Hematopoietic active bone marrow identification method, system, device and storage medium

Abstract

The present invention discloses a method, system, device and storage medium for identifying hematopoietic active bone marrow, which relates to the field of medical monitoring technology. The method for identifying hematopoietic active bone marrow includes the following steps: obtaining a CT anatomical image and a PET functional metabolic image of a subject; performing a first segmentation process on the CT anatomical image to obtain a bone structure map; performing a second segmentation process on the bone structure map to obtain a bone marrow cavity mask image; performing image interpolation and rigid registration on the PET functional metabolic image according to the bone marrow cavity mask image to obtain a bone marrow cavity functional metabolic image; normalizing the SUV value in the bone marrow cavity functional metabolic image; determining the area of ​​the bone marrow cavity functional metabolic image where the SUV value is greater than a preset SUV threshold as an active bone marrow area. The present application can improve the accuracy of identifying hematopoietic active bone marrow.

Description

Technical Field

[0001] The present invention relates to the field of medical monitoring technology, and in particular to a method, system, device and storage medium for identifying hematopoietic active bone marrow. Background Art

[0002] Bone marrow exists in the spongy cavity of bones and the medullary cavity of long bones. It is composed of various types of cells and reticular connective tissue. It is divided into red bone marrow and yellow bone marrow according to its structure. The bone marrow of an adult is one of the largest tissues in the human body, weighing about 1600 to 3700 grams, with an average of 2800 grams, accounting for 3.5 to 5.9% of body weight. Bone marrow has hematopoietic, immune and defensive functions. The composition of red bone marrow: 40% water, 40% fat, 20% protein, etc., with a rich sinusoidal system. The composition of yellow bone marrow: 15% water, 80% fat, 5% protein, etc., and the sinusoids are replaced by capillaries, venules and thin veins. Red bone marrow (i.e. hematopoietic active bone marrow, ABM) is the hematopoietic tissue of the human body, distributed in the bone marrow cavity, and also contains a small amount in the Haversian duct. It is mainly composed of sinusoids and hematopoietic tissue.

[0003] At present, the quantitative and qualitative analysis and evaluation of the bone marrow content and activity level in various parts of the body are of great clinical significance, such as auxiliary diagnosis and efficacy evaluation of aplastic anemia, leukemia, multiple myeloma, bone marrow embolism, avascular necrosis of the femoral head, myelofibrosis, polycythemia vera, and guiding the bone marrow puncture site. However, the commonly used clinical detection technology cannot show the distribution of red bone marrow and its activity level. Radionuclide bone marrow imaging, such as 18F-FDG PET-CT, can show the uptake of 18F-FDG by tissues and organs in various parts of the body, reflecting the metabolic level of glucose in these tissues and organs. However, since the uptake of normal bone marrow is uniform and diffusely distributed, and the uptake level is lower than that of the liver, it presents a low signal during PET-CT imaging. For the pelvis and abdomen, since the bladder urine contains a large amount of radioactive isopradiate 18F-FDG, its SUV value (Standard Uptake Value) is much higher than that of the surrounding normal tissues and organs (such as the pelvis, etc.). At this time, if the whole body is displayed conventionally, it is impossible to achieve the purpose of clearly displaying and accurately evaluating the state of the bone marrow in the pelvis through simple SUV threshold adjustment, and it is even more difficult to conduct quantitative and qualitative analysis of the content and activity level of active bone marrow. For the head, due to the high metabolic level of brain tissue, it also appears high-brightness during PET-CT imaging, and its SUV value is much higher than that of the surrounding normal tissues and organs (such as the skull), and it is also difficult to analyze the activity level of the skull bone marrow. Reference Figure 2During PET-CT imaging, the high activity and high display of tissues and organs such as the bladder and brain affect the fine display and evaluation of other surrounding normal tissues and organs such as the bone marrow. Summary of the invention

[0004] The present invention aims to solve at least one of the technical problems existing in the prior art. To this end, the present invention provides a method, system, device and storage medium for identifying hematopoietic active bone marrow, which can improve the accuracy of identifying hematopoietic active bone marrow.

[0005] In one aspect, an embodiment of the present invention provides a method for identifying hematopoietic active bone marrow, comprising the following steps:

[0006] Obtain CT anatomical images and PET functional metabolic images of the subjects;

[0007] Performing a first segmentation process on the CT anatomical image to obtain a bone structure map;

[0008] Performing a second segmentation process on the bone structure image to obtain a bone marrow cavity mask image;

[0009] Performing image interpolation and rigid registration on the PET functional metabolic image according to the bone marrow cavity mask image to obtain a bone marrow cavity functional metabolic image;

[0010] Normalizing the SUV value in the bone marrow cavity functional metabolic image;

[0011] The region of the bone marrow cavity functional metabolic image in which the SUV value is greater than a preset SUV threshold is determined as an active bone marrow region.

[0012] According to some embodiments of the present invention, performing a first segmentation process on the CT anatomical image to obtain a bone structure map comprises the following steps:

[0013] Get the specified analysis part;

[0014] The designated analysis portion and the CT anatomical image are input into a deep learning segmentation model to obtain the bone structure map.

[0015] According to some embodiments of the present invention, performing a second segmentation process on the bone structure image to obtain a bone marrow cavity mask image comprises the following steps:

[0016] The bony structure image is smoothed by using a Gaussian smoothing algorithm to obtain a smoothed image;

[0017] Segmenting the smoothed image based on a bone marrow cavity threshold of histogram statistics to obtain a binary image;

[0018] The binary image is processed by using a morphological method and a distance transformation method to obtain an edge marked image;

[0019] The bone marrow cavity mask image is obtained by processing the edge marker image based on a watershed algorithm.

[0020] According to some embodiments of the present invention, the normalization processing of the SUV value in the bone marrow cavity functional metabolic image comprises the following steps:

[0021] Using a post-processing algorithm to remove abnormal SUV values ​​in the bone marrow cavity functional metabolic image;

[0022] Determining a characteristic value according to the SUV value in the bone marrow cavity functional metabolic image;

[0023] The SUV value in the bone marrow cavity functional metabolic image is normalized according to the characteristic value.

[0024] According to some embodiments of the present invention, the method for identifying hematopoietic active bone marrow further comprises the following steps:

[0025] Get multiple preset SUV value ranges;

[0026] Active bone marrow distribution with different activity levels in the bone marrow cavity functional metabolic image is displayed according to the multiple SUV value intervals, wherein the display mode of the bone marrow cavity functional metabolic image includes at least one of a cross-sectional view, a coronal view, a sagittal view or an MPR view.

[0027] On the other hand, an embodiment of the present invention further provides a method for identifying blood-active bone marrow, comprising the following steps:

[0028] Acquire CT anatomical images of the subject;

[0029] The active bone marrow area of ​​the subject is obtained by registering the CT anatomical image of the subject with the prior model of the three-dimensional spatial distribution of the active bone marrow;

[0030] The three-dimensional spatial distribution prior model of active bone marrow is obtained by the following steps:

[0031] Acquire initial image data sets of multiple historical subjects, wherein the initial image data sets include CT anatomical images and PET functional metabolic images;

[0032] Performing non-rigid registration on the initial image datasets according to the bone volumes of historical subjects to obtain registered image datasets, and determining the active bone marrow area according to the registered image datasets by the blood-active bone marrow identification method as described above;

[0033] A priori model of the three-dimensional spatial distribution of active bone marrow is determined based on the multiple active bone marrow areas.

[0034] According to some embodiments of the present invention, determining a priori model of active bone marrow three-dimensional spatial distribution based on a plurality of active bone marrow areas comprises the following steps:

[0035] The point cloud spatial coordinates of the plurality of active bone marrow areas are subjected to principal component analysis and outlier removal processing to obtain a priori model of the three-dimensional spatial distribution of active bone marrow.

[0036] On the other hand, an embodiment of the present invention further provides a hematopoietic active bone marrow identification system, comprising:

[0037] The first module is used to obtain CT anatomical images and PET functional metabolic images of the subject;

[0038] The second module is used to perform a first segmentation process on the CT anatomical image to obtain a bone structure map;

[0039] A third module is used for performing a second segmentation process on the bone structure image to obtain a bone marrow cavity mask image;

[0040] A fourth module is used to perform image interpolation and rigid registration on the PET functional metabolic image according to the bone marrow cavity mask image to obtain a bone marrow cavity functional metabolic image;

[0041] A fifth module is used for normalizing the SUV value in the bone marrow cavity functional metabolic image;

[0042] The sixth module determines the area of ​​the bone marrow cavity functional metabolic image whose SUV value is greater than a preset SUV threshold as an active bone marrow area.

[0043] On the other hand, an embodiment of the present invention further provides a device for identifying hematopoietic active bone marrow, comprising:

[0044] at least one processor;

[0045] at least one memory for storing at least one program;

[0046] When the at least one program is executed by the at least one processor, the at least one processor implements the hematopoietic active bone marrow identification method as described above.

[0047] On the other hand, an embodiment of the present invention further provides a computer-readable storage medium, wherein the computer-readable storage medium stores computer-executable instructions, and the computer-executable instructions are used to enable a computer to execute the hematopoietic active bone marrow identification method as described above.

[0048] The above technical solution of the present invention has at least one of the following advantages or beneficial effects: first, the CT anatomical image and PET functional metabolic image of the subject are obtained, and then the CT anatomical image is subjected to a first segmentation process to obtain a bone structure map, and then the bone structure map is subjected to a second segmentation process to obtain a bone marrow cavity mask image. According to the bone marrow cavity mask image, the PET functional metabolic image is subjected to image interpolation and rigid registration to obtain a bone marrow cavity functional metabolic image, and after the SUV value in the bone marrow cavity functional metabolic image is normalized, the area of ​​the bone marrow cavity functional metabolic image with an SUV value greater than a preset SUV threshold is determined as an active bone marrow area. By outlining the bones and normalizing the SUV value of the bone marrow cavity, the masking and influence of the highlighted display of other high-metabolism and high-SUV value tissues and organs outside the bones on the low SUV value bone marrow is reduced, thereby improving the accuracy of identifying hematopoietic active bone marrow. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 This is a flow chart of a method for identifying hematopoietic active bone marrow provided by an embodiment of the present invention;

[0050] Figure 2 It is a schematic diagram of the display results of PET-CT human body detection provided by an embodiment of the present invention;

[0051] Figure 3 It is a schematic diagram of display results of active bone marrow areas with different activity levels provided by an embodiment of the present invention;

[0052] Figure 4 is a schematic diagram of a hematopoietic active bone marrow identification system provided by an embodiment of the present invention;

[0053] Figure 5 It is a schematic diagram of a hematopoietic active bone marrow identification device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0054] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar components or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and cannot be understood as limiting the present invention.

[0055] In the description of the present invention, it should be understood that descriptions involving orientation, such as up, down, left, right, etc., the orientations or positional relationships indicated are based on the orientations or positional relationships shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as a limitation on the present invention.

[0056] In the description of the present invention, if there is a description of first, second, etc., it is only for the purpose of distinguishing technical features, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features or implicitly indicating the sequence of the indicated technical features.

[0057] The embodiment of the present invention provides a method for identifying hematopoietic active bone marrow, referring to Figure 1 The method for identifying hematopoietic active bone marrow in an embodiment of the present invention includes but is not limited to step S110, step S120, step S130, step S140, step S150 and step S160.

[0058] Step S110, acquiring a CT anatomical image and a PET functional metabolic image of the subject;

[0059] Step S120, performing a first segmentation process on the CT anatomical image to obtain a bone structure map;

[0060] Step S130, performing a second segmentation process on the bone structure image to obtain a bone marrow cavity mask image;

[0061] Step S140, performing image interpolation and rigid registration on the PET functional metabolic image according to the bone marrow cavity mask image to obtain a bone marrow cavity functional metabolic image;

[0062] Step S150, normalizing the SUV value in the bone marrow cavity functional metabolic image;

[0063] Step S160: determining the region of the bone marrow cavity functional metabolic image whose SUV value is greater than a preset SUV threshold as an active bone marrow region.

[0064] In this embodiment, since the low resolution and false positive of PET functional metabolic images may cause positional dislocation and mis-segmentation, a relatively accurate bone marrow cavity functional metabolic image is obtained by first segmenting the bone marrow cavity mask image using the CT anatomical image, and then performing image interpolation and rigid registration morphological operations on the bone marrow cavity mask image and the PET functional metabolic image. By outlining the bones and normalizing the SUV value of the bone marrow cavity, the highlighting of other high-metabolism and high-SUV value tissues and organs outside the bones reduces the masking and influence of low-SUV value bone marrow, thereby improving the accuracy of identifying hematopoietic active bone marrow.

[0065] Furthermore, the three-dimensional contour of the active bone marrow area can be output, which meets the requirements of the radiotherapy planning system for tissue and organ contours and can be imported into the radiotherapy planning system as an organ structure for radiotherapy planning. In addition, 3D reconstruction can be performed based on the identified hematopoietic active bone marrow area of ​​the subject to plan the bone marrow puncture path to improve the accuracy and reliability of clinical bone marrow puncture.

[0066] According to some specific embodiments of the present invention, step S120 includes but is not limited to the following steps:

[0067] Get the specified analysis part;

[0068] The designated analysis part and CT anatomical image are input into the deep learning segmentation model to obtain the bone structure map.

[0069] Specifically, the designated analysis part can be the head, torso, limbs and other parts. Taking the head as an example, the skull structure of the subject can be obtained by inputting the CT anatomical images of the head and the subject into the deep learning segmentation model.

[0070] In some other embodiments, the bone structure map corresponding to the designated analysis part may be segmented based on the use of a threshold segmentation and filling algorithm.

[0071] According to some specific embodiments of the present invention, step S130 includes but is not limited to the following steps:

[0072] The bony structure image is smoothed by using Gaussian smoothing algorithm to obtain a smoothed image;

[0073] The smoothed image is segmented based on the bone marrow cavity threshold of histogram statistics to obtain a binary image.

[0074] The binary image is processed by using morphological method and distance transformation method to obtain edge marked image;

[0075] The edge marked image is processed based on the watershed algorithm to obtain the bone marrow cavity mask image.

[0076] Specifically, in order to solve the problem of abnormal distribution of bone marrow area, for example, high metabolism of bone cortex will have a certain impact on the accurate segmentation of bone marrow cavity. Therefore, the embodiment of the present invention realizes accurate segmentation of bone marrow cavity based on watershed algorithm of distance transformation. First, the bony structure image is smoothed by Gaussian smoothing algorithm to obtain smoothed image, so as to reduce the over-segmentation. Then, the initial segmented binary image is obtained by using the bone marrow cavity threshold of histogram statistics. Then, the edge marked image is created by combining morphological method and distance transformation method. Finally, the edge marked image is processed by watershed algorithm to obtain a more accurate bone marrow cavity mask image.

[0077] According to some specific embodiments of the present invention, step S150 includes but is not limited to the following steps:

[0078] Post-processing algorithms were used to remove abnormal SUV values ​​in the bone marrow cavity functional metabolic images;

[0079] Determine the characteristic value according to the SUV value in the bone marrow cavity functional metabolic image;

[0080] The SUV value in the bone marrow cavity functional metabolic image was normalized according to the characteristic value.

[0081] Specifically, the post-processing algorithm can be a histogram suppression method, which uses the histogram suppression method to remove abnormal SUV values ​​in the bone marrow cavity functional metabolic image, and then counts the SUV value and eigenvalue of each point in the bone marrow cavity functional metabolic image, and normalizes all SUV values ​​in the bone marrow cavity functional metabolic image according to the eigenvalue, and determines the area of ​​the bone marrow cavity where the normalized SUV value is greater than the preset SUV threshold as the hematopoietic active bone marrow area.

[0082] In some embodiments, the characteristic value may be the average SUV value of all SUV values ​​in the bone marrow cavity functional metabolic image, or may be the median SUV value, the maximum SUV value, or the minimum SUV value. Taking the characteristic value as the average SUV value (meanSUV) as an example, the average SUV value is calculated based on all SUV values ​​in the bone marrow cavity functional metabolic image, and each SUV value in the bone marrow cavity functional metabolic image is normalized by the average SUV value, and the area with a normalized SUV value greater than 1.0meanSUV is defined as the hematopoietic active bone marrow area.

[0083] According to some specific embodiments of the present invention, the method for identifying hematopoietic active bone marrow in the embodiments of the present invention further includes but is not limited to the following steps:

[0084] Get multiple preset SUV value ranges;

[0085] Active bone marrow distributions of different activity levels in a bone marrow cavity functional metabolic image are displayed according to a plurality of SUV value intervals, wherein the bone marrow cavity functional metabolic image is displayed in at least one of a cross-sectional view, a coronal view, a sagittal view or an MPR view.

[0086] Specifically, refer to Figure 3 The SUV value intervals can be greater than 1.0 meanSUV, greater than 1.2 meanSUV, and greater than 1.5 meanSUV, respectively. The display modes include cross-sectional images, coronal images, sagittal images, PET functional metabolic images, and three-dimensional images. Active bone marrow areas corresponding to different activity levels can be displayed using progressive pseudo-color images, or can be displayed using multiple mask modes such as contour or volume rendering images.

[0087] Furthermore, based on the active bone marrow area, its volume parameters may be calculated, and the volume parameters may include the volume size of the active bone marrow area, the volume ratio of the active bone marrow area to the bone marrow, and the like.

[0088] In addition, the distribution of active bone marrow of multiple different subjects (or the same subject, collected at different times) can also be displayed, and the display range can be displayed in corresponding cross-sectional, coronal and sagittal planes according to the designated analysis part, so as to compare the differences in the distribution of hematopoietic active bone marrow, and then give the qualitative and quantitative analysis results of the hematopoietic active bone marrow.

[0089] On the other hand, an embodiment of the present invention further provides a method for identifying blood-active bone marrow, comprising the following steps:

[0090] Acquire CT anatomical images of the subject;

[0091] The active bone marrow area of ​​the subject is obtained by registering the CT anatomical image of the subject with the prior model of the three-dimensional spatial distribution of the active bone marrow;

[0092] The prior model of the three-dimensional spatial distribution of active bone marrow is obtained by the following steps:

[0093] Acquire an initial image dataset of multiple historical subjects, wherein the initial image dataset includes CT anatomical images and PET functional metabolic images;

[0094] Performing non-rigid registration on the initial image datasets according to the bone volumes of the historical subjects to obtain registered image datasets, and determining the active bone marrow area according to the blood-active bone marrow identification method described in the previous embodiment according to the registered image datasets;

[0095] A priori model of the three-dimensional spatial distribution of active bone marrow is determined based on multiple active bone marrow areas.

[0096] Specifically, a priori model of the three-dimensional distribution of active bone marrow can be constructed based on the active bone marrow areas of a group of subjects. For the current subject, its CT anatomical image can be collected, and then the CT anatomical image of the current subject can be combined with the priori model of the three-dimensional distribution of active bone marrow through deformation registration, so as to obtain the three-dimensional distribution of the priori active bone marrow area of ​​the current subject. By constructing a priori model of the three-dimensional spatial distribution of active bone marrow, the active bone marrow area of ​​the current subject can be identified even if the current subject has not undergone PET detection, so as to more accurately obtain the dose and volume parameters of the patient's ABM with different degrees of activity during previous radiotherapy, and provide key data for evaluating related factors of blood toxicity.

[0097] The specific process of constructing the prior model of the three-dimensional spatial distribution of active bone marrow is as follows: obtaining the initial image data sets of multiple historical subjects, using the deformable registration algorithm combined with shape constraints and key point spatial distribution characteristics, normalizing and non-rigidly registering the bone volumes of the historical subjects in the evaluation range, and generating multiple normalized and registered image data sets corresponding to the historical subjects. According to the registered image data sets, the active bone marrow area is determined by the method of steps S110 to S160 above, and the point cloud spatial coordinates of the multiple active bone marrow areas are subjected to principal component analysis and post-processing, and the abnormal points are removed while retaining the main shape topological structure, so as to generate the prior model of the three-dimensional spatial distribution of active bone marrow.

[0098] Furthermore, we can compare and analyze the differences in the three-dimensional distribution of the active bone marrow area predicted based on the prior model of the three-dimensional spatial distribution of active bone marrow and the active bone marrow area directly delineated based on PET-CT, and use indicators such as the DICE coefficient, ASD distance, and HD distance to evaluate the overlapping relationship between the two predictions, thereby measuring the performance of the prior model based on the three-dimensional spatial distribution of active bone marrow.

[0099] According to some specific embodiments of the present invention, determining a priori model of active bone marrow three-dimensional spatial distribution according to multiple active bone marrow areas includes the following steps:

[0100] The point cloud spatial coordinates of multiple active bone marrow areas were subjected to principal component analysis and outlier removal to obtain a priori model of the three-dimensional spatial distribution of active bone marrow.

[0101] On the other hand, the embodiment of the present invention also provides a hematopoietic active bone marrow identification system, referring to Figure 4 ,include:

[0102] The first module is used to obtain CT anatomical images and PET functional metabolic images of the subject;

[0103] The second module is used to perform a first segmentation process on the CT anatomical image to obtain a bone structure map;

[0104] The third module is used for performing a second segmentation process on the bone structure image to obtain a bone marrow cavity mask image;

[0105] The fourth module is used to perform image interpolation and rigid registration on the PET functional metabolic image according to the bone marrow cavity mask image to obtain the bone marrow cavity functional metabolic image;

[0106] The fifth module is used to normalize the SUV value in the bone marrow cavity functional metabolic image;

[0107] The sixth module determines the area of ​​the bone marrow cavity functional metabolic image whose SUV value is greater than a preset SUV threshold as the active bone marrow area.

[0108] In this embodiment, since the low resolution and false positive of PET functional metabolic images may cause positional dislocation and mis-segmentation, a bone marrow cavity mask image is first segmented from a CT anatomical image, and then a morphological operation of image interpolation and rigid registration is performed on the bone marrow cavity mask image and the PET functional metabolic image to obtain a more accurate bone marrow cavity functional metabolic image. By outlining the bones and normalizing the SUV value of the bone marrow cavity, the highlighting of other high-metabolism and high-SUV value tissues and organs outside the bones reduces the masking and influence of low-SUV value bone marrow, and the hematopoietic active bone marrow identification system of the embodiment of the present invention improves the accuracy of hematopoietic active bone marrow identification.

[0109] Reference Figure 5 , Figure 5 Schematic diagram of a hematopoietic active bone marrow identification device provided by an embodiment of the present invention. The hematopoietic active bone marrow identification device of the embodiment of the present invention includes one or more control processors and a memory. Figure 5 A control processor and a memory are taken as an example.

[0110] The control processor and the memory can be connected via a bus or other means. Figure 5 The example of connecting through bus is taken in the following.

[0111] The memory, as a non-transient computer-readable storage medium, can be used to store non-transient software programs and non-transient computer executable programs. In addition, the memory may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some embodiments, the memory may optionally include a memory remotely disposed relative to the control processor, and these remote memories may be connected to the hematopoietic activity bone marrow identification device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0112] Those skilled in the art will understand that Figure 5 The device structure shown in the figure does not constitute a limitation on the hematopoietic active bone marrow identification device, and may include more or fewer components than shown in the figure, or a combination of certain components, or a different arrangement of components.

[0113] The non-transient software program and instructions required to implement the hematopoietic active bone marrow identification method applied to the hematopoietic active bone marrow identification device in the above embodiment are stored in the memory, and when executed by the control processor, the hematopoietic active bone marrow identification method applied to the hematopoietic active bone marrow identification device in the above embodiment is executed.

[0114] In addition, an embodiment of the present invention also provides a computer-readable storage medium, which stores computer-executable instructions. The computer-executable instructions are executed by one or more control processors, so that the one or more control processors can execute the hematopoietic active bone marrow identification method in the above method embodiment.

[0115] It will be appreciated by those skilled in the art that all or some of the steps and systems in the disclosed method above may be implemented as software, firmware, hardware and appropriate combinations thereof. Some physical components or all physical components may be implemented as software executed by a processor, such as a central processing unit, a digital signal processor or a microprocessor, or may be implemented as hardware, or may be implemented as an integrated circuit, such as an application specific integrated circuit. Such software may be distributed on a computer-readable medium, which may include a computer storage medium (or a non-transitory medium) and a communication medium (or a temporary medium). As known to those skilled in the art, the term computer storage medium includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules or other data). Computer storage media include, but are not limited to, RAM, ROM, EEPROM, flash memory or other memory technologies, CD-ROM, digital versatile disks (DVD) or other optical disk storage, magnetic cassettes, magnetic tapes, disk storage or other magnetic storage devices, or any other medium that may be used to store desired information and may be accessed by a computer. Furthermore, it is well known to those skilled in the art that communication media typically embodies computer readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave or other transport mechanism, and may include any information delivery media.

[0116] The embodiments of the present invention are described in detail above with reference to the accompanying drawings, but the present invention is not limited to the above embodiments, and various changes can be made within the knowledge scope of ordinary technicians in the relevant technical field without departing from the purpose of the present invention.

Claims

1. A method for identifying hematopoietic active bone marrow, characterized in that: The following steps are involved: Obtain CT anatomical images and PET functional metabolic images of the subjects; Performing a first segmentation process on the CT anatomical image to obtain a bone structure map; Performing a second segmentation process on the bone structure image to obtain a bone marrow cavity mask image; Performing image interpolation and rigid registration on the PET functional metabolic image according to the bone marrow cavity mask image to obtain a bone marrow cavity functional metabolic image; Normalizing the SUV value in the bone marrow cavity functional metabolic image; Determine the region of the bone marrow cavity functional metabolic image where the SUV value is greater than a preset SUV threshold as an active bone marrow region; The step of performing a second segmentation process on the bone structure image to obtain a bone marrow cavity mask image comprises the following steps: The bony structure image is smoothed by using a Gaussian smoothing algorithm to obtain a smoothed image; Segmenting the smoothed image based on a bone marrow cavity threshold of histogram statistics to obtain a binary image; The binary image is processed by using a morphological method and a distance transformation method to obtain an edge marked image; The bone marrow cavity mask image is obtained by processing the edge marker image based on a watershed algorithm.

2. The method for identifying hematopoietic active bone marrow according to claim 1, characterized in that: The step of performing a first segmentation process on the CT anatomical image to obtain a bone structure map comprises the following steps: Get the specified analysis part; The designated analysis site and the CT anatomical image are input into a deep learning segmentation model to obtain the bone structure map.

3. The method for identifying hematopoietic active bone marrow according to claim 1, characterized in that: The normalization process of the SUV value in the bone marrow cavity functional metabolic image comprises the following steps: Using a post-processing algorithm to remove abnormal SUV values ​​in the bone marrow cavity functional metabolic image; Determining a characteristic value according to the SUV value in the bone marrow cavity functional metabolic image; The SUV value in the bone marrow cavity functional metabolic image is normalized according to the characteristic value.

4. The method for identifying hematopoietic active bone marrow according to claim 1, characterized in that: The method for identifying hematopoietic active bone marrow also includes the following steps: Get multiple preset SUV value ranges; Active bone marrow distribution with different activity levels in the bone marrow cavity functional metabolic image is displayed according to the multiple SUV value intervals, wherein the display mode of the bone marrow cavity functional metabolic image includes at least one of a cross-sectional view, a coronal view, a sagittal view or an MPR view.

5. A method for identifying hematopoietic active bone marrow, characterized in that: The following steps are involved: Acquire CT anatomical images of the subject; The active bone marrow area of ​​the subject is obtained by registering the CT anatomical image of the subject with the prior model of the three-dimensional spatial distribution of the active bone marrow; The three-dimensional spatial distribution prior model of active bone marrow is obtained by the following steps: Acquire initial image data sets of multiple historical subjects, wherein the initial image data sets include CT anatomical images and PET functional metabolic images; Performing non-rigid registration on the initial image dataset according to the bone volume of the historical subjects to obtain a registered image dataset, and determining the active bone marrow area according to the hematopoietic active bone marrow identification method as claimed in claim 1 based on the registered image dataset; A priori model of the three-dimensional spatial distribution of active bone marrow is determined based on the multiple active bone marrow areas.

6. The method for identifying hematopoietic active bone marrow according to claim 5, characterized in that: Determining a priori model of active bone marrow three-dimensional spatial distribution based on a plurality of active bone marrow areas comprises the following steps: The point cloud spatial coordinates of the plurality of active bone marrow areas are subjected to principal component analysis and outlier removal processing to obtain a priori model of the three-dimensional spatial distribution of active bone marrow.

7. A hematopoietic active bone marrow identification system, characterized in that: include: The first module is used to obtain CT anatomical images and PET functional metabolic images of the subject; The second module is used to perform a first segmentation process on the CT anatomical image to obtain a bone structure map; A third module is used for performing a second segmentation process on the bone structure image to obtain a bone marrow cavity mask image; A fourth module is used to perform image interpolation and rigid registration on the PET functional metabolic image according to the bone marrow cavity mask image to obtain a bone marrow cavity functional metabolic image; A fifth module is used for normalizing the SUV value in the bone marrow cavity functional metabolic image; The sixth module determines the area of ​​the bone marrow cavity functional metabolic image whose SUV value is greater than a preset SUV threshold as an active bone marrow area; Wherein, the third module is specifically used for: The bony structure image is smoothed by using a Gaussian smoothing algorithm to obtain a smoothed image; Segmenting the smoothed image based on a bone marrow cavity threshold of histogram statistics to obtain a binary image; The binary image is processed by using a morphological method and a distance transformation method to obtain an edge marked image; The bone marrow cavity mask image is obtained by processing the edge marker image based on a watershed algorithm.

8. A device for identifying hematopoietic active bone marrow, characterized in that: include: at least one processor; at least one memory for storing at least one program; When the at least one program is executed by the at least one processor, the at least one processor implements the hematopoietic active bone marrow identification method according to any one of claims 1 to 6.

9. A computer-readable storage medium storing a program executable by a processor, characterized in that: The program executable by the processor is used to implement the method for identifying hematopoietic active bone marrow according to any one of claims 1 to 6 when executed by the processor.

Images