Method and apparatus for tumor heterogeneity analysis based on super-resolution microscopic imaging

Abstract

The application discloses a tumor heterogeneity analysis method and device based on super-resolution microscopic imaging, and relates to the field of medical image analysis. The method comprises the following steps: acquiring a tumor imaging result, wherein the tumor imaging result is generated by a super-resolution microscopic imaging technology, and the tumor imaging result comprises a tumor image; determining a target region of the tumor image; determining a quantitative analysis parameter based on the tumor imaging result, wherein the quantitative analysis parameter is used for quantifying microvessel features of at least two dimensions of tumor microvessels in the target region; and performing tumor heterogeneity analysis on different target regions based on the quantitative analysis parameter to obtain a heterogeneity analysis result, wherein the different target regions belong to the same tumor or the different target regions belong to different tumors. By using the scheme provided in the application, the quantitative analysis parameter is generated based on the tumor imaging result, quantitative analysis of the tumor microvessel features is realized, and therefore, the accuracy of tumor heterogeneity analysis is improved.

Description

Technical Field

[0001] This invention relates to the field of medical image analysis technology, and in particular to a method and device for tumor heterogeneity analysis based on super-resolution microscopy. Background Technology

[0002] Tumor heterogeneity refers to the spatial and temporal heterogeneity of tumors in terms of molecular biology, gene expression, and microenvironment characteristics, which directly affects tumor growth, metastasis, drug sensitivity, and prognosis.

[0003] Among related technologies, super-resolution microscopy can be used to image the morphology and hemodynamics of microvessels in tissues by tracking the movement of contrast agent microbubbles moving with blood flow. The imaging process of this technology can be divided into four steps: separation and extraction of moving microbubble signals, extraction of sub-pixel localization of microbubble signals, localization and tracking of moving microbubbles, and cumulative mapping imaging of the tracking trajectory.

[0004] However, super-resolution microscopy lacks quantitative methods to quantify microblood flow characteristics, resulting in a single analytical dimension and difficulty in comprehensively and accurately assessing tumor biological characteristics, leading to insufficient precision in heterogeneity assessment. Summary of the Invention

[0005] The purpose of this invention is to provide a method and device for tumor heterogeneity analysis based on super-resolution microscopy, so as to solve the problems of insufficient comprehensiveness and accuracy in tumor heterogeneity analysis.

[0006] To achieve the above-mentioned objectives, in one respect, this application provides a method for tumor heterogeneity analysis based on super-resolution microscopy, the method comprising: acquiring tumor imaging results, wherein the tumor imaging results are generated by super-resolution microscopy and the tumor imaging results include tumor images; Identify the target region of the tumor image; based on the tumor imaging results, determine quantitative analysis parameters, which are used to quantify the microvascular features of the tumor microvessels in at least two dimensions in the target region; based on the quantitative analysis parameters, perform tumor heterogeneity analysis on different target regions to obtain heterogeneity analysis results, wherein different target regions belong to the same tumor, or different target regions belong to different tumors.

[0007] As a further improvement of this application, the tumor imaging results include a density map and a velocity map, wherein the density map is used to reflect the distribution of tumor microvessels and the velocity map is used to reflect the blood flow velocity in the tumor microvessels; the method further includes: determining at least two dimensions of the quantitative analysis parameters based on the density map and the velocity map; wherein the quantitative analysis parameters of different dimensions are used to characterize microvascular features of different dimensions.

[0008] As a further improvement of this application, the quantitative analysis parameters are determined based on the density map in at least two dimensions, including: determining microvessel density distribution parameters, microvessel geometric state parameters, and microvessel perfusion parameters based on the density map; wherein, the microvessel density distribution parameters are used to quantify the microvessel density and distribution characteristics, the microvessel geometric state parameters are used to characterize the microvessel morphology and spatial structure characteristics, and the microvessel perfusion parameters are used to characterize the dynamic characteristics of blood flow perfusion within the microvessels.

[0009] As a further improvement of this application, the quantitative analysis parameters are determined based on the velocity map in at least two dimensions, including: determining blood flow velocity parameters, blood flow direction parameters, and hemodynamic parameters based on the velocity map, wherein the hemodynamic parameters are used at least to characterize the pulsation characteristics of microvessels.

[0010] As a further improvement of this application, the step of performing tumor heterogeneity analysis on different target regions based on the quantitative analysis parameters includes: performing histogram analysis on the target regions based on the quantitative analysis parameters to obtain a first heterogeneity analysis result; or, performing principal component analysis on the target regions based on the quantitative analysis parameters to obtain a second heterogeneity analysis result.

[0011] As a further improvement of this application, the step of performing tumor heterogeneity analysis on different target regions based on the quantitative analysis parameters includes: inputting the first heterogeneity analysis results and the second heterogeneity analysis results corresponding to different target regions into a heterogeneity analysis model; performing tumor heterogeneity analysis through the heterogeneity analysis model to obtain a third heterogeneity analysis result for different target regions.

[0012] As a further improvement of this application, determining the target region of the tumor image includes: acquiring a grayscale tissue image containing the tumor site; performing tumor edge recognition based on the grayscale tissue image to obtain a tumor edge recognition result; and determining the target region based on the tumor edge recognition result, wherein the target region contains at least a portion of the tumor.

[0013] As a further improvement of this application, when different target regions belong to the same tumor, determining the target region based on the tumor edge recognition result includes: determining the geometric center of the target region based on the tumor edge recognition result; generating concentric ring-shaped sub-regions by morphological erosion operation with the geometric center as the origin, wherein each concentric ring-shaped sub-region is a target region; or, determining the tumor circumscribed rectangle based on the tumor edge recognition result; and dividing the tumor circumscribed rectangle equally along the horizontal and axial directions to obtain equally divided grid regions, each equally divided grid region being a target region.

[0014] As a further improvement of this application, the method further includes: extending the target region to the tumor periphery region by morphological dilation operation or by expanding outward based on equal distance; and performing concentric ring or grid-like subdivision on the tumor periphery region to obtain a target region containing the tumor periphery region.

[0015] As a further improvement of this application, the heterogeneity analysis results include at least one of the following forms: pseudo-color mapping images, used to display differences in vascular distribution density or blood flow velocity in each sub-region; three-dimensional heatmaps, superimposed on a grayscale tumor tissue image, and marking high heterogeneity regions; statistical charts, including bar charts or line charts, used to compare the quantitative results of microvascular features in different sub-regions.

[0016] As a further improvement of this application, the method of obtaining tumor imaging results includes: tracking the movement trajectory of contrast agent microbubbles in the tumor microvessels, separating and extracting high-frequency moving microbubble signals; performing sub-pixel localization on the high-frequency moving microbubble signals, and tracking the movement of microbubbles based on the localization results to obtain a tracking trajectory; accumulating the information such as the density and velocity of the moving microbubbles reflected by the tracking trajectory, and mapping the accumulated results to the corresponding spatial location to obtain the tumor imaging results.

[0017] On the other hand, this application provides an electronic device, including: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the tumor heterogeneity analysis method based on super-resolution microscopy as described in any of the above aspects.

[0018] On the other hand, a non-transitory computer-readable storage medium storing computer instructions is provided, wherein the computer instructions are used to cause the computer to perform the tumor heterogeneity analysis method based on super-resolution microscopy as described in any of the above aspects.

[0019] On the other hand, a computer program product is provided, including a computer program that, when executed by a processor, implements the tumor heterogeneity analysis method based on super-resolution microscopy as described in any of the above aspects.

[0020] Compared with related technologies, the solution provided in this application has the following beneficial effects.

[0021] In this embodiment, based on super-resolution microscopy, tumor imaging results with micron-level resolution are obtained, and multi-dimensional quantitative analysis parameters are generated through density maps and velocity maps to achieve high-precision quantitative analysis of tumor microvascular characteristics, effectively improving the accuracy, comprehensiveness, safety, and efficiency of tumor heterogeneity analysis. Attached Figure Description

[0022] Figure 1 A flowchart illustrating a tumor heterogeneity analysis method based on super-resolution microscopy according to an illustrative embodiment of this application is shown. Figure 2 This illustration shows a schematic diagram of obtaining tumor imaging results according to an illustrative embodiment of this application; Figure 3 A schematic diagram of a density map and a velocity map provided in an illustrative embodiment of this application is shown; Figure 4 This illustration shows a schematic diagram of a density map-based method for determining quantitative analysis parameters in at least two dimensions, according to an illustrative embodiment of this application. Figure 5 This illustration shows a schematic diagram of determining quantitative analysis parameters based on a velocity map, according to an illustrative embodiment of this application. Figure 6 A schematic diagram of quantitative analysis parameters provided in an illustrative embodiment of this application is shown; Figure 7 This illustration shows a schematic diagram of the process for determining different tumor heterogeneities according to an illustrative embodiment of this application; Figure 8 This illustration shows a schematic diagram of the process for determining tumor heterogeneity in different target regions within the same tumor, provided by an illustrative embodiment of this application. Figure 9 This illustration shows a comparison diagram of target regions obtained by concentric ring subdivision and grid subdivision according to an illustrative embodiment of this application; Figure 10 A schematic diagram of an extended target area provided by an illustrative embodiment of this application is shown.

[0023] Figure 11 This illustration shows a structural block diagram of an electronic device provided in an illustrative embodiment of the present application. Detailed Implementation

[0024] The present invention will now be described in detail with reference to the specific embodiments shown in the accompanying drawings. However, these embodiments do not limit the present invention, and any structural, methodological, or functional modifications made by those skilled in the art based on these embodiments are included within the scope of protection of the present invention.

[0025] It should be noted that the term "comprising" or any other variation thereof is intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Furthermore, the terms "first," "second," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0026] Please refer to Figure 1 The illustration shows a flowchart of a tumor heterogeneity analysis method based on super-resolution microscopy provided in an illustrative embodiment of this application, which includes the following steps.

[0027] Step 101: Obtain tumor imaging results.

[0028] The tumor imaging results are generated using super-resolution microscopy, and the tumor imaging results include tumor images.

[0029] Super-resolution microscopy is an imaging method that breaks through the traditional ultrasound diffraction limit, achieving resolution down to the micrometer level and providing spatial resolution for tumor heterogeneity analysis. By tracking moving microbubbles within microvessels, super-resolution microscopy precisely images the micrometer-scale characteristics of microvessels in tumor tissue.

[0030] This embodiment employs a tumor heterogeneity analysis method based on super-resolution microscopy. By acquiring multi-dimensional feature data of tumor microvessels through ultrasound resolution microscopy, and combining it with quantitative algorithms to quantify the heterogeneity differences in different regions, it can solve the problems of feature omission and quantification deviation caused by insufficient resolution of traditional pathology or ordinary microscopy techniques.

[0031] Step 102: Determine the target region of the tumor image.

[0032] Step 103: Determine the quantitative analysis parameters based on the tumor imaging results.

[0033] Among them, the quantitative analysis parameters are used to quantify the microvascular characteristics of tumor microvessels in the target region in at least two dimensions.

[0034] Step 104: Based on the quantitative analysis parameters, tumor heterogeneity analysis is performed on different target regions to obtain the heterogeneity analysis results.

[0035] In this context, different target regions may belong to the same tumor, or different target regions may belong to different tumors.

[0036] In summary, in this embodiment of the application, tumor imaging results with micron-level resolution are obtained based on super-resolution microscopic imaging technology, and multi-dimensional quantitative analysis parameters are generated through density maps and velocity maps to achieve high-precision quantitative analysis of tumor microvascular characteristics, effectively improving the accuracy, comprehensiveness, safety and efficiency of tumor heterogeneity analysis.

[0037] The process of obtaining tumor imaging results based on super-resolution microscopy is explained below.

[0038] Please refer to Figure 2 This illustration shows a schematic diagram of obtaining tumor imaging results provided by an illustrative embodiment of this application.

[0039] After contrast agent microbubbles are injected into the patient's vein via bolus or drip, the microbubbles rapidly enter the tumor microvessels through blood circulation.

[0040] Since contrast agent microbubbles cannot penetrate the blood vessel wall and only move with the blood inside the blood vessel, high-precision imaging results reflecting the microvascular structure and blood flow dynamics inside the tumor can be obtained by accurately tracking and imaging the movement trajectory of the microbubbles.

[0041] Specifically, high-frequency moving microbubble signals are separated and extracted from ultrasound echo signals. Subpixel localization is performed on these high-frequency moving microbubble signals, and the movement of the microbubbles is tracked based on the localization results to obtain tracking trajectories. Information such as the density and velocity of the moving microbubbles reflected in the tracking trajectories is accumulated, and the accumulated results are mapped to corresponding spatial locations to obtain tumor imaging results.

[0042] Indicative, such as Figure 2 This includes the following four steps.

[0043] Step 1: Using spatiotemporal filtering techniques such as singular value decomposition filtering and frequency domain filtering, high-frequency moving microbubble signals with significant fluctuations are effectively separated and extracted from the tumor echo signals. Here, high-frequency moving microbubble signals refer to ultrasound signals with significant fluctuations in the time dimension due to the flow of contrast agent microbubbles.

[0044] Step 2: Perform subpixel localization on the extracted high-frequency moving microbubble signals. Subpixel localization refers to using centroid weighted averaging and gradient feature optimization methods to break through the diffraction limit of traditional ultrasound imaging and accurately determine the center position of a single microbubble signal.

[0045] Step 3: Based on the localization results, microbubble tracking algorithms such as Kalman filtering with added motion model constraints are used to accurately track the motion trajectory of each microbubble within the tumor microvessels to obtain its speed and direction information.

[0046] Step 4: Accumulate the information such as microbubble density and velocity reflected by the tracking trajectory, and accurately map the accumulated results to the corresponding spatial location to form tumor imaging results with micron-level resolution.

[0047] Based on the tumor imaging results obtained above, density maps and velocity maps can be further generated. The density map is used to reflect the distribution of tumor microvessels and is obtained by statistically analyzing the number of contrast agent microbubbles passing through each pixel position during the imaging process. The velocity map is used to reflect the blood flow velocity in the tumor microvessels and is obtained by statistically calculating the velocity of contrast agent microbubbles passing through each pixel position during the imaging process.

[0048] Density maps and velocity maps display the structural features and dynamic blood flow characteristics of tumor microvessels from different dimensions, thus providing an accurate and reliable data foundation for the next step of target area determination and the construction of multi-dimensional quantitative analysis parameters.

[0049] Quantitative analysis parameters in at least two dimensions are determined based on density and velocity maps. These parameters, representing different dimensions of microvascular characteristics, are used to characterize these different dimensions.

[0050] Please refer to Figure 3 This illustration shows a schematic diagram of a density map and a velocity map provided in an exemplary embodiment of this application. The density map is determined based on the microvascular morphology in tumor imaging results, with pixel amplitude representing microbubble density. The velocity map is determined based on the microblood flow velocity in tumor imaging results, with pixel amplitude representing the movement velocity of microbubbles within the blood vessels.

[0051] Please refer to Figure 4 This illustration shows a schematic diagram of a density map-based method for determining quantitative analysis parameters in at least two dimensions, according to an illustrative embodiment of this application.

[0052] In one possible implementation, microvessel density distribution parameters, microvessel geometric state parameters, and microvessel perfusion parameters are determined based on density maps.

[0053] Optional, such as Figure 4 The microvascular density distribution parameters are determined based on the amplitude information of each pixel in the density map, the microvascular geometric distribution parameters are determined based on the pixel distribution on the density map, and the microvascular perfusion parameters are determined based on the time dimension change of the density map.

[0054] Among them, the microvessel density distribution parameters are used to quantify the density and distribution characteristics of microvessels, including but not limited to the maximum, minimum, average, standard deviation, entropy, and relative blood volume of the density distribution. The amplitude of a single pixel on the density map reflects the number of contrast agent microbubbles passing through the corresponding spatial location during the acquisition period. The above-mentioned density distribution characteristic parameters can be accurately calculated by statistically analyzing the pixel amplitudes.

[0055] Microvascular geometric state parameters are used to characterize the morphology and spatial structure of microvessels. After image segmentation, the pixel distribution on the density map is converted into a binary image through binarization technology to obtain the geometric distribution information of microvessels. By performing geometric and topological analysis on the binary image, the area, area ratio, microvascular complexity, tortuosity and other indicators of the geometric figure can be quantitatively calculated.

[0056] Microvascular perfusion parameters are used to characterize the dynamic features of blood flow perfusion in microvessels. By analyzing the changes in density maps over time, dynamic characteristic parameters such as microvascular perfusion rate can be calculated.

[0057] Please refer to Figure 5 This illustrates a schematic diagram of determining quantitative analysis parameters based on a velocity map, provided in an illustrative embodiment of this application.

[0058] In one possible implementation, blood flow velocity parameters, blood flow direction parameters, and hemodynamic parameters are determined based on the velocity map. The hemodynamic parameters are used to characterize features such as microvascular pulsation.

[0059] Among them, the blood flow velocity parameter is used to quantitatively characterize the magnitude and variation of blood flow velocity in microvessels, specifically including the maximum, minimum, average, standard deviation and entropy of the velocity distribution; the blood flow direction parameter is used to characterize the spatial distribution direction of blood flow in microvessels; and the hemodynamic parameter is used to characterize the pulsation characteristics of microvessels. The pulsation index is obtained by analyzing the trend of the velocity of each pixel on the velocity map with the time dimension, so as to reflect the hemodynamic characteristics of microvessels in a more in-depth way.

[0060] Based on the above, this invention also proposes a perfusion index calculated by combining density map and velocity map. The perfusion index is a parameter calculated by combining the product of microbubble density and average velocity, which can effectively reflect the perfusion status in microvessels and further improve the accuracy of tumor heterogeneity analysis and its clinical guidance significance.

[0061] Please refer to Figure 6 This illustration shows a schematic diagram of quantitative analysis parameters provided in an illustrative embodiment of this application. The microvessel density distribution parameters, microvessel geometric distribution parameters, and microvessel perfusion characteristic parameters are obtained based on super-resolution microscopic imaging density maps. The microvessel velocity distribution parameters, microvessel orientation distribution parameters, and microvessel dynamic parameters are obtained from super-resolution microscopic imaging velocity maps. Furthermore, the microvessel perfusion index is obtained based on both the density map and the velocity map.

[0062] By constructing the above-mentioned quantitative analysis parameter package, this implementation method comprehensively quantifies the characteristics of tumor microvessels from multiple dimensions such as microvessel density, spatial structure, blood flow velocity, direction and dynamic characteristics, and achieves high-precision, quantitative and multi-dimensional assessment of tumor heterogeneity.

[0063] Based on the above-mentioned multi-dimensional quantitative analysis parameter package, this application further performs tumor heterogeneity analysis on different target regions based on the quantitative analysis parameters to obtain heterogeneity analysis results.

[0064] In one possible implementation, tumor heterogeneity analysis can be achieved using methods such as histogram analysis, principal component analysis, and deep learning model analysis.

[0065] Optionally, histogram analysis can be performed based on quantitative analysis parameters to obtain the first heterogeneity analysis results.

[0066] Histogram analysis can visually display the distribution of various characteristic parameters of microvessels in tumor tissue, helping to identify key biological features and potential heterogeneous subpopulations.

[0067] Optionally, principal component analysis can be performed on the target region based on quantitative analysis parameters to obtain the results of a second heterogeneity analysis.

[0068] Principal component analysis (PCA) reduces the dimensionality of high-dimensional data by extracting the key features that contribute most to the assessment of heterogeneity, thereby effectively reducing data complexity and accurately reflecting the main variation directions of tumor heterogeneity.

[0069] Optionally, tumor heterogeneity analysis can be performed using a deep learning model. First, histogram analysis and principal component analysis are used to obtain the first and second heterogeneity analysis results. Then, the first and second heterogeneity analysis results corresponding to different target regions are input into the heterogeneity analysis model, and tumor heterogeneity analysis is performed through the model to obtain the third heterogeneity analysis results for different target regions.

[0070] Among these methods, the heterogeneity analysis model preferably employs data-driven approaches such as deep learning and machine learning, which can automatically mine complex patterns and correlation features from large amounts of parameter data, thereby improving the accuracy and reliability of heterogeneity analysis. Features extracted from histograms and principal component analysis are used as input to the deep learning model to further optimize the accuracy of tumor classification and prediction.

[0071] In one possible implementation, the heterogeneity analysis results obtained based on the above implementation methods also include pseudo-color mapping images, three-dimensional heat maps, and statistical charts.

[0072] Pseudocolor mapping images are used to visually display the differences in vascular distribution density or blood flow velocity in each sub-region, clearly presenting the heterogeneous spatial distribution characteristics within the tumor with different colors; three-dimensional heat maps are overlaid on the grayscale tissue image of the tumor and highly heterogeneous areas are marked, revealing key biologically active areas in the tumor in a spatially visualized way; statistical charts, including bar charts or line charts, are used to quantitatively compare the microvascular characteristics of different sub-regions, enabling doctors to quickly identify and assess the differences in pathological features in different regions.

[0073] In this embodiment of the application, based on a model combining histogram analysis, principal component analysis, and heterogeneity analysis, the present invention can accurately uncover the heterogeneity of tumor regions.

[0074] After obtaining tumor imaging results, it is necessary to further determine the target region of the tumor image in order to achieve subsequent accurate heterogeneity analysis.

[0075] Optionally, a grayscale tissue image containing the tumor site is first acquired. This grayscale image visually demonstrates the significant differences in acoustic impedance across different regions within the scanned tissue. Subsequently, tumor edge recognition is performed based on the grayscale tissue image to obtain the tumor edge recognition results. Finally, based on the tumor edge recognition results, the target region containing part or all of the tumor is determined.

[0076] Specifically, tumor edge recognition employs image difference maximization, edge detection algorithms, or deep learning-based automatic recognition methods to significantly improve the accuracy and stability of tumor boundary recognition.

[0077] In this embodiment of the application, the target area may belong to the same tumor or different tumors.

[0078] When different target regions belong to different tumors, the target region may contain all tumors. Therefore, determining the tumor heterogeneity of different target regions is equivalent to determining the heterogeneity of different tumors.

[0079] Please refer to Figure 7 This illustration shows a schematic diagram of the process for determining tumor heterogeneity according to an illustrative embodiment of this application. First, super-resolution microscopic imaging results of multiple tumors are acquired. Then, a target region is determined based on the tumor imaging results. This target region may include the entire tumor. Subsequently, quantitative analysis is performed on different tumors based on quantitative analysis parameters. Then, histogram analysis or principal component analysis is performed based on the quantitative analysis results corresponding to different tumors. Alternatively, tumor heterogeneity analysis can also be achieved based on deep learning. In this case, histogram analysis and principal component analysis are performed, and the first and second heterogeneity analysis results are input into the heterogeneity analysis model to obtain the tumor heterogeneity analysis results.

[0080] Please refer to Figure 8 This illustration shows a schematic diagram of the process for determining tumor heterogeneity in different target regions within the same tumor, provided by an illustrative embodiment of this application. First, a single tumor imaging result is acquired, and tumor edge identification results are determined from it. Then, based on the tumor edge identification results, the tumor region is segmented to obtain multiple target regions. Next, quantitative analysis is performed on the multiple target regions based on quantitative analysis parameters, and the quantitative analysis results of the multiple target regions are then subjected to histogram analysis, principal component analysis, or deep learning for tumor heterogeneity analysis to obtain the tumor heterogeneity analysis results.

[0081] In the process of segmenting the tumor region based on the edge recognition results to obtain the target region, concentric ring segmentation or grid segmentation can be used.

[0082] In the process of obtaining the target region based on concentric ring subdivision, the geometric center of the target region is first determined based on the tumor edge recognition results. Then, with the geometric center as the origin, concentric ring sub-regions are generated through morphological erosion operation.

[0083] When calculating the geometric center, the distances from each pixel in a certain region to the non-zero pixels on the edge are calculated, and the position with the smallest distance sum is defined as the geometric center.

[0084] In the process of morphological erosion, the edge formed in the image after each erosion is defined as a new concentric ring sub-region. Through the above stepwise erosion process, the target region is gradually refined until it is divided into multiple concentric ring sub-regions. Each concentric ring sub-region corresponds to a target region, thereby realizing the spatial heterogeneity analysis of the tumor region from the periphery to the center.

[0085] Optionally, in the process of obtaining the target region through grid subdivision, firstly, based on the tumor edge recognition results, the tumor's circumscribed rectangle is determined. Subsequently, the tumor's circumscribed rectangle is divided equally along the horizontal and axial directions to obtain equally divided grid regions, with each equally divided grid region being a target region.

[0086] The circumscribed rectangle is the smallest rectangle that tightly surrounds the tumor edge identification area.

[0087] Please refer to Figure 9 The diagram illustrates a comparison of target regions obtained through concentric ring subdivision and mesh subdivision according to an illustrative embodiment of this application. Different target regions obtained through concentric ring subdivision have the same geometric center, while different target regions obtained through mesh subdivision are equally divided mesh regions.

[0088] In another possible implementation, considering the relationship between tumor invasiveness and the microvascular characteristics of the surrounding area, morphological dilation calculations or an outward expansion based on equal-distance division are further employed to effectively extend the target region to the surrounding area of ​​the tumor. The surrounding area is then divided into concentric rings or grids to obtain the target region encompassing the surrounding area, thereby comprehensively revealing the heterogeneous characteristics of the tumor and its surrounding region, providing more complete information support for clinical tumor invasiveness assessment and prognostic prediction.

[0089] Please refer to Figure 10 This illustrates a schematic diagram of an extended target area provided in an illustrative embodiment of this application.

[0090] In this embodiment of the application, the above-mentioned method for determining the target region enables high-precision and automated segmentation of the tumor region, ensuring the efficiency, accuracy and stability of tumor heterogeneity analysis, and significantly improving the clinical reference value of the analysis results.

[0091] Please refer to Figure 11 This diagram illustrates a structural block diagram of an electronic device provided in an illustrative embodiment of this application. The electronic device in this application may include one or more components such as a processor 1110 and a memory 1120.

[0092] Optionally, the processor 1110 executes the steps in the tumor heterogeneity analysis method based on super-resolution microscopy provided in any of the above embodiments by running or executing instructions, programs, code sets or instruction sets stored in the memory 1120, and calling data stored in the memory 1120.

[0093] In addition, the processor can also perform various functions of the device and process data. Optionally, the processor 1110 can be implemented using at least one hardware form of Digital Signal Processing (DSP), Field-Programmable Gate Array (FPGA), or Programmable Logic Array (PLA). The processor 1110 can integrate one or a combination of several of the following: Central Processing Unit (CPU), Graphics Processing Unit (GPU), Neural-network Processing Unit (NPU), and modem. The CPU primarily handles the operating system, user interface, and applications; the GPU is responsible for rendering and drawing the content required to be displayed on the touch screen; the NPU is used to implement Artificial Intelligence (AI) functions; and the modem is used to handle wireless communication. It is understood that the aforementioned modem may also not be integrated into the processor 1110 and may be implemented as a separate chip.

[0094] The memory 1120 may include random access memory (RAM) or read-only memory (ROM). Optionally, the memory 1120 may include a non-transitory computer-readable storage medium. The memory 1120 may be used to store instructions, programs, code, code sets, or instruction sets. The memory 1120 may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for at least one function (such as touch function, sound playback function, image playback function, etc.), instructions for implementing the various method embodiments described below, etc.; the data storage area may store data created according to the use of the device (such as audio data, phone book, etc.).

[0095] The device in this embodiment further includes a communication component 1130 and a display component 1140. The communication component 1130 can be a Bluetooth component, a WiFi (Wireless Fidelity) component, an NFC (Near Field Communication) component, etc., used to communicate with external devices (servers or other devices) via wired or wireless networks; the display component 1140 is used to display a graphical user interface and / or receive user interaction operations.

[0096] In addition, those skilled in the art will understand that the structure of the device shown in the above figures does not constitute a limitation on the device. The device may include more or fewer components than shown, or combine certain components, or have different component arrangements. For example, the device may also include radio frequency circuits, input units, sensors, audio circuits, speakers, power supplies, etc., which will not be described in detail here.

[0097] This application also provides a non-transitory computer-readable storage medium storing computer instructions. The storage medium stores at least one piece of program code, which is loaded and executed by a processor to implement the tumor heterogeneity analysis method based on super-resolution microscopy as described in any of the above embodiments.

[0098] This application provides a computer program product including computer instructions stored in a computer-readable storage medium. When executed by a processor, the computer program implements the tumor heterogeneity analysis method based on super-resolution microscopy as described in any of the above embodiments.

Claims

1. A method for tumor heterogeneity analysis based on super-resolution microscopy, characterized in that, The method includes: Obtain tumor imaging results, which are generated by super-resolution microscopy and include tumor images. Determine the target region of the tumor image; wherein, determining the target region of the tumor image includes: Acquire a grayscale tissue image containing the tumor site; perform tumor edge recognition based on the grayscale tissue image to obtain tumor edge recognition results; determine the geometric center of the target region based on the tumor edge recognition results; The step of determining the geometric center of the target region based on the tumor edge recognition results includes: The distances from each pixel within the tumor region to the non-zero pixels at the edge are calculated, and the position with the smallest distance sum is defined as the geometric center. Using the geometric center as the origin, multiple concentric ring-shaped sub-regions are generated through morphological erosion. Each edge formed in the image after each erosion is defined as a new concentric ring-shaped sub-region. The target region is refined through a progressive erosion process until it is divided into multiple concentric ring-shaped sub-regions, and each concentric ring-shaped sub-region serves as a target region. Based on the tumor imaging results, quantitative analysis parameters are determined. These parameters are used to quantify the microvascular characteristics of the tumor microvessels in the target region in both temporal and spatial dimensions. The quantitative analysis parameters in the temporal dimension include microvascular blood flow velocity parameters and microvascular perfusion parameters. The microvascular perfusion parameters are obtained by analyzing the changes in the density map representing the accumulation of microbubble spatial distribution in the temporal dimension and calculating the microvascular perfusion rate. Based on the quantitative analysis parameters, tumor heterogeneity analysis is performed on different target regions to obtain heterogeneity analysis results. The different target regions belong to the multiple concentric ring sub-regions of the same tumor. The heterogeneity analysis includes comparing the differences in the quantitative analysis parameters between different concentric ring sub-regions to achieve spatial heterogeneity analysis of the tumor region from the periphery to the center.

2. The method according to claim 1, characterized in that, The tumor imaging results include a density map and a velocity map, wherein the density map is used to reflect the distribution of tumor microvessels and the velocity map is used to reflect the blood flow velocity in the tumor microvessels. The method further includes: The quantitative analysis parameters in at least two dimensions are determined based on the density map and the velocity map; The quantitative analysis parameters of different dimensions are used to characterize microvascular features of different dimensions.

3. The method according to claim 2, characterized in that, The quantitative analysis parameters are determined based on the density map in at least two dimensions, including: Based on the density map, the microvessel density distribution parameters, microvessel geometric state parameters, and microvessel perfusion parameters are determined. The microvessel density distribution parameter is used to quantify the microvessel density and distribution characteristics, the microvessel geometric state parameter is used to characterize the microvessel morphology and spatial structure characteristics, and the microvessel perfusion parameter is used to characterize the dynamic characteristics of blood flow perfusion within the microvessels.

4. The method according to claim 3, characterized in that, Based on the velocity map, at least two dimensions of the quantitative analysis parameters are determined, including: Based on the velocity map, blood flow velocity parameters, blood flow direction parameters, and hemodynamic parameters are determined, wherein the hemodynamic parameters are used at least to characterize the pulsation characteristics of microvessels.

5. The method according to claim 2, characterized in that, The process of performing tumor heterogeneity analysis on different target regions based on the quantitative analysis parameters includes: Based on the quantitative analysis parameters, histogram analysis is performed on the target region to obtain the first heterogeneity analysis results; or, Based on the quantitative analysis parameters, principal component analysis is performed on the target region to obtain the second heterogeneity analysis results.

6. The method according to claim 5, characterized in that, The process of performing tumor heterogeneity analysis on different target regions based on the quantitative analysis parameters includes: Input the first heterogeneity analysis results and the second heterogeneity analysis results corresponding to different target regions into the heterogeneity analysis model; Tumor heterogeneity analysis was performed using the aforementioned heterogeneity analysis model to obtain the third heterogeneity analysis results for different target regions.

7. The method according to claim 1, characterized in that, The method further includes: The target region can be expanded to the area surrounding the tumor by morphological expansion operations or by expanding outward based on equal intervals. The area surrounding the tumor is divided into concentric rings to obtain a target region that includes the area surrounding the tumor.

8. The method according to claim 1, characterized in that, The heterogeneity analysis results include at least one of the following forms: Pseudo-color mapping images are used to display differences in vascular distribution density or blood flow velocity in different sub-regions; A three-dimensional heatmap is overlaid on a grayscale tissue image of the tumor, with highly heterogeneous areas marked. Statistical charts, including bar charts or line charts, are used to compare the quantitative results of microvascular characteristics in different sub-regions.

9. The method according to claim 1, characterized in that, The acquisition of tumor imaging results includes: Track the movement trajectory of contrast agent microbubbles in the tumor microvessels, and separate and extract high-frequency moving microbubble signals; The high-frequency moving microbubble signal is located at the subpixel level, and the movement of the microbubble is tracked based on the location result to obtain the tracking trajectory; The information such as the density and velocity of the moving microbubbles reflected by the tracking trajectory is accumulated, and the accumulated results are mapped to the corresponding spatial locations to obtain the tumor imaging results.

10. An electronic device, characterized in that, include: At least one processor; And a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the tumor heterogeneity analysis method based on super-resolution microscopy as described in any one of claims 1 to 9.

11. A non-transitory computer-readable storage medium storing computer instructions, characterized in that, The computer instructions are used to cause the computer to execute the tumor heterogeneity analysis method based on super-resolution microscopy as described in any one of claims 1 to 9.

12. A computer program product, characterized in that, The method includes a computer program that, when executed by a processor, implements the tumor heterogeneity analysis method based on super-resolution microscopy according to any one of claims 1 to 9.

Images