An automatic planning platform for minimally invasive fiber optic lesion paths for focal cortical dysplasia
Through the automatic planning platform, the use of image data and constraint parameters to generate possible paths is solved, and the trauma and location difficulties of FCD lesion treatment in traditional surgery is achieved, and the more efficient and safer minimally invasive fiber damage technology is achieved, which simplifies the clinical operation process.
Patent Information
- Application Number
- CN202210423074.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-04-21
- Publication Date
- 2024-12-17
- Estimated Expiration
- 2042-04-21
AI Technical Summary
Traditional craniotomy surgery has problems with high trauma, high postoperative risk, lesion residual and functional deficit risk, and minimally invasive fiber damage technology in lesion positioning and boundary determination , relying on manual adjustment of laser paths, is time-consuming and labor-intensive, and it is difficult to avoid important brain areas and blood vessels.
Design an automatic planning platform for focal cortical dysplasia, through neural image data acquisition and preprocessing, combined with automatic planning algorithm, use constraint parameters to generate possible paths, calculate risk coefficients, and screen out the path with the lowest risk. It also realizes automatic planning of laser fiber path through extraction of important brain areas, searching for cranial point, vascular avoidance and maximum coverage of gray matter.
It improves the accuracy and damage efficiency of FCD lesions, reduces the time and labor of manual adjustment, enhances safety, reduces the risk of lesion residues and functional deficiencies, and simplifies the operation process of clinicians.
Smart Images
Figure CN114748163B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to a relatively large scope, specifically to an automatic planning platform for minimally invasive optical fiber lesion paths for focal cortical dysplasia. Background Art
[0002] Epilepsy is a chronic disease of the nervous system that recurs and seriously threatens human health. According to WHO statistics, the prevalence of epilepsy is between 5‰ and 11.2‰. There are about 50 million epilepsy patients worldwide, about 9 million epilepsy patients in my country, including 6 million active epilepsy, and about 400,000 new cases each year. About 20% to 30% of all epilepsy patients have drug-resistant epilepsy and require preoperative evaluation and surgical treatment. Focal cortical dysplasia (FCD) is a common pathological type of epilepsy, which is highly epileptogenic and difficult to treat with drugs. The cause of FCD is the abnormal proliferation, migration, differentiation and cortical arrangement of neurons during brain development, which leads to changes in the development of cerebral cortical neurons and cortical structure. This type of abnormal cortical architecture and pathological neurons are highly epileptogenic.
[0003] FCD is not only highly epileptogenic, but also refractory to drugs. Therefore, surgical resection is the most effective treatment for FCD, both for children and adult patients. However, traditional craniotomy resection has some disadvantages that are difficult to overcome: first, craniotomy is traumatic, with a high risk of postoperative bleeding, infarction and infection, and a long postoperative recovery time; second, most FCD lesions are difficult to distinguish from normal brain tissue under a neurosurgical microscope. Even if neuronavigation technology is used, it is difficult to monitor the resection position and range in real time due to factors such as intraoperative brain drift, which can easily lead to residual lesions or damage to normal brain tissue; finally, for deep FCD lesions, the resection process will inevitably damage the superficial normal brain tissue covering it, thereby bringing unnecessary risk of functional loss. The above problems are difficult to be fundamentally solved in the era of traditional craniotomy.
[0004] With the advancement of modern surgical concepts, the advantages of minimally invasive surgery have gradually emerged in clinical practice and have become a new treatment for focal epilepsy. Magnetic resonance-guided laser interstitial thermal therapy (MRgLITT) was first clinically used in my country in 2020. Its principle is to transmit infrared laser to the scattering probe through optical fiber and evenly scatter and heat the brain tissue, thereby achieving the purpose of ablating lesions. MRgLITT has many advantages such as minimally invasive, precise and safe. It can accurately destroy lesions without craniotomy, while maximally protecting adjacent normal brain tissue. It is recognized and applied by more and more international centers. However, there are still three major challenges for MRgLITT to safely, accurately and completely destroy FCD lesions: (1) Some FCDs are hidden on MRI and difficult to detect by conventional manual visual reading, which makes it difficult to locate the lesions and determine their boundaries; (2) Intracranial EEG is another important means of diagnosing and locating FCDs. However, the recorded intracranial EEG seizure onset patterns are diverse, and manual interpretation is extremely time-consuming and has subjective bias, which seriously affects the efficiency of FCD positioning; (3) The destruction range of MRgLITT is a regular cylinder with a diameter of 20 mm centered on the laser fiber. How to make the cylinder completely cover the epileptogenic focus while avoiding blood vessels, ventricles and key structures is an important consideration in the preoperative planning of MRgLITT. When designing the laser fiber path and destruction parameters in clinical practice, in order to achieve the above goals, the surgeon often needs to manually adjust the laser path and various parameters to gradually approach the optimal solution. Therefore, the whole process is time-consuming, labor-intensive, and extremely cumbersome, and largely depends on the surgeon's experience. Summary of the invention
[0005] The object of the present invention is to provide an automatic planning platform for minimally invasive optical fiber lesion paths for focal cortical dysplasia, so as to solve the problems raised in the above-mentioned background technology.
[0006] To achieve the above object, the present invention provides the following technical solutions:
[0007] An automatic planning platform for minimally invasive optical fiber lesion paths for focal cortical dysplasia comprises the following steps:
[0008] S101. Neuroimaging data acquisition and preprocessing: Acquire intracranial structural imaging data and vascular imaging data including high-temporal-resolution three-dimensional dynamic enhanced magnetic resonance angiography sequences of magnetic resonance angiography, then perform subtraction and bone stripping on the vascular images, visualize and three-dimensionally reconstruct the preprocessed vascular image data in Freeview software, use "show as isosurface in 3D view" and set thresholds to reconstruct the blood vessels, remove redundant grids, register the reconstructed vascular images to the T1WI MPRAGE structural images, and finally check the accuracy of vascular segmentation and registration;
[0009] S201. Specific steps of the automatic planning algorithm: perform three-dimensional cortical reconstruction on the original anatomical image, obtain the gray-white matter boundary, the pia mater gray matter boundary level and brain region segmentation information, align it with the anatomical image after subtraction, manually define the intracranial target a, generate all possible paths according to the constraint parameters, and subsequently add manual judgment conditions for secondary screening, calculate the risk coefficient of each path, and screen out the three paths with the lowest risk.
[0010] As a further solution of the present invention: the constraint parameters in step S201 include the distance to important brain regions D1, the average distance to blood vessels D2, the cranial entry angle θ, the gray matter coverage G, the planning time T and the lesion damage rate L.
[0011] As a further solution of the present invention: the laser fiber path is defined as v = {T, E, R, G}, T is the intracranial target point, E is the entry point, R is the risk index, G is the gray / white matter ratio, when there are N proposed target points, the laser fiber automatic planning method is defined as V (N) = {v1 a1 ,…,vN aN}:a i ∈{1,…,M i}, i∈{1,…,N}, where M is a possible optical fiber path.
[0012] As a further solution of the present invention: if it is far away from the blood vessels, the risk index R needs to be minimized, and if it covers the gray matter as much as possible, the G needs to be maximized.
[0013] As a further solution of the present invention, it also includes extraction of important brain regions, search for entry points into the skull, avoidance of blood vessels and maximum coverage of gray matter.
[0014] As a further solution of the present invention: the extraction of important brain areas is completed by cortical reconstruction of T1WI MPRAGE anatomical images using FreeSurfer 7.0 version, and the steps include density standardization of T1WI anatomical images, radio frequency bias field inhomogeneity correction, bone stripping and image segmentation, gridding of gray matter / white matter junction and pia mater surface through hemisphere separation, surface subdivision and image deformation, and finally quality control and manual correction of cortical reconstruction results, extraction of sulcus / gyrus information from the cortical surface, and segmentation of ventricles and choroid plexus structures before path planning.
[0015] As a further solution of the present invention: the cranial entry point is defined as All brain surface meshes are considered, and most paths that do not meet the criteria are first eliminated.
[0016] As a further solution of the present invention: the path that does not meet the standard is the path Shorter than the electrode length and angle Less than 30° or greater than 90°.
[0017] As a further solution of the present invention: the vascular avoidance risk index R i，ai Used to calculate the cumulative distance from the laser fiber to the blood vessel, set d safety is the minimum distance between the path and the blood vessel, and then set d risk is the safe distance between the path and the blood vessel, and the cumulative distance is calculated as follows: where f crit (x) is the shortest distance between point x on the path and the blood vessel, and finally the risk index R i，ai Normalized to [0, 1], 0 means no risk and 1 means the highest risk, the final calculation formula is: The length is Length.
[0018] As a further solution of the present invention: the gray / white matter ratio in the maximum coverage of gray matter is calculated by the number of electrodes located in the gray matter, and for each electrode Contains a series of electrode contacts J, and the sampling range of each contact is calculated as c j :j∈{1,…,J}, with radius c r , when contact c j ±c r The records within the gray matter are considered valid, and the gray / white matter ratio is calculated as:
[0019] where f gm(·) is the sampling range c jThe distance from the gray matter surface, H[·] is the Heaviside function, which is negative when it is inside the gray matter, and J is the number of contacts;
[0020] The damage range is calculated by automatically planning the path of a cylinder with a radius of 2 cm. Here, G is the maximum and the damage coverage rate L is the maximum. The Dice index is used for calculation, which is as follows:
[0021] Where V abla is the volume of the cylinder to be damaged, V lesion is the lesion volume.
[0022] Compared with the prior art, the present invention has the following beneficial effects:
[0023] 1. The present invention targets FCD, a common pathology in epilepsy surgery. The location of FCD lesions varies greatly. The present invention utilizes the new minimally invasive MRgLITT technology with a limited damage range to achieve maximum damage. The present invention combines stereotactic technology to realize lesion localization and automatic path planning, thus saving time and effort, improving surgical efficiency, and increasing safety, efficiency, and prognosis.
[0024] 2. By building an automated platform to achieve end-to-end goals, placing the original data in the target folder can directly achieve results, reducing the learning curve for clinical physicians. It has practical value. The method of use is to directly insert the patient's corresponding images for intelligent recognition, reducing the difficulty of initial use for clinical physicians. BRIEF DESCRIPTION OF THE DRAWINGS
[0025] Figure 1 Flowchart of an automated fiber optic lesion path planning platform for minimally invasive treatment of focal cortical dysplasia.
[0026] Figure 2 This is a research diagram for automatic path planning of laser fiber placement in the present invention. DETAILED DESCRIPTION
[0027] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. 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.
[0028] See also Figures 1-2 In an embodiment of the present invention, a minimally invasive optical fiber lesion path automatic planning platform for focal cortical dysplasia comprises the following steps:
[0029] S101. Neuroimaging data acquisition and preprocessing: intracranial structural imaging data (3D T1WI MPRAGE) and vascular imaging data including high temporal resolution three-dimensional dynamic enhanced magnetic resonance angiography (TWIST) sequence of magnetic resonance angiography (MRA) were acquired, and then the vascular images were subtracted and bone stripped. The preprocessed vascular image data were visualized and three-dimensionally reconstructed in Freeview software. The blood vessels were reconstructed using "show as isosurface in 3D view" and thresholds were set, and redundant meshes were removed. The reconstructed vascular images were registered to the T1WI MPRAGE structural images, and finally the accuracy of vascular segmentation and registration was checked.
[0030] S201. Specific steps of the automatic planning algorithm: perform three-dimensional cortical reconstruction on the original anatomical image, obtain the gray-white matter junction, the pia mater gray matter junction level and brain area segmentation information, align with the anatomical image after subtraction, manually define the intracranial target a, generate all possible paths according to the constraint parameters, and subsequently add manual judgment conditions for secondary screening, calculate the risk coefficient of each path, and select the three paths with the lowest risk. The automatic planning algorithm focuses on the following directions according to clinical needs: stay away from important brain areas and ventricles, stay away from blood vessels by more than 3 to 5 mm, the angle of entry into the skull cannot be too large, cover the gray matter as much as possible, and the skull thickness at the insertion site must be greater than 2 mm. Figure 2 As shown, (A) is the original anatomical image, (B) is the three-dimensional cortical reconstruction of the anatomical image, (C) is a schematic diagram of the three-dimensional reconstruction, and (D) is an example of the preliminary work of the patient's vascular imaging.
[0031] The constraint parameters in step S201 include the distance to important brain regions D1, the average distance to blood vessels D2, the angle of entry into the skull θ, the gray matter coverage G, the planning time T and the lesion damage rate L.
[0032] The laser fiber path is defined as v = {T, E, R, G}, where T is the intracranial target, E is the entry point, R is the risk index, and G is the gray / white matter ratio. When there are N proposed targets, the laser fiber automatic planning method is defined as V(N) = {v1 a1 ,…,vN aN}:a i ∈{1,…,M i}, i∈{1,…,N}, where M is a possible optical fiber path.
[0033] In order to achieve the initial goal, the risk index R needs to be minimized if it is far away from blood vessels, and the G needs to be maximized if it covers gray matter as much as possible.
[0034] It also includes extraction of important brain areas, search for entry points, avoidance of blood vessels, and maximum coverage of gray matter.
[0035] The extraction of important brain regions was completed by cortical reconstruction of T1WI MPRAGE anatomical images using FreeSurfer version 7.0 (http: / / surfer.nmr.mgh.harvard.edu / ), and the steps included density standardization of T1WI anatomical images, radiofrequency bias field inhomogeneity correction, bone stripping and image segmentation, meshing of the gray matter / white matter junction and the pia mater surface through hemispheric separation, surface subdivision and image deformation, and finally quality control and manual correction of the cortical reconstruction results. The sulcus / gyrus information was extracted from the cortical surface, and the ventricles and choroid plexus structures were segmented before planning the path to avoid slipping or ventricular hemorrhage during needle insertion. The designed path did not pass through the sulci because the blood vessels on the brain surface are often located in the sulci.
[0036] The cranial entry point is defined as All brain surface meshes are considered, and most paths that do not meet the criteria are first eliminated.
[0037] The path that does not meet the standard is the path Shorter than the electrode length and angle Less than 30° or greater than 90°.
[0038] The vascular avoidance risk index R i，ai Used to calculate the cumulative distance from the laser fiber to the blood vessel, set d safety is the minimum distance between the path and the blood vessel. If it is lower than this, there will be a higher surgical risk. Then set d risk is the safe distance between the path and the blood vessel. If it is higher than this, it means there is no risk. The cumulative distance is calculated as follows: where f crit (x) is the shortest distance between point x on the path and the blood vessel, and finally the risk index R i，ai Normalized to [0, 1], 0 means no risk and 1 means the highest risk, the final calculation formula is: The length is Length.
[0039] Since epileptic discharges originate from gray matter, maximal coverage of gray matter is of practical significance. The gray / white matter ratio in the maximal coverage of gray matter is calculated as the number of electrodes located in the gray matter. For each electrode Contains a series of electrode contacts J, and the sampling range of each contact is calculated as c j :j∈{1,…,J}, with radius c r , when contact c j ±c r The records within the gray matter are considered valid, and the gray / white matter ratio is calculated as:
[0040] where f gm(·)is the sampling range c j The distance from the gray matter surface, H[·] is the Heaviside function, which is negative when it is inside the gray matter, and J is the number of contacts;
[0041] Consider whether the damage effect can cover the proposed epileptic lesion, and the damage range is calculated by automatically planning the path of a cylinder with a radius of 2 cm. Therefore, here G is maximized and the damage coverage rate L is maximized. The Dice index is used for calculation, which is specifically:
[0042] Where V abla is the volume of the cylinder to be damaged, V lesion is the lesion volume.
[0043] The clinical value of the constructed algorithm was evaluated using two evaluation systems: subjective evaluation and objective evaluation. ① Random sampling of the algorithm's automatically planned paths and manually planned paths was performed, and more than five epilepsy surgeons with relevant clinical experience were asked to score them with a score of 0 to 5 points in a blinded manner. ② A series of quantitative indicators were calculated, including the distance to important brain areas (D1), the average distance to blood vessels (D2), the cranial entry angle (θ), the gray matter coverage (G), the planning time (T), and the lesion destruction rate (L).
[0044] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.
Claims
1. A method for automatically planning a minimally invasive optical fiber damage path for focal cortical dysplasia, characterized in that: The steps include: S101. Neuroimaging data acquisition and preprocessing: Acquire intracranial structural imaging data and vascular imaging data including high-temporal-resolution three-dimensional dynamic enhanced MRI angiography sequence of MRI, then perform subtraction and bone stripping on the vascular images, visualize and three-dimensionally reconstruct the preprocessed vascular image data in Freeview software, reconstruct the blood vessels using "show as isosurface in 3D view" and set thresholds, remove redundant grids, register the reconstructed vascular images to the T1WI MPRAGE structural images, and finally check the accuracy of vascular segmentation and registration; S201. Specific steps of the automatic planning algorithm: perform three-dimensional cortical reconstruction on the original anatomical image, obtain the gray-white matter junction, the pia mater gray matter junction level and brain region segmentation information, perform registration with the anatomical image after subtraction, manually define the intracranial target a, generate all possible paths according to the constraint parameters, and subsequently add manual judgment conditions for secondary screening, calculate the risk coefficient of each path, and screen out the three paths with the lowest risk; It also includes extraction of important brain regions, search for entry points, avoidance of blood vessels, and maximum coverage of gray matter; The vascular avoidance risk index R i，ai Used to calculate the cumulative distance from the laser fiber to the blood vessel, set d safety is the minimum distance between the path and the blood vessel, and then set d risk is the safe distance between the path and the blood vessel, and the cumulative distance is calculated as follows: where f crit (x) is the shortest distance between point x on the path and the blood vessel, and finally the risk index R i，ai Normalized to [0, 1], 0 means no risk and 1 means the highest risk, the final calculation formula is: The length is Length.
2. The method for automatically planning a minimally invasive optical fiber damage path for focal cortical dysplasia according to claim 1, characterized in that: The constraint parameters in step S201 include the distance to important brain regions D1, the average distance to blood vessels D2, the angle of entry into the skull θ, the gray matter coverage G, the planning time T and the lesion damage rate L.
3. The method for automatic planning of minimally invasive optical fiber damage paths for focal cortical dysplasia according to claim 1, characterized in that: The laser fiber path is defined as v = {T, E, R, G}, where T is the intracranial target, E is the entry point, R is the risk index, and G is the gray / white matter ratio. When there are N proposed targets, the laser fiber automatic planning method is defined as V(N) = {v1 a1 ,…,vN aN }:a i ∈{1,…,M i }, i∈{1,…,N}, where M is a possible optical fiber path.
4. The method for automatically planning a minimally invasive optical fiber damage path for focal cortical dysplasia according to claim 3, characterized in that: Staying away from blood vessels requires the smallest risk index R, while covering gray matter as much as possible requires the largest G.
5. The method for automatic planning of minimally invasive optical fiber damage path for focal cortical dysplasia according to claim 1, characterized in that: The extraction of important brain regions was completed by cortical reconstruction of T1WI MPRAGE anatomical images using FreeSurfer 7.0 version. The steps included density standardization of T1WI anatomical images, correction of radio frequency bias field inhomogeneity, bone stripping and image segmentation, meshing of the gray matter / white matter junction and the pia mater surface through hemisphere separation, surface subdivision and image deformation, and finally quality control and manual correction of the cortical reconstruction results. The sulcus / gyrus information was extracted from the cortical surface, and the ventricles and choroid plexus structures were segmented before path planning.
6. The method for automatically planning a minimally invasive optical fiber damage path for focal cortical dysplasia according to claim 1, characterized in that: The cranial entry point is defined as All brain surface meshes are considered, and most paths that do not meet the criteria are first eliminated.
7. The method for automatically planning a minimally invasive optical fiber damage path for focal cortical dysplasia according to claim 6, characterized in that: The path that does not meet the standard is the path Shorter than the electrode length and angle Less than 30° or greater than 90°.
8. The method for automatically planning a minimally invasive optical fiber damage path for focal cortical dysplasia according to claim 1, characterized in that: The gray / white matter ratio calculation in the maximum coverage of gray matter is the number of electrodes located in the gray matter, and for each electrode Contains a series of electrode contacts J, and the sampling range of each contact is calculated as c j :j∈{1,…,J}, with radius c r , when contact c j ±c r The records within the gray matter are considered valid, and the gray / white matter ratio is calculated as: where f gm( · ) is the sampling range c j The distance from the gray matter surface, H[·] is the Heaviside function, which is negative when it is in the gray matter, and J is the number of contacts; the damage range is calculated by automatically planning the path of a cylinder with a radius of 2 cm, where G is the maximum and the damage coverage rate L is the maximum, and the Dice index is used for calculation, specifically: Where V abla is the volume of the cylinder to be damaged, V lesion is the lesion volume.