunotherapy microenvironment remodeling, characterized in that: include: A clinical data acquisition module, a magnetic resonance image acquisition module, a first prediction module and a monitoring module, wherein: The clinical data acquisition module is used to acquire clinical data information of the patient to be predicted; The magnetic resonance image acquisition module is used to acquire two sets of magnetic resonance diffusion weighted images marked with tumor regions of interest, wherein the two sets of magnetic resonance diffusion weighted images include pulse gradient spin echo sequence images and oscillating gradient spin echo sequence images obtained by magnetic resonance diffusion weighted time diffusion spectrum imaging function; The first prediction module is used to input the clinical data information obtained by the clinical data acquisition module and the two groups of magnetic resonance diffusion weighted images obtained by the magnetic resonance image acquisition module into a pre-trained prediction model, so as to predict the absolute value of TIL infiltration density and the change of TIL infiltration density of the patient to be predicted during the treatment process according to the microenvironment cell parameters in the tumor region of interest in the magnetic resonance diffusion weighted image and the clinical data information through the prediction model, wherein the microenvironment cell parameters are obtained by fitting the image features in the tumor region of interest in the two groups of magnetic resonance diffusion weighted images; The monitoring module is used to monitor the preoperative immunotherapy microenvironment remodeling of the patient to be predicted based on the absolute value of the TIL infiltration density and the change of the TIL infiltration density predicted by the first prediction module.

2. The system according to claim 1, characterized in that The magnetic resonance image acquisition module includes an image acquisition unit, a region of interest delineation unit and an image matching unit, wherein: The image acquisition unit is used to acquire two sets of magnetic resonance diffusion weighted images of the patient to be predicted through a preset storage path; The region of interest delineation unit is used to delineate and annotate the tumor contour range frame by frame on one of the two groups of magnetic resonance diffusion weighted images obtained by the image acquisition unit by an operator delineation method to obtain a tumor region of interest, so that one of the two groups of magnetic resonance diffusion weighted images is annotated with the tumor region of interest; The image matching unit is used to match another set of magnetic resonance diffusion weighted images according to a set of magnetic resonance diffusion weighted images annotated with tumor regions of interest obtained by the region of interest delineation unit, so that the two sets of magnetic resonance sequence diffusion weighted images are annotated with the same tumor regions of interest.

3. The system according to claim 1, characterized in that The system also includes: a training sample acquisition module, a model training module and a model construction module, wherein: The training sample acquisition module is used to acquire training sample data of the machine learning model before inputting the clinical data information obtained by the clinical data acquisition module and the two sets of magnetic resonance diffusion weighted images obtained by the magnetic resonance image acquisition module into the pre-trained prediction model, wherein the training sample data includes training input data and training output data; The model training module is used to input the training sample data obtained by the training sample acquisition module into the machine learning model, so as to train the mapping relationship between the training input data and the training output data through the machine learning model; The model building module is used to build a prediction model according to the mapping relationship obtained through training by the model training module when the training meets the iteration stopping condition.

4. The system according to claim 3, characterized in that The training sample acquisition module includes: an input data construction unit and an output data construction unit, wherein: The input data construction unit is used to construct training input data according to clinical information of different patient samples and two groups of magnetic resonance diffusion weighted images annotated with corresponding tumor regions of interest; The output data construction unit is used to construct training output data according to the actual absolute value of TIL infiltration density and the change of TIL infiltration density of different patient samples at different time points during the treatment process.

5. The system according to claim 4, characterized in that The system further comprises: a CD8 expression score input module and a median value calculation module, wherein: The CD8 expression score input module is used to input the CD8 expression score representing the CD8 expression level at different sites of the tumor obtained by immunohistochemical staining of multiple biopsy tissue sections of the patient sample at the time point to be monitored during immunotherapy in the training input data, so as to obtain the absolute value and change of the CD8 molecule expression score of the tumor microenvironment of the patient sample during the immunotherapy process; The median calculation module is used to calculate the median value of multiple groups of CD8 molecule expression score absolute values ​​and change values ​​relative to the initial expression score according to the absolute value and change of the tumor microenvironment CD8 molecule expression score of the patient sample obtained by the CD8 expression score input module during immunotherapy. The median value reflects the actual TIL infiltration density absolute value and TIL infiltration density change of the patient sample at a preset time point during the treatment process.

6. The system according to claim 3, characterized in that The model training module includes a training parameter acquisition unit and a model training unit, wherein: The training parameter acquisition unit is used to obtain microenvironment cell parameters and clinical factors that significantly affect the absolute value of TIL infiltration density and the change of TIL infiltration density in the immunotherapy microenvironment based on the training input data; The model training unit is used to input the microenvironment parameters and clinical factors obtained by the training parameter acquisition module into the machine learning model for training. During the training process, the absolute value of TIL infiltration density and the change of TIL infiltration density of the corresponding patient samples at different time points during immunotherapy are used as training output data to train the mapping relationship between the training input data and the training output data.

7. The system according to claim 6, characterized in that The training parameter acquisition unit is, based on the training input data, to fit the image features in the tumor region of interest in the two groups of magnetic resonance diffusion weighted images to obtain the microenvironment cell parameters in the tumor region of interest; to perform univariate regression analysis on the clinical data information of the different patient samples to determine the clinical factors that affect the absolute value of TIL infiltration density and the change of TIL infiltration density in the immunotherapy microenvironment; to perform multivariate logistic regression analysis on the microenvironment cell parameters in the tumor region of interest and the clinical factors that affect the absolute value of TIL infiltration density and the change of TIL infiltration density in the immunotherapy microenvironment to obtain the microenvironment cell parameters and clinical factors that significantly affect the absolute value of TIL infiltration density and the change of TIL infiltration density in the immunotherapy microenvironment.

8. The system according to claim 7, characterized in that The training parameter acquisition unit uses a preset fitting algorithm to fit the image features in the tumor region of interest in two sets of magnetic resonance diffusion weighted images for a preset number of times based on the training input data; after randomly initializing and repeating the fitting for a preset number of times, the microenvironment cell parameters in the tumor region of interest are obtained, and the microenvironment cell parameters include the average cell diameter, intracellular volume fraction, extracellular diffusion rate and cell density in the tumor region of interest.

9. The system according to any one of claims 1 to 8, characterized in that: The system also includes: a verification parameter acquisition module, a comparison module and an evaluation module, wherein: The verification parameter acquisition module is used to obtain the absolute value of TIL infiltration density and the change of TIL infiltration density evaluated in real time during the treatment of the patient to be verified; The comparison module is used to compare the absolute value of TIL infiltration density and the change of TIL infiltration density obtained by the verification parameter acquisition module during the treatment process as the actual measurement result with the absolute value of TIL infiltration density and the change of TIL infiltration density predicted by the prediction model to obtain a receiver operating characteristic curve predicted by the model; The evaluation module is used to evaluate the prediction accuracy of the pre-trained prediction model according to the receiver operating characteristic curve predicted by the model.

10. A method for monitoring microenvironment remodeling before immunotherapy, characterized in that: include: S1. A clinical data acquisition module acquires clinical data information of a patient to be predicted, and transmits the clinical data information to a first prediction module; S2, the magnetic resonance image acquisition module acquires two sets of magnetic resonance diffusion weighted images marked with tumor regions of interest, and transmits the two sets of magnetic resonance diffusion weighted images to the first prediction module, wherein the two sets of magnetic resonance diffusion weighted images include pulse gradient spin echo sequence images and oscillating gradient spin echo sequence images obtained by magnetic resonance diffusion weighted time diffusion spectrum imaging function; specifically including: S2.1, the image acquisition unit acquires two sets of magnetic resonance diffusion weighted images of the patient to be predicted through a preset storage path; S2.2, the region of interest delineation unit delineates and annotates the tumor contour range frame by frame on one of the two groups of magnetic resonance diffusion weighted images obtained by the image acquisition unit to obtain a tumor region of interest, so that one of the two groups of magnetic resonance diffusion weighted images is annotated with the tumor region of interest; S2.3, the image matching unit matches another set of magnetic resonance diffusion weighted images according to the set of magnetic resonance diffusion weighted images annotated with the tumor region of interest obtained by the region of interest delineation unit, so that the two sets of magnetic resonance sequence diffusion weighted images are annotated with the same tumor region of interest; S3, the first prediction module inputs the clinical data information obtained by the clinical data acquisition module and the two groups of magnetic resonance diffusion weighted images obtained by the magnetic resonance image acquisition module into a pre-trained prediction model, so as to predict the absolute value of TIL infiltration density and the change of TIL infiltration density of the patient to be predicted during the treatment process according to the microenvironment cell parameters in the tumor region of interest in the magnetic resonance diffusion weighted image and the clinical data information through the prediction model, wherein the microenvironment cell parameters are obtained by fitting the image features in the tumor region of interest in the two groups of magnetic resonance diffusion weighted images; S4, the monitoring module monitors the preoperative immunotherapy microenvironment remodeling of the patient to be predicted according to the absolute value of TIL infiltration density and the change of TIL infiltration density predicted by the first prediction module; After step S2, the method further comprises: S5, a feature fitting module performs feature fitting on the two groups of magnetic resonance diffusion weighted images obtained by the magnetic resonance image acquisition module to obtain microenvironment cell parameters in the tumor region of interest; S6. The analysis module screens out clinical factors that significantly affect the absolute value of TIL infiltration density and the change of TIL infiltration density in the immunotherapy microenvironment according to the clinical data information obtained by the clinical data acquisition module, and combines the clinical factors that significantly affect the absolute value of TIL infiltration density and the change of TIL infiltration density in the immunotherapy microenvironment with the microenvironment cell parameters to perform multivariate logistic regression analysis to obtain microenvironment cell parameters and clinical factors that significantly affect the absolute value of TIL infiltration density and the change of TIL infiltration density in the immunotherapy microenvironment; Correspondingly, S3, the first prediction module inputs the microenvironment cell parameters and clinical factors that significantly affect the absolute value of TIL infiltration density and the change of TIL infiltration density in the immunotherapy microenvironment obtained by the analysis module into the pre-trained prediction model, so as to predict the absolute value of TIL infiltration density and the change of TIL infiltration density of the patient to be predicted during the treatment process through the prediction model; Before step S3, the method further includes: S7, the training sample acquisition module acquires the training sample data of the machine learning model, wherein the training sample data includes training input data and training output data; specifically includes: S7.

1. Construct training input data based on clinical information of different patient samples and two sets of magnetic resonance diffusion-weighted images annotated with corresponding tumor regions of interest; S7.

2. Construct training output data based on the absolute value of TIL infiltration density and the change of TIL infiltration density of different patient samples at different time points during treatment; S8, the model training module inputs the training sample data obtained by the training sample acquisition module into the machine learning model, so as to train the mapping relationship between the training input data and the training output data through the machine learning model; specifically including: S8.

1. Based on the training input data, obtain microenvironment cell parameters and clinical factors that significantly affect the absolute value of TIL infiltration density and the change of TIL infiltration density in the immunotherapy microenvironment; S8.2, inputting the microenvironment parameters and significant clinical factors into the machine learning model for training, using the absolute value of TIL infiltration density and the change of TIL infiltration density of the corresponding patient samples at different time points during the treatment process as training output data, and training the mapping relationship between the training input data and the training output data; S9. When the training of the model building module meets the iteration stopping condition, a prediction model is built according to the mapping relationship obtained by the training of the model training module.

Images