Analysis method and system for multi-tracer images, storage medium, computer device

Abstract

The application discloses a kind of for the analysis method and system of multi-tracer image, storage medium, computer equipment, the method comprises: obtaining the PET image and CT image of target object under each tracer, each PET image is segmented to obtain the lesion segmentation result, each CT image is segmented to obtain organ segmentation result;By uniform CT coordinate system, CT image and PET image are registered to obtain preliminary registration result, based on each organ segmentation result, preliminary registration result is registered to obtain final registration result;Determine the visualization overlap region of final registration result, according to visualization overlap region, each lesion segmentation result and each organ segmentation result, calculate first index parameter;According to the tracer visualization region of each PET image, lesion segmentation result and matched organ segmentation result, calculate second index parameter;According to first index parameter and each second index parameter, the state analysis of target object is carried out.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of image processing, and in particular to an analysis method and system for multi-tracer images, a storage medium and a computer device. BACKGROUND

[0002] In the field of medical image diagnosis, with the rapid development of medical imaging technology, imaging technologies such as positron emission tomography (PET) and computed tomography (CT) have become important means for clinical diagnosis and disease research. PET images can reflect the metabolism, physiology and biochemical processes in the body by the distribution of tracers in the body, providing key information for early diagnosis, staging and treatment effect evaluation of diseases; CT images have high spatial resolution and can clearly show the anatomical structure of human organs and tissues. In actual clinical application, in order to more comprehensively understand the occurrence and development process of diseases and evaluate the treatment effect, multiple tracers are often used to image the same target object to obtain multi-tracer images. Different tracers have different characteristics and can reflect the pathological condition from different angles, for example, some tracers are sensitive to the metabolic activity of tumor cells, while other tracers have better tracing effect on the blood perfusion of specific tissues. Comprehensive analysis of multi-tracer images can provide doctors with more abundant and accurate diagnostic basis, which helps to improve the accuracy of disease diagnosis and the evaluation level of treatment effect.

[0003] However, the current analysis and processing of multi-tracer images still face many challenges. On the one hand, due to the body position, respiration and other factors of the target object during the imaging process of different tracers, there may be spatial position deviation between PET images and CT images, and there may also be position mismatch between PET images under different tracers. This inconsistency in spatial position between images can seriously affect the subsequent comprehensive analysis of multi-tracer images, leading to the inability to accurately obtain the visualization information of each tracer at the same anatomical position, and further affecting the accurate judgment of the disease state. On the other hand, in the image analysis process, the existing methods mostly only perform simple lesion or organ segmentation on single tracer images, lacking in-depth mining and comprehensive utilization of the correlation information between multi-tracer images, thereby directly affecting the final analysis result. SUMMARY

[0004] In view of this, this application provides a method and system for analyzing multi-tracer images, a storage medium, and a computer device. Through two registration processes—first, primary registration based on a unified CT coordinate system, and then secondary registration using organ segmentation results—the accuracy of image registration is ensured, laying the foundation for accurate subsequent analysis. By processing image data under multiple tracers, the information provided by different tracers is fully utilized to calculate indicator parameters at multiple levels, including systemic, organ, and lesion levels. These indicator parameters at different levels can complement each other, comprehensively reflecting the state of the target object from macroscopic to microscopic perspectives, improving the comprehensiveness and accuracy of the analysis, and providing doctors with more comprehensive decision-making basis. Advanced image processing and analysis algorithms are employed to automate and intelligently calculate image segmentation, registration, and indicator parameters. This not only improves analysis efficiency but also reduces interference from human factors, enhancing the reliability and repeatability of the analysis results.

[0005] According to one aspect of this application, a method for analyzing multi-tracer images is provided, comprising:

[0006] Acquire PET images of the target object under each tracer and CT images matching the PET images, and perform lesion segmentation on the PET images under each tracer to obtain the lesion segmentation results corresponding to each PET image, and perform organ segmentation on the CT images under each tracer to obtain the organ segmentation results corresponding to each CT image.

[0007] Based on the CT images under each tracer, the entire CT image and PET image are registered using a unified CT coordinate system to obtain the primary registration result. Then, based on the organ segmentation results corresponding to the CT images under each tracer, the primary registration result is further registered to obtain the final registration result.

[0008] Based on the final registration result, the overlapping regions of each tracer on the final registration result are determined, and a first index parameter of the target object is calculated according to the overlapping regions, the lesion segmentation results of each PET image, and the organ segmentation results of each CT image. The first index parameter includes a first systemic index parameter, a first organ-level index parameter, and a first lesion-level index parameter. Furthermore, for each PET image under each tracer, a second index parameter of the target object under the tracer is calculated according to the tracer-developed region of the PET image, the lesion segmentation result, and the organ segmentation result of the CT image matching the PET image. The second index parameter includes a second systemic index parameter, a second organ-level index parameter, and a second lesion-level index parameter.

[0009] Based on the first indicator parameter and the second indicator parameter corresponding to each tracer, the state of the target object is analyzed.

[0010] According to another aspect of the present application, there is provided an analysis system for multi-tracer images, comprising:

[0011] an image registration module, configured to acquire PET images of a target object under each tracer, and CT images matched with the PET images, and perform lesion segmentation on the PET images under each tracer respectively to obtain lesion segmentation results corresponding to each PET image, and perform organ segmentation on the CT images under each tracer respectively to obtain organ segmentation results corresponding to each CT image; perform registration processing on all the CT images and PET images based on the CT images under each tracer through a unified CT coordinate system to obtain a primary registration result, and perform secondary registration processing on the primary registration result based on the organ segmentation results corresponding to the CT images under each tracer to obtain a final registration result;

[0012] an image analysis module, configured to determine a visualized overlap area of each tracer on the final registration result based on the final registration result, and calculate first index parameters of the target object according to the visualized overlap area, the lesion segmentation results of each PET image and the organ segmentation results of each CT image, the first index parameters comprising first whole-body level index parameters, first organ level index parameters and first lesion level index parameters; and for the PET images under each tracer, calculate second index parameters of the target object under the tracer according to the tracer visualized area of the PET image, the lesion segmentation result and the organ segmentation result of the CT image matched with the PET image, the second index parameters comprising second whole-body level index parameters, second organ level index parameters and second lesion level index parameters; and perform state analysis on the target object according to the first index parameters and the second index parameters corresponding to each tracer.

[0013] According to yet another aspect of the present application, there is provided a storage medium having a computer program stored thereon, the program being executed by a processor to implement the above-mentioned analysis method for multi-tracer images.

[0014] According to still another aspect of the present application, there is provided a computer device comprising a storage medium, a processor and a computer program stored on the storage medium and executable on the processor, the processor implementing the above-mentioned analysis method for multi-tracer images when executing the program.

[0015] By employing the above technical solutions, this application provides a method and system for analyzing multi-tracer images, a storage medium, and a computer device. Through two registration processes—first, primary registration based on a unified CT coordinate system, and then secondary registration using organ segmentation results—the accuracy of image registration is ensured, laying the foundation for accurate subsequent analysis. By processing image data under multiple tracers, the information provided by different tracers is fully utilized to calculate indicator parameters at multiple levels, including systemic, organ, and lesion levels. These indicator parameters at different levels can complement each other, comprehensively reflecting the state of the target object from macroscopic to microscopic perspectives, improving the comprehensiveness and accuracy of the analysis, and providing doctors with more comprehensive decision-making support. Advanced image processing and analysis algorithms are adopted to automate and intelligently calculate image segmentation, registration, and indicator parameters. This not only improves analysis efficiency but also reduces interference from human factors, enhancing the reliability and repeatability of the analysis results.

[0016] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, the following are specific embodiments of this application. Attached Figure Description

[0017] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0018] Figure 1 A schematic flowchart of an analysis method for multiple tracer images provided in an embodiment of this application is shown.

[0019] Figure 2 A schematic diagram of the structure of an analysis system for multi-tracer images provided in an embodiment of this application is shown;

[0020] Figure 3 A schematic diagram of the device structure of a computer device provided in an embodiment of this application is shown. Detailed Implementation

[0021] The present application will be described in detail below with reference to the accompanying drawings and embodiments. It should be noted that, unless otherwise specified, the embodiments and features described in the embodiments of the present application can be combined with each other.

[0022] This embodiment provides a method for analyzing multi-tracer images, such as... Figure 1 As shown, the method includes:

[0023] In step 101, PET images of a target object under each tracer are acquired, and CT images matched with the PET images are acquired, and lesion segmentation is performed on the PET images under each tracer respectively to obtain lesion segmentation results corresponding to each PET image, and organ segmentation is performed on the CT images under each tracer respectively to obtain organ segmentation results corresponding to each CT image;

[0024] In step 102, based on the CT images under each tracer, all the CT images and the PET images are registered by using a unified CT coordinate system to obtain a primary registration result, and the primary registration result is subjected to secondary registration processing based on the organ segmentation results corresponding to the CT images under each tracer to obtain a final registration result;

[0025] In step 103, based on the final registration result, a developed overlap region of each tracer on the final registration result is determined, and a first index parameter of the target object is calculated according to the developed overlap region, the lesion segmentation results of each PET image and the organ segmentation results of each CT image, the first index parameter including a first whole-body level index parameter, a first organ level index parameter and a first lesion level index parameter; and for the PET image under each tracer, a second index parameter of the target object under the tracer is calculated according to a tracer developed region of the PET image, a lesion segmentation result and an organ segmentation result of the CT image matched with the PET image, the second index parameter including a second whole-body level index parameter, a second organ level index parameter and a second lesion level index parameter;

[0026] In step 104, a state of the target object is analyzed according to the first index parameter and the second index parameter corresponding to each tracer.

[0027] The method for analyzing multiple-tracer images provided by the embodiments can be applied to a comprehensive medical image processing and analysis platform, and valuable information is extracted from original image data by performing deep processing and analysis on PET images under multiple tracers and CT images matched with the PET images, so that comprehensive and accurate state evaluation of a target object is provided for clinical diagnosis, drug research and development and biomedical research. The method sequentially completes image registration and image analysis based on the registration result, and ensures that valuable information is extracted from original image data.

[0028] Specifically, first, the PET images of the target object under each tracer and the matching CT images (in the same posture) are acquired from the relevant medical imaging devices. The PET image can reflect the distribution of the tracer in the body by detecting the radioactive decay signal of the tracer in the body, thereby indirectly showing the metabolism, function, etc. of the tissue. While the CT image provides detailed anatomical structure information, including the shape, position and size of organs, etc. by using the attenuation characteristics of X-rays penetrating the human body. The combination of the two can provide more comprehensive information for subsequent analysis. For example, in tumor diagnosis, the PET image can show the metabolic activity of the tumor, while the CT image can clearly show the location of the tumor and its relationship with the surrounding tissue.

[0029] Next, the PET images under each tracer can be segmented to locate the lesion area. Advanced image segmentation algorithms can be used to achieve lesion segmentation, such as threshold-based segmentation methods, region-based segmentation methods or deep learning-based segmentation methods, etc. For example, a deep learning-based segmentation method can automatically learn the features of the lesion by training a large amount of labeled data, thereby achieving accurate segmentation of the lesion. The lesion segmentation results obtained by segmentation will be used for subsequent index parameter calculation and state analysis.

[0030] At the same time, the CT images under each tracer can also be segmented to provide anatomical reference for subsequent registration and analysis. Organ segmentation also needs to rely on image segmentation algorithms, but unlike lesion segmentation, organ segmentation needs to more accurately identify and separate different organs. For example, in a chest CT image, the lungs, heart, liver, etc. need to be accurately segmented. The quality of the organ segmentation results will directly affect the accuracy of the registration and the accuracy of the subsequent analysis.

[0031] After obtaining the segmentation results of each PET image and each CT image, further registration processing can be performed on these images. Specifically, by unifying the CT coordinate system, all CT images and PET images can be registered to obtain a primary registration result. The purpose of unifying the CT coordinate system is to ensure that different images have comparability in spatial position. Since images obtained at different times or by different devices may have differences in spatial position, unifying the coordinate system can align these images to the same reference frame, so that subsequent analysis can be accurately performed. When primary registration is performed by unifying the CT coordinate system, rigid registration algorithms, etc. can be used.

[0032] Then, based on the organ segmentation results corresponding to the CT images under each tracer, a secondary registration process is performed on the primary registration result to obtain a final registration result. The secondary registration using the organ segmentation results can further improve the accuracy of the registration. Because the organ segmentation results can more accurately reflect the position and shape of the anatomical structure, by taking the organ segmentation results as the constraint condition of registration, the errors that may exist in the primary registration can be eliminated, especially in the areas where the organ boundary or morphology changes greatly.

[0033] After obtaining the final registration result, based on the final registration result, the visualization overlap region of each tracer on the final registration result is determined. The visualization overlap region reflects the common action region of different tracers in the body, and is of great significance for understanding the mutual relationship between different tracers and the comprehensive state of the target object.

[0034] Further, according to the visualization overlap region, the lesion segmentation result of each PET image, and the organ segmentation result of each CT image, a first index parameter of the target object is calculated, and here, the first index parameter includes a first whole-body level index parameter, a first organ level index parameter, and a first lesion level index parameter. Among them, the whole-body level index parameter can reflect the overall metabolic activity distribution, tracer uptake, etc. of the target object from a macroscopic point of view; the organ level index parameter can evaluate the functional state and metabolic level of a specific organ; and the lesion level index parameter can describe the characteristics of the lesion, such as size, shape, metabolic activity, etc. These index parameters reflect the comprehensive situation of the target object under the action of multiple tracers from different levels (whole body, organ, lesion).

[0035] In addition, for the PET image under each tracer, according to the tracer visualization region of the PET image, the lesion segmentation result, and the organ segmentation result of the CT image matched with the PET image, a second index parameter of the target object under the tracer is calculated, and here, the second index parameter includes a second whole-body level index parameter, a second organ level index parameter, and a second lesion level index parameter. These index parameters focus on the situation under the action of each tracer alone, which helps to understand the characteristics of the influence of different tracers on the target object. For example, different tracers may exhibit different uptake characteristics in different organs or lesions, and by calculating the second index parameter, these differences can be understood in depth.

[0036] After that, the target object can be analyzed in terms of its state according to the first index parameter and the second index parameter corresponding to each tracer. By comprehensively analyzing these multi-level index parameters, the physiological or pathological state of the target object can be more comprehensively and accurately evaluated. For example, in tumor diagnosis, combined with whole-body level, organ level and lesion level index parameters, the severity, progression and metastasis of the tumor can be determined. The results of the state analysis can provide support for clinical decision-making. Doctors can develop personalized treatment plans, evaluate treatment effectiveness, or predict disease prognosis based on the analysis results. For example, if the analysis results show that the metabolic activity of the tumor has decreased significantly, it may indicate that the treatment plan is effective; if new lesions or enlarged lesions are found, the treatment plan may need to be adjusted.

[0037] In a specific embodiment, in order to facilitate doctors to understand and use the analysis results, the analysis results can also be visually displayed, such as generating images, charts, etc. At the same time, detailed reports can also be generated, including the values of the index parameters, the conclusions and suggestions of the state analysis, etc., to provide intuitive references for clinical decision-making.

[0038] By applying the technical solutions of the present embodiment, through two registration processes, preliminary registration is performed based on the unified CT coordinate system, and secondary registration is performed using organ segmentation results, ensuring the accuracy of image registration and laying a foundation for subsequent accurate analysis; by processing image data under multiple tracers, the information provided by different tracers is fully utilized, and multiple levels of index parameters such as whole-body level, organ level and lesion level are calculated. Different levels of index parameters can complement each other, and can comprehensively reflect the state of the target object from macro to micro, improving the comprehensiveness and accuracy of the analysis and providing doctors with more comprehensive decision-making basis; advanced image processing and analysis algorithms are used to realize the automation and intelligentization of image segmentation, registration and index parameter calculation, which not only improves the analysis efficiency, but also reduces the interference of human factors and improves the reliability and repeatability of the analysis results.

[0039] In the embodiments of the present application, the "performing registration processing on all CT images and PET images based on the CT images under each tracer through a unified CT coordinate system to obtain a primary registration result" in step 102 comprises: taking any CT image in the CT images under each tracer as a reference image, calculating the image deviation between each target CT image in the remaining CT images and the reference image through a unified CT coordinate system, and calculating the adjustment parameters corresponding to the target CT image based on the image deviation, wherein the adjustment parameters comprise translation parameters, scaling parameters and rotation parameters; for each target CT image, sequentially performing image processing on the target CT image and the PET image matched with the target CT image according to the adjustment parameters corresponding to each target CT image, and aligning the processed image with the reference image and the PET image corresponding to the reference image, and obtaining the primary registration result after the processed images corresponding to each target CT image are all aligned.

[0040] In this embodiment, when performing primary registration based on the CT images under each tracer, first, any one of the CT images is selected as a reference image. The selection of the reference image has certain flexibility, and specifically, an image with better image quality, clear anatomical structure and representative can be selected. For example, if one of the multi-tracer CT images has smaller noise, moderate contrast and can clearly display the boundaries of the main organs during the acquisition process, it can be taken as the reference image.

[0041] Then, with the reference image as a reference, the image deviation between each of the remaining CT images (each of the remaining images can be referred to as a target CT image) and the reference image is calculated in the unified CT coordinate system. This process can use a feature matching method in image registration algorithms. For example, feature points (such as corner points, edge points, etc.) in the image can be extracted, and then the position differences of these feature points in the target CT image and the reference image are calculated to obtain the image deviation. These deviations reflect the differences in the spatial position of the target CT image relative to the reference image, including changes in translation, rotation, and scaling. Based on the calculated image deviation, the adjustment parameters corresponding to the target CT image are further calculated, including translation parameters, scaling parameters, and rotation parameters. The translation parameters are used to describe the position movement amount of the target CT image in the X, Y, and Z directions; the scaling parameters are used to adjust the size proportion of the target CT image in each direction; and the rotation parameters are used to determine the rotation angle of the target CT image around different axes. In the unified CT coordinate system, each feature point has a determined coordinate position, and by comparing the coordinates of the corresponding feature points in the target CT image and the reference image, the position differences of the feature points in the X, Y, and Z directions can be calculated, which are the image deviations. For example, if the coordinates of a feature point in the reference image are (10, 20, 30), and the coordinates of the feature point in the target CT image are (12, 22, 32), then the deviation of the feature point in the X direction is 2, the deviation in the Y direction is 2, and the deviation in the Z direction is 2.

[0042] There are many methods for calculating adjustment parameters, and common ones include optimization algorithms based on least squares. This algorithm solves the optimal adjustment parameters by minimizing the differences between the target CT image and the reference image (such as the differences in pixel values, the distances between feature points, etc.). For example, when calculating the translation parameters, the position of the target CT image can be adjusted continuously to minimize the pixel value differences between the target CT image and the reference image at the corresponding positions, thereby obtaining the optimal translation parameters.

[0043] For each target CT image, the target CT image and the PET image matched therewith are processed according to the corresponding adjustment parameters in sequence. For example, the images are moved in space according to the translation parameters, resized according to the scaling parameters, and rotated according to the rotation parameters. Then, the processed target CT image and the corresponding PET image are aligned with the reference image and the PET image corresponding to the reference image. The purpose of alignment is to make the images under different tracers consistent in spatial position, so as to facilitate subsequent analysis and comparison. After all the processed target CT images corresponding to the target CT images are aligned, the primary registration result is obtained. The primary registration result eliminates the spatial position difference between different CT images to a certain extent, but there may still be some errors caused by changes in organ morphology and other factors.

[0044] Further, for each processed target CT image in the primary registration result, a mapping relationship of each pixel point in the processed target CT image relative to the corresponding pixel point in the reference image can be calculated based on the organ segmentation results of the processed target CT image and the organ segmentation results of the reference image. The organ segmentation results provide accurate boundary information of different organs in the image, and by comparing the morphology and position of the organs in the processed target CT image and the reference image, the correspondence between the pixel points can be more accurately determined. The mapping relationship can be calculated by a method based on a morphing model. The morphing model can deform the image non-rigidly according to the anatomical features and physical properties of the organs, so that the organs in the processed target CT image and the organs in the reference image are better matched. For example, by optimizing the parameters of the morphing model, the similarity of the corresponding organs in the processed target CT image and the reference image is maximized, thereby obtaining the mapping relationship of each pixel point.

[0045] According to the calculated mapping relationship, the processed target CT image and the corresponding processed PET image are re-registered. Unlike the primary registration, the secondary registration is a non-rigid registration, which can better handle cases where the organ morphology changes greatly. By non-rigidly deforming the image, the processed target CT image and the PET image are more consistent with the reference image in anatomical structure. When all the processed target CT images and the corresponding processed PET images are re-registered, the final registration result is obtained. The final registration result has higher accuracy and can more accurately reflect the spatial correspondence between images under different tracers, providing a reliable basis for subsequent image analysis and state evaluation.

[0046] The embodiment of the present application combines primary registration and secondary registration, eliminates most spatial position differences through unified CT coordinate system, and further improves the accuracy of image registration by using organ segmentation results for non-rigid registration. This multi-level registration method can better handle the complex changes between images under different tracers and reduce registration errors. In addition, this registration method can adapt to the characteristics of images under different tracers, and can correct image differences caused by different acquisition times, device differences or physiological changes through registration processing.

[0047] In the embodiment of the present application, the "determining the imaging overlap area of each tracer on the final registration result based on the final registration result" in step 103 comprises: determining the pixel set corresponding to the lesion position in the PET image after final registration under each tracer based on the final registration result, the pixel set including the coordinates of each pixel point contained in the lesion position; extracting common pixel points in each pixel set based on the pixel set corresponding to each tracer, and determining the imaging overlap area of the final registration result based on the common pixel points.

[0048] In this embodiment, the final registration result is obtained after primary registration and secondary registration, which eliminates the spatial position difference between images under different tracers, so that different images reach a high degree of consistency in anatomical structure. After obtaining the final registration result, the lesion position in the PET image after final registration under each tracer can be determined based on the final registration result. The coordinates of each pixel point contained in all lesion positions in the PET image are obtained. In a digital image, each pixel point has a unique coordinate position, and these coordinate information can be accurately obtained through image processing algorithm. Then, the coordinates of all pixel points contained in the lesion positions are collected to form a pixel set. This pixel set represents the specific position and range of the lesion in the image under this tracer. For example, for a circular lesion with a diameter of 1 centimeter, its pixel set will contain the coordinates of all pixel points in the circular region.

[0049] Then, the same pixel points are found by comparing the pixel sets corresponding to different tracers. This can be achieved by programming, for example, using the intersection operation of sets. The pixel set of each tracer is regarded as a set, and the intersection of these sets can be obtained by calculating the intersection of these sets. For example, if there are three tracers A, B and C, the corresponding pixel sets are S A , S B , S C , then the set of common pixel points is S A ∩S B ∩S CThat is, the common pixel points refer to the pixel points existing in the pixel sets corresponding to the multiple tracers. These pixel points represent the regions that show abnormal signals under different tracers, which can be the core regions of the lesion or regions with common physiological and pathological characteristics. For example, in tumor diagnosis, different tracers can reflect the metabolic activity, blood perfusion, etc. of the tumor, and the common pixel points can correspond to the regions with dense tumor cells, high metabolic activity and rich blood flow. In a specific embodiment, in order to reduce the amount of calculation, the pixel sets can also be preprocessed before comparison, such as removing some pixel points that are obviously not in the lesion region, or performing down-sampling processing on the pixel sets. Down-sampling can reduce the number of pixel points while ensuring a certain accuracy, thereby improving the comparison speed.

[0050] Subsequently, based on the extracted common pixel points, a visual overlap region of the final registration result is determined. The visual overlap region refers to the region that shows a visual signal under different tracers, which reflects the common action range of different tracers at the lesion site. For example, in tumor diagnosis, the visual overlap region can correspond to the region with high malignancy and poor prognosis of the tumor. The boundary of the visual overlap region can be determined by connecting the common pixel points. Common methods include convex hull algorithm, boundary tracking algorithm, etc. The convex hull algorithm can enclose the common pixel points in a smallest convex polygon, which is the approximate boundary of the visual overlap region. The boundary tracking algorithm gradually determines the boundary of the region by tracking the distribution of the common pixel points.

[0051] After determining the visual overlap region, further analysis can be performed on the region to extract its feature information. For example, the area, volume, average radioactivity uptake value, etc. of the region are calculated. These feature information can be used to evaluate the nature, severity, etc. of the lesion.

[0052] The embodiments of the present application reflect the physiological and pathological processes in the living body from different angles through different tracers, and by determining the visual overlap region, these information can be integrated, the core region of the lesion and the region with common characteristics can be more accurately identified, which helps to improve the accuracy of disease diagnosis.

[0053] Optionally, in the embodiment of the present application, the step 103 of "calculating the first index parameter of the target object according to the developed overlap region, the lesion segmentation result of each PET image and the organ segmentation result of each CT image" comprises: calculating the SUV mean value, SUV maximum value, FDG metabolic tumor volume and total lesion glycolysis corresponding to the developed overlap region according to the developed overlap region, and taking the SUV mean value, SUV maximum value, FDG metabolic tumor volume and total lesion glycolysis as the first whole-body level index parameter; determining the organ registration result of each CT image according to the final registration result, and performing regional division on the developed overlap region according to the organ registration result to obtain a first overlap sub-region corresponding to each organ, calculating the SUV mean value, SUV maximum value, FDG metabolic tumor volume and total lesion glycolysis corresponding to each first overlap sub-region according to each first overlap sub-region, and taking the SUV mean value, SUV maximum value, FDG metabolic tumor volume and total lesion glycolysis corresponding to each first overlap sub-region as the first organ level index parameter; determining the lesion registration result of each PET image according to the final registration result, and performing regional division on the developed overlap region according to the lesion registration result to obtain a second overlap sub-region corresponding to each lesion, calculating the SUV mean value, SUV maximum value, FDG metabolic tumor volume and total lesion glycolysis corresponding to each second overlap sub-region according to each second overlap sub-region, and taking the SUV mean value, SUV maximum value, FDG metabolic tumor volume and total lesion glycolysis corresponding to each second overlap sub-region as the first lesion level index parameter.

[0054] In this embodiment, the first whole-body level index parameter is selected from the SUV (Standardized Uptake Value) mean value, SUV maximum value, FDG metabolic tumor volume (MTV) and total lesion glycolysis (TLG) because these indexes can reflect the metabolic characteristics of the developed overlap region from different angles. The SUV (Standardized Uptake Value) is an index for measuring the degree of uptake of a tracer by tissues, the mean value reflects the average metabolic level of the entire developed overlap region, and the maximum value represents the highest metabolic activity in the region. The MTV represents the tumor volume with active metabolism in the developed overlap region, and the TLG is the product of the SUV mean value and the MTV, which comprehensively considers the metabolic intensity and tumor volume, and can more comprehensively reflect the metabolic load of the tumor.

[0055] The SUV mean is calculated as follows: First, obtain the SUV values ​​of all pixels within the overlapping development area. Then, sum these SUV values ​​and divide by the total number of pixels to obtain the SUV mean. For example, if the overlapping development area has n pixels, the corresponding SUV values ​​are SUV1, SUV2, ..., SUV2. n Then the average value of SUVs = (SUV1 + SUV2 + ... + SUV n ) / n.

[0056] SUV maximum value calculation: The maximum value of SUV is found directly in the overlapping area of ​​the developing region.

[0057] FDG-mediated tumor volume calculation: A threshold method can be used to determine the median volume (MTV). An appropriate SUV threshold is set, and pixels within the overlapping imaging region with SUV values ​​higher than this threshold are considered metabolically active tumor tissue. The number of these pixels is counted, and then the MTV is calculated based on the physical dimensions of the pixels (e.g., pixel pitch). For example, if the actual volume represented by each pixel is V, and the number of pixels within the overlapping imaging region with SUV values ​​higher than the threshold is m, then MTV = m × V.

[0058] Total lesion glycolysis calculation: TLG = mean SUV × MTV. TLG is obtained by multiplying the mean SUV by the MTV. This index comprehensively reflects the metabolic intensity and tumor volume in the contrast-overlapping region.

[0059] Furthermore, the first organ-level index parameters can be calculated. Specifically, the organ registration results for each CT image can be determined based on the final registration results. The purpose of organ registration is to ensure that the position and shape of the same organ are consistent in different CT images, providing an accurate basis for subsequent region division. For example, in chest CT images, organ registration can accurately identify the position and boundaries of organs such as the heart, lungs, and liver. Then, based on the organ registration results, the overlapping regions are divided to obtain the first overlapping sub-region corresponding to each organ. This can be achieved by determining whether the pixels within the overlapping regions belong to the registration region of a certain organ. For example, for the registration region corresponding to the heart, the pixels within the overlapping regions are assigned to the first overlapping sub-region of the heart.

[0060] Similar to the first whole-body level index parameter calculation, the SUV mean, SUV max, FDG metabolized tumor volume and total lesion glycolysis are calculated according to the first overlap sub-region corresponding to each organ, respectively. These indexes can reflect the metabolic characteristics of the overlapping regions in different organs, and help to understand the metabolic changes of different organs in the disease process. For example, in tumor patients, the overlapping regions in different organs may reflect the metastasis of the tumor, and by calculating the index parameters of each organ, the metabolic activity and load of the tumor in different organs can be evaluated.

[0061] In addition, the first lesion level index parameter calculation can also be calculated. Specifically, the lesion registration results of each PET image can be determined according to the final registration result. It should be noted that the lesion registration result does not mean that the lesions in different PET images are completely consistent, but means that after the final registration, the co-occurrence position of each lesion in the same coordinate space, which can be the union of the lesions in different PET images. Then, according to the lesion registration result, the overlapping region is regionally divided to obtain the second overlap sub-region corresponding to each lesion. This can be realized by judging whether the pixel point in the overlapping region belongs to a certain lesion region. For example, for a tumor lesion in the liver, the pixel points in the overlapping region located in the lesion registration region are divided into the second overlap sub-region of the lesion. The SUV mean, SUV max, FDG metabolized tumor volume and total lesion glycolysis are calculated according to the second overlap sub-region corresponding to each lesion, respectively. These indexes can reflect the metabolic characteristics of a single lesion, and help to evaluate the malignancy, growth activity and response to treatment of the lesion.

[0062] The embodiments of the present application calculate index parameters from three levels of whole body, organ and lesion, which can comprehensively and deeply understand the metabolic characteristics and disease conditions of the target object, and the index parameters at different levels complement each other, providing more abundant information for clinical diagnosis and treatment; by calculating specific index parameters, quantitative evaluation of the overlapping region is realized. Compared with traditional qualitative analysis, quantitative evaluation is more objective and accurate, which helps to improve the accuracy of diagnosis and evaluation of treatment effect.

[0063] In the embodiments of the present application, the PET image under each tracer is segmented based on the matched preset lesion segmentation model, and the preset lesion segmentation model under each tracer is constructed based on the following manner: a plurality of PET image samples corresponding to the target tracer are obtained, each of the PET image samples is labeled with a lesion, and each of the lesions is labeled with a lesion label; a nnU-Net framework is called, multi-dimensional feature extraction is performed on each PET image sample through the nnU-Net framework, and based on the multi-dimensional feature extraction result, the network depth, the network width, the encoder structure and the decoder structure of an initial lesion segmentation model are determined, and the initial lesion segmentation model is constructed according to the network depth, the network width, the encoder structure and the decoder structure; each of the PET image samples is subjected to down-sampling processing to obtain a down-sampling sample corresponding to each PET image sample, the initial lesion segmentation model is subjected to model training based on the down-sampling sample to obtain an intermediate lesion segmentation model; each of the PET image samples is input into the intermediate lesion segmentation model to obtain a coarse-grained lesion segmentation image corresponding to each PET image sample, and the coarse-grained lesion segmentation image is subjected to up-sampling processing to obtain an up-sampling sample corresponding to each PET image sample; the intermediate lesion segmentation model is retrained based on the up-sampling sample corresponding to each PET image sample to obtain the preset lesion segmentation model.

[0064] In this embodiment, the PET images under each tracer have unique imaging characteristics and lesion features. For example, different tracers have different distribution and metabolism in the body, leading to differences in the performance of lesions in the image (such as signal intensity, morphology, etc.). Therefore, it is necessary to construct a preset lesion segmentation model for each tracer. The preset lesion segmentation model corresponding to each tracer can be constructed by the following steps. First, obtain a plurality of PET image samples corresponding to the target tracer. These samples can cover different cases, different lesion types and lesion degrees, in order to ensure the diversity and representativeness of the samples. For example, in tumor diagnosis, the samples should include tumor lesions of different sizes, different locations, and tumor images of different pathological types. Diverse samples help the model learn more extensive features and improve the generalization ability of the model. In addition, the lesions in each PET image sample are accurately labeled to clearly define the location and range of the lesions. Specifically, professional medical experts can be used to operate, in order to ensure the accuracy of the labeling. For example, in PET images, lesions may appear as areas of increased radioactivity uptake, and experts can manually outline the boundaries of the lesions using image analysis software. Appropriate labels can also be designed for each lesion, and the labels should accurately reflect the characteristics and properties of the lesions. For example, in tumor diagnosis, the labels can include information such as the type of tumor (such as benign, malignant), grade (such as low-grade, high-grade), etc. The design of the labels should take into account the needs of subsequent model training and evaluation, ensuring that the labels have clear semantics and distinguishability.

[0065] Then, an initial lesion segmentation model can be constructed using nnU-Net. nnU-Net is a deep learning-based medical image segmentation framework that automatically adjusts network structure, optimizes training parameters, and is suitable for different types and sizes of medical image data. It can automatically learn the multi-dimensional features of the image, including spatial features, texture features, etc. Specifically, the nnU-Net framework extracts features from each PET image sample, which can perform convolution, pooling, and other operations on the image to gradually extract features at different levels. These features will be used for subsequent model structure determination and training.

[0066] Based on the multi-dimensional feature extraction results, the network depth and width of the initial lesion segmentation model can be determined. Network depth refers to the number of convolutional layers in the network. A deeper network can learn more complex features, but it may also cause problems such as gradient vanishing or overfitting. Network width refers to the number of convolutional kernels in each layer. A wider network can extract more features, but it also increases the computational load of the model. By analyzing the feature extraction results, the appropriate network depth and width can be selected to balance the performance and computational efficiency of the model. The final network depth of the initial lesion segmentation model can be determined by constructing networks of different depths, plotting the training loss and validation loss curves, and calculating the segmentation accuracy, Dice coefficient, and other indicators on the validation set to evaluate the performance of networks of different depths. Finally, the network depth that has the best performance on the validation set, stable downward trend of training loss and validation loss, high segmentation accuracy and Dice coefficient, and reasonable computational resources and training efficiency is selected as the final network depth of the initial lesion segmentation model. The final network width of the initial lesion segmentation model can be determined by constructing networks of different widths, determining the segmentation accuracy and Dice coefficient on the validation set, and selecting the network width that can extract sufficient features for accurate lesion segmentation without overfitting as the final network width of the initial lesion segmentation model.

[0067] In addition, the encoder structure and decoder structure can also be determined. The encoder is usually used to extract the features of the image, gradually compressing the input image into a low-dimensional feature representation. The decoder is used to restore the low-dimensional feature representation to the segmentation result of the original image size. The structure design of the encoder and decoder can refer to common structures such as U-Net, ResNet, etc. The appropriate encoder and decoder can be determined by verifying the lesion segmentation effect of different common structures.

[0068] After determining the network depth, network width, encoder structure, and decoder structure, the initial lesion segmentation model can be built.

[0069] Next, each PET image sample can be downsampled to obtain a corresponding downsampled sample for each PET image sample. The purpose of downsampling is to reduce the size and computational load of the image, and to speed up the model training. At the same time, the downsampled image can highlight the main features of the image, which helps the model to learn more robust features. Common downsampling methods include max pooling, average pooling, etc. During the downsampling process, an appropriate downsampling ratio can be selected to ensure that the downsampled image still retains enough information for model training.

[0070] Subsequently, model training is performed on the initial lesion segmentation model based on the down-sampled samples. During the training process, the down-sampled samples are input into the model, the model outputs the predicted lesion segmentation result, and then the predicted lesion segmentation result is compared with the real lesion segmentation result to calculate the loss function. The parameters of the model are updated through the back propagation algorithm to minimize the loss function, thereby improving the segmentation accuracy of the model. When the model loss value of the initial lesion segmentation model is less than a first preset loss threshold, an intermediate lesion segmentation model is obtained.

[0071] Then, the intermediate lesion segmentation model can be retrained. Specifically, each PET image sample is input into the intermediate lesion segmentation model to obtain a coarse-grained lesion segmentation map corresponding to each PET image sample. The coarse-grained lesion segmentation map is the prediction result of the model trained on the down-sampled samples on the original size image. Due to the influence of down-sampling, the resolution of the coarse-grained lesion segmentation map is low, and the segmentation result can not be fine enough. Therefore, the coarse-grained lesion segmentation map is further up-sampled to obtain an up-sampled sample corresponding to each PET image sample. The purpose of up-sampling is to restore the resolution of the coarse-grained lesion segmentation map to the original image size. Common up-sampling methods include deconvolution, interpolation, etc. The up-sampled sample will be used for subsequent model retraining.

[0072] Subsequently, the intermediate lesion segmentation model is retrained based on the up-sampled sample corresponding to each PET image sample. The purpose of retraining is to further improve the segmentation accuracy of the model, so that the model can better adapt to the image of the original size. By comparing the up-sampled sample with the real lesion segmentation result, the loss function is calculated and the model parameters are updated, so that the model can also obtain more accurate segmentation results on the up-sampled image. When the model loss value of the intermediate lesion segmentation model is less than a second preset loss threshold, a preset lesion segmentation model is obtained.

[0073] The embodiments of the present application automatically determine the model structure by using the nnU-Net framework, which reduces the workload of manually designing the model and can adapt to different types and sizes of PET image data, and has strong universality and adaptability. Through the hierarchical training method of first training on the down-sampled samples and then retraining on the up-sampled samples, the training efficiency is improved, and the segmentation accuracy of the model is gradually improved to obtain more accurate lesion segmentation results.

[0074] In the embodiments of the present application, the step 101 of performing lesion segmentation on the PET image of each tracer respectively to obtain the lesion segmentation result corresponding to each PET image comprises: for the PET image of each tracer, calling a preset lesion segmentation model matched with the tracer and a preset sliding window mechanism, obtaining a sub-image region of the PET image each time through the preset sliding window mechanism, inputting the sub-image region into the matched preset lesion segmentation model to obtain the lesion segmentation result corresponding to the sub-image region, and performing reliability calculation on the lesion segmentation results corresponding to each sub-image region of the PET image to obtain the lesion segmentation result corresponding to the PET image according to the reliability calculation result.

[0075] In this embodiment, when processing a new PET image, a preset lesion segmentation model matched with the tracer type used by the image is called from a preset model library. This matching mechanism ensures that the model can better adapt to the characteristics of the current image, thereby improving the accuracy of segmentation. The sliding window mechanism is used to divide a large-size PET image into multiple smaller sub-image regions. This is because directly segmenting the entire large image may face problems such as insufficient computing resources and difficulty for the model to process. By using the sliding window, the local area of the image can be processed one by one, reducing the computational complexity. Among them, the size and step length of the sliding window are two key parameters. The window size can be determined according to the typical size of the lesion and the resolution of the image. If the window is too small, it may not be able to contain complete lesion information; if the window is too large, it will increase the amount of calculation and may lead to a decrease in segmentation accuracy. The step length determines the interval at which the window slides on the image. If the step length is too small, the amount of calculation will be too large, and if the step length is too large, some image regions may be missed.

[0076] According to the preset sliding window mechanism, the window slides on the PET image by a certain step length. After each sliding, the image region within the window is extracted as a sub-image region. At the image edge, the sliding window may exceed the image range, at which time a padding method can be used to fill appropriate pixel values (such as 0 or the pixel value of the image edge) at the image edge, so that the window can completely cover the image region. Each sub-image region is input into the preset lesion segmentation model matched therewith. The model can process the sub-image region and output the lesion segmentation result of the region, which is usually a binary image in which the lesion region is marked as 1 and the non-lesion region is marked as 0. The segmentation result of each sub-image region is stored for subsequent integration processing.

[0077] After obtaining the lesion segmentation results corresponding to each sub-image region, the reliability of these lesion segmentation results can be calculated. Reliability calculation aims to evaluate the accuracy and reliability of each sub-image region segmentation result. For example, the preset lesion segmentation model can output a confidence score. The confidence score reflects the degree of confidence of the model in segmenting the sub-image region. The higher the score, the higher the reliability. According to the reliability calculation result, a reliability weight is assigned to the segmentation result of each sub-image region. The segmentation result with high reliability will have greater influence in the subsequent integration process. Finally, the segmentation results of each sub-image region can be integrated according to the reliability weight calculation result. Weighted averaging, voting and other methods can be used. For example, for each pixel position, the segmentation results of all sub-image regions containing the pixel and their reliability weights are considered comprehensively, and the final segmentation value of the pixel is obtained by weighted averaging. In addition, post-processing can be performed on the integrated segmentation result to further improve the segmentation quality. Post-processing operations can include removing small isolated regions (which are likely to be noise or missegmentation), smoothing the segmentation boundary (making the boundary more natural), and the like.

[0078] The embodiments of the present application can better adapt to the characteristics of PET images under different tracers by constructing a matching preset lesion segmentation model for each tracer, improve the accuracy of segmentation; the sliding window mechanism can be adjusted according to the image size and computing resources, so that the scheme can process images of different sizes, and to some extent, the computing efficiency and segmentation accuracy are balanced; the segmentation results of sub-image regions are integrated through reliability calculation, which can reduce the occurrence of missegmentation and improve the reliability of the final segmentation result.

[0079] In the embodiments of the present application, optionally, step 104 comprises: receiving a target analysis requirement, analyzing the target analysis requirement to obtain a to-be-analyzed state, calling a state analysis model corresponding to the to-be-analyzed state, inputting the first index parameter and the second index parameter corresponding to each tracer into the state analysis model to obtain a state analysis result corresponding to the target object, and the to-be-analyzed state comprises at least one of a total survival state, a progression-free survival state, and a disease recurrence rate state.

[0080] In this embodiment, the target object can be analyzed according to the acquired first index parameters and the second index parameters corresponding to each tracer. Specifically, a target analysis requirement can be received. The target analysis requirement can come from a clinician, a researcher or other relevant personnel. For example, in a clinical study, a doctor can want to know the overall survival of patients with a certain disease in order to evaluate the effectiveness of a treatment plan; in the process of drug development, researchers can focus on the recurrence rate of a disease in order to judge the efficacy and safety of a new drug. The requirement can be presented in various forms, such as a text description, a specific request code or an option input through a user interface, etc. For example, a user can select "analyze overall survival status" as the target analysis requirement on a graphical user interface (GUI).

[0081] Next, the target analysis requirement is analyzed. The purpose of analysis is to extract key information from the received target analysis requirement and determine the state to be analyzed. Analysis can use natural language processing (NLP) techniques to analyze text requirements and identify keywords related to the state to be analyzed; for specific format request codes or interface options, the state to be analyzed can be determined directly through pre-set rules or mapping relationships. For example, when receiving the text requirement "please analyze the overall survival of patients", the NLP algorithm can identify the keyword "overall survival" to determine that the state to be analyzed is the overall survival state.

[0082] Different states to be analyzed correspond to different state analysis models. These models are pre-built based on a large amount of clinical data and related research, and each model is optimized for a specific state analysis task. The model can be built by a machine learning algorithm (such as logistic regression, random forest, etc.). After determining the state to be analyzed, the corresponding state analysis model is retrieved from the model library according to the pre-set mapping relationship. The model library stores various pre-trained models, each of which has a unique identifier and corresponding state to be analyzed information. By querying this information, the required state analysis model can be quickly and accurately found and retrieved.

[0083] Further, the first index parameters and the second index parameters corresponding to each tracer are arranged and input according to the format required by the model. After receiving the input parameters, the state analysis model calculates according to its internal learning algorithm and parameters. Specifically, the input parameters are subjected to feature extraction, weight allocation and combination, etc. Finally, the state analysis result corresponding to the target object is output. For example, for the overall survival state analysis model, the output result can be a predicted overall survival time range or survival probability; for the disease recurrence rate state analysis model, the output result can be a predicted value of the recurrence rate.

[0084] The embodiments of the present application can more accurately perform state analysis for specific problems, improve the accuracy and reliability of the analysis results, by calling the corresponding state analysis model according to different to-be-analyzed states; comprehensively consider the first index parameters and the second index parameters corresponding to each tracer, fully utilize the information of multi-source data, and can more comprehensively reflect the health status and disease characteristics of the target object; the entire process realizes automation from receiving requirements, analyzing requirements, calling models and outputting results, reduces manual intervention, and improves analysis efficiency and consistency.

[0085] In a specific embodiment, after the to-be-analyzed state is determined, the attribute values of the remaining attributes of the target object matched with the to-be-analyzed state can also be obtained, such as gender, age, smoking history, disease history, etc., and the required target index parameters are determined from the first index parameters and the second index parameters according to the to-be-analyzed state. Then, the attribute values and the target index parameters are taken as the input of the state analysis model corresponding to the to-be-analyzed state, and finally the desired prediction result is obtained.

[0086] Further, as Figure 1 For the specific implementation of the method, the embodiments of the present application provide an analysis system for multi-tracer images, as shown in Figure 2 The system comprises:

[0087] An image registration module is configured to obtain PET images of the target object under each tracer and CT images matched with the PET images, perform lesion segmentation on the PET images under each tracer respectively to obtain lesion segmentation results corresponding to the PET images, and perform organ segmentation on the CT images under each tracer respectively to obtain organ segmentation results corresponding to the CT images; based on the CT images under each tracer, perform registration processing on all CT images and PET images through a unified CT coordinate system to obtain a primary registration result, and perform secondary registration processing on the primary registration result based on the organ segmentation results corresponding to the CT images under each tracer to obtain a final registration result.

[0088] the image analysis module is configured to determine a visual overlap area of each tracer on the final registration result based on the final registration result, and calculate a first index parameter of the target object according to the visual overlap area, a lesion segmentation result of each PET image, and an organ segmentation result of each CT image, the first index parameter including a first whole-body level index parameter, a first organ level index parameter, and a first lesion level index parameter; and for each PET image under each tracer, calculate a second index parameter of the target object under the tracer according to a tracer visual area of the PET image, a lesion segmentation result, and an organ segmentation result of a CT image matched with the PET image, the second index parameter including a second whole-body level index parameter, a second organ level index parameter, and a second lesion level index parameter; and perform state analysis on the target object according to the first index parameter and the second index parameter corresponding to each tracer.

[0089] Optionally, the image registration module is further configured to:

[0090] take any one of the CT images under each tracer as a reference image, calculate an image deviation between each target CT image in the remaining CT images and the reference image respectively through a unified CT coordinate system, and calculate an adjustment parameter corresponding to the target CT image based on the image deviation, the adjustment parameter including a translation parameter, a scaling parameter, and a rotation parameter;

[0091] for each target CT image, sequentially perform image processing on the target CT image and a PET image matched with the target CT image according to the adjustment parameter corresponding to each target CT image, and align the processed image with the reference image and a PET image corresponding to the reference image, and obtain the preliminary registration result when each processed target CT image is aligned.

[0092] The image registration module is further configured to:

[0093] for each processed target CT image in the preliminary registration result, calculate a mapping relationship of each pixel point in the processed target CT image relative to a corresponding pixel point in the reference image based on an organ segmentation result of the processed target CT image and an organ segmentation result of the reference image, and perform re-registration on the processed target CT image and a corresponding processed PET image according to the mapping relationship, and obtain the final registration result when each processed target CT image and a corresponding processed PET image are re-registered.

[0094] Optionally, the image analysis module is configured to:

[0095] determine, based on the final registration result, a pixel set corresponding to a lesion position in the final-registered PET image under each tracer, the pixel set including coordinates of each pixel point included in the lesion position;

[0096] extract, based on the pixel sets corresponding to the tracers, common pixel points in the pixel sets, and determine, based on the common pixel points, a visualization overlap region of the final registration result.

[0097] Optionally, the image analysis module is further configured to:

[0098] calculate, according to the visualization overlap region, a mean SUV value, a maximum SUV value, an FDG metabolic tumor volume, and a total lesion glycolysis corresponding to the visualization overlap region, and take the mean SUV value, the maximum SUV value, the FDG metabolic tumor volume, and the total lesion glycolysis as the first whole-body level index parameter.

[0099] determine, according to the final registration result, organ registration results of the CT images, divide the visualization overlap region according to the organ registration results to obtain a first overlap sub-region corresponding to each organ, calculate, according to each first overlap sub-region corresponding to each organ, a mean SUV value, a maximum SUV value, an FDG metabolic tumor volume, and a total lesion glycolysis corresponding to the first overlap sub-region, and take the mean SUV value, the maximum SUV value, the FDG metabolic tumor volume, and the total lesion glycolysis corresponding to each first overlap sub-region as the first organ level index parameter.

[0100] determine, according to the final registration result, lesion registration results of the PET images, divide the visualization overlap region according to the lesion registration results to obtain a second overlap sub-region corresponding to each lesion, calculate, according to each second overlap sub-region corresponding to each lesion, a mean SUV value, a maximum SUV value, an FDG metabolic tumor volume, and a total lesion glycolysis corresponding to the second overlap sub-region, and take the mean SUV value, the maximum SUV value, the FDG metabolic tumor volume, and the total lesion glycolysis corresponding to each second overlap sub-region as the first lesion level index parameter.

[0101] Optionally, the PET image under each tracer is segmented based on a matched preset lesion segmentation model, and the preset lesion segmentation model under each tracer is constructed based on the following manner:

[0102] obtain a plurality of PET image samples corresponding to the target tracer, each of the PET image samples is marked with a lesion, and each lesion is marked with a lesion label;

[0103] calling an nnU-Net framework, performing multi-dimensional feature extraction on each PET image sample through the nnU-Net framework, determining a network depth, a network width, an encoder structure and a decoder structure of an initial lesion segmentation model based on the multi-dimensional feature extraction result, and constructing the initial lesion segmentation model according to the network depth, the network width, the encoder structure and the decoder structure;

[0104] performing down-sampling processing on each PET image sample respectively to obtain a down-sampled sample corresponding to each PET image sample, performing model training on the initial lesion segmentation model based on the down-sampled sample, and obtaining an intermediate lesion segmentation model;

[0105] inputting each PET image sample into the intermediate lesion segmentation model respectively to obtain a coarse-grained lesion segmentation map corresponding to each PET image sample, and performing up-sampling processing on the coarse-grained lesion segmentation map to obtain an up-sampled sample corresponding to each PET image sample;

[0106] performing retraining on the intermediate lesion segmentation model based on the up-sampled sample corresponding to each PET image sample, and obtaining the preset lesion segmentation model.

[0107] Optionally, the image registration module is further configured to:

[0108] for a PET image under each tracer, calling a preset lesion segmentation model matched with the tracer and a preset sliding window mechanism, acquiring a sub-image region of the PET image each time through the preset sliding window mechanism, inputting the sub-image region into the matched preset lesion segmentation model, and obtaining a lesion segmentation result corresponding to the sub-image region;

[0109] performing reliability calculation on the lesion segmentation result corresponding to each sub-image region of the PET image, and obtaining a lesion segmentation result corresponding to the PET image according to the reliability calculation result.

[0110] Optionally, the image analysis module is further configured to:

[0111] receiving a target analysis requirement, analyzing the target analysis requirement to obtain a to-be-analyzed state, calling a state analysis model corresponding to the to-be-analyzed state, inputting the first index parameter and the second index parameter corresponding to each tracer into the state analysis model, and obtaining a state analysis result corresponding to the target object, the to-be-analyzed state including at least one of a total survival time state, a progression-free survival time state and a disease recurrence rate state.

[0112] It should be noted that other corresponding descriptions of each functional unit involved in the analysis system for multi-tracer images provided in the embodiments of the present application can be referred toFigure 1 The corresponding descriptions in the method will not be repeated here.

[0113] This application also provides a computer device, which may specifically be a personal computer, a server, a network device, etc. Figure 3 As shown, the computer device includes a bus, a processor, memory, and a communication interface, and may also include an input / output interface and a display device. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system, computer programs, and a 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 location information. The network interface allows communication with external terminals via a network connection. When the computer program is executed by the processor, it implements the steps in the various method embodiments.

[0114] Those skilled in the art will understand that Figure 3 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0115] In one embodiment, a computer-readable storage medium is provided, which may be non-volatile or volatile, having stored thereon a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0116] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0117] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties.

[0118] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer readable storage medium, and when the computer program is executed, the processes of the above-mentioned embodiments of the methods can be included. Any reference to memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (Read-Only Memory, ROM), magnetic tape, floppy disk, flash memory, optical storage, high-density embedded non-volatile memory, resistive memory (ReRAM), magnetoresistive random access memory (Magnetoresistive Random Access Memory, MRAM), ferroelectric memory (Ferroelectric Random Access Memory, FRAM), phase change memory (Phase Change Memory, PCM), graphene memory, etc. Volatile memory can include random access memory (Random Access Memory, RAM) or external cache memory, etc. As an illustration but not limitation, RAM can be in various forms, such as static random access memory (Static Random Access Memory, SRAM) or dynamic random access memory (Dynamic Random Access Memory, DRAM), etc. The database involved in the embodiments provided in the present application can include at least one of a relational database and a non-relational database. The non-relational database can include a distributed database based on a block chain, etc., without being limited thereto. The processor involved in the embodiments provided in the present application can be a general-purpose processor, a central processing unit, a graphics processing unit, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., without being limited thereto.

[0119] Any combination of the technical features of the above embodiments can be made. In order to make the description simple, all possible combinations of the technical features in the above embodiments are not described, however, as long as the combination of the technical features does not exist contradictory, it should be considered as the scope of the present application.

[0120] The above embodiments only express several implementation manners of the present application, and the description is more specific and detailed, but it should not be understood as a limitation on the scope of the patent of the present application. It should be pointed out that for ordinary skilled in the art, without departing from the concept of the present application, a number of modifications and improvements can be made, which are within the scope of protection of the present application. Therefore, the protection scope of the present application should be subject to the appended claims.

Claims

1. A method for analyzing multi-tracer images, characterized in that, include: Acquire PET images of the target object under each tracer and CT images matching the PET images, and perform lesion segmentation on the PET images under each tracer to obtain the lesion segmentation results corresponding to each PET image, and perform organ segmentation on the CT images under each tracer to obtain the organ segmentation results corresponding to each CT image. Based on the CT images under each tracer, the entire CT image and PET image are registered using a unified CT coordinate system to obtain the primary registration result. Then, based on the organ segmentation results corresponding to the CT images under each tracer, the primary registration result is further registered to obtain the final registration result. Based on the final registration result, the overlapping regions of each tracer on the final registration result are determined, and a first index parameter of the target object is calculated according to the overlapping regions, the lesion segmentation results of each PET image, and the organ segmentation results of each CT image. The first index parameter includes a first systemic index parameter, a first organ-level index parameter, and a first lesion-level index parameter. Furthermore, for each PET image under each tracer, a second index parameter of the target object under the tracer is calculated according to the tracer-developed region of the PET image, the lesion segmentation result, and the organ segmentation result of the CT image matching the PET image. The second index parameter includes a second systemic index parameter, a second organ-level index parameter, and a second lesion-level index parameter. Based on the first indicator parameter and the second indicator parameter corresponding to each tracer, the state of the target object is analyzed.

2. The method according to claim 1, characterized in that, The CT images based on each tracer are registered using a unified CT coordinate system, yielding a preliminary registration result, including: Using any CT image from the CT images under each of the tracers as a reference image, and through a unified CT coordinate system, the image deviation between each target CT image in the remaining CT images and the reference image is calculated respectively. Based on the image deviation, the adjustment parameters corresponding to the target CT image are calculated. The adjustment parameters include translation parameters, scaling parameters, and rotation parameters. For each target CT image, the target CT image and the PET image matching the target CT image are processed sequentially according to the adjustment parameters corresponding to each target CT image. The processed image is then aligned with the reference image and the PET image corresponding to the reference image. After the processed images corresponding to each target CT image have been aligned, the primary registration result is obtained. The organ segmentation results corresponding to the CT images under each tracer are used to perform secondary registration processing on the primary registration results to obtain the final registration results, including: For each processed target CT image in the primary registration result, based on the organ segmentation results of the processed target CT image and the organ segmentation results of the reference image, the mapping relationship between each pixel in the processed target CT image and the corresponding pixel in the reference image is calculated. According to the mapping relationship, the processed target CT image and the corresponding processed PET image are re-registered. After all processed target CT images and the corresponding processed PET images are re-registered, the final registration result is obtained.

3. The method according to claim 1, characterized in that, The step of determining the overlapping development regions of each tracer on the final registration result includes: Based on the final registration result, a set of pixels corresponding to the lesion location in the final registered PET image for each tracer is determined, wherein the set of pixels includes the coordinates of each pixel point contained in the lesion location; Based on the pixel set corresponding to each tracer, common pixels in each pixel set are extracted, and the development overlap region of the final registration result is determined based on the common pixels.

4. The method according to claim 3, characterized in that, The calculation of the first index parameter of the target object based on the overlapping imaging region, the lesion segmentation results of each PET image, and the organ segmentation results of each CT image includes: Based on the overlapping imaging region, the mean SUV, maximum SUV, FDG-metabolic tumor volume, and total lesion glycolysis corresponding to the overlapping imaging region are calculated, and the mean SUV, maximum SUV, FDG-metabolic tumor volume, and total lesion glycolysis are used as the first systemic index parameters. Based on the final registration result, the organ registration result of each CT image is determined, and the imaging overlap region is divided according to the organ registration result to obtain the first overlap sub-region corresponding to each organ. Based on the first overlap sub-region corresponding to each organ, the mean SUV, maximum SUV, FDG-metabolic tumor volume, and total lesion glycolysis corresponding to the first overlap sub-region are calculated. The mean SUV, maximum SUV, FDG-metabolic tumor volume, and total lesion glycolysis corresponding to each first overlap sub-region are used as the first organ-level index parameters. Based on the final registration result, the lesion registration result of each PET image is determined, and the overlapping region is divided according to the lesion registration result to obtain the second overlapping sub-region corresponding to each lesion. Based on the second overlapping sub-region corresponding to each lesion, the mean SUV, maximum SUV, FDG-metabolic tumor volume, and total lesion glycolysis are calculated for the second overlapping sub-region. The mean SUV, maximum SUV, FDG-metabolic tumor volume, and total lesion glycolysis corresponding to each second overlapping sub-region are used as the first lesion-level index parameters.

5. The method according to claim 1, characterized in that, For each tracer, PET images were segmented based on a pre-defined lesion segmentation model. This pre-defined lesion segmentation model for each tracer was constructed as follows: Multiple PET image samples corresponding to the target tracer are obtained, each PET image sample is marked with a lesion, and each lesion is marked with a lesion label; The nnU-Net framework is invoked to extract multidimensional features from each PET image sample. Based on the multidimensional feature extraction results, the network depth, network width, encoder structure, and decoder structure of the initial lesion segmentation model are determined. The initial lesion segmentation model is then constructed based on the network depth, network width, encoder structure, and decoder structure. Each PET image sample is downsampled to obtain a downsampled sample for each PET image sample. The initial lesion segmentation model is trained based on the downsampled sample to obtain an intermediate lesion segmentation model. Each PET image sample is input into the intermediate lesion segmentation model to obtain a coarse-grained lesion segmentation map corresponding to each PET image sample. The coarse-grained lesion segmentation map is then upsampled to obtain an upsampled sample corresponding to each PET image sample. Based on the upsampled samples corresponding to each PET image sample, the intermediate lesion segmentation model is retrained to obtain the preset lesion segmentation model.

6. The method according to claim 5, characterized in that, The step of segmenting lesions in PET images under each tracer to obtain lesion segmentation results for each PET image includes: For each tracer-based PET image, a preset lesion segmentation model and a preset sliding window mechanism matching the tracer are retrieved. Through the preset sliding window mechanism, a sub-image region of the PET image is acquired each time, and the sub-image region is input into the matching preset lesion segmentation model to obtain the lesion segmentation result corresponding to the sub-image region. The reliability of the lesion segmentation results corresponding to each sub-image region of the PET image is calculated, and the lesion segmentation results corresponding to the PET image are obtained based on the reliability calculation results.

7. The method according to any one of claims 1 to 6, characterized in that, The step of performing state analysis on the target object based on the first index parameter and the second index parameter corresponding to each tracer includes: The system receives a target analysis request, parses the target analysis request to obtain the state to be analyzed, and retrieves the state analysis model corresponding to the state to be analyzed. The first indicator parameter and the second indicator parameter corresponding to each tracer are input into the state analysis model to obtain the state analysis result corresponding to the target object. The state to be analyzed includes at least one of the following: overall survival state, progression-free survival state, and disease recurrence rate state.

8. An analysis system for multi-tracer images, characterized in that, include: The image registration module is used to acquire PET images of the target object under each tracer and CT images matching the PET images, and to perform lesion segmentation on the PET images under each tracer to obtain lesion segmentation results for each PET image, and to perform organ segmentation on the CT images under each tracer to obtain organ segmentation results for each CT image; based on the CT images under each tracer, the module performs registration processing on all CT images and PET images through a unified CT coordinate system to obtain a primary registration result, and performs secondary registration processing on the primary registration result based on the organ segmentation results corresponding to the CT images under each tracer to obtain a final registration result; The image analysis module is used to determine the overlapping regions of each tracer on the final registration result, and calculate a first index parameter of the target object based on the overlapping regions, lesion segmentation results of each PET image, and organ segmentation results of each CT image. The first index parameter includes a first systemic index parameter, a first organ-level index parameter, and a first lesion-level index parameter. Furthermore, for each tracer-treated PET image, based on the tracer-treated regions and lesion segmentation results of the PET image, and the organ segmentation results of the CT image matching the PET image, the module calculates a second index parameter of the target object under the tracer. The second index parameter includes a second systemic index parameter, a second organ-level index parameter, and a second lesion-level index parameter. Based on the first index parameter and the second index parameter corresponding to each tracer, the module performs a state analysis on the target object.

9. A storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method of any one of claims 1 to 7.

10. A computer device, comprising a storage medium, a processor, and a computer program stored on the storage medium and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method of any one of claims 1 to 7.

Images