Bone drilling force prediction method and device based on CT images

Abstract

The present invention relates to the field of bone drilling force prediction technology, and discloses a bone drilling force prediction method and device based on CT images. The method comprises step S1: generating a three-dimensional image at the drilling position; step S2: obtaining bone density m With thickness x The present invention establishes a nonlinear relationship model of the drilling force changing with the bone density based on the CT image, and predicts the change of the drilling force according to the different distribution of bone density, so as to provide a basis for personalized surgical plan and reduce the surgical risk.

Description

Technical Field

[0001] The invention relates to a method for predicting bone drilling force when drilling a bone during orthopedic surgery, in particular to a method and device for predicting bone drilling force based on CT images, belonging to the technical field of bone drilling force prediction. Background Art

[0002] In the medical field, bone drilling technology is widely used in orthopedic surgery and implant installation. The interaction between the drill bit and bone tissue will generate drilling force, and the magnitude of the drilling force will affect the mechanical damage of bone tissue during the drilling process, which is crucial to the success of the operation and the integrity of the bone tissue. Traditional drilling force prediction methods often rely on experience and experimental data, which cannot fully consider the complexity and individual differences of bone tissue.

[0003] In recent years, the advancement of computed tomography (CT) technology has enabled doctors to obtain high-resolution information about the internal structure of bone tissue, thereby more accurately evaluating bone density and geometric characteristics. By establishing a nonlinear relationship model between drilling force and bone density based on CT image data, the change in drilling force can be better predicted, which can not only reduce surgical risks, but also provide a scientific basis for personalized surgical plans, which has important clinical application value and research significance.

[0004] However, there is still a lack of research on drilling force prediction using CT imaging data, and there is almost no reference material. Summary of the invention

[0005] The purpose of the present invention is to provide a bone drilling force prediction method based on CT images to address the shortcomings of existing bone drilling force prediction technologies. The method can predict changes in bone drilling force under complex conditions, provide a basis for personalized surgical plans, reduce surgical risks, and ensure drilling quality and surgical safety. The present invention also provides a device for implementing the above method.

[0006] The bone drilling force prediction method based on CT images of the present invention comprises the following steps:

[0007] Step S1: obtaining a sample of a type of bone tissue (such as a skull, trunk bone or limb bone) of a human body, performing a computerized tomography scan, and generating a three-dimensional image of the drilling position according to the scan result;

[0008] Step S2: extracting the thickness x of the bone tissue at the drilling position and the bone density m at the thickness based on the three-dimensional image in step S1, wherein the thickness x refers to the distance from any position between the periosteum and the endosteum to the periosteum (the thickness can also be said to be the drilling depth); then obtaining the relationship between the bone density m between the periosteum and the endosteum and the thickness x (drilling depth), performing function fitting on the relationship, and obtaining a fitting function of the bone density m changing with the thickness x (drilling depth);

[0009] Step S3: Perform a drilling experiment, collect drilling force data, and establish a fitting function of drilling force changing with drilling depth (thickness) according to the change of drilling force with drilling depth (thickness), and bring the fitting function of bone density m changing with thickness x (drilling depth) obtained in step S2 into the fitting function of drilling force changing with drilling depth (thickness), and establish a function of drilling force changing with bone density;

[0010] Step S4: Based on the relationship between the bone density m between the periosteum and the endosteum and the thickness x (drilling depth) and the function of the drilling force changing with the bone density established in steps S2 and S3, a drilling force model based on the bone density measurement results is established;

[0011] Step S5: calibrating and verifying the drilling force model based on the bone density measurement result to obtain a determined drilling force model;

[0012] Step S6: The CT image of the bone tissue drilling position and the bone density information obtained before the operation are used to predict the drilling force that varies with the drilling depth using the drilling force model determined in step 5.

[0013] Further:

[0014] The number of bone tissue samples in step S1 is not less than 30.

[0015] The fitting function obtained in step S2 is as shown in the following formula:

[0016] m=1.83x -0.06 ;

[0017] Where: m is bone density, x is thickness (drilling depth).

[0018] The fitting function of the drilling force changing with the drilling depth (thickness) established in step S3 is as shown in the following formula:

[0019] y F,x =1.026x -0.06 ;

[0020] Where: y F,x is the fitting function of the drilling force changing with drilling depth, and x is the thickness (drilling depth).

[0021] The function y of drilling force versus bone density established in step S3 F,m , as shown below:

[0022] y F,m =0.56m,

[0023] Where: m is bone density.

[0024] The drilling force prediction model established in step S4 is shown in the following formula:

[0025] F z,m =0.56m×(F z,lip +F z,c )+ 0.56m×F Tind ;

[0026] Where: m is bone density, F z,m is the drilling force under different bone densities, F z,lip and F z,c are the drilling forces on the drill cutting edge and chisel edge, F Tind is the indentation force on the chisel edge.

[0027] The specific process of step S5 includes the following steps:

[0028] Step S51: randomly select at least one bone sample and obtain the bone density m between the periosteum and the endosteum through CT imaging, and import the bone density at different positions of the periosteum into the drilling force prediction model established in step S4 to obtain the drilling force prediction result;

[0029] Step S52: Perform a drilling experiment to detect changes in drilling force during the drilling process in real time and obtain drilling force test results;

[0030] Step S53: comparing the drilling force prediction result with the drilling force test result;

[0031] Step S54: If the error rate between the drilling force prediction result and the drilling force test result is not less than a set value, the drilling force prediction model is calibrated; otherwise, the drilling force prediction model is determined. The error rate setting value is 5%.

[0032] The drilling force prediction model is calibrated, specifically through a deep learning algorithm. First, collect experimental data, integrate the new experimental data with the drilling force model prediction results to form a comprehensive data set, process the data, create a feature set containing the drilling force experimental results and model prediction results, and add experimental errors as features (so that the deep learning model can learn the difference between the model prediction and the actual results); then, select a deep learning architecture (such as a multi-layer perceptron, convolutional neural network, etc.), and divide the data into a training set, a validation set, and a test set. Use activation functions and regularization methods to prevent overfitting, perform supervised learning by adding actual drilling force data, optimize model parameters, calculate and evaluate the error between the drilling force experimental results and the model prediction results, and quantitatively analyze the improvement effect of the model; finally, adjust the model architecture or parameters based on experimental feedback and comparison results, and optimize and output the calibrated drilling force model.

[0033] The above method is applicable to all types of bone tissues, and a drilling force prediction model for each type of bone tissue can be obtained according to the method.

[0034] A bone drilling force prediction device based on CT images for implementing the above method comprises: a CT imaging module, a feature extraction and analysis module and a drilling force prediction module;

[0035] The CT imaging module is used to obtain the CT image of the target bone tissue, which is completed by the computer tomography imaging system;

[0036] The feature extraction and analysis module is used to extract the thickness and bone density information of bone tissue from the CT image;

[0037] The drilling force prediction module is used to predict the drilling force based on the thickness and bone density information of bone tissue.

[0038] Further:

[0039] The image resolution of the computer tomography system is within 100 μm.

[0040] The present invention obtains a high-resolution image of the target bone tissue through CT images, which can show the internal structure and density distribution of the bone in detail, and then establishes a nonlinear relationship model of the drilling force changing with bone density based on CT images, and predicts the change of drilling force according to the different distributions of bone density. By inputting different drilling parameters, the drilling force under specific conditions can be predicted, providing a basis for personalized surgical plans. This enables doctors to adjust parameters according to the prediction results, optimize the drilling process, reduce surgical risks, ensure drilling quality and surgical safety, and has important clinical application value and research significance. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 It is a flow chart of the bone drilling force prediction method based on CT images in the present invention.

[0042] Figure 2 It is a schematic flow chart of step S1 in the method of the present invention.

[0043] Figure 3 It is a schematic flow chart of step S2 in the method of the present invention.

[0044] Figure 4 Schematic diagram of measuring bone density m and function fitting of the relationship between bone density m and thickness x (drilling depth) in step S2, wherein: (a) is a CT image of a bone sample, and (b) is a measurement result of bone density.

[0045] Figure 5 It is a schematic flow chart of step S3 in the method of the present invention.

[0046] Figure 6It is a schematic diagram of the process of establishing a drilling force prediction model based on bone density measurement results in the method of the present invention.

[0047] Figure 7 It is a schematic flow chart of step S5 in the method of the present invention.

[0048] Figure 8 It is a flow chart of drilling force model calibration.

[0049] Fig. 9 This is a comparison diagram of the predicted value of drilling force and the experimental value under a set of drilling parameters.

[0050] Fig.10 It is a schematic diagram of the structural composition of the bone drilling force prediction device based on CT images in the present invention.

[0051] Fig.11 It is a schematic diagram of the actual application state of the present invention.

[0052] In the figure: 1. CT imaging module; 2. Feature extraction and analysis module; 3. Drilling force prediction module; 4. Human body. DETAILED DESCRIPTION

[0053] The bone drilling force prediction method and device based on CT images provided by the present invention obtain high-resolution images of target bone tissue through CT images, can display the internal structure and density distribution of bone tissue in detail, and accurately predict the changes in drilling force according to the different distributions of bone density, thereby ensuring the drilling quality and surgical safety.

[0054] The bone drilling force prediction method based on CT images in the present invention includes the following steps: Figure 1 The steps shown are described in detail below.

[0055] Step S1: obtaining a plurality of samples of a type of bone tissue of a human body, performing computer tomography scanning, and generating a three-dimensional image of the drilling position according to the scanning results.

[0056] The detailed process of this step is as follows Figure 2 As shown, including:

[0057] Step S11: obtaining multiple samples of one type of bone tissue (mainly including skull, trunk bones or limb bones) of the human body; the more bone tissue samples, the better, preferably no less than 30.

[0058] Step S12: placing all the acquired bone tissue samples under a computer tomography imaging system (with an image resolution within 100 μm) to perform computer tomography (CT) to acquire a CT image of each bone tissue sample.

[0059] Step S13: generating a three-dimensional image of the drilling position of the bone tissue sample according to the CT result.

[0060] Step S2: Based on the three-dimensional image obtained in step S13, the thickness x of the bone tissue at the drilling position and the bone density m at the thickness are extracted to obtain the relationship between the bone density m between the periosteum and the endosteum and the thickness x (drilling depth). The thickness x is consistent with the drilling depth during the drilling process, and the thickness x can also be said to be the drilling depth.

[0061] The obtained relationship between the bone density m between the periosteum and the endosteum and the thickness x (drilling depth) is fitted to obtain a fitting function.

[0062] The detailed process of this step is as follows Figure 3 ,include:

[0063] Step S21: Measure the bone density m between the periosteum and the endosteum according to the three-dimensional image of the drilling position generated in step S13, see Figure 4 , Figure 4 The bone density of 7 thicknesses (distance x from the periosteum) is given in Figure 4 (a) is a CT image of a bone sample. Figure 4 (b) in the figure shows the measurement result of bone density.

[0064] Step S22: The distance x (thickness) from the measuring position to the periosteum is associated with the corresponding bone density m, and a nonlinear fitting function of the change of the bone density with the distance (drilling depth) from the measuring position to the periosteum is established, see Figure 4 , as shown below:

[0065] m=1.83x -0.06 ;

[0066] Where: m is bone density, x is thickness (the distance between the measurement position and the periosteum, which is consistent with the drilling depth).

[0067] Step S3: Conduct a cortical bone drilling experiment and establish a function of drilling force changing with bone density through drilling force data collection and analysis.

[0068] The detailed process of this step is as follows Figure 5 As shown, including:

[0069] Step S31: performing a drilling experiment on the bone sample.

[0070] Step S32: Using a dynamometer to measure the change in drilling force between the periosteum and the endosteum (drilling depth) under different drilling parameters, and establish a fitting function of the change in drilling force with drilling depth, see Figure 6 , as shown below,

[0071] yF，x =1.026x -0.06 ;

[0072] Where: y F,x is the fitting function of the drilling force changing with drilling depth, and x is the thickness (drilling depth).

[0073] Step S33: The fitting function m=1.83x -0.06 Substitute the function y of drilling force changing with drilling depth F,x , we get the function y of drilling force changing with bone density F,m , see Figure 6 , as shown below:

[0074] y F,m =0.56m.

[0075] Step S4: Based on the relationship between the bone density m between the periosteum and the endosteum and the thickness x (drilling depth) and the function of the drilling force changing with the bone density established in steps S2 and S3, a drilling force model based on the bone density measurement results is established in combination with the Merchant theory, see Figure 6 , as shown below:

[0076] F z,m =0.56m×(F z,lip +F z,c )+ 0.56m×F Tind ;

[0077] Where: m is bone density, F z,m is the drilling force under different bone densities, F z,lip and F z,c are the drilling forces on the drill cutting edge and chisel edge, F Tind F is the indentation force on the chisel edge. z,lip 、F z,c 、F Tind The Merchant theory is used to calculate the unit normal force on the cutting edge of the drill bit (dF n ) and unit friction (dF f ), as shown below:

[0078]

[0079]

[0080] Where A c is the area of ​​the cutting layer (mm 2 ), k n and k fare the specific normal force and specific friction force respectively.

[0081] Step S5: calibrate and verify the drilling force model.

[0082] The detailed process of this step is as follows Figure 7 As shown, including:

[0083] Step S51: randomly select at least one bone sample and obtain the bone density m between the periosteum and the endosteum through CT imaging, and import the bone density at different positions of the periosteum into the drilling force model established in step S4 to obtain the drilling force prediction result.

[0084] Step S52: Perform a drilling experiment to detect the real-time changes of the drilling force signal during the drilling process in real time to obtain the drilling force test results; the drilling force is measured by a dynamometer, and the force signal measurement error of the dynamometer is within 0.1%.

[0085] Step S53: comparing the drilling force prediction result with the drilling force test result.

[0086] Step S54: calibrate the drilling force model and output the calibrated drilling force model. Specifically, if the error rate between the drilling force prediction result and the drilling force test result is less than 5%, directly output the drilling force model; if the error rate is ≥5%, calibrate the drilling force model.

[0087] Specifically, the calibration of the drilling force model is completed through a deep learning algorithm. The detailed process is as follows: Figure 8 As shown. First, a large amount of experimental data is collected, and the new experimental data is integrated with the drilling force model prediction results to form a comprehensive data set. The data is processed to create a feature set containing the drilling force experimental results and the model prediction results. Experimental errors are added as features so that the deep learning model can learn the difference between the model prediction and the actual results. Then, a suitable deep learning architecture (such as multi-layer perceptron, convolutional neural network, etc.) is selected, and the data is divided into training set, validation set and test set, using appropriate activation functions and regularization methods to prevent overfitting. Supervised learning is performed by adding actual drilling force data to optimize model parameters. The error between the drilling force experimental results and the model prediction results is calculated and evaluated, and the improvement effect of the model is quantitatively analyzed. Finally, according to the experimental feedback and comparison results, the model architecture or parameters are adjusted, and the calibrated drilling force model is optimized and output.

[0088] Step S6: Before surgery, CT images and bone density information of the bone tissue at the drilling position are obtained, and the drilling force model determined after calibration is used to predict the drilling force that varies with drilling depth.

[0089] Fig. 9 A comparison between the predicted values ​​of drilling force and the experimental values ​​under a set of drilling parameters is given, and the results prove the accuracy of the drilling force prediction model established by the present invention.

[0090] The present invention also provides a bone cutting force prediction device based on CT images to implement steps S6 and S7 in the above-mentioned bone cutting force prediction method based on CT images. Fig.10 As shown, it includes: a CT imaging module 1, a feature extraction and analysis module 2, and a drilling force prediction module 3. The CT imaging module is used to obtain a CT image of the target bone tissue, which is completed by a computer tomography imaging system. The feature extraction and analysis module is used to extract the thickness and bone density information of the bone tissue from the CT image. The drilling force prediction module is used to predict the drilling force based on the bone tissue thickness and bone density information.

[0091] The actual application scenarios of the present invention are as follows Fig.11 As shown, the human body 4 is in the CT imaging module 1 , the bone density is measured by the feature extraction and analysis module 2 , and the drilling force is predicted by the drilling force prediction module 3 .

Claims

1. A bone drilling force prediction method based on CT images, characterized in that: The following steps are involved: Step S1: obtaining a sample of a type of bone tissue from a human body, performing a computerized tomography scan, and generating a three-dimensional image of the drilling position according to the scan result; Step S2: extracting the thickness x of the bone tissue at the drilling position and the bone density m at the thickness based on the three-dimensional image in step S1, wherein the thickness x refers to the distance from any position between the periosteum and the endosteum to the periosteum; and then obtaining the relationship between the bone density m between the periosteum and the endosteum and the thickness x, performing function fitting on the relationship and obtaining a fitting function of the bone density m changing with the thickness x; Step S3: Perform a drilling experiment, collect drilling force data, establish a fitting function of drilling force changing with drilling depth according to the change of drilling force with drilling depth, and bring the fitting function of bone density m changing with thickness x obtained in step S2 into the fitting function of drilling force changing with drilling depth to establish a function of drilling force changing with bone density; Step S4: Based on the relationship between the bone density m and the thickness x between the periosteum and the endosteum and the function of the drilling force changing with the bone density established in steps S2 and S3, respectively, a drilling force model based on the bone density measurement results is established; Step S5: calibrating and verifying the drilling force model based on the bone density measurement result to obtain a determined drilling force model; Step S6: The CT image of the bone tissue drilling position and the bone density information obtained before the operation are used to predict the drilling force that varies with the drilling depth using the drilling force model determined in step 5.

2. The bone drilling force prediction method based on CT images according to claim 1, characterized in that: The number of bone tissue samples in step S1 is not less than 30.

3. The bone drilling force prediction method based on CT images according to claim 1, characterized in that: The fitting function obtained in step S2 is as shown in the following formula: m=1.83x -0.06 , Where: m is bone density, x is thickness.

4. The bone drilling force prediction method based on CT images according to claim 1, characterized in that: The fitting function of the drilling force changing with the drilling depth established in step S3 is as shown in the following formula: the F,x =1.026x -0.06 , Where: y F,x is the fitting function of drilling force changing with drilling depth, and x is the thickness.

5. The bone drilling force prediction method based on CT images according to claim 1, characterized in that: The function y of drilling force versus bone density established in step S3 F,m , as shown below: and F,m =0.56m, Where: m is bone density.

6. The bone drilling force prediction method based on CT images according to claim 1, characterized in that: The drilling force prediction model established in step S4 is shown in the following formula: F z,m =0.56m×(F z,lip +F z,c )+ 0.56m×F Tind , Where: m is bone density, F z,m is the drilling force under different bone densities, F z,lip and F z,c are the drilling forces on the drill cutting edge and chisel edge, F Tind is the indentation force on the chisel edge.

7. The bone drilling force prediction method based on CT images according to claim 1, characterized in that: The specific process of step S5 includes the following steps: Step S51: randomly select at least one bone sample and obtain the bone density m between the periosteum and the endosteum through CT imaging, and import the bone density at different positions of the periosteum into the drilling force prediction model established in step S4 to obtain the drilling force prediction result; Step S52: Perform a drilling experiment to detect changes in drilling force during the drilling process in real time and obtain drilling force test results; Step S53: comparing the drilling force prediction result with the drilling force test result; Step S54: If the error rate between the drilling force prediction result and the drilling force test result is not less than a set value, the drilling force prediction model is calibrated; otherwise, the drilling force prediction model is determined.

8. The bone drilling force prediction method based on CT images according to claim 7, characterized in that: The error rate setting value in step S54 is 5%.

9. The bone drilling force prediction method based on CT images according to claim 7, characterized in that: In step S54 The calibration of the drilling force prediction model is completed through a deep learning algorithm: first, the experimental data is collected, and the new experimental data is integrated with the drilling force model prediction results to form a comprehensive data set. The data is processed to create a feature set containing the drilling force experimental results and the model prediction results, and the experimental error is added as a feature; Then, a deep learning architecture is selected, and the data is divided into training set, validation set and test set. Activation functions and regularization methods are used to prevent overfitting. Supervised learning is performed by adding actual drilling force data, and model parameters are optimized. The error between the drilling force experimental results and the model prediction results is calculated and evaluated, and the improvement effect of the model is quantitatively analyzed. Finally, based on the experimental feedback and comparison results, the model architecture or parameters are adjusted, and the calibrated drilling force model is optimized and output.

10. A bone drilling force prediction device based on CT images for implementing the bone drilling force prediction method based on CT images as claimed in any one of claims 1 to 9, characterized in that: include: CT imaging module, feature extraction and analysis module and drilling force prediction module; The CT imaging module is used to obtain the CT image of the target bone tissue, which is completed by the computer tomography imaging system; The feature extraction and analysis module is used to extract the thickness and bone density information of bone tissue from the CT image; The drilling force prediction module is used to predict the drilling force based on the thickness and bone density information of bone tissue.

Images