A method and system for predicting the efficacy of induction chemotherapy for nasopharyngeal carcinoma

The image characteristics of nasopharyngeal carcinoma were obtained through multimodal magnetic resonance imaging, and the multi-step feature selection and support vector machine classifier were used to solve the accuracy of the prediction of the efficacy of chemotherapy induced by nasopharyngeal carcinoma, and personalized medical decisions were realized. Reduces the risk of toxic side effects.

CN113793324BActive Publication Date: 2024-11-29SHANDONG NORMAL UNIV
View PDF 1 Cites 0 Cited by

Patent Information

Application Number
CN202111094412.2
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2021-09-17
Publication Date
2024-11-29
Estimated Expiration
2041-09-17

AI Technical Summary

Technical Problem

In the prior art, the accuracy of the prediction of the efficacy of nasopharyngeal carcinoma inducible chemotherapy is difficult to guarantee, and it is greatly affected by subjective factors and lacks non-invasive and non-invasive tumor heterogeneity analysis methods.

Method used

Multimodal magnetic resonance imaging is used to obtain image features, and the efficacy of chemotherapy induction by region of interest and multi-step and step-by-step feature selection is predicted through the support vector machine classifier.

Benefits of technology

Accurate prediction of the efficacy of chemotherapy induced by nasopharyngeal carcinoma has been achieved, personalized and precise medical decisions have been ensured, and the risk of toxic and side effects has been reduced.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN113793324B_ABST
    Figure CN113793324B_ABST
Patent Text Reader

Abstract

The present invention belongs to the field of medical image analysis and prediction, and provides a method and system for predicting the efficacy of induction chemotherapy for nasopharyngeal carcinoma. The method includes obtaining a multimodal magnetic resonance image before induction chemotherapy; outlining the region of interest on the multimodal magnetic resonance image, extracting the image features of the region of interest, and obtaining the imaging omics features of the region of interest; using a multi-step step-by-step feature selection method to obtain the imaging omics features that meet the set conditions as the best imaging omics features; wherein the multi-step step-by-step feature selection method is at least three-step step-by-step feature selection; using the trained support vector machine classifier to predict the best imaging omics features, and obtain the predicted efficacy results of induction chemotherapy for nasopharyngeal carcinoma.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the field of medical image analysis and prediction, and in particular relates to a method and system for predicting the efficacy of induced chemotherapy for nasopharyngeal carcinoma. Background Art

[0002] The statements in this section merely provide background information related to the present invention and do not necessarily constitute prior art.

[0003] Nasopharyngeal carcinoma (NPC) is a malignant tumor that occurs in the nasopharyngeal mucosa. Tumor cells are sensitive to radiotherapy and chemotherapy. In the late stage, a comprehensive treatment strategy of radiotherapy combined with chemotherapy is required. Induction chemotherapy can avoid severe radiation damage by reducing the tumor volume and reducing the irradiation dose of normal tissues around the target area. However, patients receiving induction chemotherapy have to endure higher toxic and side effects. Statistics show that patients have heterogeneous responses to induction chemotherapy. Short-term efficacy evaluation, that is, tumor regression, is directly related to the efficacy of induction chemotherapy. Accurately predicting the effect after chemotherapy is crucial for NPC patients to choose the medical plan of induction chemotherapy. Human evaluation is greatly affected by subjective factors and is unstable. Therefore, there is an urgent need for a non-invasive analysis of tumor heterogeneity to quickly and non-invasively evaluate tumor information, and imaging genomics came into being.

[0004] Radiomics is an emerging field that extracts high-throughput features from medical images to quantify major diseases such as tumors. It provides a non-invasive quantitative method for evaluating tumor heterogeneity, and has great advantages in tumor phenotyping, treatment selection, and prognosis analysis. It is a research hotspot in clinical medicine and biomedical engineering.

[0005] The inventors found that in existing research, the analysis process of imaging genomics has not formed a standardized processing mode, and the results obtained by selecting different feature extraction methods, feature selection methods and different machine learning classifiers are inconsistent, which leads to the problem that the accuracy of the prediction results is difficult to guarantee. Summary of the invention

[0006] In order to solve the technical problems existing in the above-mentioned background technology, the present invention provides a system for predicting the efficacy of induction chemotherapy for nasopharyngeal carcinoma, which is based on machine learning and aims to use magnetic resonance imaging to mine image features from routine clinical imaging examination data as new treatment markers to assist doctors in making clinical decisions, thereby ensuring the accuracy of the prediction of the efficacy of induction chemotherapy and realizing personalized precision medicine.

[0007] In order to achieve the above object, the present invention adopts the following technical solution:

[0008] The first aspect of the present invention provides a method for predicting the efficacy of induction chemotherapy for nasopharyngeal carcinoma, comprising:

[0009] Multimodal MRI images were obtained before induction chemotherapy;

[0010] Delineate the region of interest on the multimodal magnetic resonance images, extract the image features of the region of interest, and obtain the imaging omics features of the region of interest;

[0011] Using a multi-step step-by-step feature selection method, an imaging omics feature that meets the set conditions is obtained as the best imaging omics feature; wherein the multi-step step-by-step feature selection method is at least a three-step step-by-step feature selection method;

[0012] The trained support vector machine classifier was used to predict the optimal radiomics features and obtain the predicted efficacy results of induction chemotherapy for nasopharyngeal carcinoma.

[0013] The second aspect of the present invention provides a system for predicting the efficacy of induction chemotherapy for nasopharyngeal carcinoma, comprising:

[0014] A multimodal magnetic resonance image acquisition module, which is used to acquire multimodal magnetic resonance images before induction chemotherapy;

[0015] An imaging genomics feature acquisition module is used to delineate the region of interest on the multimodal magnetic resonance image, extract the image features of the region of interest, and obtain the imaging genomics features of the region of interest;

[0016] An optimal radiomics feature screening module, which is used to obtain radiomics features that meet set conditions as the optimal radiomics features by using a multi-step step-by-step feature selection method; wherein the multi-step step-by-step feature selection method is at least a three-step step-by-step feature selection method;

[0017] The chemotherapy efficacy prediction module is used to predict the optimal imaging genomics features using the trained support vector machine classifier to obtain the predicted efficacy results of nasopharyngeal carcinoma induction chemotherapy.

[0018] A third aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps in the method for predicting the efficacy of induction chemotherapy for nasopharyngeal carcinoma as described above.

[0019] A fourth aspect of the present invention provides a computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the steps in the method for predicting the efficacy of induction chemotherapy for nasopharyngeal carcinoma as described above are implemented.

[0020] Compared with the prior art, the present invention has the following beneficial effects:

[0021] The present invention delineates a region of interest on a multimodal magnetic resonance image, extracts image features of the region of interest, obtains radiomic features of the region of interest, and then uses a multi-step step-by-step feature selection method to obtain radiomic features that meet set conditions as optimal radiomic features; wherein the multi-step step-by-step feature selection method is at least a three-step step-by-step feature selection method, and finally, based on the optimal radiomic features and a support vector machine classifier, the efficacy results of induction chemotherapy for nasopharyngeal carcinoma are predicted, which can assist doctors in making clinical decisions, ensure the accuracy of the prediction of the efficacy of induction chemotherapy, and realize personalized precision medicine.

[0022] Advantages of additional aspects of the present invention will be given in part in the following description, and in part will become obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] The accompanying drawings in the specification, which constitute a part of the present invention, are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0024] Figure 1 A flow chart of a method for predicting the efficacy of induction chemotherapy for nasopharyngeal carcinoma provided by an embodiment of the present invention;

[0025] Figure 2 A schematic diagram of the efficacy prediction of induction chemotherapy for nasopharyngeal carcinoma provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0026] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0027] It should be noted that the following detailed descriptions are all illustrative and intended to provide further explanation of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present invention belongs.

[0028] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit exemplary embodiments according to the present invention. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprising" and / or "including" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof.

[0029] Embodiment 1

[0030] like Figure 1 and Figure 2 As shown, this embodiment provides a method for predicting the efficacy of induction chemotherapy for nasopharyngeal carcinoma, which includes:

[0031] S101: Acquire multimodal magnetic resonance images before induction chemotherapy.

[0032] S102: Delineating a region of interest on the multimodal magnetic resonance image, extracting image features of the region of interest, and obtaining imaging omics features of the region of interest.

[0033] Generally, radiologists manually outline the lesion area and lymph node area in the medical clinical data sample, and use the outlined lesion area and lymph node area as the region of interest. Feature extraction is performed on the region of interest to obtain the imaging features, and all imaging features are standardized.

[0034] In this embodiment, PyRadiomics can be used to extract features of the region of interest. Feature extraction is a prerequisite for radiomics feature selection, so the comprehensiveness of the extracted features must be ensured.

[0035] Among them, pyradiomics is an open source python package that can be used as a feature extraction tool for radiomics feature extraction of medical images. Feature extraction can be performed by specifying image categories, feature categories, and custom feature extraction parameters of specific filters.

[0036] After obtaining the imaging genomics features of the region of interest, the imaging genomics features are also standardized.

[0037] Specifically, the image features of the region of interest include shape features, first-order statistics features, texture features, and image features under different filters. This embodiment extracts 1734 features from each region of interest in the nasopharyngeal carcinoma magnetic resonance image, including 14 shape features, 18 first-order statistics features, 4 types of texture features, and also extracts image features under different filters.

[0038] The four types of texture features include: gray-level co-occurrence matrix, gray-level run-length matrix, gray-level dependence matrix, and gray-level region size matrix.

[0039] Gray-level co-occurrence matrix: The gray-level relationship between adjacent voxels in the tumor area is calculated to characterize the texture of the image.

[0040] Grayscale run-length matrix: Analyzes the relationship between pixels of the same intensity in space to describe the strip texture of the image.

[0041] Grayscale dependency matrix: Quantifies the grayscale dependency of the image. Grayscale dependency is defined as the number of connected voxels within a distance that depends on the central voxel.

[0042] Grayscale Region Size Matrix: Grayscale region size quantifies the grayscale region in an image, where grayscale region is defined as the number of connected pixels sharing the same grayscale intensity.

[0043] S103: using a multi-step step-by-step feature selection method to obtain an imaging omics feature that meets the set conditions as the best imaging omics feature; wherein the multi-step step-by-step feature selection method is at least a three-step step-by-step feature selection method.

[0044] In a specific implementation, the three-step step-by-step feature selection method is:

[0045] The T test was used for preliminary screening of radiomics features, the lasso algorithm was used for secondary screening of the remaining features after the preliminary screening, and the variance inflation factor was used for final screening of the remaining features after the secondary screening.

[0046] In this embodiment, the standardized radiomics features are screened to eliminate redundant features, and the T test, lasso algorithm and variance inflation factor are used to select important features, remove irrelevant radiomics features, and retain radiomics features that are relevant to clinical labels.

[0047] The process of preliminary screening of radiomics features using T-test is to check whether the features are significantly different in two independent groups of samples:

[0048] Step 1: Check the homogeneity of variance of the characteristics of the two independent sample groups;

[0049] Step 2: If the variance homogeneity is not satisfied, the Welch T test is performed and the t statistic is calculated;

[0050] Step 3: Use Student t to satisfy the homogeneity of variance and calculate the t statistic;

[0051] Step 4: Calculate the critical probability P value of the t statistic.

[0052] For example: screening features with P<0.05, that is, features with significant differences between two independent samples.

[0053] This embodiment first checks the homogeneity of variance of two sets of independent sample data, and the test formula is:

[0054]

[0055] in and is the two sample variance.

[0056] If the two groups of sample data do not meet the variance homogeneity, the Welch T test is used, and the formula is:

[0057]

[0058] Where μ1 and μ2 are the expected values ​​of the two samples; and is the variance of the two samples; n1 and n2 are the capacity of the two samples; t represents the t statistic, represents the mean of sample 1, represents the mean of sample 2.

[0059] If the two groups of sample data meet the variance homogeneity, the Student t test is used, and the formula is:

[0060]

[0061] in and is the variance of the two samples; n1 and n2 are the capacity of the two samples.

[0062] The Lasso algorithm is used to select the remaining features of the T test, and its loss function is:

[0063]

[0064] where y i is the prediction result; x i Corresponding to y i is the value of each feature; W is the coefficient of each feature. m is the feature capacity.

[0065] The 5-fold cross-validation method was used to select the optimal adjustment parameter λ. When the loss function was minimized, the weights of some of the filtered omics features were reduced to 0, and the features reduced to 0 were removed and the remaining features were retained, which was the result of Lasso screening in this embodiment.

[0066] The variance expansion coefficient is used to filter the features after Lasso filtering. The formula is:

[0067]

[0068] Where R i is the independent variable x i The negative correlation coefficient of the regression analysis of the remaining independent variables. The larger the VIF, the greater the possibility of collinearity between the independent variables. Generally speaking, if the variance inflation factor exceeds 10, it means that there is serious multicollinearity between the independent variables. In this embodiment, the features with variance inflation factors exceeding 10 are eliminated, and the remaining features are the final selected features.

[0069] It should be noted here that those skilled in the art can adopt a four-step step-by-step feature selection method or a five-step step-by-step feature selection method, etc., according to actual prediction accuracy requirements, to obtain the imaging omics features that meet the set conditions as the optimal imaging omics features.

[0070] Among them, after the three-step step-by-step feature selection, feature screening can be performed using any one or two methods such as lasso algorithm and recursive feature elimination.

[0071] Specifically, the recursive feature elimination operation is as follows:

[0072] Step 1: Train samples by specifying a model, such as support vector machine model, logistic regression model, etc.

[0073] Step 2: Assign a weight to each feature based on the feature importance given by the model, and remove the feature with the smallest absolute weight.

[0074] Step 3: Use the remaining features to train the model again, perform the next iteration, repeat the recursive loop, and finally select the required number of features.

[0075] S104: Use the trained support vector machine classifier to predict the optimal radiomics features and obtain the predicted efficacy results of induction chemotherapy for nasopharyngeal carcinoma.

[0076] In this embodiment, the results of the efficacy of induction chemotherapy for nasopharyngeal carcinoma are divided into two categories: an induction chemotherapy-sensitive group and an induction chemotherapy-resistant group; among them, patients with complete remission (CR) and partial remission (PR) of the disease are classified as the induction chemotherapy-sensitive group, and patients with stable disease (SD) and progressive disease (PD) are classified as the induction chemotherapy-resistant group.

[0077] It should be noted that, in other embodiments, the results of the efficacy of nasopharyngeal carcinoma induction chemotherapy can be divided into other categories according to actual conditions.

[0078] Among them, the support vector machine classifier is trained based on a multimodal magnetic resonance image dataset with known therapeutic efficacy labels.

[0079] The construction process of the multimodal magnetic resonance image dataset with known efficacy labels is as follows:

[0080] Clinical data of NPC patients receiving induction chemotherapy and magnetic resonance images of multimodal NPC before induction chemotherapy were collected to obtain medical clinical data samples that met the screening criteria.

[0081] In the specific implementation process, the clinical objectives and the inclusion and exclusion criteria of patients are first determined, and the imaging data and clinical data of patients are retrospectively collected according to the inclusion and exclusion criteria. After obtaining the medical clinical data, the collected medical clinical data are screened by data type to determine the medical clinical data samples, and the true labels of the medical clinical data samples are determined according to the clinical objective-induced chemotherapy efficacy.

[0082] The inclusion criteria of the embodiment are: patients with nasopharyngeal carcinoma who have received 2-3 cycles of induction chemotherapy and have no other malignant tumors. The efficacy of nasopharyngeal carcinoma patients after induction chemotherapy was evaluated according to the efficacy evaluation criteria for solid tumors, and patients with complete disease remission (CR) and partial disease remission (PR) were classified as the induction chemotherapy sensitive group, and patients with stable disease (SD) and progressive disease (PD) were classified as the induction chemotherapy resistant group.

[0083] In this embodiment, the features most relevant to the labels are finally screened out as input, the optimal parameters of the support vector machine model are selected using a grid search algorithm, and finally the best support vector machine model is obtained by training the support vector machine as a prediction model for the efficacy of induction chemotherapy for nasopharyngeal carcinoma. The leave-one-out cross-validation method is used to verify and evaluate the prediction model for the efficacy of induction chemotherapy for nasopharyngeal carcinoma.

[0084] Among them, the label is the classification judgment of the efficacy of induction chemotherapy: induction chemotherapy sensitive group and induction chemotherapy resistant group. Leave-one-out cross-validation is to leave one patient sample for validation and the remaining patient samples for training. If there are K patients, it will be cycled K times and the average will be taken to evaluate the model performance.

[0085] This example uses the accuracy (ACC), sensitivity (SENS), specificity (SPEC) and the area under the receiver operating characteristic (ROC) curve (AUC) to evaluate the prediction model. SENS represents the ratio of correctly predicting patients as sensitive group, and SPEC represents the ratio of correctly predicting patients as resistant group. The specific formula is as follows:

[0086]

[0087]

[0088]

[0089]

[0090] Among them, TP represents the number of NPC patients correctly judged as being sensitive to induction chemotherapy, TN represents the number of NPC patients correctly judged as being resistant to induction chemotherapy, FP represents the number of NPC patients incorrectly judged as being sensitive to induction chemotherapy, and FN represents the number of NPC patients incorrectly judged as being resistant to induction chemotherapy.

[0091] Embodiment 2

[0092] This embodiment provides a system for predicting the efficacy of induction chemotherapy for nasopharyngeal carcinoma, which specifically includes the following modules:

[0093] A multimodal magnetic resonance image acquisition module, which is used to acquire multimodal magnetic resonance images before induction chemotherapy;

[0094] An imaging genomics feature acquisition module is used to delineate the region of interest on the multimodal magnetic resonance image, extract the image features of the region of interest, and obtain the imaging genomics features of the region of interest;

[0095] An optimal radiomics feature screening module, which is used to obtain radiomics features that meet set conditions as the optimal radiomics features by using a multi-step step-by-step feature selection method; wherein the multi-step step-by-step feature selection method is at least a three-step step-by-step feature selection method;

[0096] The chemotherapy efficacy prediction module is used to predict the optimal imaging genomics features using the trained support vector machine classifier to obtain the predicted efficacy results of nasopharyngeal carcinoma induction chemotherapy.

[0097] It should be noted here that the various modules in this embodiment correspond one-to-one to the various steps in Example 1, and the specific implementation process is the same, which will not be repeated here.

[0098] Embodiment 3

[0099] This embodiment provides a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the steps in the method for predicting the efficacy of induction chemotherapy for nasopharyngeal carcinoma as described above are implemented.

[0100] Embodiment 4

[0101] This embodiment provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, the steps in the method for predicting the efficacy of induction chemotherapy for nasopharyngeal carcinoma as described above are implemented.

[0102] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of hardware embodiments, software embodiments, or embodiments combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage and optical storage, etc.) containing computer-usable program code.

[0103] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to 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 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, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing 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.

[0104] A person skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program, and the program can be stored in a computer-readable storage medium, and when the program is executed, it can include the processes of the embodiments of the above-mentioned methods. The storage medium can be a disk, an optical disk, a read-only memory (ROM) or a random access memory (RAM), etc.

[0105] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for predicting the efficacy of induction chemotherapy for nasopharyngeal carcinoma, characterized in that: include: Multimodal MRI images were obtained before induction chemotherapy; Delineate the region of interest on the multimodal magnetic resonance images, extract the image features of the region of interest, and obtain the imaging omics features of the region of interest; Using a multi-step step-by-step feature selection method, an imaging omics feature that meets the set conditions is obtained as the best imaging omics feature; wherein the multi-step step-by-step feature selection method is at least a three-step step-by-step feature selection method; The three-step step-by-step feature selection method is: The T test was used to preliminarily screen the radiomics features, the lasso algorithm was used to perform secondary screening on the remaining features of the preliminary screening, and the variance inflation factor was used to perform final screening on the remaining features of the secondary screening; irrelevant radiomics features were removed, and radiomics features that were relevant to clinical labels were retained; After the three-step step-by-step feature selection, feature screening is performed using either or both of the lasso algorithm and recursive feature elimination; The trained support vector machine classifier was used to predict the best radiomics features and obtain the predicted efficacy of induction chemotherapy for NPC. The most relevant features finally selected were used as input, and the grid search algorithm was used to select the optimal parameters of the support vector machine model. Finally, the best support vector machine model was obtained by support vector machine training as the prediction model for the efficacy of induction chemotherapy for nasopharyngeal carcinoma. The leave-one-out cross-validation method was used to verify and evaluate the prediction model for the efficacy of induction chemotherapy for nasopharyngeal carcinoma. Among them, the label is the classification judgment of the efficacy of induction chemotherapy: induction chemotherapy sensitive group and induction chemotherapy resistant group. The leave-one-out cross-validation is to leave one patient sample for verification and the remaining patient samples for training. If there are K patients, the cycle is repeated K times and the average is taken to evaluate the model performance.

2. The method for predicting the efficacy of induction chemotherapy for nasopharyngeal carcinoma according to claim 1, characterized in that: After obtaining the radiomics features of the region of interest, the radiomics features are also standardized.

3. The method for predicting the efficacy of induction chemotherapy for nasopharyngeal carcinoma according to claim 1, characterized in that: A support vector machine classifier was trained on a multimodal magnetic resonance image dataset with known therapeutic efficacy labels.

4. The method for predicting the efficacy of induction chemotherapy for nasopharyngeal carcinoma according to claim 1, characterized in that: The image features of the region of interest include shape features, first-order statistics features, texture features, and image features under different filters.

5. The method for predicting the efficacy of induction chemotherapy for nasopharyngeal carcinoma according to claim 4, characterized in that: The texture features include: a grayscale co-occurrence matrix, a grayscale run-length matrix, a grayscale dependency matrix, and a grayscale region size matrix.

6. A system for predicting the efficacy of induction chemotherapy for nasopharyngeal carcinoma based on the method according to any one of claims 1 to 5, characterized in that: include: A multimodal magnetic resonance image acquisition module, which is used to acquire multimodal magnetic resonance images before induction chemotherapy; An imaging genomics feature acquisition module is used to delineate the region of interest on the multimodal magnetic resonance image, extract the image features of the region of interest, and obtain the imaging genomics features of the region of interest; An optimal radiomics feature screening module, which is used to obtain radiomics features that meet set conditions as the optimal radiomics features by using a multi-step step-by-step feature selection method; wherein the multi-step step-by-step feature selection method is at least a three-step step-by-step feature selection method; The chemotherapy efficacy prediction module is used to predict the optimal imaging genomics features using the trained support vector machine classifier to obtain the predicted efficacy results of nasopharyngeal carcinoma induction chemotherapy.

7. The nasopharyngeal carcinoma induction chemotherapy efficacy prediction system according to claim 6, characterized in that: In the optimal radiomics feature screening module, the three-step step-by-step feature selection method is: The T test was used for preliminary screening of radiomics features, the lasso algorithm was used for secondary screening of the remaining features after the preliminary screening, and the variance inflation factor was used for final screening of the remaining features after the secondary screening.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps in the method for predicting the efficacy of induction chemotherapy for nasopharyngeal carcinoma according to any one of claims 1 to 5 are implemented.

9. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the steps in the method for predicting the efficacy of induction chemotherapy for nasopharyngeal carcinoma according to any one of claims 1 to 5 are implemented.

Citation Information

Patent Citations

  • Lung cancer prognosis auxiliary evaluation method and system based on CT radiomics

    CN112951406A