Tumor recognition and treatment decision device and method

Abstract

This invention relates to a tumor identification and treatment decision-making device and method, comprising: a scanning module for scanning and imaging suspected tumor tissue to obtain a grayscale image of the tissue, and performing pixel threshold segmentation on the grayscale image of the tissue to obtain segmented tissue regions; an extraction module for performing pixel density statistics and texture feature extraction on the segmented tissue regions to obtain corresponding region pixel density and texture feature parameters; and a determination module for determining the benign or malignant nature of the suspected tumor tissue based on the region pixel density and texture feature parameters, obtaining a determination result and corresponding tumor pathological classification information, and determining an individualized treatment path based on the determination result and tumor pathological classification information. This invention solves the problem of how to effectively integrate pixel density distribution characteristics and deep texture features from the segmented tissue regions to improve the accuracy of tumor pathological discrimination.

Description

Technical Field

[0001] This invention relates to the field of tumor identification technology, and in particular to a device and method for tumor identification and treatment decision-making. Background Technology

[0002] In the current field of tumor diagnosis and treatment, medical imaging technology is widely used for the detection and analysis of suspected tumor tissue. Existing clinical practice typically relies on modalities such as computed tomography (CT), magnetic resonance imaging (MRI), or ultrasound imaging to obtain structural information of tissues, and then combines this with the experience of radiologists to perform visual evaluation of the images to determine the benign or malignant nature of the tumor and its possible pathological type. To further improve analysis efficiency, some auxiliary systems have introduced image processing algorithms based on threshold segmentation and feature extraction. By performing simple thresholding on the intensity distribution of grayscale images, lesion areas are initially identified, and basic statistical quantities such as mean and variance are calculated as region features for subsequent classification model input, supporting doctors in formulating treatment plans.

[0003] However, existing technologies, after pixel-level segmentation of tumor tissue, often rely solely on single gray-level statistical features or coarse texture descriptors, lacking refined modeling of the spatial distribution density of pixels and their microscopic texture patterns within local regions. This leads to ambiguity in distinguishing highly heterogeneous benign lesions from early-stage malignant tumors. Particularly when dealing with tissue images with blurred boundaries and mixed densities, it struggles to accurately reflect the degree of cell arrangement disorder and matrix reconstruction characteristics, thus affecting the specificity of benign / malignant determination and the consistency of pathological classification, ultimately limiting the accuracy of personalized treatment pathway recommendations. Therefore, how to effectively integrate pixel density distribution characteristics and deep texture features from segmented tissue regions to improve the accuracy of tumor pathological discrimination has become an urgent technical problem to be solved. Summary of the Invention

[0004] The main technical problem addressed by this application is to provide a device and method for tumor identification and treatment decision-making. It solves the problem that misjudgment of nuclear overlap is prone to occur in densely populated cell areas, leading to deviations in the identification of cell nuclear morphology, which ultimately reduces the accuracy of benign and malignant determination and makes it difficult to stably output reliable individualized treatment path guidance.

[0005] To solve the above-mentioned technical problems, this application employs a tumor identification and treatment decision-making device, comprising: The scanning module is used to scan and image suspected tumor tissue to obtain a grayscale image of the tissue, and to perform pixel threshold segmentation on the grayscale image of the tissue to obtain segmented tissue regions. The extraction module is used to perform pixel density statistics and texture feature extraction on the segmented tissue region, and obtain the region pixel density and texture feature parameters accordingly. The determination module is used to determine the benign or malignant nature of the suspected tumor tissue based on the pixel density and texture feature parameters of the region, obtain the determination result and the corresponding tumor pathological classification information, and determine the individualized treatment path guidance based on the determination result and the tumor pathological classification information.

[0006] Furthermore, the step of scanning and imaging the suspected tumor tissue to obtain a grayscale image of the tissue includes: Multi-angle scanning sampling was performed on suspected tumor tissue to obtain multiple sets of raw scan data, and the overlapping areas of the multiple sets of raw scan data were registered to obtain three-dimensional tissue data; Based on the three-dimensional data of the tissue, the suspected tumor tissue is reconstructed by tomography to obtain a three-dimensional tissue model. The three-dimensional tissue model is then subjected to orthogonal projection and grayscale mapping conversion to obtain a grayscale image of the tissue.

[0007] Furthermore, the step of performing pixel thresholding on the grayscale image of the tissue to obtain segmented tissue regions includes: The grayscale image of the tissue is statistically analyzed to obtain a grayscale value distribution histogram, and the peak and trough analysis of the grayscale value distribution histogram is performed to determine the grayscale threshold range. The grayscale image of the tissue is filtered based on the grayscale threshold range, and pixels with grayscale values ​​within the grayscale threshold range are marked as candidate tissue pixels. Connectivity analysis is performed on the candidate tissue pixels to merge adjacent candidate tissue pixels into a single connected region, resulting in multiple preliminary segmentation regions. The area of ​​each preliminary segmentation region is calculated, and preliminary segmentation regions with areas smaller than a preset threshold are removed to obtain segmented tissue regions.

[0008] Furthermore, the step of performing grayscale value statistics on the grayscale image of the tissue to obtain a grayscale value distribution histogram includes: The grayscale value of each pixel in the grayscale image of the tissue is read to obtain the grayscale value of all pixels. Based on the preset grayscale value interval division rules, the grayscale values ​​of all pixels are divided into multiple grayscale value intervals. At the same time, the number of pixels contained in each grayscale value interval is counted, and a grayscale value distribution histogram is drawn with the grayscale value interval as the horizontal axis and the number of pixels as the vertical axis.

[0009] Furthermore, the step of performing pixel density statistics and texture feature extraction on the segmented tissue region to obtain the corresponding region pixel density and texture feature parameters includes: The number of pixels in the segmented tissue region is counted to obtain the total number of pixels in the region, and the area of ​​the segmented tissue region is measured to obtain the actual area of ​​the region. Based on the total number of pixels in the region and the actual area of ​​the region, the pixel density is calculated to obtain the pixel density of the region. Gray-level pairing is performed on the pixels within the segmented tissue region to obtain a gray-level co-occurrence matrix; Texture feature parameters are obtained by calculating the texture features of the segmented tissue region based on the gray-level co-occurrence matrix.

[0010] Furthermore, the step of determining the benign or malignant nature of the suspected tumor tissue based on the region's pixel density and texture feature parameters, and obtaining the determination result and corresponding tumor pathological classification information, includes: The pixel density of the region is numerically quantized to obtain a density feature vector, and the texture feature parameters are dimensionally normalized to obtain a texture feature vector. The density feature vector and the texture feature vector are concatenated to obtain a comprehensive feature vector, and the spatial distance of the comprehensive feature vector is calculated to obtain the sample difference value. The sample difference value is compared with a preset malignancy determination threshold to obtain the benign or malignant determination result; The texture feature vector is subjected to texture direction statistics and texture periodicity calculation to obtain the dominant texture direction and the spacing between texture repeating units; Based on the dominant texture direction, the spacing between texture repeating units, and the regional pixel density, the standard pathological types in the pathological feature library are matched to obtain tumor pathological classification information.

[0011] Furthermore, the step of determining individualized treatment pathway guidance based on the judgment result and tumor pathological classification information includes: Based on the benign or malignant determination results and tumor pathological classification information, combined with the current tumor diagnosis and treatment guidelines database, the tumor type and stage information that match the current determination results are retrieved, and corresponding individualized treatment path guidance is selected from the preset treatment plan library based on the retrieved information.

[0012] The present invention also provides a method for tumor identification and treatment decision-making, comprising: The suspected tumor tissue is scanned and imaged to obtain a grayscale image of the tissue, and the grayscale image of the tissue is segmented by pixel thresholding to obtain segmented tissue regions; Pixel density statistics and texture feature extraction are performed on the segmented tissue regions to obtain the corresponding region pixel density and texture feature parameters. Based on the region's pixel density and texture feature parameters, the suspected tumor tissue is judged to be benign or malignant, and the judgment result and corresponding tumor pathological classification information are obtained. Based on the judgment result and tumor pathological classification information, an individualized treatment path is determined.

[0013] The present invention also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of any of the above methods.

[0014] The present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of any of the above methods.

[0015] The above scheme includes a scanning module for scanning and imaging suspected tumor tissue to obtain a grayscale image of the tissue, and performing pixel threshold segmentation on the grayscale image to obtain segmented tissue regions; an extraction module for performing pixel density statistics and texture feature extraction on the segmented tissue regions to obtain corresponding region pixel density and texture feature parameters; and a determination module for determining the benign or malignant nature of the suspected tumor tissue based on the region pixel density and texture feature parameters, obtaining the determination result and corresponding tumor pathological classification information, and determining individualized treatment path guidance based on the determination result and tumor pathological classification information. This solves the problem of how to effectively integrate pixel density distribution characteristics and deep texture features from the segmented tissue regions to improve the accuracy of tumor pathological discrimination. It enables the determination module to comprehensively utilize region pixel density and texture feature parameters for comprehensive discrimination, enhancing the ability to identify subtle differences between benign and malignant tumors, especially improving classification specificity when distinguishing high-density benign nodules from early malignant lesions. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the 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.

[0017] Figure 1 This is a structural block diagram of a tumor identification and treatment decision-making device according to an embodiment of the present invention; Figure 2 This is a structural block diagram of a tumor identification and treatment decision-making method according to an embodiment of the present invention; Figure 3 This is a schematic block diagram of the structure of a computer device according to an embodiment of the present invention.

[0018] The objectives, features, and advantages of this invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0019] The technical solutions of the embodiments of this application 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 this application, and not all of the embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0020] Specifically, the tumor identification and treatment decision-making method of this embodiment includes the following steps: like Figure 1 As shown, Figure 1 This invention provides a tumor identification and treatment decision-making device, comprising: The scanning module is used to scan and image suspected tumor tissue to obtain a grayscale image of the tissue, and to perform pixel threshold segmentation on the grayscale image of the tissue to obtain segmented tissue regions. The extraction module is used to perform pixel density statistics and texture feature extraction on the segmented tissue region, and obtain the region pixel density and texture feature parameters accordingly. The determination module is used to determine the benign or malignant nature of the suspected tumor tissue based on the pixel density and texture feature parameters of the region, obtain the determination result and the corresponding tumor pathological classification information, and determine the individualized treatment path guidance based on the determination result and the tumor pathological classification information.

[0021] Specifically, the examination of suspected tumor tissue is first performed by the scanning module. This device uses medical imaging equipment such as CT or ultrasound systems to acquire data from the target area, obtaining a grayscale image reflecting the tissue density distribution. Different brightness values ​​in this image correspond to the X-ray absorption coefficients of different tissue types. Subsequently, pixel-level threshold segmentation is performed on this grayscale image. Specifically, regions matching tumor characteristics are selected based on a preset intensity range. For example, pixels with grayscale values ​​between 120 and 200 are marked as candidate lesion areas, while the rest are considered background or normal tissue. This generates a binary mask image to define the segmented tissue regions required for subsequent analysis. This process directly serves the "obtaining the segmented tissue region" step in the aforementioned technical solution and provides a spatial positioning basis for it.

[0022] Next, the feature extraction stage begins. The extraction module operates on the previously segmented tissue region, performing two parallel processes: First, it statistically analyzes the distribution density of all retained pixels within the region, calculating the number of effective pixels per unit area, i.e., the region pixel density. For example, if a 5mm² lesion contains approximately 380 bright pixels, its density can be expressed as 76 pixels / mm². Second, it analyzes the spatial arrangement of grayscale variations within the region, using Local Binary Pattern (LBP) or Gray-Level Co-occurrence Matrix (GLCM) methods to obtain texture feature parameters, such as contrast, correlation, energy, and entropy, to characterize the uniformity or disorder of the internal tissue structure. These operations correspond to the generation paths of "region pixel density" and "texture feature parameters" in the original scheme, respectively, constituting the data input required for judgment.

[0023] Finally, the judgment module performs comprehensive analysis. This module has a built-in classification model, such as a support vector machine or random forest algorithm. It receives the region pixel density values ​​and a set of texture feature parameters from the previous stage as input variables. Through pre-trained learning rules, it determines whether the tumor belongs to the benign or malignant category, and outputs possible pathological subtyping suggestions, such as whether it is inclined to adenocarcinoma, squamous cell carcinoma, or inflammatory pseudotumor. Based on this, it further calls the treatment guideline database and combines the subtyping results to match recommended individualized treatment paths, such as surgical resection as the priority, targeted drug intervention, or regular follow-up observation. The entire process forms a closed loop from the acquisition of the original image to the final decision output. Each module inherits the results of the previous step and refines the information expression step by step, ensuring that the conversion from image signal to clinical recommendation is traceable and feasible.

[0024] In a specific embodiment, the step of scanning and imaging suspected tumor tissue to obtain a tissue grayscale image includes: Multi-angle scanning sampling was performed on suspected tumor tissue to obtain multiple sets of raw scan data, and the overlapping areas of the multiple sets of raw scan data were registered to obtain three-dimensional tissue data; Based on the three-dimensional data of the tissue, the suspected tumor tissue is reconstructed by tomography to obtain a three-dimensional tissue model. The three-dimensional tissue model is then subjected to orthogonal projection and grayscale mapping conversion to obtain a grayscale image of the tissue.

[0025] Specifically, when performing imaging tasks, the scanning module first performs multi-angle scanning sampling on the suspected tumor tissue. For example, using a CT scanner, it collects 24 sets of raw projection data around the lesion area at 15-degree intervals. Each set of data contains attenuation information of X-rays after passing through the tissue. There are spatially overlapping areas between these raw scan data, which need to be registered to eliminate misalignment caused by slight body movements or respiratory displacement. Typically, an image alignment algorithm based on maximizing mutual information is used to adjust the overlapping parts from each viewpoint to be spatially consistent, forming a complete dataset in a unified coordinate system, thus obtaining the three-dimensional data of the tissue. This step is the basic input for subsequent three-dimensional modeling.

[0026] After obtaining the 3D tissue data, tomographic reconstruction is performed. Filtered back projection (FBP) or iterative reconstruction techniques are commonly used to reconstruct the internal structure layer by layer, generating a 3D tissue model that quantifies anatomical details. This model can represent the spatial relationship between the tumor and its surrounding tissues. Based on this, a specific direction—such as the coronal, sagittal, or transverse plane—is selected to perform orthogonal projection on the 3D model, compressing the 3D voxels along a certain axis into a 2D plane representation. During the projection process, grayscale mapping is performed simultaneously, mapping local density values ​​proportionally to 8-bit grayscale levels (0–255). For example, high-density calcification areas are set to 230, and low-density necrosis areas to 60. The final output is a grayscale image of the tissue reflecting its internal heterogeneity. The entire process begins with multi-angle sampling, and through registration, reconstruction, projection, and mapping, progressively supports the aforementioned specific implementation method of "obtaining a grayscale image of the tissue."

[0027] In a specific embodiment, performing pixel thresholding on the grayscale image of the tissue to obtain segmented tissue regions includes: The grayscale image of the tissue is statistically analyzed to obtain a grayscale value distribution histogram, and the peak and trough analysis of the grayscale value distribution histogram is performed to determine the grayscale threshold range. The grayscale image of the tissue is filtered based on the grayscale threshold range, and pixels with grayscale values ​​within the grayscale threshold range are marked as candidate tissue pixels. Connectivity analysis is performed on the candidate tissue pixels to merge adjacent candidate tissue pixels into a single connected region, resulting in multiple preliminary segmentation regions. The area of ​​each preliminary segmentation region is calculated, and preliminary segmentation regions with areas smaller than a preset threshold are removed to obtain segmented tissue regions.

[0028] Specifically, when performing pixel thresholding segmentation on a grayscale image of tissue, the grayscale values ​​of all pixels in the image are first statistically analyzed to generate a grayscale value distribution histogram. The horizontal axis of this histogram represents the grayscale levels from 0 to 255, and the vertical axis represents the frequency of occurrence of the corresponding grayscale value. For example, in lung CT images, two distinct peaks are often visible: one concentrated in the 40-60 range, representing normal lung parenchyma, and the other located between 130 and 180, corresponding to nodular areas with higher density. Next, peak and trough analysis is performed on this histogram to identify the trough locations between the main peaks, such as the minimum values ​​in the 90-110 range. These are determined as the boundaries of the grayscale threshold range, thus defining 110 to 200 as the target segmentation interval. This step directly supports the specific implementation of threshold setting in "obtaining the segmented tissue region".

[0029] Subsequently, the original image is scanned pixel by pixel based on the grayscale threshold range. Pixels with grayscale values ​​between 110 and 200 are marked as candidate tissue pixels, while the rest are discarded. These candidate points are often discretely distributed and require further processing. Next, connected component analysis is performed, setting four-neighbor or eight-neighbor connection rules to merge adjacent candidate tissue pixels into blocks, forming multiple preliminary segmentation regions, each of which is an independent connected component. Then, the area occupied by each connected component is calculated, for example, by multiplying the number of pixels by the physical size of a single pixel to obtain the actual area. If a region is smaller than a preset threshold, such as 3mm², it is considered noise or a small artifact and is removed. Only the larger regions that remain are finally confirmed as valid segmented tissue regions. The entire process starts with histogram analysis, and through threshold screening, region merging, and area filtering, it progressively completes the extraction process from the image to the effective lesion area.

[0030] In a specific embodiment, the step of performing grayscale value statistics on the grayscale image of the tissue to obtain a grayscale value distribution histogram includes: The grayscale value of each pixel in the grayscale image of the tissue is read to obtain the grayscale value of all pixels. Based on the preset grayscale value interval division rules, the grayscale values ​​of all pixels are divided into multiple grayscale value intervals. At the same time, the number of pixels contained in each grayscale value interval is counted, and a grayscale value distribution histogram is drawn with the grayscale value interval as the horizontal axis and the number of pixels as the vertical axis.

[0031] Specifically, when processing grayscale images of tissue, the system first reads the grayscale value of each pixel in the image one by one. These images are typically in 8-bit single-channel format, with each pixel's grayscale value falling between 0 and 255, representing different tissue density responses. For example, values ​​close to 0 correspond to low-density fat or gas-containing areas, while values ​​close to 255 may reflect calcification or contrast agent accumulation. This step is the basic data source for subsequently generating the grayscale value distribution histogram.

[0032] After obtaining the grayscale values ​​of all pixels, they are categorized according to a preset grayscale value interval division rule. A common approach is to divide the entire range of 0 to 255 into 32 equal intervals, each spanning 8 grayscale levels. For example, the first interval covers 0 to 7, the second covers 8 to 15, and so on. Alternatively, non-uniform division can be used based on actual imaging characteristics, especially in critical density ranges such as 80 to 180, where denser segmentation can improve resolution. The number of pixels falling within each interval is counted. For example, if 1560 pixels have grayscale values ​​between 120 and 127, this frequency is recorded in the corresponding interval. After the statistics are completed, a histogram is plotted in a two-dimensional coordinate system using these grayscale value intervals as the x-axis and the number of pixels corresponding to each interval as the y-axis, thus obtaining the grayscale value distribution histogram. This histogram visually displays the proportion of tissues of different densities in an image. For example, in lung tumor images, a bimodal distribution often appears in the low grayscale area (40–70) and the medium-to-high grayscale area (130–180), corresponding to normal lung fields and solid nodules, respectively. The above process fully realizes the technical details of "obtaining a grayscale value distribution histogram," providing quantifiable data support for subsequent peak and trough analysis.

[0033] In a specific embodiment, the step of performing pixel density statistics and texture feature extraction on the segmented tissue region to obtain the corresponding region pixel density and texture feature parameters includes: The number of pixels in the segmented tissue region is counted to obtain the total number of pixels in the region, and the area of ​​the segmented tissue region is measured to obtain the actual area of ​​the region. Based on the total number of pixels in the region and the actual area of ​​the region, the pixel density is calculated to obtain the pixel density of the region. Gray-level pairing is performed on the pixels within the segmented tissue region to obtain a gray-level co-occurrence matrix; Texture feature parameters are obtained by calculating the texture features of the segmented tissue region based on the gray-level co-occurrence matrix.

[0034] Specifically, after acquiring the segmented tissue region, the extraction module first counts all retained pixels within that region to determine the total number of pixels. For example, a lung nodule segmentation result might contain 420 effective pixels. Simultaneously, combining the physical size parameters from the imaging process, such as the actual spatial resolution of 0.5mm × 0.5mm per pixel, the actual area occupied by the region is calculated: 420 multiplied by 0.25mm², resulting in 105mm². Then, the total number of pixels is divided by the actual area to obtain the number of pixels distributed per square millimeter, i.e., the region's pixel density, which in this example is approximately 4.0 pixels / mm². This calculation process directly corresponds to the aforementioned technical implementation of "obtaining the region's pixel density," reflecting the density of signal points per unit area.

[0035] Next, texture analysis is performed on the same segmented tissue region. First, gray-level pairing is performed on the internal pixels, setting a specific direction (e.g., 0°) and a fixed step size (usually 1 pixel). All pixel pairs meeting the conditions within the region are traversed, and their gray-level combinations are recorded. For example, a pixel with a gray level of 132 is paired with a pixel to its right with a gray level of 145. These combinations are accumulated to form a frequency table. Finally, after normalization, an N×N gray-level co-occurrence matrix is ​​generated. N is commonly set to 16 or 32, mapping the original 0–255 gray levels to a coarser granular level for calculation. With this matrix, multiple texture feature parameters are derived using standard mathematical expressions. For example, contrast reflects the intensity of local gray-level differences, with the formula ∑(ij)²×P(i,j); energy reflects pattern uniformity; and entropy measures the degree of structural disorder. These parameters collectively describe the regularity or heterogeneity of the internal tissue arrangement of the tumor. For instance, high entropy values ​​often indicate disordered cell distribution and are associated with malignant tendencies. The entire process extends from the aforementioned pixel statistics to texture modeling, progressively supporting the generation path of "texture feature parameters".

[0036] In a specific embodiment, the step of determining the benign or malignant nature of the suspected tumor tissue based on the region's pixel density and texture feature parameters, and obtaining the determination result and corresponding tumor pathological classification information, includes: The pixel density of the region is numerically quantized to obtain a density feature vector, and the texture feature parameters are dimensionally normalized to obtain a texture feature vector. The density feature vector and the texture feature vector are concatenated to obtain a comprehensive feature vector, and the spatial distance of the comprehensive feature vector is calculated to obtain the sample difference value. The sample difference value is compared with a preset malignancy determination threshold to obtain the benign or malignant determination result; The texture feature vector is subjected to texture direction statistics and texture periodicity calculation to obtain the dominant texture direction and the spacing between texture repeating units; Based on the dominant texture direction, the spacing between texture repeating units, and the regional pixel density, the standard pathological types in the pathological feature library are matched to obtain tumor pathological classification information.

[0037] Specifically, after extracting the region pixel density and texture feature parameters, the system first performs numerical quantization on the region pixel density, converting it into a density feature vector that can be used for model calculation. For example, the measured 4.0 points / mm² is directly used as a one-dimensional vector input. Simultaneously, multiple texture feature parameters derived from the gray-level co-occurrence matrix, such as contrast, energy, and entropy, are normalized using methods such as min-max scaling or Z-score standardization to eliminate dimensional differences between different indicators and form a texture feature vector at a unified scale. This step is a prerequisite for subsequent feature fusion.

[0038] Next, the density feature vector and texture feature vector are concatenated sequentially to form a comprehensive feature vector containing both density and texture information, such as a four-dimensional input vector (4.0, 0.82, 5.6, 1.3). Then, in this high-dimensional space, the Euclidean distance or Mahalanobis distance between the sample and the known benign training sample set is calculated to obtain a scalar value, namely the sample difference value, which measures the degree of deviation of the current case from the typical benign pattern. Subsequently, this difference value is compared with the malignancy determination threshold determined through historical data statistics. If the former is greater than the latter, for example, the difference exceeds 1.6, then a "malignant" determination result is output; otherwise, it is determined as benign, thus obtaining the benign or malignant determination result.

[0039] Building upon this foundation, further analysis of texture structure characteristics is conducted. The gray-level co-occurrence matrix from which the texture feature vectors originate is statistically expanded along the directional dimension. For example, the average contrast in four directions—0°, 45°, 90°, and 135°—is calculated, and the direction corresponding to the maximum value is taken as the dominant texture direction, reflecting the main arrangement orientation of the internal tissue structure. Simultaneously, the autocorrelation function is used to find the secondary peak position in this direction, and the corresponding distance is the texture repeating unit spacing, used to estimate the periodic interval of cell clusters or fiber arrangement, for example, approximately 0.3 mm. Finally, combining the dominant texture direction, repeating unit spacing, and regional pixel density, pattern matching is performed in a pre-built pathological feature database to find the closest standard pathological type template. If it matches the palisade arrangement and moderate spacing characteristics of lung adenocarcinoma, the corresponding tumor pathological classification information is output. The entire process, from feature integration to classification decision, progressively advances to achieve accurate discrimination.

[0040] In a specific embodiment, determining the individualized treatment pathway guidance based on the determination result and tumor pathology classification information includes: Based on the benign or malignant determination results and tumor pathological classification information, combined with the current tumor diagnosis and treatment guidelines database, the tumor type and stage information that match the current determination results are retrieved, and corresponding individualized treatment path guidance is selected from the preset treatment plan library based on the retrieved information.

[0041] Specifically, after obtaining the benign / malignant determination result and tumor pathological classification information, the system uses these two outputs as key search conditions and connects to the built-in tumor diagnosis and treatment guidelines database. This database integrates the latest clinical guidelines published by authoritative institutions such as NCCN and CSCO, covering staging standards and recommended treatment strategies for different tumor types. For example, when the determination result is "malignant" and the pathological classification information points to "lung adenocarcinoma," the system first matches the classification entries for this type in the guidelines, and combines auxiliary information such as the size of the lesion assessable by imaging and the extent of boundary invasion to preliminarily infer the possible clinical stage, such as stage IIA or stage IIIB.

[0042] Based on this matching result, a linked screening is performed in a pre-set treatment plan library. This library stores structured treatment recommendations, including indication rules for various modalities such as surgical resection, radiotherapy, chemotherapy, targeted therapy, and immunotherapy. The system extracts corresponding recommended pathways based on the current case's subtype and stage combination. For example, for a patient with stage IIA lung adenocarcinoma, the system outputs the guidance "anatomical lobectomy plus mediastinal lymph node dissection as the first choice, followed by adjuvant chemotherapy." If specific gene mutation information exists (although not directly input, an interface can be reserved in the extended fields), it can be further refined to "EGFR status testing is recommended; those who are positive can choose osimertinib as adjuvant therapy." The entire process is driven by the judgment result and subtype information, achieving automatic connection from diagnosis to treatment recommendations through rule matching. This ensures that the output personalized treatment pathway guidance conforms to current clinical practice standards, while also retaining room for manual review and adjustment.

[0043] Please see Figure 2 , Figure 2 This is a schematic diagram of the framework of an embodiment of the disease early warning method for heterogeneous data integration in this application; as shown... Figure 2 As shown, a tumor identification and treatment decision-making method includes: The suspected tumor tissue is scanned and imaged to obtain a grayscale image of the tissue, and the grayscale image of the tissue is segmented by pixel thresholding to obtain segmented tissue regions; Pixel density statistics and texture feature extraction are performed on the segmented tissue regions to obtain the corresponding region pixel density and texture feature parameters. Based on the region's pixel density and texture feature parameters, the suspected tumor tissue is judged to be benign or malignant, and the judgment result and corresponding tumor pathological classification information are obtained. Based on the judgment result and tumor pathological classification information, an individualized treatment path is determined.

[0044] Reference Figure 3 This invention also provides a computer device whose internal structure can be as follows: Figure 3As shown, the computer device includes a processor, memory, display screen, input device, network interface, and database connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores the operating system, computer programs, and database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The database stores the data corresponding to this embodiment. The network interface is used to communicate with external terminals via a network connection. When the computer program is executed by the processor, it implements the above-described method.

[0045] Those skilled in the art will understand that Figure 3 The structures shown are merely block diagrams of some structures related to the present invention and do not constitute a limitation on the computer devices on which the present invention is applied.

[0046] An embodiment of the present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-described method. It is understood that the computer-readable storage medium in this embodiment can be a volatile readable storage medium or a non-volatile readable storage medium.

[0047] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the present invention and embodiments can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual-rate SDRAM (SSRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM, etc.

[0048] In some embodiments, the functions or modules of the apparatus provided in this disclosure can be used to perform the methods described in the above method embodiments. The specific implementation can be referred to the description of the above method embodiments, and for the sake of brevity, it will not be repeated here.

[0049] The description of the various embodiments above tends to emphasize the differences between the various embodiments. The similarities or similarities between them can be referred to, and for the sake of brevity, they will not be repeated here.

[0050] In the several embodiments provided in this application, it should be understood that the disclosed methods and apparatus can be implemented in other ways. For example, the apparatus implementations described above are merely illustrative. For instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the mutual coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection of devices or units may be electrical, mechanical, or other forms.

[0051] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0052] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0053] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) or processor to execute all or part of the steps of the methods of various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0054] If the technical solution of this application involves personal information, the product using this technical solution has clearly informed the user of the personal information processing rules and obtained the user's voluntary consent before processing the personal information. If the technical solution of this application involves sensitive personal information, the product using this technical solution has obtained the user's separate consent before processing the sensitive personal information, and also meets the requirement of "express consent". For example, at personal information collection devices such as cameras, clear and prominent signs are set up to inform users that they have entered the scope of personal information collection and that personal information will be collected. If an individual voluntarily enters the collection scope, it is deemed that they have agreed to the collection of their personal information; or on the personal information processing device, with clear signs / information informing users of the personal information processing rules, authorization is obtained from the individual through pop-up information or by asking the individual to upload their personal information; wherein, the personal information processing rules may include information such as the personal information processor, the purpose of personal information processing, the processing method, and the types of personal information processed.

Claims

1. A tumor recognition and treatment decision device, characterized by, include: The scanning module is used to scan and image suspected tumor tissue to obtain a grayscale image of the tissue, and to perform pixel threshold segmentation on the grayscale image of the tissue to obtain segmented tissue regions. The extraction module is used to perform pixel density statistics and texture feature extraction on the segmented tissue region, and obtain the region pixel density and texture feature parameters accordingly. The determination module is used to determine the benign or malignant nature of the suspected tumor tissue based on the pixel density and texture feature parameters of the region, obtain the determination result and the corresponding tumor pathological classification information, and determine the individualized treatment path guidance based on the determination result and the tumor pathological classification information.

2. The tumor recognition and treatment decision device of claim 1, wherein, The process of scanning and imaging suspected tumor tissue to obtain a grayscale image of the tissue includes: Multi-angle scanning sampling was performed on suspected tumor tissue to obtain multiple sets of raw scan data, and the overlapping areas of the multiple sets of raw scan data were registered to obtain three-dimensional tissue data; Based on the three-dimensional data of the tissue, the suspected tumor tissue is reconstructed by tomography to obtain a three-dimensional tissue model. The three-dimensional tissue model is then subjected to orthogonal projection and grayscale mapping conversion to obtain a grayscale image of the tissue.

3. The tumor recognition and treatment decision device of claim 1, wherein, The step of performing pixel thresholding on the grayscale image of the tissue to obtain segmented tissue regions includes: The grayscale image of the tissue is statistically analyzed to obtain a grayscale value distribution histogram, and the peak and trough analysis of the grayscale value distribution histogram is performed to determine the grayscale threshold range. The grayscale image of the tissue is filtered based on the grayscale threshold range, and pixels with grayscale values ​​within the grayscale threshold range are marked as candidate tissue pixels. Connectivity analysis is performed on the candidate tissue pixels to merge adjacent candidate tissue pixels into a single connected region, resulting in multiple preliminary segmentation regions. The area of ​​each preliminary segmentation region is calculated, and preliminary segmentation regions with areas smaller than a preset threshold are removed to obtain segmented tissue regions.

4. The tumor identification and treatment decision apparatus of claim 3, wherein, The step of performing grayscale value statistics on the grayscale image of the tissue to obtain a grayscale value distribution histogram includes: The grayscale value of each pixel in the grayscale image of the tissue is read to obtain the grayscale value of all pixels. Based on the preset grayscale value interval division rules, the grayscale values ​​of all pixels are divided into multiple grayscale value intervals. At the same time, the number of pixels contained in each grayscale value interval is counted, and a grayscale value distribution histogram is drawn with the grayscale value interval as the horizontal axis and the number of pixels as the vertical axis.

5. The tumor identification and treatment decision-making device according to claim 1, characterized in that, The step of performing pixel density statistics and texture feature extraction on the segmented tissue region to obtain the corresponding region pixel density and texture feature parameters includes: The number of pixels in the segmented tissue region is counted to obtain the total number of pixels in the region, and the area of ​​the segmented tissue region is measured to obtain the actual area of ​​the region. Based on the total number of pixels in the region and the actual area of ​​the region, the pixel density is calculated to obtain the pixel density of the region. Gray-level pairing is performed on the pixels within the segmented tissue region to obtain a gray-level co-occurrence matrix; Texture feature parameters are obtained by calculating the texture features of the segmented tissue region based on the gray-level co-occurrence matrix.

6. The tumor identification and treatment decision-making device according to claim 1, characterized in that, The process of determining the benign or malignant nature of the suspected tumor tissue based on the region's pixel density and texture feature parameters, yielding the determination result and corresponding tumor pathological classification information, includes: The pixel density of the region is numerically quantized to obtain a density feature vector, and the texture feature parameters are dimensionally normalized to obtain a texture feature vector. The density feature vector and the texture feature vector are concatenated to obtain a comprehensive feature vector, and the spatial distance of the comprehensive feature vector is calculated to obtain the sample difference value. The sample difference value is compared with a preset malignancy determination threshold to obtain the benign or malignant determination result; The texture feature vector is subjected to texture direction statistics and texture periodicity calculation to obtain the dominant texture direction and the spacing between texture repeating units; Based on the dominant texture direction, the spacing between texture repeating units, and the regional pixel density, the standard pathological types in the pathological feature library are matched to obtain tumor pathological classification information.

7. The tumor identification and treatment decision-making device according to claim 1, characterized in that, The process of determining individualized treatment pathways based on the judgment results and tumor pathological classification information includes: Based on the benign or malignant determination results and tumor pathological classification information, combined with the current tumor diagnosis and treatment guidelines database, the tumor type and stage information that match the current determination results are retrieved, and corresponding individualized treatment path guidance is selected from the preset treatment plan library based on the retrieved information.

8. A method for tumor identification and treatment decision-making, characterized in that, include: The suspected tumor tissue is scanned and imaged to obtain a grayscale image of the tissue, and the grayscale image of the tissue is segmented by pixel thresholding to obtain segmented tissue regions; Pixel density statistics and texture feature extraction are performed on the segmented tissue regions to obtain the corresponding region pixel density and texture feature parameters. Based on the region's pixel density and texture feature parameters, the suspected tumor tissue is judged to be benign or malignant, and the judgment result and corresponding tumor pathological classification information are obtained. Based on the judgment result and tumor pathological classification information, an individualized treatment path is determined.

9. A computer device, characterized in that, The method includes a memory and a processor coupled to each other, wherein the memory stores program instructions and the processor executes the program instructions to implement the tumor identification and treatment decision method of claim 8.

10. A computer-readable storage medium, characterized in that, The system stores program instructions that can be executed by a processor, the program instructions being used to implement the tumor identification and treatment decision method of claim 8.

Images