Color Doppler ultrasound imaging omics feature extraction method and application

Abstract

The invention discloses a color Doppler ultrasound image omics feature extraction method and application, and belongs to the technical field of medical image processing, and the method comprises the steps: determining a tumor solid region image from a color sampling frame of a color Doppler ultrasound image, and determining a blood vessel distribution density as blood flow distribution based on the tumor solid region image; after the tumor solid region image is processed into RGB three-channel data, an obvious red pixel position and an obvious blue pixel position are found from the tumor solid region image, the blood flow direction and the blood flow velocity are determined based on the positions, and a processed tumor real-time region image containing the blood flow direction, the blood flow velocity and blood flow distribution is obtained; according to the method, the color Doppler imaging characteristics are extracted from the processed tumor real-time region image, so that the extracted imaging characteristics have more comprehensive and abundant tissue structure information, meanwhile, the application of the color Doppler imaging characteristics in clinical medical diagnosis is also provided, and the comprehensive capability of disease assessment is improved in combination with clinical data.

Description

Technical Field

[0001] The present invention belongs to the technical field of image processing, and in particular relates to a method and application for extracting color Doppler ultrasound imaging features. Background Art

[0002] Ultrasound color Doppler is an imaging method that uses Doppler technology to describe the relative motion characteristics of tissues and body fluids (usually blood) and ultrasonic transducers. The development range of its signal represents the distribution of blood flow in the region of interest, the red and blue colors represent the direction of blood flow, and the brightness represents the flow velocity, thereby evaluating the characteristics of vascular blood flow and solid tumor neovascularization. Taking ovarian tumors as an example, according to the expert consensus definition of the International Ovarian Tumor Analysis (IOTA), the description of the richness of blood vessels (signal development range) is semi-quantitatively scored as no blood flow, a small amount, a moderate amount, and a large amount, respectively, but it is subjective to a certain extent by ultrasound doctors. Spectral Doppler technology quantifies the blood flow velocity at the sampling point, but due to factors such as the morphology of vascular blood flow (laminar flow, eddy flow, turbulent flow, etc.) and the heterogeneity within the tumor, the blood flow velocity at the sampling point is difficult to reflect the overall flow velocity characteristics.

[0003] Radiomics is a quantitative image analysis technology that uses computer image processing to extract features of medical images at high throughput. In recent years, a large number of papers have proven that radiomics can be used for early screening and accurate identification of tumors, and its diagnostic efficiency is close to the gold standard of pathological examination. However, the current ultrasound radiomics technology mainly revolves around grayscale ultrasound images or monochrome red coding of Doppler signals. For example, Chen Hui and Feng Weiwei from Ruijin Hospital affiliated to Shanghai Jiaotong University School of Medicine encoded the extracted blood flow signals in the lesion area into a single color red, and the rest of the lesion area into white. This monochrome processing only shows the distribution of blood vessels, and loses the blood flow direction and some flow velocity information. To date, there is no radiomics analysis method for extracting the signal features of the three colors and brightness of color Doppler. Summary of the invention

[0004] In view of the above, the purpose of the present invention is to provide a color Doppler ultrasound imaging genomics feature extraction method, which aims to describe blood flow information in a high-throughput quantitative manner, including blood flow distribution, direction, flow velocity and other characteristics displayed by ultrasound color Doppler, and to extract imaging genomics features based on these characteristics describing blood flow, and to provide clinical applications based on imaging genomics features.

[0005] To achieve the above-mentioned purpose of the invention, the embodiment provides a method for extracting color Doppler ultrasound imaging features, comprising the following steps: After the color Doppler ultrasound image is standardized, the solid area image of the tumor is determined from the color sampling frame, and the color area in the color sampling frame is divided by the solid area image of the tumor to obtain the vascular distribution density as the blood flow distribution; After processing the solid area image of the tumor into RGB three-channel data, the solid area image of the tumor is cropped, and then obvious red pixel positions and blue pixel positions are found from the cropped image, and the blood flow direction and blood flow velocity are determined based on the blue pixel position and the red pixel position, so as to obtain a processed real-time area image of the tumor including the blood flow direction, blood flow velocity, and blood flow distribution; Radiomic features were extracted from the real-time regional images of the processed tumors, and the average layer features of the radiomic features corresponding to the multi-frame images were calculated as candidate radiomic features. The candidate radiomic features were then standardized and screened to obtain color Doppler radiomic features.

[0006] Preferably, cropping the solid area image of the tumor includes: In the solid area image of the tumor, according to the position rule of the color scale, the image is cropped to ensure that the image contains the color scale and the flow rate text, and the flow rate text in the cropped image is extracted, wherein the flow rate text represents the maximum blood flow velocity.

[0007] Preferably, determining the blood flow direction and blood flow velocity based on the blue pixel position and the red pixel position includes: The red boundary information and blue boundary information of the color scale are extracted and quantified according to the blue pixel position and the red pixel position. These boundary information represent the RGB values ​​of the pixels corresponding to the maximum blood flow velocity in the red direction and the blue direction of each image. The relative blood flow velocity of each pixel of the blood flow signal in the solid area of ​​the tumor is calculated based on the RGB value and the maximum blood flow velocity to obtain the normalized blood flow velocity. The red direction corresponds to the blood flow direction towards the probe, and the blue reverse direction corresponds to the blood flow direction away from the probe.

[0008] Preferably, the method further comprises: processing the color Doppler ultrasound image into RGB three-channel data, and simultaneously extracting imaging omics features from the processed real-time tumor regional image and the RGB three-channel color Doppler ultrasound image to participate in the subsequent calculation of average layer features.

[0009] Preferably, the candidate radiomics features are standardized, including: Using the Z-score standardization method, the candidate imaging features were subtracted from the feature mean and then divided by the standard deviation of the feature to obtain the standardized processing results.

[0010] Preferably, feature screening includes: The univariate logistic regression method was used to analyze the correlation between the candidate radiomics features and disease labels corresponding to each standardized color Doppler image. The candidate radiomics features with correlation less than 0.1 were retained. The maximum correlation and minimum redundancy feature selection method was then used to rank the importance of the candidate radiomics features, and the color Doppler radiomics features were determined based on the ranking.

[0011] The embodiment of the present invention further provides a target classification device based on color Doppler ultrasound imaging genomics features, comprising: A first color Doppler imaging omics feature module, which is used to extract color Doppler imaging omics features based on the color Doppler ultrasound imaging omics feature extraction method; A first grayscale ultrasound imaging omics feature module, which is used to extract grayscale ultrasound imaging omics features from grayscale ultrasound images; The target classification module is used to classify the target by combining the color Doppler imaging genomics features alone or by fusing the color Doppler imaging genomics features and the grayscale ultrasound imaging genomics features simultaneously to obtain the target classification results, which include borderline or malignant ovarian tumors.

[0012] The embodiment of the present invention further provides a device for classifying the grade of serous ovarian malignant tumors based on color Doppler ultrasound imaging features, comprising: A second color Doppler imaging genomics feature module, which is used to extract color Doppler imaging genomics features based on the above-mentioned color Doppler ultrasound imaging genomics feature extraction method; A second grayscale ultrasound imaging omics feature module, which is used to extract grayscale ultrasound imaging omics features from grayscale ultrasound images; The grade classification module is used to classify the grade of serous ovarian malignant tumors by combining color Doppler imaging genomics features alone, or by fusing color Doppler imaging genomics features and grayscale ultrasound imaging genomics features simultaneously, to obtain a grade classification result, which includes high grade or non-high grade.

[0013] The embodiment of the present invention further provides a device for predicting clinical staging of ovarian malignant tumors based on color Doppler ultrasound imaging features, comprising: A third color Doppler imaging genomics feature module, which is used to extract color Doppler imaging genomics features based on the color Doppler ultrasound imaging genomics feature extraction method; A third grayscale ultrasound imaging omics feature module is used to extract grayscale ultrasound imaging omics features from grayscale ultrasound images; The level classification module is used to predict the clinical staging of ovarian malignant tumors by combining color Doppler imaging genomics features alone or by fusing color Doppler imaging genomics features and grayscale ultrasound imaging genomics features simultaneously to obtain staging prediction results, which include I-IV.

[0014] Compared with the prior art, the present invention has the following beneficial effects: The present invention extracts RGB three-channel information from solid tumor area images of color Doppler ultrasound images to analyze the color Doppler blood flow distribution, blood flow direction, and blood flow velocity characteristics, and extracts color Doppler imaging features based on these features. The imaging features extracted in this way have more comprehensive and rich tissue structure information, and are verified by clinical application based on imaging features, indicating that these imaging features can be more accurately used in clinical tumor differential diagnosis, neovascularization evaluation, pathological classification, heterogeneity, clinical staging and other tumor diseases, and combined with clinical data to improve the comprehensive ability of disease assessment. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. 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.

[0016] Figure 1 is a flow chart of a method for extracting color Doppler ultrasound imaging features provided in an embodiment; Figure 2 is a schematic diagram of the structure of a target classification device provided in an embodiment; Figure 3 is a workflow diagram of a target classification device provided in an embodiment; Figure 4 is a schematic diagram of the structure of a device for classifying the grade of serous ovarian malignant tumors provided in an embodiment; Figure 5 It is a schematic diagram of the structure of the device for predicting clinical staging of ovarian malignant tumors provided in the embodiment. DETAILED DESCRIPTION

[0017] To make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific implementation methods described herein are only used to explain the present invention and do not limit the scope of protection of the present invention.

[0018] The inventive concept of the present invention is: in view of the technical problem that the existing imaging omics features mainly come from the analysis of grayscale images and monochrome-coded color Doppler images, and lack important information on the blood flow velocity and direction of tumor blood vessels, the present invention provides a color Doppler ultrasound imaging omics feature extraction method, which extracts blood flow distribution, blood flow direction, and blood flow velocity information from RGB three-channel color Doppler images, and extracts imaging omics features based on this. The imaging omics features extracted in this way contain more comprehensive and accurate tissue structure information. The imaging omics features provided by the present invention are also clinically verified to illustrate that the extraction of imaging omics features can extract classification, recognition, and evaluation results for clinical applications.

[0019] like Figure 1 As shown, the color Doppler ultrasound imaging feature extraction method provided in the embodiment includes the following steps: S1, after standardizing the color Doppler ultrasound image, determine the solid area image of the tumor from the color sampling frame, and divide the color area in the color sampling frame by the solid area image of the tumor to obtain the vascular distribution density as the blood flow distribution.

[0020] In an embodiment, a color Doppler ultrasound image is obtained, and each frame of the color Doppler ultrasound image in the image is standardized, including size adjustment, etc., as a basic image for subsequent feature extraction, and stored in DICOM (Digital Imaging and Communications in Medicine) format, and then a color sampling frame in the color Doppler ultrasound image after the standardized processing is determined, and the solid area of ​​the tumor is determined as the region of interest (ROI area) based on the color sampling frame. The solid area of ​​the tumor is the intracystic nipple, cystic wall and diaphragm, etc., and the solid area of ​​the tumor is saved, and other areas are set to blank, and a tumor solid area image with the same size as the initial color Doppler ultrasound image is obtained, and stored as an Nrrd (Nearly Raw Raster Data) file. The color area in the color sampling frame is divided by the solid area image of the tumor to obtain the vascular distribution density as the blood flow distribution. For multiple color Doppler ultrasound images, the vascular distribution density of each image is calculated, and the average is recorded as the average vascular density value to exclude the influence of the cardiac cycle on the blood flow display.

[0021] S2, after processing the solid area image of the tumor into RGB three-channel data, the solid area image of the tumor is cropped, and obvious red pixel positions and blue pixel positions are found from the cropped image, and the blood flow direction and blood flow velocity are determined based on the blue pixel position and the red pixel position, to obtain a processed real-time tumor area image including the blood flow direction, blood flow velocity, and blood flow distribution.

[0022] In an embodiment, Matlab can be used to unify the DICOM file of the color Doppler image and the Nrrd file of the tumor solid area image into R / G / B two-dimensional three-channel data, and the pixel value sampling of each channel is normalized to 1 to 64. Then the solid area image of the tumor is cropped, specifically, according to the position rule of the color scale bar in the solid area image of the tumor, the image is cropped to ensure that the image contains the color scale bar and the flow velocity text, and the OCR tool is used to extract the flow velocity text in the cropped image, wherein the flow velocity text represents the maximum blood flow velocity, which is used for the subsequent calculation of the flow velocity in other areas.

[0023] In the embodiment, obvious red pixel positions and blue pixel positions are found from the cropped image, for example, the upper limit of red pixels is set to [25, 255, 255], and the lower limit is set to [0, 30, 30]; the upper limit of blue pixels is set to [140, 255, 255], and the lower limit is set to [100, 50, 50]. Since the image contains certain impurities, such as flow rate values ​​and images other than the color scale bar, the maximum connected area is extracted. Since the shape of the maximum connected area is irregular, the linear structure with a smaller area is removed, and the structure that is as close to a rectangle as possible is retained.

[0024] Then, the blood flow direction and blood flow velocity are determined based on the blue pixel position and the red pixel position. Specifically, the red boundary information and the blue boundary information of the color ruler are extracted and quantified based on the blue pixel position and the red pixel position. These boundary information represent the RGB values ​​of the pixels corresponding to the maximum blood flow velocities in the red direction and the blue direction of each image. The relative blood flow velocity of each pixel of the blood flow signal in the solid area of ​​the tumor is calculated based on the RGB value and the maximum blood flow velocity to obtain the normalized blood flow velocity. At the same time, the blood flow direction corresponding to the red direction is determined to be the direction toward the probe, and the blood flow direction corresponding to the blue reverse direction is the direction away from the probe. Based on this, the blood flow direction is determined, where the positive sign "+" represents red and the negative sign "-" represents blue, thereby obtaining a processed real-time regional image of the tumor including the blood flow direction, blood flow velocity, and blood flow distribution.

[0025] S3, extracting radiomics features from the real-time regional images of the processed tumor and calculating the average layer features of the radiomics features corresponding to the multi-frame images as candidate radiomics features, and then standardizing and feature screening the candidate radiomics features to obtain color Doppler radiomics features.

[0026] In an embodiment, open source Matlab can be used to extract radiomics features from real-time regional images of treated tumors, including shape features, first-order statistical features, texture features, and high-order features. For wavelet radiomics features, considering all possible combinations of applying high-pass filters (H) or low-pass filters (L) on the x, y, and z axes, eight types of wavelet features can be generated, including: "-LLL", "-LLH", "-LHL", "-LHH", "-HLL", "-HLH", "-HHL", and "-HHH" features. Texture features include gray level cooccurence matrix (GLCM), gray level run length matrix (GLLRLM), gray level size zone matrix (GLSZM), neighboring gray tone difference matrix (NGTDM), and gray level dependence matrix (GLDM) features.

[0027] Then, the average values ​​of the corresponding imaging features of multiple color Doppler ultrasound images in the color Doppler ultrasound images are calculated to obtain the average layer features as candidate imaging features to eliminate the influence of the cardiac cycle on blood flow display.

[0028] In the embodiment, the candidate imaging genomics features are also standardized and feature screened to obtain color Doppler imaging genomics features. Specifically, the Z-score standardization method is used to subtract the feature mean from the candidate imaging genomics features, and then divided by the standard deviation of the feature to obtain the standardized processing result; then, when screening the features, the single-factor logical review method is first used to analyze the correlation between the candidate imaging genomics features and the disease labels corresponding to each standardized color Doppler image, and the candidate imaging genomics features with a correlation less than 0.1 are retained, and then the maximum correlation and minimum redundancy (mRMR, Max-Relevance and Min-Redundancy) feature selection method is used to rank the importance of the candidate imaging genomics features, and determine the appropriate color Doppler imaging genomics features based on the ranking. Among them, the maximum correlation constraint is to select the feature with the largest mutual information value between a single feature and the label. The minimum redundancy constraint is to remove redundant features that are highly dependent on other features.

[0029] In the embodiment, a disease prediction model based on color Doppler ultrasound imaging omics features is constructed, and the model performance is evaluated using ROC (Receiver Operating Characteristic Curve) and AUC (Area Under Curve), and the corresponding sensitivity, specificity, PPV (positive predictive value), NPV (negative predictive value), and accuracy are calculated to verify the model effectiveness. At the same time, the ultrasound diagnostic capability is improved from the two dimensions of ultrasound grayscale and color by combining the grayscale ultrasound imaging omics features of the grayscale image.

[0030] Next, taking ovarian cancer as an example, it is demonstrated that the above-mentioned color Doppler ultrasound imaging feature extraction method can be applied to, including but not limited to: distinguishing borderline ovarian tumors from malignant tumors, distinguishing high-grade from non-high-grade serous ovarian malignant tumors, and predicting the clinical stage of ovarian malignant tumors.

[0031] For the task of distinguishing borderline ovarian tumors from malignant tumors, accurate diagnosis of borderline ovarian tumors is crucial for optimizing the surgical plan to preserve fertility. Therefore, the current preliminary recommendation for intraoperative frozen section analysis (FSA) for borderline ovarian tumors is only 44.8-81.0%. Figure 2 and Figure 3 As shown, the embodiment of the present invention constructs a target classification device 20 based on color Doppler ultrasound imaging genomics features, including: a first color Doppler imaging genomics feature module 31, a first grayscale ultrasound imaging genomics feature module 22, and a target classification module 23.

[0032] The first color Doppler imaging genomics feature module 21 is used to extract color Doppler imaging genomics features using the color Doppler ultrasound imaging genomics feature extraction method. The specific extraction process is detailed in the above extraction method.

[0033] The first grayscale ultrasound image omics feature module 22 is used to extract grayscale ultrasound image omics features from grayscale ultrasound images. Specifically, grayscale ultrasound image omics features are extracted from grayscale ultrasound images, including shape features, first-order statistical features, texture features and high-order features. The extraction method is not limited.

[0034] The target classification module 33 is used to combine the grayscale ultrasound imaging genomics features alone, the color Doppler imaging genomics features alone, or the color Doppler imaging genomics features and the grayscale ultrasound imaging genomics features at the same time for target classification, and obtain the target classification result, which includes ovarian borderline tumors or malignant tumors. Specifically, in the target classification module 23, the diagnostic performance of the grayscale ultrasound imaging genomics model (Gray-scale Radomics, GS1), the color Doppler imaging genomics model (Color- Doppler Radomics, CD1) and the fusion model (Gray-scale Color-Doppler Radiomics, GSCD1) in distinguishing ovarian borderline tumors from malignant tumors is constructed and compared. When fusing the color Doppler imaging genomics features and the grayscale ultrasound imaging genomics features, the color Doppler imaging genomics features and the grayscale ultrasound imaging genomics features are combined, and then the features with the same number as the color Doppler imaging genomics features are selected and input into the fusion model for target classification to ensure that the feature input size of the three models is consistent. Each model uses the random forest algorithm.

[0035] In the data set, the AUCs of the three models GS1, CD1, and GSCD1 for the classification of ovarian borderline tumors and malignant tumors are 0.825, 0.899, and 0.908, respectively. This shows that compared with GS1, the performance of CD1 and GSCD1 is enhanced, and the comparison between the two models GSCD1 and GS1 is P The value is 0.005. The smaller the value, the more obvious the statistical difference between the two models. Therefore, the difference between the two models is statistically significant. The diagnostic accuracy of GSCD1 (90.8%) is better than the current clinical preoperative FSA diagnosis of borderline ovarian tumors (44.8% -81%), indicating that color Doppler blood flow signals can effectively improve the diagnostic performance of the model (Table 1).

[0036] Table 1 Diagnostic performance of GS1, CD1, and GSCD1

[0037] The model included 209 cases. The values ​​in brackets of each model indicator are the corresponding 95% confidence intervals, which are used to estimate the interval range of the corresponding population parameter. For example, AUC 0.825 (0.767, 0.874) means that by repeatedly sampling and calculating intervals from the same population data, approximately 95% of the intervals will contain the true population AUC value.

[0038] For the task of distinguishing high-grade and non-high-grade serous ovarian malignancies, the precision treatment of ovarian cancer is closely related to the histological subtype of the tumor. Currently, the tissue type of ovarian epithelial malignancies is determined clinically by puncture biopsy and pathological examination before surgery. However, domestic and foreign guidelines point out that fine needle aspiration is prohibited for early-stage and cystic lesions to avoid artificially increasing the clinical stage of the tumor. To this end, the embodiment of the present invention uses preoperative ultrasound image analysis, such as Figure 4 As shown, a serous ovarian malignant tumor grade classification device 40 based on color Doppler ultrasound imaging genomics features is constructed, including: a second color Doppler imaging genomics feature module 41, a second grayscale ultrasound imaging genomics feature module 42, and a grade classification module 43.

[0039] Among them, the second color Doppler imaging omics feature module 41 is used to extract color Doppler imaging omics features based on the above-mentioned color Doppler ultrasound imaging omics feature extraction method; the second grayscale ultrasound imaging omics feature module 42 is used to extract grayscale ultrasound imaging omics features from grayscale ultrasound images. The specific process is detailed in the above-mentioned color Doppler ultrasound imaging omics feature extraction method, which will not be repeated here.

[0040] The level classification module 43 is used to classify the level of serous ovarian malignant tumors by combining grayscale ultrasound imaging omics features alone, color Doppler imaging omics features alone, or by fusing color Doppler imaging omics features and grayscale ultrasound imaging omics features at the same time, and obtain a level classification result, which includes high level or non-high level. Specifically, in the level classification module 43, the diagnostic performance of three models GS2, CD2 and GSCD2 in distinguishing high-level from non-high-level serous ovarian malignant tumors is also constructed. The AUC values ​​of the three models in the data set are 0.760, 0.792, and 0.861, respectively. It can be seen that compared with the GS2 model, the performance of GSCD2 is significantly enhanced ( P <0.0001), and maintained a stable and high sensitivity, as shown in Table 2.

[0041] Table 2 Diagnostic performance of GS2, CD2, and GSCD2

[0042] The number of cases included in the model was 428, and the values ​​in brackets of each model indicator were their corresponding 95% confidence intervals.

[0043] For the task of predicting the clinical staging of ovarian malignant tumors, the commonly used ADNEX model in clinical practice has a certain predictive value for the FIGO clinical staging (International Federation of Gynecology and Obstetrics) of ovarian malignant tumors, but the prediction accuracy of ovarian borderline tumors and FIGO stage I is relatively low (AUC0.75). Figure 5 As shown, the embodiment of the present invention constructs a clinical staging prediction device 50 for ovarian malignant tumors based on color Doppler ultrasound imaging features, including: a third color Doppler imaging feature module 51, a third grayscale ultrasound imaging feature module 52, and a staging module 53.

[0044] Among them, the third color Doppler imaging omics feature module 51 is used to extract color Doppler imaging omics features based on the above-mentioned color Doppler ultrasound imaging omics feature extraction method; the third grayscale ultrasound imaging omics feature module 52 is used to extract grayscale ultrasound imaging omics features from grayscale ultrasound images. The specific process is detailed in the above-mentioned color Doppler ultrasound imaging omics feature extraction method, which will not be repeated here.

[0045] The staging module 53 is used to predict the clinical staging of ovarian malignant tumors by combining grayscale ultrasound imaging genomics features alone, color Doppler imaging genomics features alone, or by fusing color Doppler imaging genomics features and grayscale ultrasound imaging genomics features at the same time, to obtain the staging prediction results. Specifically, in the staging module 53, three models GS3, CD3 and GSCD3 are also constructed to distinguish the diagnostic performance of ovarian borderline tumors and FIGO stage I epithelial malignant tumors and FIGO stage II-I epithelial malignant tumors. In the task of distinguishing ovarian borderline tumors from FIGO stage I epithelial malignant tumors, the AUC values ​​of the three models are 0.803, 0.857, and 0.871, respectively, as shown in Table 3.

[0046] Table 3 Diagnostic efficacy of GS3, CD3, and GSCD3

[0047] The number of cases included in the model was 106, and the values ​​in brackets of each model indicator were their corresponding 95% confidence intervals.

[0048] In the task of distinguishing ovarian borderline tumors from FIGO stage II-IV epithelial malignant tumors, the AUC values ​​of the three models were 0.853, 0.918, and 0.924, respectively, as shown in Table 4.

[0049] Table 4 Diagnostic efficacy of GS4, CD4, and GSCD4

[0050] The number of cases included in the model was 170, and the values ​​in brackets of each model indicator were their corresponding 95% confidence intervals.

[0051] GSCD has the value of predicting the differentiation of borderline ovarian tumors and FIGO clinical staging, especially in the task of differentiating borderline ovarian tumors from FIGO stage I epithelial malignant tumors, and its AUC value is higher than the ADNEX model commonly used in clinical ovarian malignant tumor staging.

[0052] Through the above clinical application comparison, the fusion model GSCD that fuses ultrasound grayscale images and color Doppler images has been tested. Compared with the model GS constructed only based on grayscale images, it has significantly improved the ability to predict and differentially diagnose the pathological properties of preoperative ovarian masses, indicating the feasibility and practicality of the color Doppler ultrasound imaging genomics characteristics of the present invention.

[0053] The specific implementation methods described above provide a detailed description of the technical solutions and beneficial effects of the present invention. It should be understood that the above is only the most preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, supplements and equivalent substitutions made within the scope of the principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for extracting color Doppler ultrasound imaging features, characterized in that: The following steps are involved: After the color Doppler ultrasound image is standardized, the solid area image of the tumor is determined from the color sampling frame, and the color area in the color sampling frame is divided by the solid area image of the tumor to obtain the vascular distribution density as the blood flow distribution; After processing the solid area image of the tumor into RGB three-channel data, the solid area image of the tumor is cropped, and then obvious red pixel positions and blue pixel positions are found from the cropped image, and the blood flow direction and blood flow velocity are determined based on the blue pixel position and the red pixel position, so as to obtain a processed real-time area image of the tumor including the blood flow direction, blood flow velocity, and blood flow distribution; Radiomic features were extracted from the real-time regional images of the processed tumors, and the average layer features of the radiomic features corresponding to the multi-frame images were calculated as candidate radiomic features. The candidate radiomic features were then standardized and screened to obtain color Doppler radiomic features.

2. The color Doppler ultrasound imaging feature extraction method according to claim 1, characterized in that: Crop the solid area image of the tumor, including: In the solid area image of the tumor, according to the position rule of the color scale, the image is cropped to ensure that the image contains the color scale and the flow rate text, and the flow rate text in the cropped image is extracted, wherein the flow rate text represents the maximum blood flow velocity.

3. The color Doppler ultrasound imaging feature extraction method according to claim 2, characterized in that: The blood flow direction and blood flow velocity are determined based on the blue pixel position and the red pixel position, including: The red boundary information and blue boundary information of the color scale are extracted and quantified according to the blue pixel position and the red pixel position. These boundary information represent the RGB values ​​of the pixels corresponding to the maximum blood flow velocity in the red direction and the blue direction of each image. The relative blood flow velocity of each pixel of the blood flow signal in the solid area of ​​the tumor is calculated based on the RGB value and the maximum blood flow velocity to obtain the normalized blood flow velocity. The red direction corresponds to the blood flow direction towards the probe, and the blue reverse direction corresponds to the blood flow direction away from the probe.

4. The color Doppler ultrasound imaging feature extraction method according to claim 2, characterized in that: Also includes: The color Doppler ultrasound image is processed into RGB three-channel data, and the imaging omics features are simultaneously extracted from the processed real-time tumor regional image and the RGB three-channel color Doppler ultrasound image to participate in the subsequent calculation of the average layer features.

5. The color Doppler ultrasound imaging feature extraction method according to claim 2, characterized in that: Standardization of candidate radiomics features including: Using the Z-score standardization method, the candidate imaging features were subtracted from the feature mean and then divided by the standard deviation of the feature to obtain the standardized processing results.

6. The method for extracting color Doppler ultrasound imaging features according to claim 2, characterized in that: Feature screening, including: The univariate logistic regression method was used to analyze the correlation between the candidate radiomics features and disease labels corresponding to each standardized color Doppler image. The candidate radiomics features with correlation less than 0.1 were retained. The maximum correlation and minimum redundancy feature selection method was then used to rank the importance of the candidate radiomics features, and the color Doppler radiomics features were determined based on the ranking.

7. A target classification device based on color Doppler ultrasound imaging features, characterized in that: include: A first color Doppler imaging omics feature module, which is used to extract color Doppler imaging omics features based on the color Doppler ultrasound imaging omics feature extraction method according to any one of claims 1 to 6; A first grayscale ultrasound imaging omics feature module, which is used to extract grayscale ultrasound imaging omics features from grayscale ultrasound images; The target classification module is used to perform target classification by combining color Doppler imaging genomics features alone or by fusing color Doppler imaging genomics features and grayscale ultrasound imaging genomics features simultaneously to obtain target classification results, which include ovarian borderline tumors or malignant tumors.

8. A device for classifying the grade of serous ovarian malignant tumors based on color Doppler ultrasound imaging features, characterized in that: include: A second color Doppler imaging omics feature module, which is used to extract color Doppler imaging omics features based on the color Doppler ultrasound imaging omics feature extraction method according to any one of claims 1 to 6; A second grayscale ultrasound imaging omics feature module, which is used to extract grayscale ultrasound imaging omics features from grayscale ultrasound images; The grade classification module is used to classify the grade of serous ovarian malignant tumors by combining color Doppler imaging genomics features alone, or by fusing color Doppler imaging genomics features and grayscale ultrasound imaging genomics features simultaneously, to obtain a grade classification result, which includes high grade or non-high grade.

9. A device for predicting clinical staging of ovarian malignant tumors based on color Doppler ultrasound imaging features, characterized in that: include: A third color Doppler imaging omics feature module, which is used to extract color Doppler imaging omics features based on the color Doppler ultrasound imaging omics feature extraction method according to any one of claims 1 to 6; A third grayscale ultrasound imaging omics feature module is used to extract grayscale ultrasound imaging omics features from grayscale ultrasound images; The staging module is used to predict the clinical staging of ovarian malignant tumors by combining color Doppler imaging genomics features alone or by simultaneously fusing color Doppler imaging genomics features and grayscale ultrasound imaging genomics features to obtain staging prediction results, which include I-IV.

Images