Method and system for dividing tumor perfusion sub-regions based on dynamic ultrasound contrast data

Abstract

The application provides a method and system for dividing tumor perfusion sub-regions based on dynamic ultrasound contrast data, the method comprising obtaining pre-processed dynamic ultrasound contrast images and pixelizing, so as to obtain time-intensity curves of each pixel point of tumor tissue and average time-intensity curves of normal tissue, obtain perfusion characteristic parameter ratios of each pixel point of tumor tissue, generate perfusion sub-regions of the tumor through clustering results of clustering of all pixel points in the interval, and map information thereof back to the original image. The application maximizes the use of dynamic image data through a non-invasive diagnostic method, comprehensively captures perfusion characteristics of the tumor from the pixel level, visualizes and quantifies tumor perfusion heterogeneity, improves the readability of the dynamic contrast video, helps to explore the correlation between tumor sub-regions and benignity and malignancy and construct a diagnostic model accordingly, improves the diagnostic accuracy of breast cancer, and provides more information support for personalized treatment of patients.

Description

Technical Field

[0001] This invention belongs to the field of ultrasound contrast imaging technology, specifically relating to a method and system for dividing tumor perfusion subregions based on dynamic ultrasound contrast imaging data. Background Technology

[0002] Breast cancer is one of the most common cancers among women worldwide, and its rising incidence poses a significant threat to women's health. Despite significant advancements in medical technology for the early diagnosis and treatment of breast cancer, its high degree of heterogeneity results in substantial differences in clinical presentation, treatment response, and disease progression. This heterogeneity presents new challenges for precision medicine; therefore, in-depth research into the heterogeneity of breast cancer is crucial for developing efficient diagnostic models and improving cure rates. Currently, although biopsy is a common method for diagnosing benign and malignant breast tumors, its accuracy and comprehensiveness are limited by the complex heterogeneity and potential invasiveness of breast cancer.

[0003] Currently, ultrasound contrast-enhanced perfusion assessment of breast tumors exhibits significant subjectivity and experience dependence. Although ultrasound contrast-enhanced imaging provides physicians with important diagnostic information, the overlapping perfusion characteristics of benign and malignant breast tumors make it challenging to differentiate between these lesions solely through imaging techniques. Furthermore, the available ultrasound contrast-enhanced indicators are primarily qualitative, such as high enhancement during the arterial phase, rapid decay, and heterogeneous perfusion. While these indicators are helpful for diagnosis to some extent, their application and accuracy are limited, failing to comprehensively reflect the complexity of the tumor.

[0004] Furthermore, current techniques for analyzing contrast-enhanced ultrasound data often only consider the characteristics of the entire tumor. However, breast tumors exhibit high internal heterogeneity, and analyses based solely on overall characteristics neglect the differences between different subregions within the tumor. Most studies have over-relying on static images and intensity-based parameters for tumor subregion segmentation, while ignoring the temporal dimension of the tumor. This results in insufficient assessment of the dynamic biological behavior of the tumor, limiting the comprehensive evaluation of tumor heterogeneity and the comprehensiveness and accuracy of diagnosis, thus affecting more accurate diagnosis and the development of personalized treatment plans. Summary of the Invention

[0005] The applicant discovered that dividing tumor perfusion subregions based on pixel-level data resulted in different proportions of benign and malignant tumor subregions. Through analysis of data from multiple patients with benign and malignant tumors, significant differences were found in the percentages of certain subregions, with statistical significance. Therefore, this invention proposes a method for dividing tumor perfusion subregions based on dynamic ultrasound contrast imaging data. This method aims to visualize and quantify dynamic information from ultrasound contrast imaging, improving the diagnostic efficacy for benign and malignant tumors. The method includes:

[0006] A preprocessed dynamic ultrasound contrast image is obtained and pixelated to determine a first region representing tumor tissue and a second region representing normal tissue in the dynamic ultrasound contrast image; the preprocessing is used to remove image noise, compensate for motion displacement, and achieve image registration;

[0007] Obtain the temporal intensity curve of each pixel in the first interval and the average temporal intensity curve in the second interval, and calculate the perfusion feature parameter ratio.

[0008] Based on the perfusion feature parameters of each pixel in the first interval and the average perfusion feature parameters of all pixels in the second interval, the ratio of perfusion feature parameters of each pixel in the first interval is obtained; the perfusion feature parameters of each pixel in the first interval are obtained from the time intensity curve of each pixel in the first interval, and the average perfusion feature parameters of all pixels in the second interval are obtained from the average time intensity curve of all pixels in the second interval.

[0009] Based on the ratio of the perfusion feature parameters of each pixel in the first interval, the pixels in the first interval are clustered.

[0010] Based on the clustering results, perfusion subregions of the tumor are generated, and the information of the perfusion subregions is mapped back to the original grayscale image of the ultrasound contrast imaging.

[0011] Specifically, the preprocessing includes noise removal and / or motion correction.

[0012] Further, the step of "obtaining the time intensity curve of each pixel in the first interval and obtaining the average time intensity curve in the second interval" includes:

[0013] The feature values ​​of the dynamic ultrasound contrast image deleted by the preprocessing are supplemented by interpolation. Based on the ultrasound contrast image with supplemented features by interpolation, the time intensity curve of each pixel in the first interval and the average time intensity curve in the second interval are obtained.

[0014] Optionally, the perfusion characteristic parameters include any one or more of the following: time to peak, peak intensity, rise slope, half-fall slope, time to half peak, and area under the curve.

[0015] Specifically, the step of "obtaining the ratio of the perfusion feature parameters of each pixel in the first interval based on the perfusion feature parameters of each pixel in the first interval and the average perfusion feature parameters of all pixels in the second interval" includes:

[0016] The median or average value of the perfusion feature parameters of all pixels in the second interval is taken as the average perfusion feature parameter;

[0017] Based on the perfusion feature parameters of each pixel in the first interval, the perfusion feature parameters of each pixel in the first interval are obtained, and the perfusion feature parameters of each pixel in the first interval are divided by the average perfusion feature parameter to obtain the ratio of the perfusion feature parameters of each pixel in the first interval.

[0018] Optionally, the clustering of pixels in the first interval is achieved through principal component analysis, K-means clustering, mean shift, spectral clustering, DBSCAN, hierarchical clustering, or machine learning clustering.

[0019] Specifically, the clustering results include at least two different sets of pixels.

[0020] Specifically, the step of "generating tumor perfusion subregions based on clustering results and mapping the information of the perfusion subregions back to the original grayscale image of the ultrasound contrast imaging image" includes:

[0021] The perfusion sub-region information generated based on the clustering results is mapped back to the original grayscale image of the ultrasound contrast image, and the number and / or percentage of pixels in each perfusion sub-region are calculated to describe the tissue characteristics and distribution of different regions of the tumor in the original grayscale image in a visual and / or quantitative manner.

[0022] This invention also proposes a system for segmenting tumor perfusion subregions based on dynamic ultrasound contrast imaging data, the system comprising:

[0023] A labeling module is used to obtain and pixelate a preprocessed dynamic ultrasound contrast image, and to determine a first region representing tumor tissue and a second region representing normal tissue in the dynamic ultrasound contrast image; the preprocessing is used to remove image noise, compensate for motion displacement, and achieve image registration.

[0024] The acquisition module is used to acquire the time intensity curve of each pixel in the first interval and to acquire the average time intensity curve in the second interval.

[0025] The calculation module is used to obtain the ratio of the perfusion feature parameters of each pixel in the first interval based on the perfusion feature parameters of each pixel in the first interval and the average perfusion feature parameters of all pixels in the second interval; the perfusion feature parameters of each pixel in the first interval are obtained from the time intensity curve of each pixel in the first interval, and the average perfusion feature parameters of all pixels in the second interval are obtained from the average time intensity curve of all pixels in the second interval.

[0026] The clustering module is used to cluster the pixels in the first interval based on the ratio of the infusion feature parameters of each pixel in the first interval;

[0027] The mapping module is used to generate perfusion subregions of the tumor based on the clustering results, and to map the information of the perfusion subregions back to the original grayscale image of the ultrasound contrast image.

[0028] The present invention also proposes a computer-readable storage medium storing executable instructions that, when executed by a processor, implement the method described above for dividing tumor perfusion subregions based on dynamic ultrasound contrast imaging data.

[0029] The present invention has at least the following beneficial effects:

[0030] The proposed solution utilizes non-invasive imaging technology. By performing more detailed analysis and processing of tumor perfusion images and employing computer algorithms to process and analyze the images, automated processing can be achieved, reducing the workload of medical personnel and more accurately delineating tumor perfusion subregions. This provides a visual representation of the tumor perfusion status, helping doctors make more precise diagnoses and develop targeted, personalized treatment plans to improve treatment outcomes.

[0031] Furthermore, by acquiring keyframes and performing motion correction, this approach ensures the spatial and temporal continuity and accuracy of the obtained dynamic ultrasound contrast imaging image set, which is beneficial for subsequent image processing and analysis. By clustering pixels using methods such as principal component analysis or K-means, tumor perfusion subregions can be effectively divided, helping to distinguish different tissue characteristics and distributions. By mapping the perfusion feature parameters of pixels in different regions back to the original dynamic ultrasound contrast imaging images, the tissue characteristics and distribution of different regions can be quantitatively described. This helps medical personnel determine the perfusion heterogeneity within the tumor, providing more information for personalized treatment and guiding clinical decision-making and treatment plan development.

[0032] Therefore, this invention provides a method and system for segmenting tumor perfusion subregions based on dynamic ultrasound contrast imaging data. This invention maximizes the use of dynamic image data through a non-invasive diagnostic method, comprehensively captures the perfusion characteristics of the tumor at the pixel level, visualizes and quantifies tumor perfusion heterogeneity, improves the readability of dynamic contrast imaging videos, and helps to explore the correlation between tumor subregions and benign or malignant tumors and construct diagnostic models accordingly. While improving the diagnostic accuracy of breast cancer, it also provides more information support for personalized treatment of patients. Attached Figure Description

[0033] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0034] Figure 1 This is a schematic diagram of the overall process of the method for dividing tumor perfusion subregions based on dynamic ultrasound contrast imaging data provided in Example 1;

[0035] Figure 2 Example of time-intensity curves for pixels in the first interval;

[0036] Figure 3 This is an example diagram of clustering and mapping pixels based on infusion feature parameters;

[0037] Figures 4(a)-(f) show examples of the results of mapping back to the original image based on the clustering results and performing benign and malignant identification. Figures 4(a)-(c) show examples of images identified as benign tumors, and Figures 4(d)-(f) show examples of images identified as malignant tumors.

[0038] Figure 5 This is an example diagram illustrating the differentiation of benign and malignant breast tumors through perfusion subregions;

[0039] Figure 6 A flowchart illustrating the method for acquiring and preprocessing dynamic ultrasound contrast images;

[0040] Figure 7 This is a schematic diagram of the module structure of the system for dividing tumor perfusion subregions based on dynamic ultrasound contrast imaging data, as provided in Example 2.

[0041] Figure Labels

[0042] 10 - Labeling module; 11 - Acquisition unit; 12 - Filtering unit; 13 - Correction unit; 20 - Generation module; 30 - Calculation module; 40 - Clustering module; 50 - Mapping module. Detailed Implementation

[0043] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0044] Various embodiments of the invention will be described more fully below. The invention may have various embodiments, and adjustments and changes may be made therein. However, it should be understood that there is no intention to limit the various embodiments of the invention to the specific embodiments disclosed herein, but rather the invention should be understood to cover all modifications, equivalents, and / or alternatives falling within the spirit and scope of the various embodiments of the invention.

[0045] In the following, the terms “comprising” or “may include” as used in various embodiments of the invention indicate the presence of the disclosed functions, operations, or elements, and do not limit the addition of one or more functions, operations, or elements. Furthermore, as used in various embodiments of the invention, the terms “comprising,” “having,” and their cognates are intended only to indicate a specific feature, number, step, operation, element, component, or combination of the foregoing, and should not be construed as primarily excluding the presence of one or more other features, numbers, steps, operations, elements, components, or combinations of the foregoing, or the possibility of adding one or more combinations of the foregoing.

[0046] In various embodiments of the invention, the expression "or" or "at least one of A and / or B" includes any combination or all combinations of the words listed simultaneously. For example, the expression "A or B" or "at least one of A and / or B" may include A, may include B, or may include both A and B.

[0047] The expressions used in the various embodiments of the present invention (such as "first," "second," etc.) may modify various constituent elements in the various embodiments, but do not limit the corresponding constituent elements. For example, the above expressions do not limit the order and / or importance of the elements. The above expressions are only used for the purpose of distinguishing one element from other elements. For example, a first user device and a second user device refer to different user devices, although both are user devices. For example, a first element may be referred to as a second element without departing from the scope of the various embodiments of the present invention, and similarly, a second element may also be referred to as a first element.

[0048] It should be noted that, in this invention, unless otherwise explicitly specified and defined, terms such as "installation," "connection," and "fixation" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0049] In this invention, those skilled in the art should understand that the terms indicating orientation or positional relationship in the text are based on the orientation or positional relationship shown in the accompanying drawings, and are only for the purpose of facilitating the description of this invention and simplifying the description, and are not intended to indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of this invention.

[0050] The terminology used in the various embodiments of the invention is for the purpose of describing particular embodiments only and is not intended to limit the various embodiments of the invention. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Unless otherwise defined, all terms used herein (including technical and scientific terms) have the same meaning as commonly understood by one of ordinary skill in the art to which the various embodiments of the invention pertain. The terms (such as those defined in a generally used dictionary) are to be interpreted as having the same meaning as in the context of the relevant technical field and are not to be interpreted as having an idealized or overly formal meaning, unless clearly defined in the various embodiments of the invention.

[0051] Example 1

[0052] This embodiment proposes a method for delineating tumor perfusion subregions based on dynamic ultrasound contrast imaging data. This method can effectively improve the diagnostic efficacy of benign and malignant breast tumors and can be applied to the diagnosis and treatment evaluation of other hypervascular or hypovascular benign and malignant tumors. See [link to relevant documentation]. Figure 1 The method includes:

[0053] S100: Obtain and pixelate a preprocessed dynamic ultrasound contrast image to determine a first region representing tumor tissue and a second region representing normal tissue in the dynamic ultrasound contrast image.

[0054] In this embodiment, regions of interest are marked in the dynamic ultrasound contrast-enhanced images by manual labeling. Specifically, the marked regions of interest may include a first region representing tumor tissue and a second region representing normal tissue.

[0055] S200: Obtain the time intensity curve of each pixel in the first interval, and obtain the average time intensity curve in the second interval.

[0056] See Figure 2 Specifically, step S200 uses square pixel blocks as the unit to divide the interval into multiple pixel units, and calculates the signal intensity of each pixel unit to obtain the time-intensity curve of each interval based on the pixel level, thereby constructing a time-intensity curve matrix of pixel points in each interval whose signal intensity is related to the time change.

[0057] S300: Based on the perfusion feature parameters of each pixel in the first interval and the average perfusion feature parameters of all pixels in the second interval, the ratio of perfusion feature parameters of each pixel in the first interval is obtained.

[0058] Specifically, the perfusion feature parameters of each pixel in the first interval are obtained from the time intensity curve of each pixel in the first interval, and the average perfusion feature parameters of all pixels in the second interval are obtained from the average time intensity curve of all pixels in the second interval.

[0059] In this embodiment, the perfusion characteristic parameters may specifically include parameters such as PT (time to peak), PI (peak intensity), AS (ascending slope), DS (descending slope), HDT (half time of descent) and / or AUC (area under curve). The time-intensity curve can be displayed in the form of a time-intensity curve matrix.

[0060] Specifically, in step S200, after obtaining the average injection feature parameter of all pixels in the second interval, the injection feature parameter of each pixel in the first interval is divided by the average injection feature parameter to obtain the injection feature parameter ratio of each pixel in the first interval, such as R-PT, R-PI, R-AS, ​​etc. In this embodiment, the median or average value of the injection feature parameters of all pixels in the second interval is used as the average injection feature parameter.

[0061] S400: Cluster the pixels in the first interval based on the ratio of the infusion feature parameters of each pixel in the first interval.

[0062] It should be noted that in step S400, the calculated perfusion feature parameters are used to perform cluster analysis on the pixels, and pixels with similar perfusion features are classified into the same category. Specifically, step S400 uses clustering methods such as principal component analysis (PCA) or K-means to cluster into multiple categories based on the ratio R of tumor tissue perfusion parameters. In this embodiment, the clustering results include at least two different sets of pixels, and at least one set of pixels corresponds to the tumor type in the dynamic ultrasound contrast imaging image.

[0063] In this process, PCA can help reduce the dimensionality of the data and extract the principal components that best represent the original data information, thereby reducing the complexity of the data. The K-means clustering algorithm can divide the samples into K clusters based on their similarity. By inputting the ratio of tumor tissue perfusion parameters into PCA for dimensionality reduction, and then using the K-means algorithm to divide the dimensionality-reduced data into multiple classes, multiple tumor perfusion patterns with similar characteristics can be obtained.

[0064] Different types of tumors may have different growth patterns, metabolic characteristics, or responses to treatment. By classifying tumors according to their blood perfusion, step S400 can help doctors better understand the blood flow inside the tumor, providing more information and reference for clinical diagnosis and treatment. Medical personnel can adopt targeted treatment strategies for different types of tumors to improve treatment outcomes and develop more individualized treatment plans.

[0065] In other embodiments, step S400 may also employ clustering methods such as mean-shift clustering, spectral clustering, DBSCAN, hierarchical clustering, or machine learning.

[0066] S500: Generates the perfusion subregion of the tumor based on the clustering results, and maps the information of the perfusion subregion back to the original grayscale image of the ultrasound contrast imaging.

[0067] See Figure 3 Figures 4(a)-(f) Figure 5 It should be noted that each pixel after clustering in step S400 has its own clustering result. The perfusion sub-region information generated based on the clustering result is mapped back to the original grayscale image of the ultrasound contrast imaging image, that is, mapped back to the original tumor. The number and / or percentage of pixels in each perfusion sub-region of the tumor in each patient are calculated. This can realize the quantitative description of the tissue characteristics and distribution of different regions of the tumor in the original grayscale image in a visual and / or quantitative manner.

[0068] In the field of medical imaging, perfusion subregions refer to the division of blood perfusion within a tumor or tissue. By processing and analyzing dynamic ultrasound contrast images using the method proposed in this embodiment, the tumor can be divided into different regions, which are called perfusion subregions. Each perfusion subregion has specific perfusion characteristics. The perfusion characteristics of each perfusion subregion can help doctors to have a more comprehensive understanding of the tumor dynamics and to make more accurate assessments and diagnoses of lesions.

[0069] Specifically, perfusion subregion analysis plays a crucial role in clinical medicine, providing physicians with more imaging information to help them make more accurate diagnostic and treatment decisions. By dividing perfusion subregions, physicians can better understand the blood flow conditions in different areas, such as blood supply, velocity, and density. Therefore, the method proposed in this embodiment is of great significance for tumor diagnosis, staging, and treatment planning.

[0070] Specifically, preprocessing of dynamic ultrasound contrast images can include, but is not limited to, operations such as denoising, enhancement, and contrast adjustment. Preprocessing of dynamic ultrasound contrast images can remove image noise, compensate for motion displacement, and achieve image registration, making the structures in the images clearer and more realistic, improving image correlation and structural similarity, and facilitating observation, comparison, and analysis. In this embodiment, the preprocessing of dynamic ultrasound contrast images includes motion correction. Through motion correction, the correlation and structural similarity of each frame of the dynamic ultrasound contrast image video are improved, achieving basic registration of each frame.

[0071] Specifically, see Figure 6 The specific steps for acquiring dynamic ultrasound contrast images and preprocessing them are as follows:

[0072] S110: Acquire ultrasound contrast imaging video including grayscale images and contrast images.

[0073] S120: Calculate the similarity between a frame of the ultrasound contrast video and other frames, select the frame with the highest similarity to other frames in the nodule region, and use it as the key frame.

[0074] S130: Crops grayscale images and contrast images of the same size and position, performs motion correction on the grayscale images and contrast images based on keyframes, and obtains a set of dynamic ultrasound contrast images.

[0075] Specifically, step S130 uses a motion correction algorithm to perform motion correction on the grayscale image and the angiographic image. The motion correction algorithm may include, but is not limited to, motion compensation to remove motion or motion correction based on deep learning.

[0076] Furthermore, in step S200, after obtaining the time-intensity curves of each tumor tissue and normal tissue at the pixel level, an interpolation method can be applied to compensate for the feature values ​​of some poorly correlated time point images that were deleted due to preprocessing. Thus, based on the ultrasound contrast-enhanced image with features supplemented by interpolation, the time-intensity curves of each pixel in the first interval and the average time-intensity curve in the second interval are obtained.

[0077] Preferably, the method proposed in this embodiment can also resample the image as needed to improve the computational efficiency of image processing while standardizing the image.

[0078] Example 2

[0079] This embodiment proposes a system for segmenting tumor perfusion subregions based on dynamic ultrasound contrast imaging data, used to implement the method for segmenting tumor perfusion subregions based on dynamic ultrasound contrast imaging data proposed in Embodiment 1. (See also...) Figure 7 The system includes:

[0080] The labeling module 10 is used to obtain and pixelate a preprocessed dynamic ultrasound contrast image to determine a first region representing tumor tissue and a second region representing normal tissue in the dynamic ultrasound contrast image.

[0081] The acquisition module 20 is used to acquire the time intensity curve of each pixel in the first interval and the average time intensity curve in the second interval.

[0082] The calculation module 30 is used to obtain the ratio of the perfusion feature parameters of each pixel in the first interval based on the perfusion feature parameters of each pixel in the first interval and the average perfusion feature parameters of all pixels in the second interval.

[0083] Clustering module 40 is used to cluster pixels in the first interval based on the ratio of the infusion feature parameters of each pixel in the first interval;

[0084] The mapping module 50 is used to generate the perfusion subregion of the tumor based on the clustering results and map the information of the perfusion subregion back to the original grayscale image of the ultrasound contrast imaging image.

[0085] In this embodiment, the marking module 10 provides a marking interface for the user, allowing the user to manually mark regions in the dynamic ultrasound contrast-enhanced image. Specifically, the marked regions may include a first region representing tumor tissue and a second region representing normal tissue. Specifically, the generation module 20 uses preset square pixel blocks as units to divide the user-marked regions into multiple pixel units.

[0086] Specifically, the perfusion feature parameters of each pixel in the first interval are obtained from the time intensity curve of each pixel in the first interval, and the average perfusion feature parameters of all pixels in the second interval are obtained from the average time intensity curve of all pixels in the second interval.

[0087] In this embodiment, the perfusion characteristic parameters may specifically include parameters such as PT (time to peak), PI (peak intensity), AS (ascending slope), DS (descending slope), HDT (half time of descent) and / or AUC (area under curve). The time-intensity curve can be displayed in the form of a time-intensity curve matrix.

[0088] Specifically, the calculation module 30 obtains the average perfusion feature parameter of all pixels in the second interval, divides the perfusion feature parameter of each pixel in the first interval by the average perfusion feature parameter to obtain the perfusion feature parameter ratio of each pixel in the first interval, such as R-PT, R-PI, R-AS, ​​etc. In this embodiment, the median or average value of the perfusion feature parameter of all pixels in the second interval is used as the average perfusion feature parameter.

[0089] The clustering module 40 uses the calculated perfusion feature parameters to perform cluster analysis on the pixels, classifying pixels with similar perfusion features into the same category. Specifically, the clustering module 40 uses clustering methods such as principal component analysis (PCA) or K-means to cluster into multiple categories based on the ratio of tumor tissue perfusion parameters. In this embodiment, the clustering results include at least two different sets of pixels, and at least one set of pixels corresponds to the tumor type in the dynamic ultrasound contrast imaging image.

[0090] After clustering by the clustering module 40, each pixel has its own clustering result. The mapping module 50 maps the perfusion subregion information generated based on the clustering results back to the original grayscale image of the ultrasound contrast imaging image, that is, back to the original tumor, and calculates the number and / or percentage of pixels in each perfusion subregion of the tumor for each patient. This can realize the quantitative description of the tissue characteristics of different regions of the tumor in the original grayscale image in a visual and / or quantitative manner.

[0091] Specifically, the marking module 10 includes:

[0092] Acquisition unit 11 is used to acquire ultrasound contrast imaging video including grayscale images and contrast images;

[0093] The filtering unit 12 is used to calculate the similarity between a frame of the ultrasound contrast video and other frames, filter out the frame with the highest similarity to other frames in the nodule region, and use it as the key frame.

[0094] The correction unit 13 is used to crop grayscale images and contrast images of the same size and position. After motion correction of the grayscale images and contrast images based on key frames, a set of dynamic ultrasound contrast images is obtained.

[0095] Specifically, the correction unit 13 uses a motion correction algorithm to perform motion correction on the grayscale image and the angiographic image. The motion correction algorithm may include, but is not limited to, motion compensation to remove motion or deep learning-based motion correction.

[0096] Furthermore, after obtaining the time-intensity curves of each tumor tissue and normal tissue at the pixel level, the generation module 20 can apply interpolation to compensate for the feature values ​​of some poorly correlated time point images that were deleted due to preprocessing. Thus, based on the ultrasound contrast-enhanced image with features supplemented by interpolation, the generation module 20 obtains the time-intensity curves of each pixel in the first interval and the average time-intensity curve in the second interval.

[0097] Preferably, the system proposed in this embodiment can also resample the image as needed to improve the computational efficiency of image processing while standardizing the image.

[0098] Example 3

[0099] This embodiment proposes a computer-readable storage medium storing executable instructions that, when executed by a processor, implement the method for dividing tumor perfusion subregions based on dynamic ultrasound contrast imaging data as proposed in Embodiment 1.

[0100] In summary, this invention provides a method and system for segmenting tumor perfusion subregions based on dynamic ultrasound contrast imaging data. This invention maximizes the use of dynamic image data through a non-invasive diagnostic method, comprehensively capturing the perfusion characteristics of tumors at the pixel level, visualizing and quantifying tumor perfusion heterogeneity, improving the readability of dynamic contrast imaging videos, and helping to explore the correlation between tumor subregions and benign or malignant tumors and construct diagnostic models accordingly. While improving the diagnostic accuracy of breast cancer, it also provides more information support for personalized treatment of patients.

[0101] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for delineating tumor perfusion subregions based on dynamic ultrasound contrast imaging data, characterized in that, The method includes: A preprocessed dynamic ultrasound contrast image is obtained and pixelated to determine a first region representing tumor tissue and a second region representing normal tissue in the dynamic ultrasound contrast image; the preprocessing is used to remove image noise, compensate for motion displacement, and achieve image registration; Obtain the temporal intensity curve of each pixel in the first interval, and obtain the average temporal intensity curve of all pixels in the second interval; Based on the perfusion feature parameters of each pixel in the first interval and the average perfusion feature parameters of all pixels in the second interval, the ratio of perfusion feature parameters of each pixel in the first interval is obtained; the perfusion feature parameters of each pixel in the first interval are obtained from the time intensity curve of each pixel in the first interval, and the average perfusion feature parameters of all pixels in the second interval are obtained from the average time intensity curve of all pixels in the second interval. Based on the ratio of the perfusion feature parameters of each pixel in the first interval, the pixels in the first interval are clustered. Based on the clustering results, perfusion subregions of the tumor are generated, and the information of the perfusion subregions is mapped back to the original grayscale image of the ultrasound contrast imaging.

2. The method for dividing tumor perfusion subregions based on dynamic ultrasound contrast imaging data according to claim 1, characterized in that, The preprocessing includes noise removal and / or motion correction.

3. The method for dividing tumor perfusion subregions based on dynamic ultrasound contrast imaging data according to claim 1 or 2, characterized in that, The step of obtaining the time intensity curve of each pixel in the first interval and obtaining the average time intensity curve in the second interval includes: The feature values ​​of the dynamic ultrasound contrast image deleted by the preprocessing are supplemented by interpolation. Based on the ultrasound contrast image with supplemented features by interpolation, the time intensity curve of each pixel in the first interval and the average time intensity curve in the second interval are obtained.

4. The method for dividing tumor perfusion subregions based on dynamic ultrasound contrast imaging data according to claim 1, characterized in that, The perfusion characteristic parameters include any one or more of the following: time to peak, peak intensity, rise slope, half-fall slope, time to half peak, and area under the curve.

5. The method for dividing tumor perfusion subregions based on dynamic ultrasound contrast imaging data according to claim 1 or 4, characterized in that, The step of obtaining the ratio of the perfusion feature parameters of each pixel in the first interval based on the perfusion feature parameters of each pixel in the first interval and the average perfusion feature parameters of all pixels in the second interval includes: The median or average value of the perfusion feature parameters of all pixels in the second interval is taken as the average perfusion feature parameter; Based on the perfusion feature parameters of each pixel in the first interval, the perfusion feature parameters of each pixel in the first interval are obtained, and the perfusion feature parameters of each pixel in the first interval are divided by the average perfusion feature parameter to obtain the ratio of the perfusion feature parameters of each pixel in the first interval.

6. The method for dividing tumor perfusion subregions based on dynamic ultrasound contrast imaging data according to claim 1, characterized in that, The clustering of pixels in the first interval is achieved through principal component analysis, K-means clustering, mean shift, spectral clustering, DBSCAN, hierarchical clustering, or machine learning clustering methods.

7. The method for dividing tumor perfusion subregions based on dynamic ultrasound contrast imaging data according to claim 1 or 6, characterized in that, The clustering results include at least two different sets of pixels.

8. The method for dividing tumor perfusion subregions based on dynamic ultrasound contrast imaging data according to claim 1, characterized in that, The step of generating tumor perfusion subregions based on clustering results and mapping the information of the perfusion subregions back to the original grayscale image of the ultrasound contrast imaging includes: The perfusion sub-region information generated based on the clustering results is mapped back to the original grayscale image of the ultrasound contrast image, and the number and / or percentage of pixels in each perfusion sub-region are calculated to describe the tissue characteristics and distribution of different regions of the tumor in the original grayscale image in a visual and / or quantitative manner.

9. A system for delineating tumor perfusion subregions based on dynamic ultrasound contrast imaging data, characterized in that, The system includes: A labeling module is used to obtain and pixelate a preprocessed dynamic ultrasound contrast image, and to determine a first region representing tumor tissue and a second region representing normal tissue in the dynamic ultrasound contrast image; the preprocessing is used to remove image noise, compensate for motion displacement, and achieve image registration. The acquisition module is used to acquire the time intensity curve of each pixel in the first interval and to acquire the average time intensity curve in the second interval. The calculation module is used to obtain the ratio of the perfusion feature parameters of each pixel in the first interval based on the perfusion feature parameters of each pixel in the first interval and the average perfusion feature parameters of all pixels in the second interval; the perfusion feature parameters of each pixel in the first interval are obtained from the time intensity curve of each pixel in the first interval, and the average perfusion feature parameters of all pixels in the second interval are obtained from the average time intensity curve of all pixels in the second interval. The clustering module is used to cluster the pixels in the first interval based on the ratio of the infusion feature parameters of each pixel in the first interval; The mapping module is used to generate perfusion subregions of the tumor based on the clustering results, and to map the information of the perfusion subregions back to the original grayscale image of the ultrasound contrast image.

10. A computer-readable storage medium, characterized in that, It stores executable instructions for implementing the method as described in any one of claims 1-8 when executed by a processor.

Images