MRI (Magnetic Resonance Imaging) image laryngeal tumor three-dimensional reconstruction system for preoperative auxiliary planning

Abstract

The invention relates to the technical field of three-dimensional image data processing, in particular to an MRI (Magnetic Resonance Imaging) image laryngeal tumor three-dimensional reconstruction system for preoperative auxiliary planning. According to the system, for each voxel point in an initial three-dimensional model of the throat, a preoperative three-dimensional diagram result of the throat is constructed through tissue image features. In the intraoperative stage, expected tissue change characteristics are obtained according to information of an instrument coordinate track in a laryngeal three-dimensional diagram structure, the expected tissue change characteristics are compared with actual tissue change characteristics embodied by an electrotome power feedback data sequence, then the tissue anomaly type under the coordinate track is determined, and labeling updating is carried out in an initial three-dimensional model. According to the method, the abnormal type of the detail tissue feature region can be effectively identified by comparing the expected tissue change feature with the actual tissue change feature, and then real-time updating and rendering are performed on the three-dimensional model through a labeling method, so that effective response of intraoperative features is realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of three-dimensional image data processing, in particular to a preoperative auxiliary planning MRI image laryngeal tumor three-dimensional reconstruction system. BACKGROUND

[0002] Laryngeal cancer is a common malignant tumor of the head and neck, and surgical resection is the main treatment method. The core goal of surgery is to completely remove the tumor while maximizing the preservation of key functions such as voice and respiration in the larynx. In order to overcome the limitations of two-dimensional reading, a preoperative planning system based on three-dimensional reconstruction has been proposed. Existing technology can convert two-dimensional image sequences into three-dimensional geometric models based on two-dimensional magnetic resonance imaging (MRI), and intuitively display the spatial adjacency relationship between the tumor and the surrounding organs. However, in existing technology, the three-dimensional model constructed from MRI images focuses more on the macroscopic geometric shape of the tissue, which makes it difficult to effectively distinguish between tissues with similar shapes but different histological characteristics, such as distinguishing between postoperative fibrotic scars and residual micro-tumor lesions on the model. And the existing three-dimensional model of preoperative auxiliary planning is static, and the surgical navigation information it provides is completely based on preoperative data, and it cannot respond to dynamic processes during surgery. When the actual situation of the surgery deviates from the preoperative planning, the three-dimensional model cannot provide effective updates and responses. SUMMARY

[0003] In order to solve the technical problems of the existing technology of constructing a preoperative auxiliary planning laryngeal three-dimensional model that cannot accurately identify the details of the tissue feature area, and cannot provide effective responses, the purpose of the present application is to provide a preoperative auxiliary planning MRI image laryngeal tumor three-dimensional reconstruction system, and the technical solution adopted is as follows:

[0004] The present application proposes a preoperative auxiliary planning MRI image laryngeal tumor three-dimensional reconstruction system, which comprises:

[0005] A preoperative three-dimensional image atlas construction module is used to construct an initial three-dimensional model of the laryngeal surgery region of interest according to the MRI image. For each voxel point in the initial three-dimensional model, the tissue image characteristics of each voxel point are obtained according to the spatial T2 signal intensity variation characteristics and the ADC value variation characteristics. The difference in tissue image characteristics between adjacent voxel points is used as an edge weight value to construct a preoperative laryngeal three-dimensional graph structure.

[0006] An intraoperative information acquisition module is configured to acquire a coordinate trajectory of an intraoperative high-frequency electrotome tip in the three-dimensional graph structure and a sequence of electrotome power feedback data in real time according to a surgical navigation system; the coordinate trajectory corresponds to a voxel point sequence formed by corresponding voxel points in the three-dimensional graph structure; an edge weight value sequence corresponding to the voxel point sequence is obtained; a change feature of the edge weight value sequence is taken as an expected tissue change feature; and a change feature of the sequence of electrotome power feedback data is taken as an actual tissue change feature.

[0007] A three-dimensional model auxiliary updating module is configured to compare the expected tissue change feature and the actual tissue change feature, determine an abnormal type of tissue under the coordinate trajectory according to a comparison result, and label the initial three-dimensional model according to the abnormal type.

[0008] Further, the tissue image feature is a two-dimensional feature composed of a T2 signal intensity change rate and an ADC value dispersion degree of a neighborhood space.

[0009] Further, the T2 signal intensity change rate is a spatial gradient module length of a voxel point.

[0010] Further, the ADC value dispersion degree is a standard deviation of an ADC value in a neighborhood space of a voxel point.

[0011] Further, the edge weight value is an Euclidean distance of tissue image features between adjacent voxel points.

[0012] Further, the expected tissue change feature is a standard deviation of the edge weight value sequence after normalization processing.

[0013] Further, the actual tissue change feature is a standard deviation of the sequence of electrotome power feedback data after normalization.

[0014] Further, the system further comprises a warning module configured to take a difference between the expected tissue change feature and the actual tissue change feature as a deviation index; acquire a definition index of an endoscope video in the surgical navigation system; obtain a navigation reliability index of the surgical navigation system according to the deviation index and the definition index; and feed back a warning signal according to the navigation reliability index.

[0015] Further, the method for obtaining the navigation reliability index comprises:

[0016] The deviation index is negatively correlated and mapped, and then multiplied by the definition index to obtain the navigation reliability index of the surgical navigation system.

[0017] Further, the method for determining the abnormal type of tissue under the coordinate trajectory according to the comparison result comprises:

[0018] If the actual tissue change feature is greater than the preset expected tissue change feature, and the difference between the expected tissue change feature and the actual tissue change feature is greater than the preset difference error threshold, it indicates that the abnormal type of the laryngeal region corresponding to the coordinate trajectory is unexpected high feedback tissue.

[0019] If the actual tissue change feature is less than the preset expected tissue change feature, and the difference between the expected tissue change feature and the actual tissue change feature is greater than the preset difference error threshold, it indicates that the abnormal type of the laryngeal region corresponding to the coordinate trajectory is unexpected low feedback tissue.

[0020] The present application has the following beneficial effects:

[0021] Based on the initial three-dimensional model, in order to reflect the tissue characteristics of the candidate tissue region, and to effectively compare and respond in the intraoperative stage, a quantitative reference derived from the image itself is provided by constructing a graph structure. Further, in the intraoperative information acquisition module, the expected tissue change feature is determined by real-time monitoring of the coordinate trajectory in the surgical navigation system combined with the preoperative candidate three-dimensional graph structure, and the actual tissue change feature is further obtained based on the real-time feedback of the electrotome power feedback data sequence. By quantifying the two features, the expected data and the actual data can be effectively compared, and the strong and weak performances are analyzed based on the comparison results of the two, the abstract deviation value is converted into the actual abnormal type, and then the abnormal type is rendered to the initial three-dimensional model in real time, so as to realize the labeling of the model, that is, the model makes an effective response to the electrotome feedback in the operation. The present application can effectively identify the abnormal type of the detailed tissue feature region by comparing the expected tissue change feature and the actual tissue change feature, and then update and render in real time in the three-dimensional model through the labeling method, so as to realize the effective response of the intraoperative feature. BRIEF DESCRIPTION OF DRAWINGS

[0022] In order to more clearly illustrate the technical solutions and advantages of the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or the prior art description. Obviously, the drawings in the following description are only some embodiments of the present application, and those skilled in the art can obtain other drawings according to these drawings without creative labor.

[0023] Figure 1 A preoperative auxiliary planning MRI image laryngeal tumor three-dimensional reconstruction system block diagram provided by an embodiment of the present application. DETAILED DESCRIPTION

[0024] In order to further clarify the technical means and effects taken by the present application to achieve the predetermined inventive purpose, the following describes in detail the specific implementation, structure, features and effects of a preoperative auxiliary planning MRI image laryngeal tumor three-dimensional reconstruction system according to the present application, combined with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, the specific features, structures or characteristics in one or more embodiments can be combined in any suitable form.

[0025] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.

[0026] The specific scheme of the preoperative auxiliary planning MRI image laryngeal tumor three-dimensional reconstruction system provided by the present application is described in detail below in combination with the accompanying drawings.

[0027] Please refer to Figure 1 which shows a block diagram of a preoperative auxiliary planning MRI image laryngeal tumor three-dimensional reconstruction system according to an embodiment of the present application, which includes a preoperative three-dimensional image atlas construction module 101, an intraoperative information acquisition module 102, and a three-dimensional model assisted updating module 103.

[0028] In the embodiment of the present application, the basic three-dimensional modeling method is still obtained by using the prior art, that is, in the preoperative three-dimensional image atlas construction module 101, an initial three-dimensional model of the laryngeal surgery region of interest needs to be constructed according to the MRI image. The initial three-dimensional model is the basic processing object of the embodiment of the present application, and the subsequent modules can be regarded as further analysis and updating thereof.

[0029] In the embodiment of the present application, the initial three-dimensional model acquisition method of the laryngeal surgery region of interest includes: acquiring three-dimensional data of T2 weighted imaging sequence, and calculating the corresponding apparent diffusion coefficient (ADC) atlas from the diffusion weighted imaging sequence data. Subsequently, three-dimensional rigid registration is performed on the two three-dimensional data sets. The purpose of this operation is to eliminate the spatial position deviation caused by the slight movement of the patient during different sequence scanning, so as to ensure that the voxels with the same index coordinates in the two data sets can correspond to the same anatomical position after the registration is completed. After registration, the three-dimensional region of interest related to the surgery is outlined by the operator on the T2 weighted image, which should completely contain the visible tumor, the key anatomical structure of the larynx and the potential surgical range. It should be noted that the specific image registration method is a well-known technical means to those skilled in the art, and will not be described here.

[0030] The T2 signal intensity is an important parameter in nuclear magnetic resonance imaging for evaluating tissue characteristics, and the intensity change is closely related to the water content and lesion properties of the tissue; the ADC value (apparent diffusion coefficient) is a core parameter for quantifying the diffusion ability of water molecules in nuclear magnetic resonance imaging (MRI), and therefore the two parameters can represent the tissue characteristics of each voxel point in the initial three-dimensional model. The preoperative three-dimensional image atlas construction module 101 further considers that the change characteristics can reflect the tissue characteristics in space, and therefore uses the T2 signal intensity change characteristics in space and the ADC value change characteristics to obtain the tissue image characteristics of each voxel point.

[0031] The preoperative three-dimensional image atlas construction module 101 further regards each voxel point in the initial three-dimensional model as a node of a graph, and for each node, all the spatially adjacent nodes are connected to form edges of the graph, and the edge weight of each edge in the graph is the difference in tissue image characteristics between adjacent voxel points. The edge weight directly and quantitatively describes the difference in the comprehensive imaging characteristics that needs to be crossed when moving from one tissue element to another tissue element adjacent thereto, and a larger edge weight means that the instrument is crossing a boundary where the tissue characteristics change dramatically. Thus, the preoperative laryngeal three-dimensional graph structure is constructed, which not only contains the geometric characteristics of the tissue region, but also contains the tissue characteristics at each position on the tissue region. The atlas encapsulates the positions of all the voxel points in the region of interest, the tissue characteristics of the voxel points, and the tissue characteristic difference relationship between the voxel points as a unified object, which facilitates the processing of subsequent modules in the system.

[0032] Preferably, in the embodiment of the present application, the tissue image characteristics are a two-dimensional feature composed of the T2 signal intensity change rate and the ADC value dispersion in the neighborhood space. The T2 signal intensity change rate can reflect the spatial change intensity of the water content of the target voxel point; and the ADC value dispersion is used to reflect the spatial non-uniformity of the tissue cell density in the local microenvironment of the point, and the local ADC value dispersion of tumor tissue is different from that of normal and uniform structure tissue due to the proliferation of cells and the disordered arrangement.

[0033] The T2 signal intensity change rate is the spatial gradient module length of the voxel point, and the embodiment of the present application uses the central difference method to estimate the T2 signal intensity change rate of the target voxel point in three spatial axes, and then calculates the square root of the sum of squares to obtain the spatial gradient module length; and the ADC value dispersion is the standard deviation of the ADC value in the neighborhood space of the voxel point. The standard deviation is a well-known technical feature for those skilled in the art, and the specific acquisition method is not described again. In the embodiment of the present application, the neighborhood space is set as a 3x3x3 cubic region in space.

[0034] Preferably, the edge weight in the embodiment of the present application is the Euclidean distance of the tissue image features between adjacent voxel points.

[0035] In order to realize that the initial three-dimensional model of preoperative auxiliary planning can effectively respond in the operation, the real-time motion state of the surgical instrument in the surgical navigation system should be analyzed. The intraoperative information acquisition module 102 can generate a contrast feature index for real-time and quantitative evaluation of the degree of coincidence between the current operation and the preoperative image prediction by fusing and analyzing the real-time spatial position information of the surgical instrument and the real-time power feedback signal of the energy instrument, and comparing them with the preoperative laryngeal three-dimensional graph structure. Therefore, the intraoperative information acquisition module 102 can first obtain the coordinate trajectory of the high-frequency electrotome tip in the three-dimensional graph structure according to the surgical navigation system in real time, as well as the electrotome power feedback data sequence.

[0036] The method for obtaining the coordinate trajectory includes: connecting with the standard interface of the surgical navigation system to continuously obtain the three-dimensional spatial coordinates of the high-frequency electrotome tip at a preset fixed frequency, and the three-dimensional spatial coordinates are obtained based on the surgical navigation system, so that the coordinate system in the preoperative laryngeal three-dimensional graph structure needs to be registered, and then the coordinate trajectory is obtained. In the embodiment of the present application, the preset fixed frequency is set to 30 times per second. In the embodiment of the present application, the coordinate trajectory is obtained with 1 second as a time window to be analyzed.

[0037] The method for obtaining the electrotome power feedback data sequence includes: connecting the power output port of the high-frequency electrotome energy platform host through the power feedback acquisition device, continuously collecting the instantaneous output power value of the electrotome at the same collection frequency, and obtaining the electrotome power feedback data sequence in the time window to be analyzed. In the embodiment of the present application, the collection frequency of the electrotome power feedback data sequence is set to 100 times per second, and the coordinate trajectory is obtained with 1 second as a time window to be analyzed.

[0038] To achieve the type identification of the intraoperative work organization area and the real-time response in the three-dimensional model according to the intraoperative information, the embodiment of the present application adopts the method of comparing the expected tissue change characteristics with the actual tissue change characteristics. Through the comparison between the expected and the actual, the tissue characteristics of the work area at the real-time intraoperative stage can be determined. Therefore, the intraoperative information acquisition module 102 forms a voxel point sequence by the corresponding voxel points of the coordinate trajectory in the three-dimensional graph structure, and obtains the edge weight value sequence corresponding to the voxel point sequence, that is, each element in the edge weight value sequence is the edge weight value between the adjacent two voxel points in the voxel point sequence. It should be noted that because the coordinate trajectory is continuous, the voxel points in the voxel point sequence are also continuous, and there is an edge weight value between the adjacent two voxel points in the sequence. The edge weight value sequence can directly reflect the expected result before the operation, that is, the expected tissue characteristic difference crossed by the electrotome in this small movement. The embodiment of the present application further quantifies the change characteristics of the edge weight value sequence to obtain the preoperative expected tissue change characteristics. The preoperative expected tissue change characteristics objectively reflect the fluctuation degree of the tissue image characteristic change predicted by the three-dimensional model constructed by the preoperative auxiliary planning under the work path of the electrotome. The higher the preoperative expected tissue change characteristics, the more the work path passes through the region with the dramatic and uneven change of the tissue characteristics; otherwise, it means that it passes through a relatively uniform region.

[0039] To achieve effective comparison with the preoperative expected tissue change characteristics, the intraoperative information acquisition module 102 further takes the change characteristics of the electrotome power feedback data sequence as the actual tissue change characteristics. The actual tissue change characteristics objectively reflect the fluctuation degree of the tissue electrical impedance change embodied by the energy feedback signal on the work path.

[0040] It should be noted that the embodiment of the present application considers that there will be a dimensional influence problem when comparing data, so the edge weight value sequence and the electrotome power feedback data sequence need to be normalized before the change characteristic acquisition process. The embodiment of the present application adopts range standardization for normalization processing, determines the maximum value and the minimum value under the respective dimension, and realizes the normalization under the respective dimension based on the maximum value and the minimum value, thereby obtaining the normalized edge weight value sequence and the normalized electrotome power feedback data sequence. Through the normalization processing, the distribution form of the original data can be preserved, and the influence of the dimension can be eliminated, and the data value range is limited between 0 and 1.

[0041] The three-dimensional model assisted updating module 103 can compare the expected tissue change characteristics with the actual tissue change characteristics based on the information of the above-mentioned modules, determine the abnormal type of the tissue under the coordinate trajectory according to the comparison result, that is, determine whether the abnormal type belongs to the high feedback tissue or the low feedback tissue. And mark in the initial three-dimensional model according to the abnormal type.

[0042] Preferably, in the embodiments of the present application, the system further comprises a warning module, the warning module is configured to take the difference between the expected tissue change feature and the actual tissue change feature as a deviation index. That is, the greater the deviation index, the greater the difference between the expected and the actual, and the less reliable the surgical navigation system is in judging the actual candidate tissue at this time. Further, an image clarity index of the endoscope video in the surgical navigation system is obtained, and the image clarity index is used to further represent the reliability of the surgical navigation system; a navigation reliability index of the surgical navigation system is obtained according to the deviation index and the image clarity index; and a warning signal is fed back according to the navigation reliability index.

[0043] In the embodiments of the present application, in order to effectively compress the dynamic range of the data, so that the deviation coefficient is sensitive to both small and large original values, the embodiments of the present application select a logarithmic transformation method to obtain the deviation index, which is expressed by the formula:

[0044] A = |log2(B+a)-log2(C+a)|; wherein A is the deviation index, B is the expected tissue change feature, C is the actual tissue change feature, and a is a hyperparameter. The hyperparameter is a very small normal number, and the purpose is to prevent the logarithmic calculation error that occurs when B and C result in 0, and to ensure the numerical stability of the algorithm, and the embodiments of the present application are set to 0.1.

[0045] The deviation index is a dimensionless, non-negative scalar. The size of its value directly and robustly quantifies the difference between the expected tissue change feature and the actual tissue change feature. A deviation index close to zero indicates that the two are highly consistent, that is, the surgical operation conforms to the expectation of the preoperative image; and a value significantly greater than zero indicates that there is a significant deviation between the two, which suggests that an unexpected tissue structure or lesion may be encountered.

[0046] In the embodiments of the present application, the image clarity index can be obtained by the color saturation and the gradient sum of each frame of image in the endoscope video, wherein the color saturation is used to detect the red screen caused by bleeding, and the gradient sum is used to detect the blur caused by delay or defocus. Therefore, the sum of the normalized color saturation and the normalized gradient sum of each frame of image is taken as the initial image clarity index, and the average initial image clarity index of all frames of image in the endoscope video is taken as the image clarity index. It should be noted that the method for obtaining the color saturation and the gradient sum is a technical means known to those skilled in the art, and will not be described in detail. The normalization method can also use the range standardization, which will not be described in detail.

[0047] Further, in the embodiment of the present application, the navigation reliability index of the surgical navigation system is obtained by multiplying the negative correlation mapping of the deviation index and the definition index. That is, the smaller the deviation index and the higher the definition index, the more the surgical process under the surgical navigation system meets the expectation, and the safer the surgical process. The embodiment of the present application can divide the early warning signals into three levels, i.e., safety, vigilance and danger, according to the size of the navigation reliability index, and indicate different levels through different sounds (such as different frequencies and rhythms of sound) and different color signal lights (such as green, yellow and red three colors) in the system. The navigation reliability threshold interval of the early warning signals of the three levels can be set according to the actual implementation scene, and the embodiment of the present application will not be described again.

[0048] It should be noted that the method of negative correlation mapping and normalization in the embodiment of the present application adopts an exponential function mapping method, and the reciprocal of the deviation index is taken as the power of the exponential function with the natural constant as the base number. The output result of the exponential function is the result of negative correlation mapping and normalization.

[0049] Preferably, in the embodiment of the present application, the abnormal type of the tissue under the coordinate track is determined according to the comparison result, comprising:

[0050] If the actual tissue change feature is greater than the preset expected tissue change feature, and the difference between the expected tissue change feature and the actual tissue change feature is greater than the preset difference error threshold, it is indicated that the instrument path is displayed as a region with relatively uniform tissue characteristics on the preoperative image, but the power feedback of the high-frequency electrotome appears a sharp fluctuation. This usually corresponds to cutting into small structures with significantly different electrical impedance, such as small nutrient blood vessels, nerve bundles or dense fibrotic scar tissue, which are not shown or underestimated on the image, so the abnormal type of the laryngeal region corresponding to the coordinate track is unexpected high feedback tissue. In the embodiment of the present application, the coordinate track in the initial three-dimensional model is rendered and marked with a specific highlight color (such as blue) on the surgical navigation interface, and the interface will pop up a text prompt "unexpected high feedback tissue is encountered", and the rendering update of the three-dimensional model can help the doctor to judge the surgical trend.

[0051] If the actual tissue variation feature is less than the preset expected tissue variation feature, and the difference between the expected tissue variation feature and the actual tissue variation feature is greater than the preset difference error threshold, it indicates that the instrument path is passing through a boundary identified on the preoperative image as a boundary of a dramatic change in tissue characteristics (e.g., the junction of a tumor and normal muscle), but the power feedback of the high-frequency electrotome is abnormally smooth. This usually corresponds to entering a region of highly homogeneous histological characteristics, such as a large mass of necrotic tissue, a cystic region, or a liquefied area, which lacks effective electrical impedance changes, and the abnormal type of the laryngeal region corresponding to the coordinate trajectory is unexpected low feedback tissue. Similarly, a specific highlight color (e.g., purple) can be selected for rendering and marking, and a pop-up text prompt "Tissue feedback is lower than expected, may enter necrotic / cystic area" is displayed.

[0052] It should be noted that the absolute value of the difference between the expected tissue variation feature and the actual tissue variation feature is obtained, and the difference error threshold is set to 0.6 after normalization. That is, when the normalized absolute value of the difference is greater than 0.6, it is determined that one feature is much greater than the other feature.

[0053] In summary, the embodiment of the present application responds to intraoperative information in real time by real-time, classified, and dynamic color labeling of a three-dimensional model. The quantified one-dimensional feature is mapped back to the three-dimensional model and becomes part of the three-dimensional reconstruction model. The embodiment of the present application is a dynamic supplement to the original static model. The doctor can directly see the abnormal area position and determine the abnormal type, thereby making more accurate judgments and operation adjustments. For example, the trajectory of the blue mark is selected to protect the blood vessels, or it is determined whether the purple mark area needs to be cleaned more thoroughly. The embodiment of the present application not only provides risk warnings, but also feeds back the energy feedback information obtained in real time during the operation to the preoperative planning model in an intuitive and interpretable manner, realizes the closed loop of preoperative auxiliary planning and intraoperative reality, and greatly improves the practical value and decision support capability of the three-dimensional reconstruction system in complex cases.

[0054] It should be noted that the above-mentioned embodiment of the present application is only for description, and does not represent the advantages and disadvantages of the embodiment. The processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are possible or can be advantageous.

[0055] Each embodiment in the specification is described in a progressive manner, and the same or similar parts between each embodiment can be referred to each other. Each embodiment mainly describes the differences from other embodiments.

Claims

1. A three-dimensional reconstruction system for laryngeal tumors based on MRI images for preoperative planning, characterized in that, The system includes: The preoperative 3D image atlas construction module is used to construct an initial 3D model of the region of interest for laryngeal surgery based on MRI images. For each voxel in the initial 3D model, the tissue image features of each voxel are obtained based on the spatial T2 signal intensity variation characteristics and ADC value variation characteristics. The differences in tissue image features between adjacent voxel points are used as edge weights to construct the preoperative 3D map structure of the larynx. The intraoperative information acquisition module is used to acquire, in real time, the coordinate trajectory of the high-frequency electrosurgical tip in the three-dimensional graph structure and the electrosurgical power feedback data sequence based on the surgical navigation system; the voxel points corresponding to the coordinate trajectory in the three-dimensional graph structure constitute a voxel point sequence, and the edge weight sequence corresponding to the voxel point sequence is obtained; the change characteristics of the edge weight sequence are used as the expected tissue change characteristics before surgery; the change characteristics of the electrosurgical power feedback data sequence are used as the actual tissue change characteristics; The 3D model-assisted update module is used to compare the expected tissue change characteristics with the actual tissue change characteristics, determine the anomaly type of the tissue under the coordinate trajectory based on the comparison results, and annotate it in the initial 3D model according to the anomaly type.

2. The three-dimensional reconstruction system for laryngeal tumors based on MRI images for preoperative planning as described in claim 1, characterized in that, The tissue imaging features are two-dimensional features consisting of the T2 signal intensity change rate and the ADC value dispersion in the neighborhood space.

3. The three-dimensional reconstruction system for laryngeal tumors based on MRI images for preoperative planning according to claim 2, characterized in that, The rate of change of the T2 signal intensity is the spatial gradient modulus of the voxel point.

4. The three-dimensional reconstruction system for laryngeal tumors based on MRI images for preoperative planning according to claim 2, characterized in that, The dispersion of the ADC value is the standard deviation of the ADC value in the neighborhood space of the voxel point.

5. The three-dimensional reconstruction system for laryngeal tumors based on MRI images for preoperative planning according to claim 1, characterized in that, The edge weight is the Euclidean distance between adjacent voxel points representing tissue image features.

6. The three-dimensional reconstruction system for laryngeal tumors based on MRI images for preoperative planning according to claim 1, characterized in that, The expected preoperative tissue changes are the standard deviation of the normalized edge weight sequence.

7. The three-dimensional reconstruction system for laryngeal tumors based on MRI images for preoperative planning according to claim 1, characterized in that, The actual tissue change characteristics are the standard deviation of the normalized electrosurgical power feedback data sequence.

8. The three-dimensional reconstruction system for laryngeal tumors based on MRI images for preoperative planning as described in claim 1, characterized in that, The system also includes an early warning module, which is used to use the difference between the expected tissue change characteristics and the actual tissue change characteristics as a deviation index; to obtain the clarity index of the endoscopic video in the surgical navigation system; to obtain the navigation reliability index of the surgical navigation system based on the deviation index and the clarity index; and to provide a feedback early warning signal based on the navigation reliability index.

9. A three-dimensional reconstruction system for laryngeal tumors based on MRI images for preoperative planning, as described in claim 8, is characterized in that... The method for obtaining the navigation reliability index includes: The navigation reliability index of the surgical navigation system is obtained by mapping the negative correlation of the deviation index and multiplying it by the clarity index.

10. A three-dimensional reconstruction system for laryngeal tumors based on MRI images for preoperative planning according to claim 1, characterized in that, The process of determining the anomaly type of the tissue under the coordinate trajectory based on the comparison results includes: If the actual tissue change characteristics are greater than the preset expected tissue change characteristics, and the difference between the expected tissue change characteristics and the actual tissue change characteristics is greater than the preset difference error threshold, then it indicates that the anomaly type of the throat region corresponding to the coordinate trajectory is an unexpected high feedback tissue. If the actual tissue change characteristics are less than the preset expected tissue change characteristics, and the difference between the expected tissue change characteristics and the actual tissue change characteristics is greater than the preset difference error threshold, then it indicates that the anomaly type of the throat region corresponding to the coordinate trajectory is an unexpected low-feedback tissue.

Images