Spinal malformation orthopedic prediction method based on artificial intelligence

Abstract

The invention discloses a spinal malformation orthopedic prediction method based on artificial intelligence, and belongs to the technical field of spinal malformation orthopedic. The method comprises the steps that influence factors for spinal malformation orthopedic result prediction are determined; performing digital analysis on the DR image of the spinal malformation patient; constructing a prediction model for a spinal malformation orthopedic result; optimizing the performance of the spinal malformation orthopedic result prediction model; carrying out characterization processing on input parameters of the prediction model; evaluating a spinal malformation orthopedic result predicted by the model; and storing a spinal malformation orthopedic prediction result to the cloud. According to the method, an orthopedic result can be predicted by constructing the model, the accuracy of model prediction can be improved by setting constraint conditions and removing edge parameters, input parameters can be divided according to weights by processing the input parameters of the model, and the accuracy of model prediction is improved. And evaluation and training are carried out according to a prediction result of the model, so that a subsequent treatment scheme of a patient can be optimized.

Description

Technical Field

[0001] The present invention relates to the technical field of spinal deformity correction, and in particular to a spinal deformity correction prediction method based on artificial intelligence. Background Art

[0002] Spinal deformity: A normal spine has a physiological curvature and a balance in the sagittal and coronal positions. If the normal physiological curvature is exceeded, resulting in abnormal coronal, sagittal and transverse shapes, it is called spinal deformity.

[0003] Common spinal deformities include adolescent idiopathic scoliosis (AIS) and adult spinal deformity (ASD). AIS is the most common spinal disease in adolescents, with a global prevalence of 0.5% to 5.2%. AIS is a complex three-dimensional spinal deformity with an unknown etiology. Severe AIS can be accompanied by abnormal cardiopulmonary function, and patients with a scoliotic Cobb angle of >40° require corrective surgery. With the development of an aging population, the prevalence of ASD is gradually increasing. Compared with AIS, the pathophysiological changes of ASD are more complex, the clinical manifestations are more diverse, and the treatment is more difficult.

[0004] In the traditional spinal deformity correction treatment process, DR images are often used to observe the patient's correction status. Not only is the process cumbersome, but it also frequently exposes patients to radiation. With the development of modern medicine, research on optimizing spinal deformity correction treatment plans based on predictive models has demonstrated the great potential of AI and ML technologies in improving surgical success rates, reducing complications, and optimizing patient selection. These technologies not only provide clinicians with powerful decision-making support tools, but also bring more accurate and personalized treatment plans to patients.

[0005] Chinese invention patent CN111341450B discloses a spinal deformity correction prediction method, device and terminal based on artificial intelligence, including obtaining training sample set data, the training sample set data includes: Cobb angle information of thoracic and lumbar curvatures before and after surgery, coronal offset, apical vertebra offset, surgical screw placement position and patient basic information, wherein the coordinates of each spine, the front and back screw placement positions, and the coordinates of other positions are used as input, and the postoperative spine is output instead of the entire spine as the input sequence to predict the postoperative sequence, and the screw placement position includes: according to the preoperative image, by manually marking different segments for screw placement, and giving the corresponding postoperative simulation image when the screw is placed at different positions; through the training sample set data, the preset multiple prediction models are trained separately to obtain training results; based on the training results, the best prediction model is determined; based on the best prediction model, the spinal deformity correction of the target patient is predicted. This invention can improve the accuracy of the model by selecting a suitable learning model according to the patient's spinal deformity, but it does not optimize the model and has certain defects. Summary of the invention

[0006] The present invention provides a spinal deformity correction prediction method based on artificial intelligence. By constructing a support vector machine regression model, it can predict the correction results of spinal deformity patients and provide patients with more accurate and personalized treatment plans.

[0007] In order to solve the above technical problems, the present invention provides the following technical solutions: The present invention provides a spinal deformity correction prediction method based on artificial intelligence, comprising: S1 Establish the influencing factors for predicting the results of spinal deformity correction; S2 digitally analyzes DR images of patients with spinal deformity; S3 constructs a prediction model for spinal deformity correction outcomes; S4 optimizes the performance of the prediction model for spinal deformity correction outcomes; S5 characterizes the input parameters of the prediction model; S6 evaluates the spinal deformity correction results predicted by the model; S7 stores the results of spinal deformity correction prediction in the cloud.

[0008] The beneficial effects brought about by the technical solution provided by the present invention include at least: The present invention establishes influencing factors for predicting spinal deformity correction results; digitally analyzes DR images of patients with spinal deformity; constructs a prediction model for spinal deformity correction results; optimizes the performance of the spinal deformity correction result prediction model; characterizes the input parameters of the prediction model; evaluates the spinal deformity correction results predicted by the model; and stores the spinal deformity correction prediction results in the cloud. The method of the present invention can predict the correction results according to the patient's spinal deformity by constructing a support vector machine regression model, and can provide patients with more accurate and personalized treatment plans. The accuracy of the model prediction can be improved by setting constraints and removing marginal parameters. By characterizing the input parameters of the model, the input parameters can be divided according to the size of the weight, so that the prediction results of the model are more in line with the patient's treatment conditions. By evaluating and training the prediction results of the model, the treatment plans for subsequent patients can be optimized. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0010] Figure 1 A flowchart of a spinal deformity correction prediction method based on artificial intelligence provided by an embodiment of the present invention; Figure 2 A schematic diagram of the Cobb angle measurement steps provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0011] In order to make the objectives, technical solutions and advantages of the present invention more clear, the embodiments of the present invention will be further described in detail below with reference to the accompanying drawings.

[0012] Example An artificial intelligence-based prediction method for spinal deformity correction.

[0013] Please refer to Figure 1 It is a flow chart of a spinal deformity correction prediction method based on artificial intelligence provided by an embodiment of the present invention.

[0014] S1 Establish the influencing factors for predicting the results of spinal deformity correction; S101, the influencing factors include the influencing factors before orthopedics and the influencing factors after orthopedics; S102, the pre-correction influencing factors include the patient's spinal development status and the type of spinal deformity. The patient's spinal development status can be divided into three intervals according to age: under 10 years old, 10 to 16 years old, and over 16 years old. The types of spinal deformity include congenital deformity, degenerative deformity, neuromuscular deformity, and genetic syndrome-related deformity; It should be noted that the age division is used to judge the patient's spinal development status. According to the patient's development status, it can be further subdivided into 0-3 years old, 3-9 years old, 10-13 years old, 14-16 years old, 16-18 years old, and over 18 years old. In the actual prediction process, the finer the age division, the more accurate the factors affecting the patient's spinal development.

[0015] S103, the post-orthopedic influencing factors include the type of orthopedic surgery and rehabilitation conditions. The type of orthopedic surgery includes drug treatment, rehabilitation treatment, and surgical treatment. The rehabilitation conditions include the patient's bone recovery speed, whether the patient has underlying diseases, and the patient's treatment psychology.

[0016] It should be noted that rehabilitation therapy includes exercise therapy and wearing orthopedic gear, underlying diseases include diabetes and hypertension, and the patient's treatment psychology refers to whether the patient is willing to cooperate with the treatment plan.

[0017] S2 digitally analyzes DR images of patients with spinal deformity; S201, the digital analysis process includes dividing the measurement of the sagittal vertical axis and the Cobb angle; It should be noted that in the actual prediction process, the coronal vertical axis can also be added for analysis.

[0018] Please refer to Figure 2 A schematic diagram of the Cobb angle measurement steps provided in an embodiment of the present invention.

[0019] S202, the sagittal vertical axis is divided by measuring the position of the C7 plumb line and the posterior edge of the S1 sacrum. If the C7 plumb line is in front of the posterior superior edge of the S1 sacrum, it is defined as a positive value, and if it is in the rear, it is defined as a negative value. When the sagittal vertical axis is a positive value, it is defined as trunk forward tilt, and when the sagittal vertical axis is a negative value, it is defined as backward tilt. When the absolute value of the sagittal vertical axis exceeds 50 mm, it is defined as sagittal plane decompensation.

[0020] S203, the steps of measuring the Cobb angle are to determine the scoliotic end vertebra, draw a horizontal line at the upper edge of the vertebral body of the upper end vertebra and a horizontal line at the lower edge of the vertebral body of the lower end vertebra, and draw a vertical line on the basis of the horizontal lines of the upper and lower end vertebrae respectively. The intersection angle of the two vertical lines is the Cobb angle. If the Cobb angle is greater than 10 degrees, it is defined as spinal deformity.

[0021] S3 constructs a prediction model for spinal deformity correction outcomes; S301, the result of spinal deformity correction is predicted by constructing a support vector machine regression model, and the regression model infers the training data by constructing a regression function, as shown in formula (S3-1): S3-1 In the formula, Expressed as Function of Represents the input value; represents a matrix vector; express and Multiply the corresponding elements of and then sum them; Represents the bias term, which is used to adjust the output of the function; S302, after the training data inference is completed, a regression equation of the support vector machine regression model is constructed, as shown in Formula (S3-2): S3-2 In the formula, represents the number of support vectors; and denote the Lagrange multiplier and its corresponding dual variable respectively; Represents the kernel function, which is used to calculate the input and support vector The similarity between Represents the bias term.

[0022] It should be noted that in the actual prediction process, It can usually be omitted and expressed as a constant.

[0023] S4 optimizes the performance of the prediction model for spinal deformity correction outcomes; S401, the performance of the model is optimized by setting constraints and establishing penalty factors, as shown in formulas (S4-1 to S4-3): S4-1 S4-2 S4-3 In the formula, represents the optimal weight vector With vector The dot product of represents the penalty factor of the representative model; and Represent the slack variables of positive examples and negative examples, respectively, which are used to measure the distance from the sample point to the classification hyperplane; Represents the distance from the i-th sample point to the decision boundary; represents the optimal weight vector With the eigenvector The dot product of represents the slack variable in the soft interval; Represents the feature vector of the kth sample point; Represents the value of minimizing the objective function; It should be noted that The value of is positively correlated with the complexity of the model. Worth increasing for its own sake, The value of is negatively correlated with the model calculation error. The model error increases with The value increases and decreases.

[0024] S402, after the constraint condition is established, the edge parameters are removed by normalization to reduce the influence of the edge threshold on the prediction result, as shown in formula (S4-4): S4-4 In the formula, represents the normalized data, Indicates the current value; Indicates the maximum value in the overall data; Represents the minimum value in the overall data.

[0025] S5 characterizes the input parameters of the prediction model; S501, the model parameter processing performs corresponding weight calculation on each feature item according to its feature degree, and the size of the weight is used as the influence degree of the prediction model, as shown in formula (S5-1): S5-1 In the formula, represents the i-th sample, Represents the weight of the nth feature item in the i-th sample; It should be noted that It can also be expressed as , The weight can be calculated using the TF-IDF algorithm.

[0026] S501, the characteristic screening divides the probability of occurrence of various parameters by information gain method, as shown in formula (S5-2): S5-2 In the formula, Represents categorical data Probability of occurrence; Representing feature data Probability of occurrence; Indicates that when the feature data Category data when it appears Probability of occurrence; Indicates that no feature data appears probability; Indicates that no feature data appears But it belongs to category data probability.

[0027] S6 evaluates the spinal deformity correction results predicted by the model; S601, the evaluation method is evaluated by using a K-fold cross-validation method, wherein the K-fold cross-validation method divides the original sample into K groups to obtain K models, and uses the average accuracy of the final test set of the K models as the evaluation performance indicator; S602, the average accuracy is characterized by the minimum mean square error, and the parameter value with the minimum mean square error is found in the parameter space, as shown in formula (S6-1): S6-1 In the formula, represents mean square error; represents the number of samples; Represents the actual sample value; Indicates an estimated value.

[0028] S7 stores the results of spinal deformity correction prediction in the cloud; S701, the prediction result is judged by the physician according to the Cobb angle. If the Cobb angle error is kept within 3 degrees, the prediction result is accurate, and the prediction data is stored in the cloud. If the Cobb angle error is greater than 3 degrees, the prediction result is not recorded, and the cause of the prediction error is analyzed; S702, after the prediction results are stored in the cloud, the sample data is trained by machine learning to optimize the support vector machine regression model.

[0029] In addition, it should be noted that the present invention can be provided as a method, an apparatus or a computer program product. Therefore, the embodiments of the present invention can take the form of a complete hardware embodiment, a complete software embodiment or an embodiment combining software and hardware. Moreover, the embodiments of the present invention can take the form of a computer program product implemented on one or more computer-usable storage media containing computer-usable program code.

[0030] The embodiments of the present invention are described with reference to the flowcharts and / or block diagrams of the methods, terminal devices (systems), and computer program products according to the embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of the processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, an embedded processor, or other programmable data processing terminal device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing terminal device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0031] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing terminal device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce a manufactured product including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 These computer program instructions can also be loaded onto a computer or other programmable data processing terminal device, so that a series of operation steps are executed on the computer or other programmable terminal device to produce a computer-implemented process, so that the instructions executed on the computer or other programmable terminal device provide for implementing the process in the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0032] It should also be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. The terms "include", "comprises" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or terminal device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or terminal device. In the absence of further restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, article or terminal device including the elements.

[0033] Finally, it should be noted that the above is a preferred embodiment of the present invention. It should be pointed out that although the preferred embodiment of the present invention has been described, for those skilled in the art, once the basic creative concept of the present invention is known, several improvements and modifications can be made without departing from the principles of the present invention, and these improvements and modifications should also be regarded as the protection scope of the present invention. Therefore, the attached claims are intended to be interpreted as including the preferred embodiment and all changes and modifications that fall within the scope of the embodiments of the present invention.

Claims

1. A spinal deformity correction prediction method based on artificial intelligence, characterized in that: include: S1 Establish the influencing factors for predicting the results of spinal deformity correction; S2 digitally analyzes DR images of patients with spinal deformity; S3 constructs a prediction model for spinal deformity correction outcomes; S4 optimizes the performance of the spinal deformity correction outcome prediction model; S5 characterizes the input parameters of the prediction model; S6 evaluates the spinal deformity correction results predicted by the model; S7 stores the results of spinal deformity correction prediction in the cloud.

2. The method for predicting spinal deformity correction based on artificial intelligence according to claim 1, characterized in that: S1 establishes the influencing factors for predicting spinal deformity correction results, among which: S101, the influencing factors include the influencing factors before orthopedics and the influencing factors after orthopedics; S102, the pre-correction influencing factors include the patient's spinal development status and the type of spinal deformity. The patient's spinal development status can be divided into three intervals according to age: under 10 years old, 10 to 16 years old, and over 16 years old. The types of spinal deformity include congenital deformity, degenerative deformity, neuromuscular deformity, and genetic syndrome-related deformity; S103, the post-orthopedic influencing factors include the type of orthopedic surgery and rehabilitation conditions. The type of orthopedic surgery includes drug treatment, rehabilitation treatment, and surgical treatment. The rehabilitation conditions include the patient's bone recovery speed, whether the patient has underlying diseases, and the patient's treatment psychology.

3. The method for predicting spinal deformity correction based on artificial intelligence according to claim 1, characterized in that: The S2 digitally analyzes the DR images of patients with spinal deformity, wherein: S201, the digital analysis process includes dividing the measurement of the sagittal vertical axis and the Cobb angle; S202, the sagittal vertical axis is divided by measuring the position of the C7 plumb line and the posterior edge of the S1 sacrum. If the C7 plumb line is in front of the posterior superior edge of the S1 sacrum, it is defined as a positive value, and if it is in the rear, it is defined as a negative value. When the sagittal vertical axis is a positive value, it is defined as a forward tilt of the trunk, and when the sagittal vertical axis is a negative value, it is defined as a backward tilt. When the absolute value of the sagittal vertical axis exceeds 50 mm, it is defined as sagittal plane decompensation; S203, the steps of measuring the Cobb angle are to determine the scoliotic end vertebra, draw a horizontal line at the upper edge of the vertebral body of the upper end vertebra and a horizontal line at the lower edge of the lower end vertebra, and draw a vertical line on the basis of the horizontal lines of the upper and lower end vertebrae respectively. The intersection angle of the two vertical lines is the Cobb angle. If the Cobb angle is greater than 10 degrees, it is defined as spinal deformity.

4. The method for predicting spinal deformity correction based on artificial intelligence according to claim 1, characterized in that: The S3 constructs a prediction model for spinal deformity correction results, wherein: S301, the result of spinal deformity correction is predicted by constructing a support vector machine regression model, and the regression model infers the training data by constructing a regression function, as shown in formula (S3-1): S3-1 In the formula, Expressed as Function of Represents the input value; represents a matrix vector; express and Multiply the corresponding elements of and then sum them; Represents the bias term, which is used to adjust the output of the function; S302, after the training data inference is completed, a regression equation of the support vector machine regression model is constructed, as shown in Formula (S3-2): S3-2 In the formula, represents the number of support vectors; and denote the Lagrange multiplier and its corresponding dual variable respectively; Represents the kernel function, which is used to calculate the input and support vector The similarity between Represents the bias term.

5. The method for predicting spinal deformity correction based on artificial intelligence according to claim 1, characterized in that: The S4 optimizes the performance of the spinal deformity correction result prediction model, wherein: S401, the performance of the model is optimized by setting constraints and establishing penalty factors, as shown in formulas (S4-1 to S4-3): S4-1 S4-2 S4-3 In the formula, represents the optimal weight vector With vector The dot product of represents the penalty factor of the representative model; and Represent the slack variables of positive examples and negative examples, respectively, which are used to measure the distance from the sample point to the classification hyperplane; Represents the distance from the i-th sample point to the decision boundary; represents the optimal weight vector With the eigenvector The dot product of represents the slack variable in the soft interval; Represents the feature vector of the kth sample point; Represents the value of minimizing the objective function; S402, after the constraint condition is established, the edge parameters are removed by normalization to reduce the influence of the edge threshold on the prediction result, as shown in formula (S4-4): S4-4 In the formula, represents the normalized data, Indicates the current value; Indicates the maximum value in the overall data; Represents the minimum value in the overall data.

6. The method for predicting spinal deformity correction based on artificial intelligence according to claim 1, characterized in that: S5 characterizes the input parameters of the prediction model, wherein: S501, the model parameter processing performs corresponding weight calculation on each feature item according to its feature degree, and the size of the weight is used as the influence degree of the prediction model, as shown in formula (S5-1): S5-1 In the formula, represents the i-th sample, Represents the weight of the nth feature item in the i-th sample; S501, the characteristic screening divides the probability of occurrence of various parameters by information gain method, as shown in formula (S5-2): S5-2 In the formula, Represents categorical data Probability of occurrence; Representing feature data Probability of occurrence; Indicates that when the feature data Category data when it appears Probability of occurrence; Indicates that no feature data appears probability; Indicates that no feature data appears But it belongs to category data probability.

7. The method for predicting spinal deformity correction based on artificial intelligence according to claim 1, characterized in that: S6 evaluates the spinal deformity correction results predicted by the model, wherein: S601, the evaluation method is evaluated by using a K-fold cross-validation method, wherein the K-fold cross-validation method divides the original sample into K groups to obtain K models, and uses the average accuracy of the final test set of the K models as the evaluation performance indicator; S602, the average accuracy is characterized by the minimum mean square error, and the parameter value with the minimum mean square error is found in the parameter space, as shown in formula (S6-1): S6-1 In the formula, represents mean square error; represents the number of samples; Represents the actual sample value; Indicates an estimated value.

8. The method for predicting spinal deformity correction based on artificial intelligence according to claim 1, characterized in that: The S7 stores the result of spinal deformity correction prediction in the cloud, wherein: S701, the prediction result is judged by the physician according to the Cobb angle. If the Cobb angle error is kept within 3 degrees, the prediction result is accurate, and the prediction data is stored in the cloud. If the Cobb angle error is greater than 3 degrees, the prediction result is not recorded, and the cause of the prediction error is analyzed; S702, after the prediction results are stored in the cloud, the sample data is trained by machine learning to optimize the support vector machine regression model.

Images