Multi-focal epilepsy positioning method and device based on MRI (Magnetic Resonance Imaging)

Abstract

The invention discloses a multi-focal epilepsy positioning method and device based on MRI (Magnetic Resonance Imaging). The MRI-based multi-focal epilepsy positioning method comprises the following steps: acquiring a trained focus-cortex epilepsy network; according to the trained classification model; obtaining to-be-predicted image information; extracting image features of the to-be-predicted image information; inputting the image features into the trained lesion-cortex epilepsy network so as to obtain a lesion-cortex epilepsy network result; inputting the image features into the trained classification model so as to obtain a prediction result; and obtaining a final prediction result according to the lesion-cortex epilepsy network result and the prediction result. According to the MRI-based multi-focal epilepsy positioning method, an epilepsy structure network and a function network are considered, a fusion model is established in combination with multiple lesion features of a patient, and the non-invasive, simple and easy multi-focal epilepsy positioning method purely based on MRI is provided.

Description

Technical Field

[0001] The present application relates to the field of image recognition technology, and in particular to an MRI-based multifocal epilepsy localization method and an MRI-based multifocal epilepsy localization device. Background Art

[0002] Epilepsy is a common and frequently occurring disease of the nervous system. In addition to frequent epileptic seizures, these patients often suffer from cognitive and mental disorders, which seriously affect their physical and mental health, and also create huge economic pressure on society and families, becoming one of the prominent social problems.

[0003] Most focal epilepsy has one origin area, but some epilepsy patients have multiple pathological foci, such as tuberous sclerosis complex (TSC) (this application takes tuberous sclerosis complex as an example), multiple encephalomalacia and multiple cavernous hemangiomas. It is extremely important to find and locate a single epileptogenic focus from multiple pathological foci to control epilepsy.

[0004] At present, the localization of epileptic focus mainly relies on clinical symptoms, pathological focus location (MRI) and electroencephalogram (including scalp electroencephalogram and intracranial electrode electroencephalogram, etc.). However, more than 60% of patients cannot be located by non-invasive methods, 20% of patients can be located by invasive methods, and 40% of patients cannot locate the epileptic focus, which seriously affects surgical treatment.

[0005] Therefore, it is hoped that a technical solution can be provided to solve or at least alleviate the above-mentioned deficiencies of the prior art. Summary of the invention

[0006] The purpose of the present invention is to provide an MRI-based multifocal epilepsy localization method to solve at least one of the above-mentioned technical problems.

[0007] The present invention provides the following scheme:

[0008] According to one aspect of the present invention, a method for localizing multifocal epilepsy based on MRI is provided, comprising:

[0009] Obtaining a trained lesion-cortical epilepsy network;

[0010] According to the trained classification model;

[0011] Obtaining image information to be predicted;

[0012] Extracting image features of the image information to be predicted;

[0013] Inputting the image features into the trained lesion-cortical epilepsy network to obtain lesion-cortical epilepsy network results;

[0014] Inputting the image features into the trained classification model to obtain prediction results;

[0015] A final prediction result is obtained according to the lesion-cortical epilepsy network result and the prediction result.

[0016] Optionally, the lesion-cortical epilepsy network is established by the following method:

[0017] Obtain the sensitivity T-Map of the epileptogenic lesion;

[0018] Obtain specific T-Map of epileptogenic lesions;

[0019] The sensitivity T-Map and the specificity T-Map were overlapped, and the overlapping area was retained. The overlapping voxels were divided into positive connection areas and negative connection areas according to the positive and negative T values. The positive functional connections between the positively connected voxels and all cortical voxels, and the negative functional connections between the negatively connected voxels and all cortical voxels were calculated respectively. The positive functional connections were superimposed with the negative functional connections to obtain the lesion-cortical epilepsy network.

[0020] Optionally, obtaining a sensitivity T-Map of the epileptogenic lesion includes:

[0021] Acquire a T1 image group, wherein the T1 image group includes a plurality of T1 images, and one T1 image corresponds to one patient, wherein the patient includes a TSC-related epilepsy patient who has completed resective surgery and a patient who has been seizure-free for 6 months after surgery;

[0022] Acquire a T2-Flair image group, wherein the T2-Flair image group includes a plurality of T2-Flair images, and one T2-Flair image corresponds to one patient, wherein the patient includes a TSC-related epilepsy patient who has completed resective surgery and a patient who has been seizure-free for 6 months after surgery, and each T2-Flair image includes an epileptogenic focus ROI;

[0023] Acquire an EPI image group, wherein the EPI image group includes a plurality of EPI images, and one EPI image corresponds to one healthy person;

[0024] Registering the T1 image group with the T2-Flair image group to obtain a first MNI template;

[0025] Normalizing the EPI image group to the first MNI template, thereby obtaining a second MNI template;

[0026] The epileptogenic lesion ROI in the second MNI template was used as the seed region, and the average time series of each seed region in the EPI image was calculated. The functional connectivity between the seed region and all cortical voxels was calculated using the Pearson correlation coefficient, and the Fisher Z transformation was used to ensure that the Pearson correlation coefficient conformed to the normal distribution. Among them, each epileptogenic lesion ROI would obtain a functional connectivity map with the same number as the EPI image. The above functional connectivity maps were integrated into a network map using a one-sample T test, which was superimposed and the voxels that appeared in 50% of the patients were calculated. The average brain functional connectivity of the overlapping area was retained to obtain the sensitivity T-Map of the epileptogenic lesion.

[0027] Optionally, the step of obtaining a specific T-Map of an epileptic lesion comprises:

[0028] Obtaining the functional connectivity map of the epileptogenic nodule group and the functional connectivity map of the non-epileptic nodule group;

[0029] The two-sample T test was used to compare the functional connectivity maps of the epileptogenic nodule group and the non-epileptic nodule group, and FDR correction was performed. Voxels with p < 0.05 were selected to establish the T-Map.

[0030] Optionally, the epileptogenic lesion ROI on the T2-Flair image is acquired by the following method:

[0031] Acquire the first cortical lesion region of interest and the second cortical lesion region of interest of each T2-Flair image;

[0032] Calculating the overlap in cortical lesion identification of the first cortical lesion region of interest and the second cortical lesion region of interest of each T2-Flair image;

[0033] The overlap degree in the cortical lesion identification of each T2-Flair image is judged as follows: whether the overlap degree is greater than a preset threshold is judged, and if so, the overlapping area is regarded as the epileptogenic lesion ROI.

[0034] Optionally, registering the T1 image group with the T2-Flair image group to obtain a first MNI template includes:

[0035] The T2-Flair image of each patient was registered to the T1 image, and then each epileptogenic lesion ROI was registered to the T1 image and finally standardized to the MNI template.

[0036] Optionally, the trained classification model is obtained by the following method:

[0037] Acquire a training set, wherein the training set includes images of epileptic patients and images of non-epileptic patients;

[0038] The average time series of the lesion area was calculated based on the images of non-epileptic patients in the training set. Subsequently, the functional connectivity between the lesion area and the whole brain voxels was calculated, and then Fisher z transformation was performed to obtain the functional connectivity matrix.

[0039] The functional connections of the sub-patients whose lesions were epileptogenic lesions were designated as positive samples, while those of the sub-patients whose lesions were non-epileptogenic lesions were labeled as negative samples;

[0040] The functional connection matrix is ​​reduced in dimension by principal component analysis to obtain characteristic information;

[0041] Get a random forest classifier;

[0042] The random forest classifier is trained through feature information to obtain hyperparameters of the random forest model.

[0043] The present application also provides a multifocal epilepsy localization device based on MRI, and the multifocal epilepsy localization device based on MRI comprises:

[0044] A lesion-cortical epilepsy network acquisition module, wherein the lesion-cortical epilepsy network acquisition module is used to acquire a trained lesion-cortical epilepsy network;

[0045] A classification model acquisition module, wherein the classification model acquisition module is used to obtain a classification model according to a trained classification model;

[0046] A module for acquiring information of an image to be predicted, wherein the module is used to acquire information of an image to be predicted;

[0047] An image feature acquisition module, the image feature acquisition module is used to extract image features of the image information to be predicted;

[0048] A lesion-cortical epilepsy network result acquisition module, wherein the lesion-cortical epilepsy network result acquisition module is used to input the image features into the trained lesion-cortical epilepsy network, thereby acquiring a lesion-cortical epilepsy network result;

[0049] A prediction result acquisition module, wherein the prediction result acquisition module is used to input the image features into the trained classification model to obtain a prediction result;

[0050] A final prediction result acquisition module is used to acquire a final prediction result according to the lesion-cortical epilepsy network result and the prediction result.

[0051] The multifocal epilepsy localization method based on MRI in this application takes into account both the epileptic structural network and the functional network, and at the same time combines the patient's multi-lesion characteristics to establish a fusion model, providing a non-invasive, simple and easy multifocal epilepsy localization method based solely on MRI. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 It is a flowchart of a multifocal epilepsy localization method based on MRI in one embodiment of the present application.

[0053] Figure 2 It is a schematic diagram of a multifocal epilepsy localization device based on MRI in one embodiment of the present application.

[0054] Figure 3 It is a schematic diagram of the overall process of the MRI-based multifocal epilepsy localization method in one embodiment of the present application. DETAILED DESCRIPTION

[0055] The technical solution of the present invention will be described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0056] Figure 1 It is a flowchart of a multifocal epilepsy localization method based on MRI in one embodiment of the present application.

[0057] like Figure 1 as well as Figure 3 The MRI-based methods for localizing multifocal epilepsy shown include:

[0058] Obtaining a trained lesion-cortical epilepsy network;

[0059] According to the trained classification model;

[0060] Obtaining image information to be predicted;

[0061] Extracting image features of the image information to be predicted;

[0062] Inputting the image features into the trained lesion-cortical epilepsy network to obtain lesion-cortical epilepsy network results;

[0063] Inputting the image features into the trained classification model to obtain prediction results;

[0064] A final prediction result is obtained according to the lesion-cortical epilepsy network result and the prediction result.

[0065] In this embodiment, the lesion-cortical epilepsy network is established by the following method:

[0066] Obtain the sensitivity T-Map of the epileptogenic lesion;

[0067] Obtain specific T-Map of epileptogenic lesions;

[0068] The sensitivity T-Map and the specificity T-Map were overlapped, and the overlapping area was retained. The overlapping voxels were divided into positive connection areas and negative connection areas according to the positive and negative T values. The positive functional connections between the positively connected voxels and all cortical voxels, and the negative functional connections between the negatively connected voxels and all cortical voxels were calculated respectively. The positive functional connections were superimposed with the negative functional connections to obtain the lesion-cortical epilepsy network.

[0069] In this embodiment, obtaining a sensitivity T-Map of the epileptogenic lesion includes:

[0070] Acquiring a T1 image group, wherein the T1 image group includes a plurality of T1 images (T1 weighted images), and one T1 image corresponds to one patient, wherein the patient includes a TSC-related epilepsy patient who has completed resective surgery and has no seizures for 6 months after the surgery;

[0071] Acquire a T2-Flair image group, wherein the T2-Flair image group includes a plurality of T2-Flair images, and one T2-Flair image (T2-weighted image) corresponds to one patient, wherein the patient includes a TSC-related epilepsy patient who has completed resective surgery and has no seizures for 6 months after the surgery, and each T2-Flair image includes an epileptogenic focus ROI;

[0072] Acquire an EPI image group, wherein the EPI image group (resting state functional magnetic resonance imaging) includes a plurality of EPI images, and one EPI image corresponds to one healthy person;

[0073] Registering the T1 image group and the T2-Flair image group, and performing standardization to obtain a first MNI template;

[0074] Normalizing the EPI image group to the first MNI template, thereby obtaining a second MNI template;

[0075] The epileptogenic lesion ROI in the second MNI template was used as the seed region, and the average time series of each seed region in the EPI image was calculated. The functional connectivity between the seed region and all cortical voxels was calculated using the Pearson correlation coefficient, and the Fisher Z transformation was used to ensure that the Pearson correlation coefficient conformed to the normal distribution. Among them, each epileptogenic lesion ROI would obtain a functional connectivity map with the same number as the EPI image. The above functional connectivity maps were integrated into a network map using a one-sample T test, which was superimposed and the voxels that appeared in 50% of the patients were calculated. The average brain functional connectivity of the overlapping area was retained to obtain the sensitivity T-Map of the epileptogenic lesion.

[0076] In this embodiment, obtaining a specific T-Map of the epileptogenic focus includes:

[0077] Obtaining the functional connectivity map of the epileptogenic nodule group and the functional connectivity map of the non-epileptic nodule group;

[0078] The two-sample T test was used to compare the functional connectivity maps of the epileptogenic nodule group and the non-epileptic nodule group, and FDR correction was performed. Voxels with p < 0.05 were selected to establish the T-Map.

[0079] Specifically, it is also necessary to obtain the functional connection map related to the epileptic nodules first. This method is the same as the establishment of the functional connection map related to the epileptic nodules in the process of constructing the sensitive T-Map. When constructing the T-Map, a non-epileptic nodule group is also required. In this embodiment, the non-epileptic nodule group is similar. The average time series in the seed area of ​​the non-epileptic nodule mapping is calculated with the functional connection of all cortical voxels of healthy subjects in the normal control group, and the z-scores (Z scores are z-transformed by the previously obtained Pearson correlation coefficient, i.e., r value) of the functional connection standardization of each nodule with the normal data set are merged to obtain the functional connection map related to the non-epileptic nodules. The functional connection map of the epileptic nodule group and the functional connection map of the non-epileptic nodule group are compared between groups using a two-sample T test, and FDR correction is performed, and voxels with p<0.05 are selected to establish a T-Map map.

[0080] In this embodiment, the epileptogenic focus ROI on the T2-Flair image is obtained by the following method:

[0081] Acquire the first cortical lesion region of interest and the second cortical lesion region of interest of each T2-Flair image;

[0082] Calculating the overlap in cortical lesion identification of the first cortical lesion region of interest and the second cortical lesion region of interest of each T2-Flair image;

[0083] The overlap degree in the cortical lesion identification of each T2-Flair image is judged as follows: whether the overlap degree is greater than a preset threshold is judged, and if so, the overlapping area is regarded as the epileptogenic lesion ROI.

[0084] In this embodiment, registering the T1 image group with the T2-Flair image group to obtain a first MNI template includes:

[0085] The T2-Flair image of each patient was registered to the T1 image, and then each epileptogenic lesion ROI was registered to the T1 image and finally standardized to the MNI template.

[0086] In this embodiment, the trained classification model is obtained by the following method:

[0087] Acquire a training set, wherein the training set includes images of epileptic patients and images of non-epileptic patients;

[0088] The average time series of the lesion area is calculated based on the images of non-epileptic patients in the training set. Subsequently, the functional connectivity between the lesion area and the whole brain voxels is calculated, and then the Fisher z transform is performed to obtain the functional connectivity matrix. In order to reduce the dimension, the connections with zero values ​​are excluded. For example, it can be understood as a large cube composed of 100 cubes, each cube has a time series. In order to calculate the functional connectivity (i.e., correlation coefficient) between the large cube and the seed point, the values ​​of all the small cubes are averaged, and then the functional connectivity (correlation coefficient) is calculated with other small cubes in the brain.

[0089] The functional connections of the sub-patients whose lesions were epileptogenic lesions were designated as positive samples, while those of the sub-patients whose lesions were non-epileptogenic lesions were labeled as negative samples;

[0090] The functional connection matrix is ​​reduced in dimension by principal component analysis to obtain characteristic information;

[0091] Get a random forest classifier;

[0092] The random forest classifier is trained through feature information to obtain hyperparameters of the random forest model.

[0093] In this embodiment, obtaining the final prediction result according to the lesion-cortical epilepsy network result and the prediction result includes:

[0094] The risk score, which is influenced by the overlap and location of epileptic network lesions, was combined with the classification probability generated by the prediction model, which reflects the functional properties of the lesion area. This fusion method takes into account both structural and functional characteristics and improves the prediction accuracy. Initially, to standardize the risk score, we selected 97.5% of its distribution and truncated any value exceeding 1 to convert it to a standardized score in the range of 0 to 1. The two models were then fused based on the consistency of the standardized scores and classification probabilities.

[0095] The present application is further described in detail below by way of examples. It should be understood that the examples do not constitute any limitation to the present application.

[0096] Data source acquisition:

[0097] Based on the existing TSC-related epilepsy patients

[0098] Inclusion criteria: no age limit, male or female; MRI shows >3 intracranial cortical nodules; clinically meets the 2021 version of the TSC international diagnostic criteria, with or without TSC gene mutations; clear diagnosis of epilepsy, and has taken more than 3 anti-epileptic seizure drugs; has or has not been treated with mTOR inhibitors or ketogenic diet; after strict preoperative examination and evaluation, meets the surgical indications identified in the "Chinese Expert Consensus on Surgical Treatment of TSC-Related Epilepsy"; undergoes single epileptogenic nodule resection and completes 6 months of follow-up; family members agree and sign the consent form.

[0099] Exclusion criteria: postoperative MRI confirmed that the preoperatively identified epileptogenic nodule was not completely removed; obvious subependymal giant cell astrocytoma; patients with severe complications after surgery or brain trauma within six months, history of cranial surgery, or vagus nerve stimulation treatment that affect the prognosis; patients whose family members refused to participate in the study or did not complete the 6-month follow-up after surgery.

[0100] (2) Preoperative evaluation and data collection of patients with TSC-related epilepsy

[0101] MRI: MRI data of patients were collected, including 3D-T1 sequence scanning parameters: repetition time = 2300ms, echo time = 2.98ms, flip angle 9°, FOV 256×256mm 2 , voxel 0.5×0.5×1mm 3 , no-spacing scanning; Fluid-attenuated inversion recovery sequence (Flair) scanning parameters: TR = 11000ms, TE = 140ms, TI = 2850ms, voxel 1×1×1mm 3 , no spacing scanning.

[0102] Clinical data: Patients’ gender, age at surgery, age at onset, course of disease, frequency of epileptic seizures, duration of epileptic seizures, number of anti-epileptic drugs, and number of epileptogenic nodules were collected.

[0103] Scalp EEG and seizure symptom collection: EEG is a non-invasive examination method to measure brain electrical activity. By placing electrodes on the scalp and recording the electrical activity of brain neurons, it helps determine the type of epilepsy, the epileptogenic focus, the conduction of epileptic discharges, etc., and guides the internal and external treatment of epilepsy. EEG is an indispensable and necessary examination in the preoperative evaluation of epilepsy. In-depth analysis of EEG is of great significance to improving the clinical diagnosis rate.

[0104] 18F-FDG PET-CT scan: PET-CT examination is of great significance in the preoperative evaluation of TSC-related epilepsy. It reflects the abnormal metabolic activity of the epileptic seizure starting area and other brain areas during the epileptic seizure by detecting the areas with abnormal glucose metabolism in the brain, especially the low metabolic areas during the interictal period. It helps to understand the impact of epileptic lesions on surrounding brain areas and evaluate the severity and scope of epilepsy.

[0105] SEEG: Patients with TSC-related epilepsy who underwent comprehensive evaluation including clinical history collection, neurological examination, neuropsychological testing, high-resolution MRI, and scalp EEG to confirm the epileptogenic nodule hypothesis (including possible primary and secondary epileptogenic nodules, etc.) were selected.

[0106] (3) Surgical treatment and follow-up of patients with TSC-related epilepsy

[0107] Epileptic nodule resection: The surgical plan is determined based on the epileptic nodule identified by SEEG, and the surgical treatment is performed under general anesthesia. The resection range includes the epileptic nodule and the adjacent cerebral cortex.

[0108] Postoperative follow-up and criteria for determining epileptic nodules: Epilepsy seizures were followed up 6 months after surgery and classified using the ILAE six-level classification method. Patients were divided into a postoperative no epileptic seizure group (ILAE-I grade) and a postoperative epileptic seizure group (ILAE II-VI grade) according to whether or not there was an epileptic seizure. In the postoperative no epileptic seizure group, the nodules removed were accurately identified as epileptic nodules, and the nodules other than the epileptic nodules in these patients were identified as non-epileptic nodules. Patients who had epileptic seizures 6 months after surgery were considered to have an incorrect determination of epileptic nodules.

[0109] (4) Retrospective collection of normal subjects’ individual data database

[0110] Inclusion criteria: no age limit, male or female; no abnormal signals on 3T-MRI structural images; no history of epileptic seizures or neurological diseases; completed rs-fMRI scans with good image quality that can be used for subsequent analysis.

[0111] Exclusion criteria: motion artifacts in MRI data; lack of 3D-MPRAGE in 3T-MRI structural image data.

[0112] (5) rs-fMRI scanning parameters and data preprocessing of normal subjects

[0113] Scanning parameters: A 3.0T MR750 GE scanner was used to acquire the data set using a gradient echo-planar imaging (GRE-EPI) sequence: TR = 2000 ms, TE = 30 ms, FA = 90°, FOV = 240 × 240 mm 2 , matrix = 64 × 64, voxel = 3.75 × 3.75 × 4 mm 3 All subjects were instructed to remain awake, close their eyes naturally, and avoid head movement during the scanning process.

[0114] Data preprocessing: The Matlab toolkit DPABI (http: / / rfmri.org / DPABI) was used for resting-state data preprocessing, which included the following steps: removal of the first 10 time points and head motion correction; rigid registration of T1 images to the EPI average image; normalization of the EPI image to the Montreal Neurological Institute (MNI) standard space using the T1 image and subsequent resampling to 3 × 3 × 3 mm 3 ; Remove noise, including whole-brain signals, head motion, and linear trends; Temporal filter (0.01–0.08 Hz).

[0115] This section divides each patient into several "sub-patients" with only a single cortical lesion, that is, each patient can be divided into several sub-patients according to the number of lesions.

[0116] Construction of lesion-cortex related epilepsy network based on rs-fMRI of normal subjects and MRI of patients: establishment of risk scoring model

[0117] Discovery set and validation set: Among the enrolled patients with TSC (tuberous sclerosis complex)-related epilepsy who had completed resective surgery, 50 cases were randomly selected as the validation set according to the digital random method, and the remaining patients who were seizure-free 6 months after surgery were selected as the discovery set.

[0118] Extraction of cortical lesion regions of interest: Two clinicians used ITK-SNAP software to draw cortical lesion regions of interest (ROI) on the T2-Flair weighted images of each patient in the discovery set and validation set (V4.2.0, http: / / www.itksnap.org), and calculated the overlap in cortical lesion identification based on the nilearn library in Python (https: / / nilearn.github.io / stable / index.html). That is, the percentage of non-overlapping lesion identifications of the same patient by two neurologists. If the percentage is less than 5%, the overlapping area is considered as a cortical lesion and used for subsequent analysis. When the cortical lesion identification of the two clinicians was inconsistent (>5%), the data was abandoned or the selection of the cortical lesion ROI was determined by a third senior clinician. Then, using FSL software (https: / / fsl.fmrib.ox.ac.uk / ), the T2-Flair image of each patient was registered to the T1 image, and then each annotated lesion was registered to the T1 image, and finally standardized to the MNI template.

[0119] Voxel-based cortical lesion-seizure symptom mapping: To verify that a single cortical voxel in patients with TSC-related epilepsy is not associated with epileptic seizures, a general linear model was used to establish a voxel-based cortical lesion-seizure symptom mapping. The general linear model is in the form of Y = βX + ε. Y represents the lesion area of ​​each voxel, 1 means the voxel is within the lesion area, and 0 means the voxel is outside the lesion area. X represents the relevance of the cortical lesion to epilepsy, 1 represents the epileptogenic lesion, and 0 represents the non-epileptic lesion. The model parameter β is estimated, and ε is the estimated residual. Finally, the p value of the T test for β (based on the FDR correction for each brain region, p < 0.05) indicates the sensitivity of the brain region to epileptic seizures.

[0120] Sensitivity T-Map construction: In the discovery set of cases, the epileptogenic lesion ROI after MNI template registration was used as the seed region, and the average time series of each seed region was calculated in 400 healthy subjects in the normal control group. The functional connectivity between the corresponding region and all cortical voxels was calculated using the Pearson correlation coefficient, and the Fisher Z transformation was used to ensure that the Pearson correlation coefficient conformed to the normal distribution. Finally, 400 sets of functional connection maps were obtained for each epileptogenic lesion. The above functional connection maps were integrated into a network map using a single-sample T test, which was superimposed and the voxels that appeared in 50% of the patients were calculated. The average brain functional connectivity of the overlapping area was retained to obtain the sensitivity T-Map of the epileptogenic lesion.

[0121] Specific T-Map construction: Similar to the sensitive T-Map construction, the functional connectivity map related to the epileptogenic lesion was obtained in the discovery set. Similarly, in the non-epileptic lesion group, the functional connectivity was calculated by the average time series in the seed area of ​​the non-epileptic lesion map and the whole cortical voxels of healthy subjects in the normal control group, and the z-scores of the functional connectivity standardization of each lesion with the normal data set were merged to obtain the functional connectivity map related to the non-epileptic lesion. The functional connectivity map of the epileptogenic lesion group was compared with that of the non-epileptic lesion group using a two-sample T test, and FDR correction was performed. Voxels with p < 0.05 were selected to establish the T-Map.

[0122] Construction of lesion-cortical epilepsy network: overlap the specific T-Map and the sensitive T-Map, retain the overlapping area, divide the overlapping voxels into positive connection area and negative connection area according to the positive and negative T values, and calculate the positive functional connection between the positively connected voxels and all cortical voxels and the negative functional connection between the negatively connected voxels and all cortical voxels based on a large number of healthy subjects. Superimpose the two to obtain the lesion-cortical epilepsy network. In the standard space, according to the distribution of the lesion-cortical epilepsy network, calculate the positive epilepsy network weight and the negative epilepsy network weight of the overlapping part of the lesion and this network respectively, and add them to obtain the damage intensity, that is, the risk score. The higher the score, the greater the possibility that the cortical lesion is an epileptogenic lesion. The risk score can be simply understood as the intensity of the damaged voxels overlapping with the epilepsy network.

[0123] Establishment of prediction model based on functional network

[0124] Based on the epileptic and non-epileptic patients in the training dataset, a binary classification model was developed using machine learning techniques to predict the classification probability of each lesion. Initially, the average time series of the lesion area was calculated based on the fMRI data of the rs fMRI dataset. Subsequently, the functional connectivity between the lesion area and the whole-brain voxels was calculated and then Fisher z-transformed. To reduce the dimensionality, connections with zero values ​​were excluded.

[0125] The functional connections of sub-patients whose lesions were epileptogenic lesions were designated as positive samples, while the connections of sub-patients whose lesions were non-epileptic lesions were labeled as negative samples. Principal component analysis (PCA) was used to further reduce the dimensionality. PCA uses an orthogonal transformation to transform the correlated variables into a set of linearly uncorrelated variables, called principal components. This transformation combines multiple original indicators into a new set of independent comprehensive indicators, retaining as much information as possible from the original variables. PCA helps reduce computational complexity, eliminate noise and irrelevant features, and improve the usability and interpretability of the dataset.

[0126] The features obtained after PCA dimensionality reduction were used as input to a random forest classifier to predict the classification probability of each lesion. Random forest is a classifier composed of multiple decision trees, which was chosen for its efficiency, robustness to high-dimensional data, avoidance of overfitting, and ability to effectively handle unbalanced datasets. Subsequently, for each individual lesion, the model generated a probability score indicating the probability of an epileptogenic focus based on the functional brain network.

[0127] Establishment of a fusion model for predicting cortical epileptogenic lesions

[0128] To build a more accurate probability model of epileptogenic lesions, we combined the risk score influenced by the overlap and location of epileptic network lesions with the classification probability generated by the prediction model, which reflects the functional characteristics of the lesion area. This fusion method takes into account both structural and functional characteristics and improves the prediction accuracy. Initially, to standardize the risk score, we selected 97.5% of its distribution and truncated any value exceeding 1 to convert it to a standardized score in the range of 0 to 1. The two models were then fused based on the consistency of the standardized scores and classification probabilities. If both the standardized score and the classification probability were below the threshold, the minimum value was selected as the final probability.

[0129] Validation of a fusion model for predicting cortical epileptogenic lesions

[0130] The cortical lesion ROIs of 50 patients with TSC-related epilepsy in the validation set were used to calculate the cortical lesion risk score according to the distribution of the established lesion-cortical epilepsy network. Furthermore, the relevant clinical, PET-CT, EEG and other characteristics of the patients were input into the selected machine learning model to obtain the epileptogenicity score of each lesion. The results obtained based on the lesion-cortical epilepsy network and the machine learning model were input into the fusion model to evaluate the epileptogenicity of the cortical lesions and determine the epileptogenic lesions. The actual outcomes of no seizures after surgery in the 50 patients in the validation set and 15 patients who still had seizures after surgery were used as the gold standard for the accuracy of epileptogenic lesion determination. The actual resected cortical lesions of patients without epileptic seizures after surgery were true epileptogenic lesions, while the unresected lesions were true non-epileptic lesions; while the resected lesions of patients with seizures after surgery were true non-epileptic lesions, and the true epileptogenicity of the unresected lesions could not be determined.

[0131] (1) We divided each patient with multifocal epilepsy into multiple single-lesion “sub-patients” for further treatment.

[0132] (2) We localize the epileptogenic zone (epileptic lesion) based on MRI alone, without relying on clinical symptoms or electroencephalogram (EEG). Clinically, the determination of the epileptogenic lesion requires an ictal electroencephalogram (EEG), which is usually difficult to obtain. This method evaluates the epileptogenic lesion based solely on structural imaging information, which can localize the epileptogenic lesion early and intervene in epilepsy at an early stage.

[0133] (3) We established a risk score for epileptogenic foci based on lesion-related networks, established a prediction model using brain networks and machine learning, and finally formed a fusion model.

[0134] (4) This fusion model takes into account both the epilepsy structural network and the functional network, while also incorporating the patient's multi-lesion characteristics.

[0135] (5) The application of this model is not restricted by unit, MRI equipment type, etc., and has a wide range of applications.

[0136] This method is based on the existing MRI data of TSC-related epilepsy patients and the rs-fMRI data of 400 healthy subjects. According to the postoperative epilepsy control, the epileptogenic lesions and non-epileptic lesions are extracted and distinguished. The lesion network mapping technology, machine learning, ensemble learning and other methods are used to construct the lesion-cortical epilepsy network, machine learning model and epileptogenic lesion prediction fusion model; the epileptogenic lesions are currently determined by the epilepsy control outcomes of patients after resection surgery, and the accuracy of the fusion model in predicting epileptogenic lesions is verified.

[0137] This study aims to establish a fusion model and a non-invasive, simple, and accurate method for determining epileptogenic lesions based on conventional MRI, which can be used for multifocal epilepsy such as multiple encephalomalacia and multiple cavernous hemangiomas. It can reduce the surgical treatment gap for TSC-related epilepsy and other multifocal epilepsy, reduce the use of invasive examinations and invasive treatments, and improve the clinical efficacy and quality of life of patients.

[0138] The present application also provides a multifocal epilepsy localization device based on MRI, and the multifocal epilepsy localization device based on MRI comprises:

[0139] A lesion-cortical epilepsy network acquisition module, wherein the lesion-cortical epilepsy network acquisition module is used to acquire a trained lesion-cortical epilepsy network;

[0140] A classification model acquisition module, wherein the classification model acquisition module is used to obtain a classification model according to a trained classification model;

[0141] A module for acquiring information of an image to be predicted, wherein the module is used to acquire information of an image to be predicted;

[0142] An image feature acquisition module, the image feature acquisition module is used to extract image features of the image information to be predicted;

[0143] A lesion-cortical epilepsy network result acquisition module, wherein the lesion-cortical epilepsy network result acquisition module is used to input the image features into the trained lesion-cortical epilepsy network, thereby acquiring a lesion-cortical epilepsy network result;

[0144] A prediction result acquisition module, wherein the prediction result acquisition module is used to input the image features into the trained classification model to obtain a prediction result;

[0145] A final prediction result acquisition module, which is used to acquire a final prediction result based on the lesion-cortical epilepsy network result and the prediction result.

[0146] The above explanation of the method also applies to the explanation of the device.

[0147] Figure 2 It is a structural block diagram of an electronic device provided by one or more embodiments of the present invention.

[0148] like Figure 2 As shown, the present application also discloses an electronic device (i.e., the master controller in the present application), comprising: a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other via the communication bus; a computer program is stored in the memory, and when the computer program is executed by the processor, the processor executes the steps of the MRI-based multifocal epilepsy localization method.

[0149] The present application also provides a computer-readable storage medium storing a computer program executable by an electronic device. When the computer program runs on the electronic device, the steps of the MRI-based multifocal epilepsy localization method can be implemented.

[0150] The communication bus mentioned in the above electronic device can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. The communication bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, only one thick line is used in the figure, but it does not mean that there is only one bus or one type of bus.

[0151] The electronic device includes a hardware layer, an operating system layer running on the hardware layer, and an application layer running on the operating system. The hardware layer includes hardware such as a central processing unit (CPU), a memory management unit (MMU), and a memory. The operating system can be any one or more computer operating systems that implement electronic device control through a process, such as a Linux operating system, a Unix operating system, an Android operating system, an iOS operating system, or a Windows operating system. In addition, in an embodiment of the present invention, the electronic device can be a handheld device such as a smart phone or a tablet computer, or can be an electronic device such as a desktop computer or a portable computer, which is not particularly limited in the embodiment of the present invention.

[0152] The execution subject of the electronic device control in the embodiment of the present invention may be an electronic device, or a functional module in the electronic device that can call and execute a program. The electronic device may obtain the firmware corresponding to the storage medium. The firmware corresponding to the storage medium is provided by the supplier. The firmware corresponding to different storage media may be the same or different, which is not limited here. After the electronic device obtains the firmware corresponding to the storage medium, the firmware corresponding to the storage medium may be written into the storage medium, specifically, the firmware corresponding to the storage medium may be burned into the storage medium. The process of burning the firmware into the storage medium may be implemented using existing technology, which will not be described in detail in the embodiment of the present invention.

[0153] The electronic device may also obtain a reset command corresponding to the storage medium. The reset command corresponding to the storage medium is provided by the supplier. The reset commands corresponding to different storage media may be the same or different, and are not limited here.

[0154] At this time, the storage medium of the electronic device is a storage medium in which the corresponding firmware is written, and the electronic device can respond to the reset command corresponding to the storage medium in the storage medium in which the corresponding firmware is written, so that the electronic device resets the storage medium in which the corresponding firmware is written according to the reset command corresponding to the storage medium. The process of resetting the storage medium according to the reset command can be implemented by the existing technology and will not be described in detail in the embodiments of the present invention.

[0155] For the convenience of description, the above devices are described in terms of functions and are divided into various units and modules. Of course, when implementing the present application, the functions of each unit and module can be implemented in the same or multiple software and / or hardware.

[0156] Those skilled in the art will appreciate that, unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as those generally understood by those skilled in the art in the art to which the present invention belongs. It should also be understood that terms such as those defined in general dictionaries should be understood to have meanings consistent with the meanings in the context of the prior art, and will not be interpreted with idealized or overly formal meanings unless specifically defined.

[0157] For the method embodiments, for the sake of simplicity, they are all described as a series of action combinations, but those skilled in the art should know that the embodiments of the present invention are not limited by the order of the actions described, because according to the embodiments of the present invention, some steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification are all preferred embodiments, and the actions involved are not necessarily required by the embodiments of the present invention.

[0158] It can be known from the description of the above implementation modes that those skilled in the art can clearly understand that the present application can be implemented by means of software plus a necessary general hardware platform. Based on such an understanding, the technical solution of the present application can be essentially or partly contributed to the prior art in the form of a software product, which can be stored in a storage medium such as ROM / RAM, a magnetic disk, an optical disk, etc., and includes several 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 the various implementation modes of the present application or certain parts of the implementation modes.

[0159] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A multifocal epilepsy localization method based on MRI, characterized in that: The MRI-based multifocal epilepsy localization method comprises: Obtaining a trained lesion-cortical epilepsy network; According to the trained classification model; Obtaining image information to be predicted; Extracting image features of the image information to be predicted; Inputting the image features into the trained lesion-cortical epilepsy network to obtain lesion-cortical epilepsy network results; Inputting the image features into the trained classification model to obtain prediction results; A final prediction result is obtained according to the lesion-cortical epilepsy network result and the prediction result.

2. The MRI-based multifocal epilepsy localization method according to claim 1, characterized in that: The lesion-cortical epilepsy network is established by the following method: Obtain the sensitivity T-Map of the epileptogenic lesion; Obtain specific T-Map of epileptogenic lesions; The sensitivity T-Map and the specificity T-Map were overlapped, and the overlapping area was retained. The overlapping voxels were divided into positive connection areas and negative connection areas according to the positive and negative T values. The positive functional connections between the positively connected voxels and all the cortical voxels, and the negative functional connections between the negatively connected voxels and all the cortical voxels were calculated respectively. The positive functional connections and the negative functional connections were superimposed to obtain the lesion-cortical epilepsy network.

3. The MRI-based multifocal epilepsy localization method according to claim 2, characterized in that: The method of obtaining the sensitivity T-Map of the epileptogenic lesion comprises: Acquire a T1 image group, wherein the T1 image group includes a plurality of T1 images, and one T1 image corresponds to one patient, wherein the patient includes a TSC-related epilepsy patient who has completed resective surgery and a patient who has been seizure-free for 6 months after surgery; Acquire a T2-Flair image group, wherein the T2-Flair image group includes a plurality of T2-Flair images, and one T2-Flair image corresponds to one patient, wherein the patient includes a TSC-related epilepsy patient who has completed resective surgery and a patient who has been seizure-free for 6 months after surgery, and each T2-Flair image includes an epileptogenic focus ROI; Acquire an EPI image group, wherein the EPI image group includes a plurality of EPI images, and one EPI image corresponds to one healthy person; Registering the T1 image group with the T2-Flair image group to obtain a first MNI template; Normalizing the EPI image group to the first MNI template, thereby obtaining a second MNI template; The epileptogenic lesion ROI in the second MNI template was used as the seed region, and the average time series of each seed region in the EPI image was calculated. The functional connectivity between the seed region and all cortical voxels was calculated using the Pearson correlation coefficient, and the Fisher Z transformation was used to ensure that the Pearson correlation coefficient conformed to the normal distribution. Among them, each epileptogenic lesion ROI would obtain a functional connectivity map with the same number as the EPI image. The above functional connectivity maps were integrated into a network map using a one-sample T test, which was superimposed and the voxels that appeared in 50% of the patients were calculated. The average brain functional connectivity of the overlapping area was retained to obtain the sensitivity T-Map of the epileptogenic lesion.

4. The MRI-based multifocal epilepsy localization method according to claim 3, characterized in that: The method of obtaining a specific T-Map of an epileptogenic focus comprises: Obtaining the functional connectivity map of the epileptogenic nodule group and the functional connectivity map of the non-epileptic nodule group; The two-sample T test was used to compare the functional connectivity maps of the epileptogenic nodule group and the non-epileptic nodule group, and FDR correction was performed. Voxels with p < 0.05 were selected to establish the T-Map.

5. The MRI-based multifocal epilepsy localization method according to claim 4, characterized in that: The epileptogenic lesion ROI on the T2-Flair image is obtained by the following method: Acquire the first cortical lesion region of interest and the second cortical lesion region of interest of each T2-Flair image; Calculating the overlap in cortical lesion identification of the first cortical lesion region of interest and the second cortical lesion region of interest of each T2-Flair image; The overlap degree in the cortical lesion identification of each T2-Flair image is judged as follows: whether the overlap degree is greater than a preset threshold is judged, and if so, the overlapping area is regarded as the epileptogenic lesion ROI.

6. The MRI-based multifocal epilepsy localization method according to claim 5, characterized in that: The registering the T1 image group with the T2-Flair image group to obtain a first MNI template includes: The T2-Flair image of each patient was registered to the T1 image, and then each epileptogenic lesion ROI was registered to the T1 image and finally standardized to the MNI template.

7. The MRI-based multifocal epilepsy localization method according to claim 1, characterized in that: The trained classification model is obtained by the following method: Acquire a training set, wherein the training set includes images of epileptic patients and images of non-epileptic patients; The average time series of the lesion area was calculated based on the images of non-epileptic patients in the training set. Subsequently, the functional connectivity between the lesion area and the whole brain voxels was calculated, and then Fisher z transformation was performed to obtain the functional connectivity matrix. The functional connections of the sub-patients whose lesions were epileptogenic lesions were designated as positive samples, while those of the sub-patients whose lesions were non-epileptogenic lesions were labeled as negative samples; The functional connection matrix is ​​reduced in dimension by principal component analysis to obtain characteristic information; Get a random forest classifier; The random forest classifier is trained through feature information to obtain hyperparameters of the random forest model.

8. A multifocal epilepsy localization device based on MRI, characterized in that: The MRI-based multifocal epilepsy localization device comprises: A lesion-cortical epilepsy network acquisition module, wherein the lesion-cortical epilepsy network acquisition module is used to acquire a trained lesion-cortical epilepsy network; A classification model acquisition module, wherein the classification model acquisition module is used to obtain a classification model according to a trained classification model; A module for acquiring information of an image to be predicted, wherein the module is used to acquire information of an image to be predicted; An image feature acquisition module, the image feature acquisition module is used to extract image features of the image information to be predicted; A lesion-cortical epilepsy network result acquisition module, wherein the lesion-cortical epilepsy network result acquisition module is used to input the image features into the trained lesion-cortical epilepsy network, thereby acquiring a lesion-cortical epilepsy network result; A prediction result acquisition module, wherein the prediction result acquisition module is used to input the image features into the trained classification model to obtain a prediction result; A final prediction result acquisition module is used to acquire a final prediction result according to the lesion-cortical epilepsy network result and the prediction result.

Images