A preoperative condylar bone resorption risk prediction method and system
By preprocessing and feature extraction of preoperative CBCT image data, a model for predicting the risk of condylar bone resorption was constructed, which solved the problem of inaccurate preoperative assessment in existing technologies and achieved accurate prediction of the risk of preoperative condylar bone resorption.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- PEKING UNIV SCHOOL OF STOMATOLOGY
- Filing Date
- 2026-03-30
- Publication Date
- 2026-06-26
AI Technical Summary
Existing methods for assessing temporomandibular joint condylar bone resorption rely on postoperative imaging data, making it difficult to accurately assess risks preoperatively. Furthermore, inconsistent follow-up periods lead to inaccurate assessments.
By preprocessing, segmenting, and extracting features from CBCT images of the temporomandibular joint region of patients before surgery, a risk prediction model for condylar bone resorption is constructed. The model is then used to predict the risk using radiomics features and generate patient risk grading information.
It enables the prediction of condylar bone resorption risk based on preoperative imaging data, improves the ability to characterize early risk features, and allows for risk assessment before treatment without relying on postoperative imaging.
Smart Images
Figure CN122289204A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of oral medicine technology, and more specifically to a method and system for predicting the risk of preoperative condylar bone resorption. Background Technology
[0002] Currently, temporomandibular joint condylar bone resorption is a common and serious complication after orthognathic surgery. Early changes can first occur at the level of cancellous bone structure inside the condyle, and conventional imaging observation is difficult to identify in time before there are obvious changes in the cortical bone or overall morphology.
[0003] However, most existing methods for assessing temporomandibular joint condylar bone resorption use the following two approaches: (1) assessment based solely on postoperative images; (2) assessment based on the characteristic changes between preoperative and postoperative images. Both of these methods are inaccurate and rely solely on postoperative image data, with inconsistent follow-up times, making it difficult to complete risk assessment in the preoperative stage.
[0004] Therefore, how to provide a method for predicting the risk of preoperative condylar bone resorption that can solve the above problems is a problem that urgently needs to be solved by those skilled in the art. Summary of the Invention
[0005] In view of this, the present invention provides a method and system for predicting the risk of preoperative condylar bone resorption. By analyzing and modeling the radiomics characteristics of the preoperative condylar cancellous bone region of the patient, the risk of postoperative condylar bone resorption can be predicted without using any postoperative images as predictive input.
[0006] To achieve the above objectives, the present invention adopts the following technical solution: A method for predicting the risk of preoperative condylar bone resorption includes the following steps: Preoperative small-field CBCT image data of the temporomandibular joint region of the patient were acquired, and the small-field CBCT image data were preprocessed to obtain the corresponding region of interest in the condylar cancellous bone region. The region of interest in the condylar cancellous bone region is segmented and its features are extracted to obtain the corresponding radiomics features; A risk prediction model for condylar bone resorption was constructed, and the radiomics features were input into the risk prediction model for condylar bone resorption to obtain the corresponding risk prediction results; Based on the risk prediction results, patient risk grading information is generated to assist clinicians in making patient management decisions.
[0007] Preferably, the specific processing steps for obtaining the corresponding radiomics features include: The small field-of-view CBCT image data is preprocessed, including image resampling, grayscale normalization, and noise reduction. The preprocessed small field-of-view CBCT image data is segmented into regions of interest (ROIs) to obtain the corresponding segmentation results, i.e., masks. Feature extraction is performed on the segmentation results to obtain corresponding radiomics features, which include 107 features in total, including first-order statistical features, morphological features, and texture features.
[0008] Preferably, the specific processing steps to obtain the corresponding extraction results include: Multiplanar reconstruction (MPR) technology was used to spatially locate the preprocessed small-field-of-view CBCT image data, determine the corresponding anatomical boundaries, and combine gray-level gradient changes to obtain the corresponding extraction results. The specific implementation process is as follows: A three-dimensional anatomical coordinate system was established in the coronal, sagittal, and axial planes using MPR technology to accurately locate the long axis of the condyle. Using the line from the uppermost point of the condyle to the sigmoid notch as the anatomical boundary, a digital contour was drawn along the inner edge of the high-brightness cortical layer on a three-dimensional voxel level, referencing the significant gray-level gradient between cortical and cancellous bone. Through this dual constraint of "anatomical boundary limitation + gray-level feature recognition," the outer dense cortical bone was accurately removed, achieving standardized extraction of the three-dimensional region of interest (ROI) of the internal condylar cancellous bone. The ROI is used to characterize the trabecular structure of the cancellous bone within the condyle.
[0009] Preferably, the condylar bone resorption risk prediction model includes a feature input module, a risk assessment module, and a result output module connected in sequence; The feature input module is used to receive preoperative condylar cancellous bone radiomics features, the risk assessment module is used to generate corresponding condylar bone resorption risk assessment results based on the image features, and the result output module is used to output the risk level or probability of the patient developing condylar bone resorption.
[0010] This invention also provides a preoperative condylar bone resorption risk prediction system, comprising: The acquisition module is used to acquire small-field CBCT image data of the temporomandibular joint region of the patient before surgery, and to preprocess the small-field CBCT image data to obtain the corresponding region of interest of the condylar cancellous bone region. The feature extraction module is used to segment and extract features from the region of interest in the condylar cancellous bone region to obtain the corresponding radiomics features; The risk prediction module is used to construct a risk prediction model for condylar bone resorption, and input the radiomics features into the risk prediction model for condylar bone resorption to make predictions and obtain the corresponding risk prediction results. The decision-making module is used to generate patient risk grading information based on the risk prediction results, assisting clinicians in making patient management decisions.
[0011] The present invention also provides a system for predicting the risk of preoperative condylar bone resorption using any of the above embodiments, comprising: The acquisition module is used to acquire small field-of-view CBCT image data of the temporomandibular joint region of the patient before surgery, and to extract features from the small field-of-view CBCT image data to obtain corresponding radiomics features. The prediction module is used to construct a risk prediction model for condylar bone resorption, and input the radiomics features into the risk prediction model for condylar bone resorption to make predictions and obtain the corresponding risk prediction results. The decision-making module is used to classify and manage patients based on the risk prediction results.
[0012] As can be seen from the above technical solution, compared with the prior art, the present invention discloses a method and system for predicting the risk of preoperative condylar bone resorption. Using condylar cancellous bone as the analysis object, it can more fully reflect the internal texture and morphological characteristics of the bone, and improve the ability to characterize early risk characteristics. The prediction process is entirely based on preoperative small field CBCT images, without relying on postoperative image data. It can achieve risk assessment before treatment and predict the risk of postoperative condylar bone resorption. Attached Figure Description
[0013] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.
[0014] Figure 1 A flowchart illustrating the overall process of a method for predicting the risk of preoperative condylar bone resorption provided by this invention; Figure 2 A structural principle block diagram of the condylar bone resorption risk prediction model provided in this embodiment of the invention; Figure 3 The structural principle block diagram of a preoperative condylar bone resorption risk prediction system provided by the present invention. Detailed Implementation
[0015] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0016] See Figure 1 As shown in the figure, this invention discloses a method for predicting the risk of preoperative condylar bone resorption, including the following steps: Preoperative small-field CBCT image data of the temporomandibular joint region of the patient were acquired, and the small-field CBCT image data were preprocessed to obtain the corresponding region of interest in the condylar cancellous bone region. The region of interest in the condylar cancellous bone region is segmented and its features are extracted to obtain the corresponding radiomics features; A risk prediction model for condylar bone resorption was constructed, and the radiomics features were input into the risk prediction model for condylar bone resorption to obtain the corresponding risk prediction results; Based on the risk prediction results, patient risk grading information is generated to assist clinicians in making patient management decisions.
[0017] In a specific embodiment, the specific processing steps for obtaining the corresponding radiomics features include: The small field-of-view CBCT image data is preprocessed, including image resampling, grayscale normalization, and noise reduction. The preprocessed small field-of-view CBCT image data is segmented into regions of interest (ROIs) to obtain the corresponding segmentation results, i.e., masks. Feature extraction is performed on the segmentation results to obtain corresponding radiomics features. These radiomics features include 107 features in total, including first-order statistical features, morphological features, and texture features. Texture features include gray-level co-occurrence matrix (GLCM), gray-level run-length matrix (GLRLM), gray-level size region matrix (GLSZM), gray-level dependency matrix (GLDM), and neighborhood gray-level difference matrix (NGTDM). The specific implementation of feature extraction can use standardized radiomics feature extraction algorithms (such as the PyRadiomics library). Gray-level distribution is characterized by calculating the skewness and kurtosis of voxel distribution. Spatial texture features are extracted using GLCM and GLRLM to quantify the arrangement complexity of trabecular bone. Since bone resorption is accompanied by subtle changes in bone density and trabecular thinning, the extracted gray-level non-uniformity and energy indicators can directly reflect the porosity of the internal bone structure, thereby establishing a correlation between imaging phenotype and pathological state.
[0018] In a specific embodiment, the specific processing steps to obtain the corresponding extraction results include: Multiplanar reconstruction (MPR) technology was used to spatially locate the preprocessed small-field-of-view CBCT image data, determine the corresponding anatomical boundaries, and combine gray-level gradient changes to obtain the corresponding extraction results. The specific implementation process is as follows: A three-dimensional anatomical coordinate system was established in the coronal, sagittal, and axial planes using MPR technology to accurately locate the long axis of the condyle. Using the line from the uppermost point of the condyle to the sigmoid notch as the anatomical boundary, a digital contour was drawn along the inner edge of the high-brightness cortical layer on a three-dimensional voxel level, referencing the significant gray-level gradient between cortical and cancellous bone. Through this dual constraint of "anatomical boundary limitation + gray-level feature recognition," the outer dense cortical bone was accurately removed, achieving standardized extraction of the three-dimensional region of interest (ROI) of the internal condylar cancellous bone. The ROI is used to characterize the trabecular structure of the cancellous bone within the condyle.
[0019] In one specific embodiment, see Figure 2 As shown, the condylar bone resorption risk prediction model includes a feature input module, a risk assessment module, and a result output module connected in sequence. The feature input module is used to receive preoperative condylar cancellous bone radiomics features, the risk assessment module is used to generate corresponding condylar bone resorption risk assessment results based on the image features, and the result output module is used to output the risk level or probability of the patient developing condylar bone resorption.
[0020] The risk assessment module includes a risk prediction unit, a correction unit, a comparison unit, a first management plan output unit, and a second management plan output unit. The risk prediction unit is connected to the correction unit and the comparison unit, respectively, and the comparison unit is connected to the first management plan output unit and the second management plan output unit. The risk prediction unit preferably employs a machine learning-based risk prediction model. In a preferred embodiment, a logistic regression model is used to predict the risk of extracted radiomics features, obtaining the corresponding risk prediction result and the corresponding actual risk threshold. The correction unit can correct the risk prediction result based on the actual risk of historical condylar bone resorption and the corresponding actual occurrence result, and output the corrected risk prediction result and the actual risk threshold. The comparison unit compares the corrected risk prediction result with the actual risk threshold. When they match, it is determined to be low risk, and the management process is recorded by the first management plan output unit, which can manage according to the routine follow-up process. When they do not match, it is determined to be high risk, and the second management plan output unit can perform high-frequency management and output a plan to assist subsequent surgical decisions.
[0021] See Figure 3 As shown, this embodiment of the invention also provides a preoperative condylar bone resorption risk prediction system, comprising: The acquisition module is used to acquire small-field CBCT image data of the temporomandibular joint region of the patient before surgery, and to preprocess the small-field CBCT image data to obtain the corresponding region of interest of the condylar cancellous bone region. The feature extraction module is used to segment and extract features from the region of interest in the condylar cancellous bone region to obtain the corresponding radiomics features; The risk prediction module is used to construct a risk prediction model for condylar bone resorption, and input the radiomics features into the risk prediction model for condylar bone resorption to make predictions and obtain the corresponding risk prediction results. The decision-making module is used to generate patient risk grading information based on the risk prediction results, assisting clinicians in making patient management decisions.
[0022] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to the method section.
[0023] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims
1. A method for predicting the risk of preoperative condylar bone resorption, characterized in that, Includes the following steps: Preoperative small-field CBCT image data of the temporomandibular joint region of the patient were acquired, and the small-field CBCT image data were preprocessed to obtain the corresponding region of interest in the condylar cancellous bone region. The region of interest in the condylar cancellous bone region is segmented and its features are extracted to obtain the corresponding radiomics features; A risk prediction model for condylar bone resorption was constructed, and the radiomics features were input into the risk prediction model for condylar bone resorption to obtain the corresponding risk prediction results; Based on the risk prediction results, patient risk grading information is generated to assist clinicians in making patient management decisions.
2. The method for predicting the risk of preoperative condylar bone resorption according to claim 1, characterized in that, The specific processing steps to obtain the corresponding radiomics features include: The small field-of-view CBCT image data is preprocessed, including image resampling, grayscale normalization, and noise reduction. The preprocessed small-field-of-view CBCT image data is segmented into regions of interest to obtain the corresponding segmentation results. Feature extraction is performed on the segmentation results to obtain corresponding radiomics features, which include first-order statistical features, morphological features, and texture features.
3. The method for predicting the risk of preoperative condylar bone resorption according to claim 2, characterized in that, The specific processing steps to obtain the corresponding extraction results include: Using multiplanar reconstruction (MPR) technology, the preprocessed small-field CBCT image data is spatially located to determine the corresponding anatomical boundaries, and the corresponding extraction results are obtained by combining grayscale gradient changes.
4. The method for predicting the risk of preoperative condylar bone resorption according to claim 1, characterized in that, The condylar bone resorption risk prediction model includes a feature input module, a risk assessment module, and a result output module connected in sequence. The feature input module is used to receive preoperative condylar cancellous bone radiomics features, the risk assessment module is used to generate corresponding condylar bone resorption risk assessment results based on the image features, and the result output module is used to output the risk level or probability of the patient developing condylar bone resorption.
5. A system utilizing the preoperative condylar bone resorption risk prediction method according to any one of claims 1-4, characterized in that, include: The acquisition module is used to acquire small-field CBCT image data of the temporomandibular joint region of the patient before surgery, and to preprocess the small-field CBCT image data to obtain the corresponding region of interest of the condylar cancellous bone region. The feature extraction module is used to segment and extract features from the region of interest in the condylar cancellous bone region to obtain the corresponding radiomics features; The risk prediction module is used to construct a risk prediction model for condylar bone resorption, and input the radiomics features into the risk prediction model for condylar bone resorption to make predictions and obtain the corresponding risk prediction results. The decision-making module is used to generate patient risk grading information based on the risk prediction results, assisting clinicians in making patient management decisions.