EGFR mutation state prediction method and device

By combining radiomics and clinical features to create an EGFR mutation prediction model, the high cost and low sensitivity of existing detection methods have been addressed, achieving high accuracy and low cost in predicting EGFR mutation status, thus optimizing the allocation of medical resources and the patient's treatment experience.

CN119049547APending Publication Date: 2024-11-29GENERAL HOSPITAL OF NUCLEAR IND
View PDF 7 Cites 0 Cited by

Patent Information

Application Number
CN202410925116.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-11
Publication Date
2024-11-29

AI Technical Summary

Technical Problem

Existing methods for detecting EGFR mutation status suffer from high costs, long processing times, difficult sampling, and poor sensitivity and specificity, especially ctDNA-based liquid biopsy methods.

Method used

Combining radiomics and clinical features, this study acquires patient imaging and clinical data and uses a pre-built EGFR mutation prediction model for prediction. The model uses the LASSO algorithm to screen features and employs multivariate logistic regression to improve prediction accuracy.

Benefits of technology

It significantly improves the accuracy of predicting EGFR mutation status, reduces testing costs and time, optimizes the allocation of medical resources, and enhances patient experience and compliance.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119049547A_ABST
    Figure CN119049547A_ABST
Patent Text Reader

Abstract

The invention discloses an EGFR mutation state prediction method and device. The method comprises the following steps: acquiring image data and clinical data of a to-be-detected lung cancer patient; extracting the image data to obtain image omics features; extracting the clinical data to obtain clinical features; inputting the radiomics features and the clinical features into a pre-constructed EGFR mutation prediction model; wherein a corresponding relation among radiomics characteristics, clinical characteristics and EGFR mutation states is learned in the EGFR mutation prediction model; and determining an EGFR mutation state prediction result according to the output of the EGFR mutation prediction model. By combining radiomics characteristics and clinical characteristics, the EGFR mutation state prediction accuracy is improved by using the method provided by the invention.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application belongs to the field of medical imaging and bioinformatics, and relates to a method and device for predicting EGFR mutation status, and specifically to a method and device for predicting EGFR mutation status in patients with non-small cell lung cancer based on imaging genomics features and clinical features. Background Art

[0002] The main mutated gene in non-small cell lung cancer (NSCLC) is epidermal growth factor receptor (EGFR). In recent years, its treatment has emphasized the importance of identifying its mutation status, which requires timely evaluation before treatment.

[0003] The current methods for detecting EGFR mutations mainly involve obtaining tissue specimens through invasive procedures, including surgery or puncture biopsy. However, these methods are costly and time-consuming, and may also have problems such as sampling difficulties, unstable tissue availability, and lack of description of internal tumor heterogeneity. Recently, non-invasive detection methods have gradually emerged. For example, liquid biopsy, especially circulating tumor DNA (ctDNA), can provide rapid and comprehensive EGFR detection. However, as a representative method of liquid biopsy, ctDNA has poor sensitivity and specificity. Summary of the invention

[0004] Relevant features identified from clinical images such as CT can be used as biomarkers to reveal diagnostic and prognostic correlations. Although there is evidence supporting the effectiveness of imaging omics in predicting EGFR mutation status in NSCLC patients, many related studies still face limitations.

[0005] Objective: In view of at least one of the above technical problems, the present application provides a method and device for predicting EGFR mutation status, which combines imaging genomics features and clinical characteristics to improve the accuracy of EGFR mutation status prediction.

[0006] The technical solution adopted in this application is: In a first aspect, the present application provides a method for predicting EGFR mutation status, comprising: Obtain imaging data and clinical data of lung cancer patients to be tested; Extracting radiomics features from the image data; Extracting clinical characteristics from the clinical data; Inputting the radiomics features and clinical features into a pre-constructed EGFR mutation prediction model; wherein the EGFR mutation prediction model learns the correspondence between the radiomics features, clinical features and EGFR mutation status; According to the output of the EGFR mutation prediction model, the EGFR mutation status prediction result is determined.

[0007] In some embodiments, the method for obtaining the EGFR mutation prediction model comprises: Acquire a lung cancer patient dataset, wherein the lung cancer patient dataset includes multiple samples, each sample includes imaging data, gene sequencing data, and clinical data related to lung cancer of a lung cancer patient; Annotating and extracting the image data to obtain a first radiomics feature; Extracting a first clinical feature from the clinical data; Labeling samples according to the gene sequencing data, wherein the labels are divided into EGFR wild type and EGFR mutant type; The least absolute shrinkage and selection operator LASSO algorithm was used to screen the radiomic features and clinical features for EGFR mutation prediction from the first radiomic features and the first clinical features; Based on the screened radiomics features and clinical features, determining training samples from the lung cancer patient dataset; Using the radiomics features, clinical characteristics and labels corresponding to the training samples, a multifactorial logistic regression was used to construct an EGFR mutation prediction model.

[0008] In some embodiments, the first radiomics feature comprises at least one of a morphological feature, a first-order statistical feature, a second-order and a higher-order texture feature.

[0009] Further, in some embodiments, the morphological features are used to describe the size of the region of interest ROI and the degree of similarity between the region of interest ROI and a sphere, including volume, surface area, maximum diameter, area percentage, density, eccentricity, and sphericity; these features reflect the geometric morphology of the ROI; The first-order statistical features are used to describe the statistical characteristics of the voxel intensity distribution within the region of interest ROI, including mean, median, minimum, maximum, cluster prominence, difference standard deviation, difference variance, skewness and kurtosis; these features reflect the symmetry and uniformity of the voxel intensity distribution and the change of local intensity distribution; The second-order and high-order texture features are used to describe the characteristics of the intensity level of the spatial distribution of voxels, including the intensity histogram distribution pattern, low grayscale highlight area, neighborhood grayscale difference matrix density, neighborhood grayscale difference matrix complexity, highlight area, gray level co-occurrence matrix (GLCM), gray level long matrix (GLRLM), gray level band matrix (GLSZM) and neighborhood grayscale difference matrix (NGTDM); these features reflect the image texture, that is, the perceptible or measurable spatial changes in intensity levels.

[0010] In some embodiments, the imaging genomics features include: area percentage, intensity histogram distribution pattern, low grayscale highlight area, neighborhood grayscale difference matrix complexity, skewness, highlight area, difference variance, and cluster prominence.

[0011] In some embodiments, the clinical characteristics include: age, gender, clinical stage, and thyroid transcription factor expression level.

[0012] In some embodiments, the data processing process of the EGFR mutation prediction model includes: The imaging score is calculated according to the imaging omics features; Obtaining the probability of EGFR mutation according to the imaging score and the clinical characteristics; The EGFR mutation probability is compared with the EGFR mutation threshold to obtain an EGFR mutation status prediction result.

[0013] Further, in some embodiments, the imaging score is calculated according to the imaging omics features, including: Radiological score = area percentage*1+intensity histogram distribution mode*0.304+low grayscale highlight area*0.112-neighborhood grayscale difference matrix complexity*0.017-skewness*0.47-highlight area*0.485-difference variance*0.978-cluster prominence*1.

[0014] In a second aspect, the present application provides an EGFR mutation status prediction device, including a processor and a storage medium; The storage medium is used to store instructions; The processor is configured to operate according to the instructions to execute the method according to the first aspect.

[0015] In a third aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the method described in the first aspect.

[0016] In a fourth aspect, the present application provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the method described in the first aspect when executing the computer program.

[0017] In a fifth aspect, the present application provides a computer program product, comprising a computer program, which implements the steps of the method described in the first aspect when executed by a processor.

[0018] Beneficial effects: The EGFR mutation status prediction method and device provided in this application have the following advantages: The AUC of the model of the method of this application reaches 0.84, showing extremely high prediction accuracy; by integrating imaging genomics features and clinical pathological predictors, the model overcomes the limitations of a single data type and provides a more comprehensive and accurate EGFR mutation prediction method. In addition, by explaining the biological function of the model through genomic data, the reliability and application prospects of the model are further enhanced, laying a solid foundation for the development of precision medicine.

[0019] It significantly reduces the extra expenses caused by difficult tissue sampling and repeated testing, avoids the pain and risks of traditional biopsy methods for patients, and significantly improves the diagnosis and treatment experience and compliance of patients. The TRANS model predicts EGFR mutation status through quantitative and visual nomograms, greatly reducing the cost and time of testing, allowing more patients to receive accurate diagnosis and treatment in a timely manner. It optimizes the allocation of medical resources, reduces the overall burden on the medical system, and thus saves social medical expenses. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 Schematic diagram of a process for predicting EGFR mutation status according to an embodiment of the present application; Figure 2 Schematic diagram of a method for obtaining an EGFR mutation prediction model according to an embodiment of the present application; Figure 3 It is a schematic diagram of an EGFR mutation prediction model according to one embodiment of the present application. DETAILED DESCRIPTION

[0021] The present application is further described below in conjunction with the accompanying drawings and embodiments. The following embodiments are only used to more clearly illustrate the technical solution of the present application, and cannot be used to limit the protection scope of the present application.

[0022] In the description of this application, "several" means more than one, "more" means more than two, "greater than", "less than", "exceed", etc. are understood to exclude the number itself, and "above", "below", "within", etc. are understood to include the number itself. If there is a description of "first" or "second", it is only used for the purpose of distinguishing technical features, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features or implicitly indicating the order of the indicated technical features.

[0023] In the description of the present application, the description with reference to the terms "one embodiment", "some embodiments", "illustrative embodiments", "examples", "specific examples", or "some examples" means that the specific features, structures, materials, or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.

[0024] Example 1: This example provides a method for predicting EGFR mutation status. Figure 1 As shown, including: Obtain imaging data and clinical data of lung cancer patients to be tested; Extracting radiomics features from the image data; Extracting clinical characteristics from the clinical data; Inputting the radiomics features and clinical features into a pre-constructed EGFR mutation prediction model; wherein the EGFR mutation prediction model learns the correspondence between the radiomics features, clinical features and EGFR mutation status; According to the output of the EGFR mutation prediction model, the EGFR mutation status prediction result is determined.

[0025] In some embodiments, Figure 2 As shown, the method for obtaining the EGFR mutation prediction model includes: Acquire a lung cancer patient dataset, wherein the lung cancer patient dataset includes multiple samples, each sample includes imaging data, gene sequencing data, and clinical data related to lung cancer of a lung cancer patient; Annotating and extracting the image data to obtain a first radiomics feature; Extracting a first clinical feature from the clinical data; Labeling samples according to the gene sequencing data, wherein the labels are divided into EGFR wild type and EGFR mutant type; The least absolute shrinkage and selection operator LASSO algorithm was used to screen the radiomic features and clinical features for EGFR mutation prediction from the first radiomic features and the first clinical features; Based on the screened radiomics features and clinical features, determining training samples from the lung cancer patient dataset; Using the radiomics features, clinical features and labels corresponding to the training samples, a multi-factor logistic regression was used to construct an EGFR mutation prediction model. Figure 3 Shown is a schematic diagram of the EGFR mutation prediction model in this embodiment.

[0026] In some embodiments, the first radiomics feature comprises at least one of a morphological feature, a first-order statistical feature, a second-order and a higher-order texture feature.

[0027] Further, in some embodiments, the morphological features are used to describe the size of the region of interest ROI and the degree of similarity between the region of interest ROI and a sphere, including volume, surface area, maximum diameter, area percentage, density, eccentricity, and sphericity; these features reflect the geometric morphology of the ROI; The first-order statistical features are used to describe the statistical characteristics of the voxel intensity distribution within the region of interest ROI, including mean, median, minimum, maximum, cluster prominence, difference standard deviation, difference variance, skewness and kurtosis; these features reflect the symmetry and uniformity of the voxel intensity distribution and the change of local intensity distribution; The second-order and high-order texture features are used to describe the characteristics of the intensity level of the spatial distribution of voxels, including the intensity histogram distribution pattern, low grayscale highlight area, neighborhood grayscale difference matrix density, neighborhood grayscale difference matrix complexity, highlight area, gray level co-occurrence matrix (GLCM), gray level long matrix (GLRLM), gray level band matrix (GLSZM) and neighborhood grayscale difference matrix (NGTDM); these features reflect the image texture, that is, the perceptible or measurable spatial changes in intensity levels.

[0028] In some embodiments, the imaging genomics features include: area percentage, intensity histogram distribution pattern, low grayscale highlight area, neighborhood grayscale difference matrix complexity, skewness, highlight area, difference variance, and cluster prominence.

[0029] In some embodiments, the clinical characteristics include: age, gender, clinical stage, and thyroid transcription factor expression level.

[0030] In some embodiments, the data processing process of the EGFR mutation prediction model includes: The imaging score is calculated according to the imaging omics features; Obtaining the probability of EGFR mutation according to the imaging score and the clinical characteristics; The EGFR mutation probability is compared with the EGFR mutation threshold to obtain an EGFR mutation status prediction result.

[0031] Further, in some embodiments, the imaging score is calculated according to the imaging omics features, including: Radiological score = area percentage*1+intensity histogram distribution mode*0.304+low grayscale highlight area*0.112-neighborhood grayscale difference matrix complexity*0.017-skewness*0.47-highlight area*0.485-difference variance*0.978-cluster prominence*1.

[0032] Example 2: Based on Example 1, this example provides an EGFR mutation status prediction device, including a processor and a storage medium; The storage medium is used to store instructions; The processor is used to operate according to the instructions to execute the steps of the method according to embodiment 1.

[0033] Embodiment 3: Based on Embodiment 1, this embodiment provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps of the method described in Embodiment 1 are implemented.

[0034] Embodiment 4: Based on Embodiment 1, this embodiment provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the method described in Embodiment 1 when executing the computer program.

[0035] Embodiment 5: Based on Embodiment 1, this embodiment provides a computer program product, including a computer program, which implements the steps of the method described in Embodiment 1 when executed by a processor.

[0036] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0037] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes 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 generate 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 1A device that provides the functions specified in a block or multiple blocks.

[0038] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0039] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0040] The above is only a preferred implementation of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present application. These improvements and modifications should also be regarded as the scope of protection of the present application.

Claims

1. A method for predicting EGFR mutation status, characterized in that: include: Obtain imaging data and clinical data of lung cancer patients to be tested; Extracting radiomics features from the image data; Extracting clinical characteristics from the clinical data; Inputting the radiomics features and clinical features into a pre-constructed EGFR mutation prediction model; wherein the EGFR mutation prediction model learns the correspondence between the radiomics features, clinical features and EGFR mutation status; According to the output of the EGFR mutation prediction model, the EGFR mutation status prediction result is determined.

2. The method according to claim 1, characterized in that The method for obtaining the EGFR mutation prediction model comprises: Acquire a lung cancer patient dataset, wherein the lung cancer patient dataset includes multiple samples, each sample includes imaging data, gene sequencing data, and clinical data related to lung cancer of a lung cancer patient; Annotating and extracting the image data to obtain a first radiomics feature; Extracting the clinical data to obtain a first clinical feature; Labeling samples according to the gene sequencing data, wherein the labels are divided into EGFR wild type and EGFR mutant type; The least absolute shrinkage and selection operator LASSO algorithm was used to screen the radiomic features and clinical features for EGFR mutation prediction from the first radiomic features and the first clinical features; Based on the screened radiomics features and clinical features, determining training samples from the lung cancer patient dataset; Using the radiomics features, clinical characteristics and labels corresponding to the training samples, a multifactorial logistic regression was used to construct an EGFR mutation prediction model.

3. The method according to claim 1 or 2, characterized in that: The first radiomics feature includes at least one of a morphological feature, a first-order statistical feature, and a second-order and higher-order texture feature.

4. The method according to claim 3, characterized in that The morphological features are used to describe the size of the region of interest ROI and the degree of similarity between the region of interest ROI and a sphere, including volume, surface area, maximum diameter, area percentage, density, eccentricity, and sphericity; The first-order statistical features are used to describe the statistical characteristics of the voxel intensity distribution in the region of interest ROI, including mean, median, minimum, maximum, cluster prominence, difference standard deviation, difference variance, skewness and kurtosis; The second-order and high-order texture features are used to describe the characteristics of the intensity level of the spatial distribution of voxels, including intensity histogram distribution pattern, low grayscale highlight area, neighborhood grayscale difference matrix density, neighborhood grayscale difference matrix complexity, highlight area, grayscale co-occurrence matrix, grayscale level length matrix, grayscale band matrix and neighborhood grayscale difference matrix.

5. The method according to claim 1 or 2, characterized in that: The radiomics features include: area percentage, intensity histogram distribution pattern, low grayscale highlight area, neighborhood grayscale difference matrix complexity, skewness, highlight area, difference variance, and cluster prominence.

6. The method according to claim 1 or 2, characterized in that: The clinical characteristics include: age, gender, clinical stage, and thyroid transcription factor expression level.

7. The method according to claim 1, characterized in that The data processing process of the EGFR mutation prediction model includes: The imaging score is calculated according to the imaging omics features; Obtaining the probability of EGFR mutation according to the imaging score and the clinical characteristics; The EGFR mutation probability is compared with the EGFR mutation threshold to obtain an EGFR mutation status prediction result.

8. The method according to claim 7, characterized in that The imaging score is calculated based on the imaging features, including: Radiological score = area percentage*1+intensity histogram distribution mode*0.304+low grayscale highlight area*0.112-neighborhood grayscale difference matrix complexity*0.017-skewness*0.47-highlight area*0.485-difference variance*0.978-cluster prominence*1.

9. An EGFR mutation status prediction device, characterized in that: including processor and storage medium; The storage medium is used to store instructions; The processor is configured to operate according to the instructions to execute the steps of the method according to any one of claims 1 to 8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 8 are implemented.

Citation Information

Patent Citations

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

    CN112951406A

  • EGFR gene mutation detection method and system based on chest CT image

    CN113017674A

  • Hepatoma extrahepatic metastasis prediction model based on radiomics and construction method and application thereof

    CN113555112A

  • Artificial intelligence system for noninvasive prediction of EGFR / TP53 co-mutation lung cancer patient

    CN115312126A

  • Prediction of immunotherapy effect of advanced cancer patient based on weighted multi-lesion radiomics characteristics

    CN115440386A