Image analysis method, device and computer equipment

Abstract

The application relates to an image analysis method, device and computer equipment. The method comprises the following steps: acquiring historical PET images, historical CT images, current PET images and current CT images of a subject; determining a historical analysis result according to the historical PET images, the historical CT images and a baseline model; and determining a current analysis result according to the current PET images, the current CT images, the historical analysis result and a follow-up model. The current analysis result comprises a third region of interest of the current CT images and / or a fourth region of interest of the current CT images corresponding to a first region of interest; the third region of interest corresponds to a fifth region of interest of the current PET images, and a first parameter of the fifth region of interest is greater than or equal to a first preset threshold; the fourth region of interest corresponds to the first region of interest, and the fourth region of interest corresponds to a sixth region of interest of the current PET images, and a first parameter of the sixth region of interest is less than the first preset threshold.

Description

Technical Field

[0001] This application relates to the field of medical technology, and in particular to an image analysis method, apparatus, and computer equipment. Background Technology

[0002] Doctors typically observe metabolic activity based on PET (Positron Emission Computed Tomography) images of patients, and then use this metabolic data from the PET images to identify regions of interest (ROIs) in the corresponding CT (Computed Tomography) images. For example, doctors observe the standard uptake value (SUV) in the PET images and map regions with an SUV greater than or equal to 2.5 from the PET images onto the corresponding CT images to determine the ROIs in the CT images.

[0003] However, current image analysis methods have limitations in their application scope. Summary of the Invention

[0004] Therefore, it is necessary to provide an image analysis method, apparatus, and computer equipment that can expand the application scope in response to the above-mentioned technical problems.

[0005] Firstly, this application provides an image analysis method. The method includes:

[0006] Acquire the subject's historical PET images, historical CT images, current PET images, and current CT images;

[0007] Based on the historical PET image, the historical CT image and the baseline model, the historical analysis results are determined; the historical analysis results include a first region of interest in the historical CT image, the first region of interest corresponds to a second region of interest in the historical PET image, and the first parameter of the second region of interest is greater than or equal to a first preset threshold.

[0008] Based on the current PET image, the current CT image, the historical analysis results, and the follow-up model, the current analysis result is determined; the current analysis result includes a third region of interest in the current CT image, and / or includes a fourth region of interest in the current CT image corresponding to the first region of interest; the third region of interest corresponds to a fifth region of interest in the current PET image, and the first parameter of the fifth region of interest is greater than or equal to the first preset threshold; the fourth region of interest corresponds to the first region of interest, and the fourth region of interest corresponds to a sixth region of interest in the current PET image, and the first parameter of the sixth region of interest is less than the first preset threshold.

[0009] In one embodiment, the determination of historical analysis results based on the historical PET images, the historical CT images, and the baseline model includes:

[0010] Based on the historical PET image and the historical CT image, the registered historical PET image is obtained;

[0011] The historical analysis results are determined based on the historical CT images, the registered historical PET images, and the baseline model.

[0012] In one embodiment, the determination of the historical analysis results based on the historical CT images, the registered historical PET images, and the baseline model includes:

[0013] Preprocessing the historical CT images yields preprocessed historical CT images.

[0014] The registered historical PET image is preprocessed to obtain the preprocessed historical PET image.

[0015] The preprocessed historical CT images and the preprocessed historical PET images are input into the baseline model to obtain the historical analysis results.

[0016] In one embodiment, the method further includes:

[0017] Obtain historical PET image samples and historical CT image samples;

[0018] The initial baseline model was trained based on historical PET image samples, historical CT image samples, and the corresponding first segmentation gold standard to obtain the baseline model.

[0019] The first segmentation gold standard includes a first region determined in the historical CT image sample, which corresponds to a second region in the historical PET image sample, and the first parameter of the second region is greater than or equal to the first preset threshold.

[0020] In one embodiment, the determination of the current analysis result based on the current PET image, the current CT image, the historical analysis results, and the follow-up model includes:

[0021] Based on the current PET image and the current CT image, the registered current PET image is obtained;

[0022] Based on the current CT image and the historical analysis results, the registered historical analysis results are obtained;

[0023] The current analysis result is determined based on the current CT image, the current PET image after registration, the historical analysis results after registration, and the follow-up model.

[0024] In one embodiment, the determination of the current analysis result based on the current CT image, the registered current PET image, the registered historical analysis results, and the follow-up model includes:

[0025] Preprocess the current CT image, the current PET image after registration, and the historical analysis results after registration to obtain the preprocessed current CT image, the preprocessed current PET image, and the preprocessed historical analysis results.

[0026] The preprocessed current CT image, the preprocessed current PET image, and the preprocessed historical analysis results are input into the follow-up model to obtain the current analysis result.

[0027] In one embodiment, the method further includes:

[0028] Obtain historical PET image samples, historical CT image samples, current PET image samples, and current CT image samples;

[0029] Based on the historical PET image sample, the historical CT image sample, and the baseline model, a historical CT segmentation result sample is determined; the historical CT segmentation result sample includes a first region in the historical CT image sample, the first region corresponds to a second region in the historical PET image sample, and the first parameter of the second region is greater than or equal to a first preset threshold.

[0030] Based on the historical CT segmentation result sample, the current PET image sample, the current CT image sample, and the second segmentation gold standard, the initial follow-up model is trained to obtain the follow-up model; the second segmentation gold standard includes a third region in the current CT image sample, and / or includes a fourth region in the current CT image sample corresponding to the first region; the third region corresponds to a fifth region in the current PET image sample, and the first parameter of the fifth region is greater than or equal to the first preset threshold; the fourth region corresponds to the first region, and the fourth region corresponds to a sixth region in the current PET image sample, and the first parameter of the sixth region is less than the first preset threshold.

[0031] In one embodiment, the current analysis result is determined based on the current CT image, the registered current PET image, the registered historical analysis results, and the follow-up model. Determining the current analysis result includes:

[0032] Based on the historical PET image and the current CT image, obtain the historical PET image registered to the current CT image;

[0033] The current analysis result is obtained based on the current CT image, the current PET image after registration, the historical analysis results after registration, the historical PET images registered to the current CT image, and the follow-up model.

[0034] In one embodiment, the method further includes:

[0035] Based on the historical analysis results and the current analysis results, the analysis results for the first parameter are determined.

[0036] In one embodiment, the method further includes:

[0037] The historical analysis results are mapped onto the historical PET image to obtain the first target region in the historical PET image.

[0038] Map the current analysis result to the current PET image to obtain the second target region in the current PET image;

[0039] Based on the first target region and the second target region, determine the analysis results of the second parameter.

[0040] In one embodiment, the method further includes:

[0041] Based on the historical analysis results and the current analysis results, the analysis results for the first parameter are determined.

[0042] In one embodiment, the method further includes:

[0043] The historical analysis results are mapped onto the historical PET image to obtain the first target region in the historical PET image.

[0044] Map the current analysis result to the current PET image to obtain the second target region in the current PET image;

[0045] Based on the first target region and the second target region, determine the analysis results of the second parameter.

[0046] Secondly, this application also provides an image analysis device. The device includes:

[0047] The acquisition module is used to acquire the subject's historical PET images, historical CT images, current PET images, and current CT images.

[0048] The first determining module is used to determine the historical analysis results based on the historical PET image, the historical CT image and the baseline model; the historical analysis results include a first region of interest in the historical CT image, the first region of interest corresponds to a second region of interest in the historical PET image, and a first parameter of the second region of interest is greater than or equal to a first preset threshold.

[0049] The second determining module is used to determine the current analysis result based on the current PET image, the current CT image, the historical analysis results, and the follow-up model; the current analysis result includes a third region of interest in the current CT image, and / or includes a fourth region of interest in the current CT image corresponding to the first region of interest; the third region of interest corresponds to a fifth region of interest in the current PET image, and the first parameter of the fifth region of interest is greater than or equal to the first preset threshold; the fourth region of interest corresponds to the first region of interest, and the fourth region of interest corresponds to a sixth region of interest in the current PET image, and the first parameter of the sixth region of interest is less than the first preset threshold.

[0050] Thirdly, this application also provides a computer device. The computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the steps of any of the methods described above.

[0051] Fourthly, this application also provides a computer-readable storage medium. The computer-readable storage medium stores a computer program thereon, which, when executed by a processor, implements the steps of any of the methods described above.

[0052] Fifthly, this application also provides a computer program product. The computer program product includes a computer program that, when executed by a processor, implements the steps of any of the methods described above.

[0053] The aforementioned image analysis method, apparatus, and computer equipment acquire the subject's historical PET images, historical CT images, current PET images, and current CT images. Based on the historical PET images, historical CT images, and a baseline model, historical analysis results are determined. Then, based on the current PET images, current CT images, historical analysis results, and a follow-up model, current analysis results are determined. Since the historical analysis results include a first region of interest in the historical CT images, and this first region of interest corresponds to a second region of interest in the historical PET images, with a first parameter greater than or equal to a first preset threshold for the second region of interest, the computer equipment can first determine the first region of interest in the subject's historical examinations where the first parameter is greater than or equal to the first preset threshold, for example, determining a first region of interest with an SUV ≥ 2.5 on the subject's historical CT images. Furthermore, since the current analysis result includes the third region of interest in the current CT image, and / or includes the fourth region of interest in the current CT image corresponding to the first region of interest, and the third region of interest corresponds to the fifth region of interest in the current PET image, the first parameter of the fifth region of interest is greater than or equal to the first preset threshold; the fourth region of interest corresponds to the first region of interest, and the fourth region of interest corresponds to the sixth region of interest in the current PET image, the first parameter of the sixth region of interest is less than the first preset threshold, therefore, even if the first parameter of the first region of interest is less than the first preset threshold in the current examination, the computer device can still determine the fourth region of interest. For example, if in the current examination, the SUV of the first region of interest with an SUV≥2.5 in the patient's historical CT images is less than 2.5, the computer device will still determine the fourth region of interest with an SUV<2.5. Thus, the method provided in this embodiment is not only applicable to the case where the first parameter is greater than or equal to the first preset threshold in every examination, but also applicable to the case where the first parameter is greater than or equal to the first preset threshold in the first examination, but the first parameter of the region with the first parameter greater than or equal to the first preset threshold is less than the first preset threshold in the second examination, thus expanding the scope of application. Attached Figure Description

[0054] Figure 1 This is a diagram illustrating the application environment of the image analysis method in the embodiments of this application;

[0055] Figure 2 This is a flowchart illustrating the image analysis method in an embodiment of this application;

[0056] Figure 3 This is a schematic diagram illustrating the usage or training process of a baseline model.

[0057] Figure 4 This is a schematic diagram illustrating the usage or training process of a follow-up model.

[0058] Figure 5 This is a schematic diagram of a process for determining historical analysis results in an embodiment of this application;

[0059] Figure 6 This is a schematic diagram of another process for determining historical analysis results in an embodiment of this application;

[0060] Figure 7 This is a schematic diagram of a process for obtaining a baseline model in an embodiment of this application;

[0061] Figure 8 This is a schematic diagram of a process for determining the current analysis result in an embodiment of this application;

[0062] Figure 9 This is a schematic diagram of another process for determining the current analysis result in an embodiment of this application;

[0063] Figure 10 This is a schematic diagram of a process for obtaining the current analysis result in an embodiment of this application;

[0064] Figure 11 This is a schematic diagram of a sample in an embodiment of this application;

[0065] Figure 12 This is a schematic diagram of another process for determining the current analysis result in an embodiment of this application;

[0066] Figure 13 This is a schematic diagram of a process for determining the analysis and results of the second parameter in an embodiment of this application;

[0067] Figure 14 This is a schematic diagram of the overall process of an image analysis method in an embodiment of this application;

[0068] Figure 15 This is a schematic diagram illustrating the usage process of an image analysis method in an embodiment of this application;

[0069] Figure 16 This is a structural block diagram of the image analysis device in the embodiments of this application. Detailed Implementation

[0070] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0071] Follow-up refers to the process by which hospitals periodically monitor changes in the condition of patients who have previously received treatment at that hospital; it is also known as follow-up examination. In follow-up analysis, taking a combination of PET and CT images as an example, doctors need to observe not only the mapping of regions with an SUV greater than or equal to 2.5 in the PET images to the regions of interest in the CT images, but also the metabolic changes in the regions of interest corresponding to the previous examination in the current examination. For example, aggressive lymphomas generally have high metabolic rates, and it is very likely that the SUV values ​​of regions of interest with an SUV ≥ 2.5 before this follow-up examination will be less than 2.5.

[0072] In other words, if an area with an SUV ≥ 2.5 in the previous PET image has an SUV less than 2.5 in the current re-examination, indicating a decrease in metabolism, but as long as the residual metabolism in that area is higher than the background metabolism of the surrounding area, the doctor still needs to combine the corresponding CT images and describe its size, SUV value, etc. in the report for comparative analysis with the results of the previous examination.

[0073] However, current image analysis techniques analyze each examination result independently. For example, the same analysis algorithm is run on PET and CT images from the previous examination and those from the current examination. Therefore, based on the re-examination PET and CT images, only regions with an SUV ≥ 2.5 in the corresponding PET image can be identified on the CT image. In other words, if a region with an SUV ≥ 2.5 in the previous PET image has an SUV < 2.5 in the current re-examination, it is impossible to determine the region with an SUV < 2.5 in the PET image and map it to the region of interest in the CT image. Therefore, the current analysis method has a limited scope of application.

[0074] Therefore, it is necessary to propose an image analysis method, which will be introduced below.

[0075] Figure 1 This is an application environment diagram of the image analysis method in the embodiments of this application. In this embodiment, a computer device is provided, which can be a terminal. Its internal structure diagram can be as follows: Figure 1As shown, the computer device includes a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, NFC (Near Field Communication), or other technologies. When the computer program is executed by the processor, it implements an image analysis method. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad mounted on the computer device casing, or an external keyboard, touchpad, or mouse.

[0076] Those skilled in the art will understand that Figure 1 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.

[0077] This embodiment illustrates the method by applying it to a terminal. It is understood that this method can also be applied to a server, and to a system that includes both a terminal and a server, and is implemented through interaction between the terminal and the server. The terminal can be, but is not limited to, various personal computers, laptops, smartphones, and tablets. The server can be a standalone server or a server cluster consisting of multiple servers.

[0078] Figure 2 This is a flowchart illustrating the image analysis method in an embodiment of this application. This method can be applied to... Figure 1 In one embodiment of the computer device shown, such as Figure 2 As shown, it includes the following steps:

[0079] S201, acquire the subject's historical PET images, historical CT images, current PET images, and current CT images.

[0080] PET images are images containing functional information. For example, doctors can use PET images to determine the SUV (Uniform Scale Value) of a target area, thereby determining the metabolic status of that area. PET (Positron Emission Computed Tomography) equipment can determine the corresponding SUV value based on parameters such as the patient's medication dosage, height, and weight when acquiring PET images, and then visualize this value on the PET image. The higher the SUV value on the PET image, the darker the image.

[0081] However, PET images cannot clearly show anatomical structures. Therefore, using PET images alone is not sufficient to accurately analyze areas where the first parameter is greater than or equal to a first preset threshold. CT (Computed Tomography) images, on the other hand, contain detailed anatomical information, thus allowing for image analysis by combining PET and CT images.

[0082] Therefore, in this embodiment, the computer device first acquires the subject's historical PET images, historical CT images, current PET images, and current CT images.

[0083] In this context, "current" and "historical" represent two relatively sequential time points, and this embodiment does not limit the time interval between them. Current PET images and current CT images represent images obtained during the current follow-up examination of the subject; historical PET images and historical CT images represent images obtained during examinations prior to the current follow-up examination of the subject. For example, current PET images and current CT images could be images obtained from the examination on the same day, while historical PET images and historical CT images could be images obtained from the examination one month prior.

[0084] It is understandable that the subject's historical PET images, historical CT images, current PET images, and current CT images all include at least one identical body part. To ensure the effectiveness of the analysis, historical PET images and historical CT images can be images acquired at the same historical time, and current PET images and current CT images can be images acquired at the same current time.

[0085] Optionally, the computer device can acquire historical PET images, historical CT images, current PET images, and current CT images collected by medical equipment, and can also receive historical PET images, historical CT images, current PET images, and current CT images sent by other electronic or medical devices. The computer device can also pre-store the above images for direct retrieval when the image analysis method needs to be executed.

[0086] S202, Based on historical PET images, historical CT images and baseline models, determine the historical analysis results; the historical analysis results include a first region of interest in historical CT images, the first region of interest corresponds to a second region of interest in historical PET images, and the first parameter of the second region of interest is greater than or equal to a first preset threshold.

[0087] In this embodiment, after the computer device acquires the subject's historical PET images, historical CT images, current PET images, and current CT images, it first determines the historical analysis results based on the historical PET images, historical CT images, and baseline model.

[0088] The following examples will all use SUV as the first parameter and 2.5 as the first preset threshold. It is understood that the first parameter can be other parameters that represent the body's metabolic status, and the first preset threshold can also be set to other values. This embodiment is not limited to these.

[0089] In other words, the computer equipment determines the second region of interest (ROI) with an SUV ≥ 2.5 in the historical PET images, and then maps the second ROI to the historical CT images to obtain the first ROI on the historical CT images.

[0090] Figure 3 This is a schematic diagram illustrating the usage or training process of a baseline model, such as... Figure 3 As shown, during use, Figure 3 (a) can represent historical CT images. Figure 3 (b) can represent historical PET images. Figure 3 (c) can indicate according to Figure 3 (a) and Figure 3 (b) Determined historical analysis results. It can be seen that... Figure 3 (b) shows that the darker developed area on the left side of the historical PET image has an SUV ≥ 2.5, which is the second region of interest in the historical PET image. Figure 3 (b) The boundaries and structure of the second region of interest are not clearly visible; therefore, the computer device will... Figure 3 (b) The second region of interest for SUVs ≥ 2.5 corresponds to Figure 3 In (a), thus determining Figure 3 (c) The first region of interest is shown in white within a white circular dashed box.

[0091] Optionally, the computer can directly input historical PET and CT images into the baseline model, from which the baseline model outputs historical analysis results. Alternatively, the computer can optimize the historical PET and CT images before inputting them into the baseline model, allowing the baseline model to output historical analysis results. Optimization processing can include registration, normalization, compression, etc.

[0092] Understandably, a baseline model needs to be trained on the computer before it can be used. This baseline model can be a target classification model, a convolutional neural network (CNN), a recurrent neural network (RNN), or other deep learning networks, machine learning networks, etc.

[0093] One way to obtain a baseline model is as follows: a computer trains an initial baseline model using historical PET image samples, historical CT image samples, and corresponding sample labels. Training can be stopped and the baseline model determined when a training stopping condition is met. The training stopping condition is that the difference between the analysis results output by the baseline model and the actual results in the historical PET and CT image samples is less than a preset threshold. The training process of the baseline model will be described in detail below and will not be repeated here.

[0094] S203, based on the current PET image, the current CT image, historical analysis results, and the follow-up model, determine the current analysis result; the current analysis result includes the third region of interest in the current CT image, and / or, includes the fourth region of interest in the current CT image corresponding to the first region of interest; the third region of interest corresponds to the fifth region of interest in the current PET image, and the first parameter of the fifth region of interest is greater than or equal to the first preset threshold; the fourth region of interest corresponds to the first region of interest, and the fourth region of interest corresponds to the sixth region of interest in the current PET image, and the first parameter of the sixth region of interest is less than the first preset threshold.

[0095] In this embodiment, once the computer device has determined the historical analysis results, it can determine the current analysis results based on the current PET image, the current CT image, the historical analysis results, and the follow-up model.

[0096] Continuing with the SUV example, there are three possibilities for the current analysis results. The first possibility is that the current analysis results include the third region of interest on the current CT image, which corresponds to the fifth region of interest in the current PET image with an SUV ≥ 2.5. In other words, it is still necessary to focus on the region with an SUV ≥ 2.5 in the patient's current follow-up images.

[0097] The second possibility is that the current analysis results include the fourth region of interest on the current CT image corresponding to the first region of interest, the fourth region of interest corresponding to the first region of interest with an SUV ≥ 2.5 in historical PET images, and the fourth region of interest corresponding to the sixth region of interest with an SUV < 2.5 in the current PET image. In other words, if the subject had a region with an SUV ≥ 2.5 when examined before the current follow-up, but the SUV in that region is now less than 2.5 at the current follow-up, then the portion of that region with an SUV < 2.5 still needs to be considered.

[0098] The third possibility is that the current analysis results include both the third and fourth regions of interest. For example, some areas of the subject have an SUV ≥ 2.5 in the current follow-up images, while other areas had an SUV ≥ 2.5 when examined before the current follow-up, but an SUV < 2.5 in the current follow-up images.

[0099] Figure 4 This is a schematic diagram illustrating the usage or training process of a follow-up model, such as... Figure 4 As shown, during use, Figure 4 (a) can represent historical CT images. Figure 4 (b) can represent historical PET images. Figure 4 (c) can represent based on the current CT images. Figure 4 (d) can represent the current PET image. Figure 4 (e) can represent a definite current analysis result.

[0100] It can be seen that the darker area in the lower left of the subject's historical PET images is the second region of interest in that historical PET image with an SUV ≥ 2.5. Figure 4 (b) The boundaries and structure of the second region of interest are not clearly visible; therefore, the historical analysis results determined by the computer equipment are... Figure 4 (a) The first region of interest is indicated by a white circular dashed box. The first region of interest is also the region that will be... Figure 4 (b) The second region of interest for SUVs ≥ 2.5 is mapped to Figure 4 (a) The region obtained after that.

[0101] Furthermore, combined Figure 4(d) It can be seen that the contrast in the lower left corner of the subject's current PET image is no longer obvious, with an SUV value of 1.7. This means that when the subject is examined again after a period of time, the SUV value of the sixth region of interest (ROI) in the current PET image, corresponding to the second ROI in the historical PET image, has decreased. Similarly, the fourth region of interest (ROI) in the current CT image, corresponding to the first ROI in the historical PET image, has also decreased. Although the regions have shrunk and metabolism has decreased, doctors still need to pay attention.

[0102] Therefore, the computer equipment can still map the first region of interest (ROI) with an SUV ≥ 2.5 in the historical PET images from the previous examination to the current examination, thus identifying the sixth region of interest (ROI) where the SUV was ≥ 2.5 in the previous examination but < 2.5 in the current PET image. This sixth ROI is then mapped onto the current CT image to obtain the fourth ROI. In other words, the computer equipment can determine... Figure 4 (e) The white fourth region of interest is indicated by a white circular dashed box.

[0103] Optionally, the computer device can directly input the current PET images, current CT images, and historical analysis results into the follow-up model, so that the follow-up model can output the current analysis results. Alternatively, the computer device can optimize the current PET images, current CT images, and historical analysis results before inputting them into the follow-up model, so that the follow-up model can output the historical analysis results.

[0104] Similarly, before using a follow-up model, it must first be trained. The follow-up model can be a target classification model, such as a Convolutional Neural Network (CNN), a Recurrent Neural Network (RNN), or other deep learning networks, machine learning networks, etc.

[0105] The process of training the follow-up model is similar to that of obtaining the baseline model. The training process of the follow-up model will be described in detail below, and will not be repeated here.

[0106] In this way, doctors can compare and contrast historical and current analysis results. In some embodiments, the computer device can register and highlight the historical and current analysis results to facilitate comparison and analysis by doctors.

[0107] The image analysis method provided in this embodiment acquires the subject's historical PET images, historical CT images, current PET images, and current CT images. Based on the historical PET images, historical CT images, and a baseline model, it determines the historical analysis results. Then, based on the current PET images, current CT images, historical analysis results, and a follow-up model, it determines the current analysis results. Since the historical analysis results include a first region of interest in the historical CT images, and the first region of interest corresponds to a second region of interest in the historical PET images, with a first parameter greater than or equal to a first preset threshold, the computer device can first determine the first region of interest in the subject's historical examinations where the first parameter is greater than or equal to the first preset threshold, for example, determining a first region of interest with an SUV ≥ 2.5 on the subject's historical CT images. Furthermore, since the current analysis result includes the third region of interest in the current CT image, and / or includes the fourth region of interest in the current CT image corresponding to the first region of interest, and the third region of interest corresponds to the fifth region of interest in the current PET image, the first parameter of the fifth region of interest is greater than or equal to the first preset threshold; the fourth region of interest corresponds to the first region of interest, and the fourth region of interest corresponds to the sixth region of interest in the current PET image, the first parameter of the sixth region of interest is less than the first preset threshold, therefore, even if the first parameter of the first region of interest is less than the first preset threshold in the current examination, the computer device can still determine the fourth region of interest. For example, if in the current examination, the SUV of the first region of interest with an SUV≥2.5 in the patient's historical CT images is less than 2.5, the computer device will still determine the fourth region of interest with an SUV<2.5. Thus, the method provided in this embodiment is not only applicable to the case where the first parameter is greater than or equal to the first preset threshold in every examination, but also applicable to the case where the first parameter is greater than or equal to the first preset threshold in the first examination, but the first parameter of the region with the first parameter greater than or equal to the first preset threshold is less than the first preset threshold in the second examination, thus expanding the scope of application.

[0108] Figure 5 This is a flowchart illustrating a process for determining historical analysis results in an embodiment of this application. (Refer to...) Figure 5 This embodiment relates to an optional implementation method for determining historical analysis results. Based on the above embodiment, step S202, which determines historical analysis results based on historical PET images, historical CT images, and a baseline model, includes the following steps:

[0109] S501, based on historical PET images and historical CT images, obtains the registered historical PET images.

[0110] In some applications, PET-CT dual-mode imaging can be used to obtain historical PET and CT images. In PET-CT dual-mode imaging, the PET and CT detectors are mounted on the same gantry and share the same scanning bed, allowing the patient to undergo a PET scan immediately after a CT scan. Although PET-CT dual-mode imaging can acquire images in two different modalities, the scan times differ because the two detectors still scan separately, resulting in inconsistencies in the spatial location of the same tissue or organ in the two images.

[0111] Therefore, in this embodiment, in order to improve the accuracy of image analysis, after the computer device acquires historical PET images and historical CT images, it first obtains the registered historical PET images based on the historical PET images and historical CT images.

[0112] Registration refers to mapping images from different imaging methods to the same coordinate system through spatial transformation, so that the images of corresponding tissues and organs achieve spatial consistency. Registration methods include, but are not limited to, rigid registration algorithms or flexible registration methods.

[0113] Optionally, the computer equipment uses historical CT images as a reference to register historical PET images onto historical CT images, so that the registered historical PET images have the same spatial resolution and position as the historical CT images.

[0114] S502 determines the historical analysis results based on historical CT images, registered historical PET images, and baseline models.

[0115] In this embodiment, after the computer device obtains the registered historical PET images, it determines the historical analysis results based on the historical CT images, the registered historical PET images, and the baseline model.

[0116] Optionally, the computer device can directly input historical CT images and registered historical PET images into the baseline model, so that the baseline model can output historical analysis results. Alternatively, the computer device can preprocess the historical CT images and registered historical PET images before inputting them into the baseline model, so that the baseline model can output historical analysis results.

[0117] In this embodiment, the registered historical PET images are obtained based on historical PET images and historical CT images, thus improving the matching degree between the registered historical PET images and historical CT images. In this way, based on historical CT images, registered historical PET images, and baseline models, more accurate historical analysis results can be determined.

[0118] Figure 6This is a schematic diagram of another process for determining historical analysis results in the embodiments of this application, referred to... Figure 6 This embodiment relates to an optional implementation of how to determine historical analysis results. Based on the above embodiment, the determination of historical analysis results includes the following steps:

[0119] S601, preprocessing historical CT images to obtain preprocessed historical CT images.

[0120] In this embodiment, when the computer device needs to determine the historical analysis results based on historical CT images, registered historical PET images, and baseline models, it will preprocess the historical CT images to obtain preprocessed historical CT images.

[0121] Preprocessing includes normalization, compression, and standardization. This embodiment does not limit the specific method of preprocessing, as long as it can reduce the amount of data.

[0122] For example, taking preprocessing including normalization as an example. The computer device sets the window level (wl) and window width (ww), and uses the window level and window width to truncate the gray values ​​of the historical CT images according to the following formula (1), so that the values ​​of each pixel in the historical CT images are limited to [wl-ww / 2, wl+ww / 2].

[0123]

[0124] Furthermore, the computer device normalizes the value of each pixel in the truncated historical CT image to [-1, 1] according to the following formula (2).

[0125] Among them, ww=400, wl=40, and ww and wl can also be adjusted according to needs.

[0126]

[0127] S602, preprocess the registered historical PET images to obtain preprocessed historical PET images.

[0128] In this embodiment, the computer device further preprocesses the registered historical PET images to obtain preprocessed historical PET images. The preprocessing process can be referred to in S601, and will not be repeated here. When normalizing the registered historical PET images, ww = 5 and wl = 2.5, which can also be adjusted as needed.

[0129] S603 inputs the preprocessed historical CT images and preprocessed historical PET images into the baseline model to obtain the historical analysis results.

[0130] In this embodiment, after obtaining the preprocessed historical CT images and preprocessed historical PET images, the computer device can input the preprocessed historical CT images and preprocessed historical PET images into the baseline model to obtain the historical analysis results output by the baseline model.

[0131] This embodiment requires preprocessing historical CT images to obtain preprocessed historical CT images, and preprocessing registered historical PET images to obtain preprocessed historical PET images. Therefore, it can reduce the amount of data input to the baseline model, thereby improving the efficiency of determining historical analysis results.

[0132] The process of training to obtain the baseline model will be described below. Figure 7 This is a schematic diagram of a process for obtaining a baseline model in an embodiment of this application. (Refer to...) Figure 7 This embodiment relates to an optional implementation of how to obtain a baseline model. Based on the above embodiment, the image analysis method further includes the following steps:

[0133] S701, acquire historical PET image samples and historical CT image samples.

[0134] In this embodiment, during the training phase to obtain the baseline model, the computer device acquires historical PET image samples and historical CT image samples. It is understood that, to ensure the accuracy of the baseline model, the historical PET image samples and historical CT image samples should include PET and CT images of different subjects at different historical times.

[0135] Furthermore, there is a correspondence between historical PET image samples and historical CT image samples. For example, if historical PET image sample 1 is a PET image of subject 1 three months ago, then the historical CT image sample 1 corresponding to historical PET image sample 1 also needs to be a CT image of subject 1 three months ago, and historical PET image sample 1 and historical CT image sample 1 include at least one of the same locations.

[0136] S702, the initial baseline model is trained based on historical PET image samples, historical CT image samples and the corresponding first segmentation gold standard to obtain the baseline model; wherein, the first segmentation gold standard includes a first region determined by historical CT image samples, the first region corresponds to a second region in historical PET image samples, and the first parameter of the second region is greater than or equal to a first preset threshold.

[0137] In this embodiment, after acquiring historical PET image samples and historical CT image samples, the computer device also needs to determine the corresponding first segmentation gold standard. The first segmentation gold standard is also known as the segmentation mask based on the standard annotation of the baseline model.

[0138] Please continue to refer to this. Figure 3 During the training process, Figure 3 (a) can be a historical CT image sample. Figure 3 (b) If it can be a historical PET image sample, then the first segmentation gold standard is as follows: Figure 3 As shown in (d). That is, the computer device determines a second region with an SUV ≥ 2.5 in the historical PET image sample, maps the second region to the historical CT image sample, or annotates the historical CT image sample according to the second region to obtain a first region on the historical CT image sample, and uses the first region as the first segmentation gold standard.

[0139] Furthermore, the computer equipment uses historical PET and CT image samples as samples and the corresponding first segmentation gold standard as labels to train the initial baseline model and obtain the baseline model. For example, the computer equipment can stop training to obtain the baseline model if the difference between the historical analysis results predicted by the initial baseline model and the actual first segmentation gold standard is less than a preset difference.

[0140] Optionally, the computer equipment can obtain registered historical PET image samples based on historical PET image samples and historical CT image samples, and train the initial baseline model based on historical CT image samples, registered historical PET image samples, and the corresponding first segmentation gold standard.

[0141] Optionally, the computer device can preprocess the historical CT image samples to obtain preprocessed historical CT image samples, and preprocess the registered historical PET image samples to obtain preprocessed historical PET image samples. Then, based on the preprocessed historical CT image samples, the preprocessed historical PET image samples, and the corresponding first segmentation gold standard, the initial baseline model is trained to obtain the baseline model.

[0142] Taking VB-Net as an example, the computer can obtain historical CT and PET image samples. Based on these samples, it can generate registered historical PET image samples, normalize both samples, and then input them into VB-Net to train the initial baseline model, resulting in the final baseline model. The first segmentation gold standard serves as the learning label during the training process of the initial baseline model.

[0143] This embodiment acquires historical PET image samples and historical CT image samples, and trains an initial baseline model based on the historical PET image samples, historical CT image samples, and the corresponding first segmentation gold standard to obtain a baseline model. Since the first segmentation gold standard includes a first region determined in the historical CT image samples, and the first region corresponds to a second region in the historical PET image samples, and the first parameter of the second region is greater than or equal to a first preset threshold, the first region of interest on the historical CT image with the first parameter greater than or equal to the first preset threshold can be determined based on the baseline model.

[0144] Let's continue with the process of determining the current analysis results. Figure 8 This is a schematic diagram of a process for determining the current analysis result in an embodiment of this application. (Refer to...) Figure 8 This embodiment relates to an optional implementation of how to determine the current analysis result. Based on the above embodiment, step S203, which determines the current analysis result according to the current PET image, current CT image, historical analysis results, and follow-up model, includes the following steps:

[0145] S801: Based on the current PET image and the current CT image, obtain the registered current PET image.

[0146] S802, based on the current CT images and historical analysis results, obtains the historical analysis results after registration.

[0147] S803 determines the current analysis results based on the current CT images, the current PET images after registration, the historical analysis results after registration, and the follow-up model.

[0148] In this embodiment, following the same principle as S501, the computer device, after determining the current analysis result based on the current PET image, the current CT image, historical analysis results, and the follow-up model, first obtains the registered current PET image based on the current PET image and the current CT image, and then obtains the registered historical analysis result based on the current CT image and the historical analysis results.

[0149] Optionally, the computer equipment uses the current CT image as a reference to register the current PET image onto the current CT image, and registers the historical analysis results onto the current CT image, so that the registered current PET image and the current CT image have the same spatial resolution and position, and the historical analysis results and the current CT image have the same spatial resolution and position.

[0150] This embodiment improves the matching degree among the current CT image, the registered current PET image, and the registered historical analysis result by obtaining the registered current PET image based on the current CT image and the current CT image and historical analysis results. Thus, based on the current CT image, the registered current PET image, the registered historical analysis result, and the follow-up model, the current analysis result can be determined more accurately.

[0151] Figure 9 This is a schematic diagram of another process for determining the current analysis result in an embodiment of this application. (Refer to...) Figure 9 This embodiment relates to an optional implementation of how to determine the current analysis result. Based on the above embodiment, step S803, which determines the current analysis result based on the current CT image, the registered current PET image, the registered historical analysis results, and the follow-up model, includes the following steps:

[0152] S901 preprocesses the current CT image, the registered current PET image, and the registered historical analysis results to obtain the preprocessed current CT image, the preprocessed current PET image, and the preprocessed historical analysis results.

[0153] S902 inputs the preprocessed current CT image, the preprocessed current PET image, and the preprocessed historical analysis results into the follow-up model to obtain the current analysis results.

[0154] In this embodiment, when the computer device obtains the current analysis result, it needs to preprocess the current CT image, the registered current PET image, and the registered historical analysis result respectively to obtain the preprocessed current CT image, the preprocessed current PET image, and the preprocessed historical analysis result.

[0155] Preprocessing can be referred to in S601, and will not be repeated here. For example, when normalizing the current CT image, ww = 400, wl = 40; when normalizing the registered current PET image, ww = 5, wl = 2.5; when normalizing the registered historical analysis results, ww = 1, wl = 0.5.

[0156] Furthermore, by inputting the preprocessed current CT images, preprocessed current PET images, and preprocessed historical analysis results into the follow-up model, the historical analysis results output by the follow-up model can be obtained. Moreover, since preprocessing is required before inputting into the follow-up model, the amount of data input can be reduced, thereby improving the efficiency of determining the current analysis results.

[0157] The following will describe the process of training and obtaining the follow-up model. Figure 10 This is a schematic diagram of a process for obtaining the current analysis result in an embodiment of this application. (Refer to...) Figure 10 This embodiment relates to an optional implementation of how to obtain the current analysis result. Based on the above embodiment, S802, according to the current CT image, the registered current PET image, the registered historical analysis results, and the follow-up model, determines the current analysis result, including the following steps:

[0158] S1001, acquire historical PET image samples, historical CT image samples, current PET image samples, and current CT image samples.

[0159] In this embodiment, during the training phase to obtain the follow-up model, the computer device acquires historical PET image samples, historical CT image samples, current PET image samples, and current CT image samples.

[0160] It is understandable that historical PET and CT image samples are earlier than current PET and CT image samples.

[0161] Furthermore, there are corresponding relationships between historical PET image samples, historical CT image samples, current PET image samples, and current CT image samples. For example, if historical PET image sample 1 is a PET image of subject 1 from three months ago, then the corresponding historical CT image sample 1 must also be a CT image of subject 1 from three months ago; if current PET image sample 1 is a PET image of subject 1 from yesterday, then the corresponding current CT image sample 1 must also be a CT image of subject 1 from yesterday. Moreover, historical PET image sample 1, historical PET image sample 1, current PET image sample 1, and current CT image sample 1 all include at least one identical body part.

[0162] S1002, Based on historical PET image samples, historical CT image samples and baseline model, determine historical CT segmentation result samples; historical CT segmentation result samples include a first region in historical CT image samples, the first region corresponds to a second region in historical PET image samples, and the first parameter of the second region is greater than or equal to a first preset threshold.

[0163] In this embodiment, a baseline model needs to be trained before training the follow-up model. Referring to the description of the baseline model in the above embodiment, after obtaining historical PET image samples, historical CT image samples, current PET image samples, and current CT image samples, the computer device will first determine the historical CT segmentation result samples based on the historical PET image samples, historical CT image samples, and the baseline model.

[0164] Figure 11 This is a sample illustration from an embodiment of this application, such as... Figure 11 As shown, Figure 11 (a) is a sample of historical CT images. Figure 11 (b) is a historical PET image sample. Figure 11 (c) is the current CT image sample. Figure 11 (d) Current PET image sample.

[0165] because Figure 11 (b) The black-developed SUV on the right side is ≥2.5. Figure 11 (b) The second region corresponding to SUVs ≥ 2.5 corresponds to Figure 11 In (a), it can be determined that Figure 11 (a) shows the first area indicated by the white circular dashed box. Further, the computer device will... Figure 11 The first region in (a) is used as a historical CT segmentation result sample. That is, the first region is the part of the historical CT image sample where the boundary can be clearly seen, corresponding to the second region with SUV≥2.5 in the current PET image sample.

[0166] S1003, the initial follow-up model is trained based on historical CT segmentation result samples, current PET image samples, current CT image samples, and the second segmentation gold standard to obtain the follow-up model; the second segmentation gold standard includes the third region in the current CT image sample, and / or includes the fourth region in the current CT image sample corresponding to the first region; the third region corresponds to the fifth region in the current PET image sample, and the first parameter of the fifth region is greater than or equal to the first preset threshold; the fourth region corresponds to the first region, and the fourth region corresponds to the sixth region in the current PET image sample, and the first parameter of the sixth region is less than the first preset threshold.

[0167] In this embodiment, the computer device can determine the corresponding second segmentation gold standard based on the first segmentation gold standard corresponding to the historical CT image sample, the current PET image sample, and the current CT image sample.

[0168] The second segmentation gold standard includes not only the third region with an SUV ≥ 2.5 in the current PET image sample corresponding to the current CT image sample, but also the fourth region, which is a region with slight metabolism in the current PET image sample corresponding to the first segmentation gold standard based on historical CT image samples, and whose boundary can be clearly seen in the current CT image sample.

[0169] In other words, on the one hand, the computer equipment will determine the fifth region with an SUV ≥ 2.5 on the current PET image sample and map the fifth region to the current CT image sample to obtain the third region with a clearly visible boundary in the current CT image sample; on the other hand, the computer equipment will also map the first region in the first segmentation gold standard corresponding to the historical CT image sample to the current PET image sample to determine the sixth region with an SUV < 2.5 on the current PET image sample and map the sixth region to the current CT image sample to obtain the fourth region with a clearly visible boundary in the current CT image sample.

[0170] Thus, the computer equipment not only identified a third region with an SUV ≥ 2.5 on the current CT image sample, but also a fourth region with an SUV ≥ 2.5 on historical CT image samples, but an SUV less than 2.5 on the current CT image sample. Furthermore, the computer equipment used the third and fourth regions as the second segmentation gold standard.

[0171] Please continue to refer to this. Figure 11 , combined Figure 11 (b) and Figure 11 (d) It can be seen that, Figure 11 (d) The portion of the original SUV with a displacement of ≥2.5 corresponds to Figure 11 (d) The SUV value decreases, that is, Figure 11 In (d), the SUV of the area visible on the right is less than 2.5. However, because the doctor still needs to monitor this area, the computer equipment will... Figure 11 The visualized portion in (d) is taken as the sixth region, and the sixth region is mapped onto the current CT image sample to obtain... Figure 11 (e) The white area enclosed by the white dotted circle is designated as the fourth region. Furthermore, the computer device uses this fourth region as the second segmentation gold standard, such as... Figure 11 As shown in (f).

[0172] Similarly, please refer to Figure 4 During the training process, Figure 4 (a) is a sample of historical CT images. Figure 4 (b) is a historical PET image sample. Figure 4 (c) is the current CT image sample. Figure 4 (d) Current PET image sample.

[0173] like Figure 4 As shown in the lower left developed portion of (b), a second region with an SUV ≥ 2.5 exists in historical PET image samples, while the SUV in the current PET image sample is < 2.5. Therefore, the second region becomes smaller after being mapped to the current PET image sample, as shown below. Figure 4 As shown in (d).

[0174] Furthermore, when training the follow-up model, the computer equipment will consider SUV ≥ 2.5 in Figure (b), but... Figure 4 The portion of SUV < 2.5 in (d) corresponds to the current CT image sample to obtain Figure 4 The fourth region shown in (e) is used as the second dividing gold standard, such as Figure 4 As shown in (f).

[0175] Furthermore, the computer device uses historical CT segmentation results samples, current PET image samples, and current CT image samples as samples, and the corresponding second segmentation gold standard as labels, to train the initial follow-up model and obtain the follow-up model. For example, the computer device can stop training to obtain the follow-up model if the difference between the current analysis result predicted by the initial follow-up model and the actual second segmentation gold standard is less than a preset difference.

[0176] Optionally, based on the current PET image sample and the current CT image sample, a registered current PET image sample is obtained, and based on the current CT image sample and the historical CT segmentation result sample, a registered historical CT segmentation result sample is obtained. Thus, based on the current CT image sample, the registered current PET image sample, the registered historical CT segmentation result sample, and the corresponding second segmentation gold standard, the initial follow-up model is trained to obtain the follow-up model.

[0177] Optionally, the computer device can preprocess the current CT image sample, the registered current PET image sample, and the registered historical CT segmentation result sample respectively to obtain the preprocessed current CT image sample, the preprocessed current PET image sample, and the preprocessed historical CT segmentation result sample. The preprocessed current CT image sample, the preprocessed current PET image sample, the preprocessed historical CT segmentation result sample, and the corresponding second segmentation gold standard are used to train the follow-up model for the initial follow-up model.

[0178] Furthermore, the computer device can obtain a registered current PET image sample based on the current PET image sample and the current CT image sample, and obtain a registered historical CT segmentation result sample based on the current CT image sample and historical CT segmentation result samples. After normalizing the current CT image sample, the registered current PET image sample, and the registered historical CT segmentation result sample respectively, these are input into the initial follow-up model for training to obtain the final required follow-up model. In some embodiments, the initial follow-up model can also use VB-Net. The second segmentation gold standard serves as the learning label during the training process of the initial follow-up model.

[0179] This embodiment acquires historical PET image samples, historical CT image samples, current PET image samples, and current CT image samples. Based on these samples and a baseline model, it determines historical CT segmentation result samples. Then, based on these samples, the current PET image samples, the current CT image samples, and a second segmentation gold standard, it trains an initial follow-up model to obtain a follow-up model. Since the second segmentation gold standard includes a third region in the current CT image sample and / or a fourth region in the current CT image sample corresponding to the first region, wherein the third region corresponds to a fifth region in the current PET image sample, and the first parameter of the fifth region is greater than or equal to a first preset threshold; and the fourth region corresponds to the first region and also to a sixth region in the current PET image sample, and the first parameter of the sixth region is less than the first preset threshold, the third region of interest and / or the fourth region of interest on the current CT image can be determined based on the follow-up model.

[0180] Figure 12 This is a schematic diagram of another process for determining the current analysis result in an embodiment of this application. (Refer to...) Figure 12 This embodiment relates to an optional implementation of how to determine the current analysis result. Based on the above embodiment, step S803, which determines the current analysis result based on the current CT image, the registered current PET image, the registered historical analysis results, and the follow-up model, may further include the following steps:

[0181] S1201: Based on historical PET images and current CT images, obtain historical PET images registered to the current CT images.

[0182] In this embodiment, the computer device can also obtain a historical PET image registered to the current CT image based on historical PET images and the current CT image. That is, the computer device uses the current CT image as a reference to register the historical PET image to the current CT image to obtain a historical PET image registered to the current CT image.

[0183] S1202: Based on the current CT image, the current PET image after registration, the historical analysis results after registration, the historical PET images registered to the current CT image, and the follow-up model, the current analysis results are obtained.

[0184] In this embodiment, after S1201, the computer device can obtain the current analysis result based on the current CT image, the current PET image after registration, the historical analysis result after registration, the historical PET images registered to the current CT image, and the follow-up model.

[0185] Optionally, the computer device can directly input the current CT image, the registered current PET image, the registered historical analysis results, and the historical PET images registered to the current CT image into the follow-up model to obtain the current analysis results output by the follow-up model. Alternatively, the computer device can preprocess the current CT image, the registered current PET image, the registered historical analysis results, and the historical PET images registered to the current CT image before inputting them into the follow-up model to obtain the current analysis results output by the follow-up model.

[0186] Understandably, registering historical PET images to the current CT images can further improve the accuracy of the current analysis results output by the follow-up model. Of course, in this case, historical PET image samples registered to the current CT image samples also need to be used when training the follow-up model. For example... Figure 12 As shown, when training the follow-up model, it is necessary not only to normalize the current CT image sample, the registered current PET image sample, the registered historical CT segmentation result sample, and the corresponding second segmentation gold standard before inputting them into the initial follow-up model, but also to normalize the historical PET image sample registered to the current CT image sample before inputting it into the initial follow-up model, so as to train the initial follow-up model to obtain the final follow-up model.

[0187] In this embodiment, it is necessary to obtain historical PET images registered to the current CT image based on historical PET images and the current CT image. Combining historical PET images allows analysis of the first parameter in previous examinations. Therefore, registering historical PET images to the current CT image helps the follow-up model determine the current analysis result. Thus, the current analysis result obtained based on the current CT image, the registered current PET image, the registered historical analysis result, the historical PET image registered to the current CT image, and the follow-up model will be more accurate.

[0188] In one embodiment, optionally, the above image analysis method further includes the following steps:

[0189] Based on the historical and current analysis results, the analysis result of the first parameter is determined.

[0190] The first parameter analysis result represents the comparison between historical analysis results and current analysis results. It is used to reflect the changes in the size, shape and other structural information of the area of ​​concern to the doctor in the patient's historical CT images and current CT images. It includes, but is not limited to, at least one of the following forms: text, image, and chart.

[0191] Optionally, the computer device can display historical analysis results, current analysis results, and first parameter analysis results, so that doctors can perform subsequent analysis based on CT images in a timely manner.

[0192] Because the first parameter analysis result can be determined based on historical and current analysis results, doctors can easily compare and analyze the structural information such as size and shape between the historical and current analysis results.

[0193] Figure 13 This is a flowchart illustrating the process of determining the analysis and results of the second parameter in an embodiment of this application. (Refer to...) Figure 13 This embodiment relates to an optional implementation of how to determine the current analysis result. Based on the above embodiment, the image analysis method may further include the following steps:

[0194] S1301, Map the historical analysis results to historical PET images to obtain the first target region in the historical PET images.

[0195] S1302, map the current analysis results to the current PET image to obtain the second target region in the current PET image.

[0196] S1303, Determine the analysis results of the second parameter based on the first target region and the second target region.

[0197] In this embodiment, the computer device can employ a rigid registration algorithm or a flexible registration method to map historical analysis results to historical PET images and current analysis results to current PET images. The mapping principle can refer to the registration process in the above embodiments.

[0198] Furthermore, continuing with the first parameter as the SUV example, after mapping historical analysis results to historical PET images and current analysis results to current PET images, the first target region represents the area with an SUV ≥ 2.5 in the subject's historical PET images; the second target region represents the area with an SUV ≥ 2.5 in the subject's current PET images, and / or the area with an SUV ≥ 2.5 in historical PET images but an SUV < 2.5 in the current PET image. In other words, both historical and current PET images only contain high metabolic information from abnormal areas of interest to the physician, and do not contain redundant information such as high metabolic information from normal areas. Therefore, it is beneficial to compare the first and second target regions to determine the results of the second parameter analysis.

[0199] Furthermore, the second parameter analysis results represent a comparative analysis of the first and second target regions, used to reflect changes in metabolic information in the areas of interest to the physician in the subject's historical PET images and current PET images. Similarly, the second parameter analysis results may include, but are not limited to, at least one of the following formats: text, image, or chart.

[0200] Optionally, the computer device can display the first target area, the second target area, and the results of the second parameter analysis, so that doctors can perform subsequent analysis based on the PET images in a timely manner.

[0201] In this embodiment, historical analysis results are mapped onto historical PET images to obtain a first target region in the historical PET images, and current analysis results are mapped onto current PET images to obtain a second target region in the current PET images. Therefore, the first and second target regions can accurately describe the areas of interest to the doctor. Thus, the second parameter analysis results further facilitate timely and convenient comparison and analysis of the metabolic status of the first and second target regions by the doctor.

[0202] Of course, in some embodiments, the computer device may also display the results of the first parameter analysis and the second parameter analysis simultaneously, so that doctors can follow up by combining PET images and CT images.

[0203] To more clearly illustrate the image analysis method in this application, this paper combines... Figure 14 and Figure 15 Please provide an explanation. Figure 14 This is a schematic diagram of the overall process of an image analysis method according to an embodiment of this application. Figure 15 This is a schematic diagram illustrating the usage flow of an image analysis method according to an embodiment of this application. Figure 14 As shown, the computer device executes the image analysis method according to the following procedure.

[0204] S1401, acquire historical PET image samples, historical CT image samples, current PET image samples, and current CT image samples.

[0205] S1402, based on historical PET image samples, historical CT image samples and the corresponding first segmentation gold standard, the initial baseline model is trained to obtain the baseline model.

[0206] S1403, based on historical PET image samples, historical CT image samples and baseline model, determine historical CT segmentation result samples.

[0207] S1404. Based on historical CT segmentation results samples, current PET image samples, current CT image samples, and the second segmentation gold standard, the initial follow-up model is trained to obtain the follow-up model.

[0208] S1405, acquire the subject's historical PET images, historical CT images, current PET images, and current CT images.

[0209] S1406: Based on historical PET images and historical CT images, obtain the registered historical PET images.

[0210] S1407, preprocess the historical CT images to obtain the preprocessed historical CT images.

[0211] S1408, preprocess the registered historical PET images to obtain preprocessed historical PET images.

[0212] S1409: Input the preprocessed historical CT images and preprocessed historical PET images into the baseline model to obtain the historical analysis results.

[0213] S1410: Obtain the registered current PET image based on the current PET image and the current CT image.

[0214] S1411, based on the current CT images and historical analysis results, obtain the historical analysis results after registration.

[0215] S1412: Based on historical PET images and current CT images, obtain historical PET images registered to the current CT images.

[0216] S1413 preprocesses the current CT image, the current PET image after registration, the historical analysis results after registration, and the historical PET images registered to the current CT image, respectively, to obtain the preprocessed current CT image, the preprocessed current PET image, the preprocessed historical analysis results, and the preprocessed registered historical PET images.

[0217] S1414: Input the preprocessed current CT image, the preprocessed current PET image, the preprocessed historical analysis results, and the preprocessed registered historical PET images into the follow-up model to obtain the current analysis results.

[0218] S1415, Based on the historical analysis results and the current analysis results, determine the analysis result of the first parameter.

[0219] S1416, Map the historical analysis results to historical PET images to obtain the first target region in the historical PET images.

[0220] S1417, Map the current analysis results to the current PET image to obtain the second target region in the current PET image.

[0221] S1418, Determine the analysis results of the second parameter based on the first target region and the second target region.

[0222] S1401–S1404 represent the training process, and S1405–S1418 represent the usage process. Please refer to… Figure 15 During use, the computer equipment acquires historical PET images and historical CT images, registers the historical PET images to the historical CT images to obtain the registered historical PET images, and then normalizes the registered historical PET images and historical CT images respectively before inputting them into the baseline model to obtain the historical analysis results.

[0223] Furthermore, the computer equipment acquires the current PET image and the current CT image, and registers the current PET image, historical analysis results, and historical PET images to the current CT image, respectively, to obtain the registered current PET image, the registered historical analysis results, and the historical PET images registered to the current CT image. Then, the computer equipment performs normalization processing on the current CT image, the registered current PET image, the registered historical analysis results, and the historical PET images registered to the current CT image, and inputs them into the follow-up model to obtain the current analysis results.

[0224] In this way, the computer equipment can determine the results of the first parameter analysis and the second parameter analysis, so that doctors can carry out subsequent follow-up work. Figure 14 and Figure 15 The steps described in the above embodiments can be referred to, and will not be repeated here.

[0225] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0226] Based on the same inventive concept, this application also provides an image analysis apparatus for implementing the image analysis method described above. The solution provided by this apparatus is similar to the implementation scheme described in the above method; therefore, the specific limitations in one or more image analysis apparatus embodiments provided below can be found in the limitations of the image analysis method described above, and will not be repeated here.

[0227] Figure 16 This is a structural block diagram of the image analysis device in the embodiments of this application, such as... Figure 16 As shown, this application provides an image analysis device 1600, including: an acquisition module 1601, a first determination module 1602, and a second determination module 1603, wherein:

[0228] The acquisition module 1601 is used to acquire the subject's historical PET images, historical CT images, current PET images, and current CT images.

[0229] The first determining module 1602 is used to determine the historical analysis results based on historical PET images, historical CT images and a baseline model; the historical analysis results include a first region of interest in the historical CT images, the first region of interest corresponds to a second region of interest in the historical PET images, and the first parameter of the second region of interest is greater than or equal to a first preset threshold.

[0230] The second determining module 1603 is used to determine the current analysis result based on the current PET image, the current CT image, historical analysis results, and the follow-up model; the current analysis result includes the third region of interest in the current CT image, and / or includes the fourth region of interest in the current CT image corresponding to the first region of interest; the third region of interest corresponds to the fifth region of interest in the current PET image, and the first parameter of the fifth region of interest is greater than or equal to the first preset threshold; the fourth region of interest corresponds to the first region of interest, and the fourth region of interest corresponds to the sixth region of interest in the current PET image, and the first parameter of the sixth region of interest is less than the first preset threshold.

[0231] The image analysis device provided in this embodiment acquires the subject's historical PET images, historical CT images, current PET images, and current CT images. Based on the historical PET images, historical CT images, and a baseline model, it determines the historical analysis results. Then, based on the current PET images, current CT images, historical analysis results, and a follow-up model, it determines the current analysis results. Since the historical analysis results include a first region of interest in the historical CT images, and the first region of interest corresponds to a second region of interest in the historical PET images, with a first parameter greater than or equal to a first preset threshold, the computer device can first determine the first region of interest in the subject's historical examinations where the first parameter is greater than or equal to the first preset threshold, for example, determining a first region of interest with an SUV ≥ 2.5 on the subject's historical CT images. Furthermore, since the current analysis result includes the third region of interest in the current CT image, and / or includes the fourth region of interest in the current CT image corresponding to the first region of interest; the third region of interest corresponds to the fifth region of interest in the current PET image, and the first parameter of the fifth region of interest is greater than or equal to the first preset threshold; the fourth region of interest corresponds to the first region of interest, and the fourth region of interest corresponds to the sixth region of interest in the current PET image, and the first parameter of the sixth region of interest is less than the first preset threshold, therefore, even if the first parameter of the first region of interest is less than the first preset threshold in the current examination, the computer device can still determine the fourth region of interest. For example, if in the current examination, the SUV of the first region of interest with an SUV≥2.5 in the CT image of the examinee in a previous examination is less than 2.5, the computer device will still determine the fourth region of interest with an SUV<2.5. Thus, the device provided in this embodiment is not only applicable to the case where the first parameter is greater than or equal to the first preset threshold in every examination, but also applicable to the case where the first parameter is greater than or equal to the first preset threshold in the first examination, but the first parameter of the region with the first parameter greater than or equal to the first preset threshold is less than the first preset threshold in the second examination, thus expanding the scope of application.

[0232] Optionally, the first determining module 1602 includes:

[0233] The first determining unit is used to obtain the registered historical PET image based on historical PET images and historical CT images.

[0234] The second determination unit is used to determine the historical analysis results based on historical CT images, registered historical PET images, and the baseline model.

[0235] Optionally, the second determining unit includes:

[0236] The first preprocessing subunit is used to preprocess historical CT images to obtain preprocessed historical CT images.

[0237] The second preprocessing subunit is used to preprocess the registered historical PET images to obtain preprocessed historical PET images.

[0238] The first determination sub-unit is used to input preprocessed historical CT images and preprocessed historical PET images into the baseline model to obtain historical analysis results.

[0239] Optionally, the image analysis device 1600 also includes:

[0240] The first acquisition module is used to acquire historical PET image samples and historical CT image samples.

[0241] The first training module is used to train the initial baseline model based on historical PET image samples, historical CT image samples and the corresponding first segmentation gold standard to obtain the baseline model; wherein, the first segmentation gold standard includes a first region determined by historical CT image samples, the first region corresponds to a second region in historical PET image samples, and the first parameter of the second region is greater than or equal to a first preset threshold.

[0242] Optionally, the second determining module 1603 includes:

[0243] The third determining unit is used to obtain the registered current PET image based on the current PET image and the current CT image.

[0244] The fourth determination unit is used to obtain the registered historical analysis results based on the current CT images and historical analysis results.

[0245] The fifth determination unit is used to determine the current analysis result based on the current CT image, the current PET image after registration, the historical analysis results after registration, and the follow-up model.

[0246] Optionally, the fifth determining unit includes:

[0247] The third preprocessing subunit is used to preprocess the current CT image, the registered current PET image, and the registered historical analysis results respectively, to obtain the preprocessed current CT image, the preprocessed current PET image, and the preprocessed historical analysis results.

[0248] The second determining subunit is used to input the preprocessed current CT image, the preprocessed current PET image, and the preprocessed historical analysis results into the follow-up model to obtain the current analysis results.

[0249] Optionally, the image analysis device 1600 also includes:

[0250] The second acquisition module is used to acquire historical PET image samples, historical CT image samples, current PET image samples, and current CT image samples.

[0251] The third determining module is used to determine the historical CT segmentation result sample based on the historical PET image sample, the historical CT image sample and the baseline model. The historical CT segmentation result sample includes a first region in the historical CT image sample, the first region corresponds to a second region in the historical PET image sample, and the first parameter of the second region is greater than or equal to a first preset threshold.

[0252] The second training unit is used to train the initial follow-up model based on historical CT segmentation result samples, current PET image samples, current CT image samples, and the second segmentation gold standard to obtain the follow-up model. The second segmentation gold standard includes the third region in the current CT image sample and / or includes the fourth region in the current CT image sample corresponding to the first region. The third region corresponds to the fifth region in the current PET image sample, and the first parameter of the fifth region is greater than or equal to the first preset threshold. The fourth region corresponds to the first region and the fourth region corresponds to the sixth region in the current PET image sample, and the first parameter of the sixth region is less than the first preset threshold.

[0253] Optionally, the fifth determining unit may also include:

[0254] The third determining subunit is used to obtain historical PET images registered to the current CT images based on historical PET images and current CT images.

[0255] The fourth determination subunit is used to obtain the current analysis results based on the current CT image, the current PET image after registration, the historical analysis results after registration, the historical PET images registered to the current CT image, and the follow-up model.

[0256] Optionally, the image analysis device 1600 also includes:

[0257] The fourth determining module is used to determine the first parameter analysis result based on the historical analysis results and the current analysis results.

[0258] Optionally, the image analysis device 1600 also includes:

[0259] The first mapping module is used to map the historical analysis results to the historical PET image to obtain the first target region in the historical PET image.

[0260] The second mapping module is used to map the current analysis result to the current PET image to obtain the second target region in the current PET image.

[0261] The fifth determining module is used to determine the second parameter analysis result based on the first target region and the second target region.

[0262] Each module in the aforementioned image analysis device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of a computer device in hardware form or independent of it, or stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0263] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0264] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps in the above method embodiments.

[0265] 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.

[0266] 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.

[0267] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0268] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0269] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. An image analysis method, characterized in that, The method includes: Acquire the subject's historical PET images, historical CT images, current PET images, and current CT images; Based on the historical PET images, the historical CT images, and the baseline model, historical analysis results are determined; the historical analysis results include a first region of interest in the historical CT images, the first region of interest corresponds to a second region of interest in the historical PET images, and a first parameter of the second region of interest is greater than or equal to a first preset threshold; wherein, the first parameter is a parameter representing the body's metabolic status; The current analysis result is determined based on the current PET image, the current CT image, the historical analysis results, and the follow-up model, including: obtaining a registered current PET image based on the current PET image and the current CT image; obtaining a registered historical analysis result based on the current CT image and the historical analysis results; and determining the current analysis result based on the current CT image, the registered current PET image, the registered historical analysis results, and the follow-up model. The current analysis result includes a third region of interest in the current CT image, and / or includes a fourth region of interest in the current CT image corresponding to the first region of interest. The third region of interest corresponds to a fifth region of interest in the current PET image, and the first parameter of the fifth region of interest is greater than or equal to a first preset threshold. The fourth region of interest corresponds to the first region of interest, and the fourth region of interest corresponds to a sixth region of interest in the current PET image, and the first parameter of the sixth region of interest is less than the first preset threshold. Based on the historical analysis results and the current analysis results, the analysis result of the first parameter is determined.

2. The method according to claim 1, characterized in that, The determination of historical analysis results based on the historical PET images, the historical CT images, and the baseline model includes: Based on the historical PET images and the historical CT images, the registered historical PET images are obtained; The historical analysis results are determined based on the historical CT images, the registered historical PET images, and the baseline model.

3. The method according to claim 2, characterized in that, The step of determining the historical analysis results based on the historical CT images, the registered historical PET images, and the baseline model includes: The historical CT images are preprocessed to obtain preprocessed historical CT images. The registered historical PET images are preprocessed to obtain preprocessed historical PET images; The preprocessed historical CT images and the preprocessed historical PET images are input into the baseline model to obtain the historical analysis results.

4. The method according to claim 1, characterized in that, The step of determining the current analysis result based on the current CT image, the registered current PET image, the registered historical analysis results, and the follow-up model includes: The current CT image, the registered current PET image, and the registered historical analysis results are preprocessed respectively to obtain the preprocessed current CT image, the preprocessed current PET image, and the preprocessed historical analysis results. The preprocessed current CT image, the preprocessed current PET image, and the preprocessed historical analysis results are input into the follow-up model to obtain the current analysis results.

5. The method according to claim 1, characterized in that, The step of determining the current analysis result based on the current CT image, the registered current PET image, the registered historical analysis results, and the follow-up model includes: Based on the historical PET images and the current CT images, a historical PET image registered to the current CT image is obtained; The current analysis result is obtained based on the current CT image, the registered current PET image, the registered historical analysis results, the historical PET images registered to the current CT image, and the follow-up model.

6. The method according to any one of claims 1-3, characterized in that, The analysis result of the first parameter represents the comparison between the historical analysis results and the current analysis results, and is used to reflect the changes in the structural information of the region of interest in the subject's historical CT images and current CT images.

7. The method according to any one of claims 1-3, characterized in that, The method further includes: The historical analysis results are mapped onto the historical PET images to obtain the first target region in the historical PET images; The current analysis result is mapped to the current PET image to obtain the second target region in the current PET image; The analysis results of the second parameter are determined based on the first target region and the second target region.

8. The method according to any one of claims 1-3, characterized in that, The method further includes: The historical PET images and the historical CT images were obtained using PET-CT dual-mode imaging.

9. An image analysis device, characterized in that, The device includes: The acquisition module is used to acquire the subject's historical PET images, historical CT images, current PET images, and current CT images. The first determining module is used to determine historical analysis results based on the historical PET images, the historical CT images, and the baseline model; the historical analysis results include a first region of interest in the historical CT images, the first region of interest corresponds to a second region of interest in the historical PET images, and a first parameter of the second region of interest is greater than or equal to a first preset threshold; wherein, the first parameter is a parameter representing the body's metabolic status; The second determining module is used to determine the current analysis result based on the current PET image, the current CT image, the historical analysis results, and the follow-up model, including: obtaining a registered current PET image based on the current PET image and the current CT image; obtaining a registered historical analysis result based on the current CT image and the historical analysis results; determining the current analysis result based on the current CT image, the registered current PET image, the registered historical analysis results, and the follow-up model; the current analysis result includes a third region of interest in the current CT image, and / or includes a fourth region of interest in the current CT image corresponding to the first region of interest; the third region of interest corresponds to a fifth region of interest in the current PET image, and the first parameter of the fifth region of interest is greater than or equal to a first preset threshold; the fourth region of interest corresponds to the first region of interest, and the fourth region of interest corresponds to a sixth region of interest in the current PET image, and the first parameter of the sixth region of interest is less than the first preset threshold; The fourth determining module is used to determine the first parameter analysis result based on the historical analysis results and the current analysis results.

10. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 8.

Images