A method and system for analyzing muscle tissue in sarcopenic patients based on CT images

Abstract

This invention proposes a method for analyzing muscle tissue in sarcopenia patients based on CT images. The method includes acquiring abdominal CT images of the same subject; performing pixel-level segmentation on the preprocessed CT images to distinguish and locate muscle, subcutaneous fat, and visceral fat regions; calculating changes in area and / or volume of each tissue and changes in HU value based on the output results of the same subject at different time points; performing longitudinal quantitative analysis and comparison based on a timeline; and outputting visualized labeled results. This invention assesses the tissue regions and changes of three core components: muscle, subcutaneous fat, and visceral fat. It quantifies the assessment results using multiple dimensions such as pixel count, actual physical area or volume, grayscale density, and standard deviation. The assessment is comprehensive, objective, and repeatable, meeting the needs of precise clinical assessment. It can intuitively identify key pathological changes such as muscle loss and fat infiltration, providing a reliable basis for early intervention.

Description

Technical Field

[0001] This invention relates to the field of medical image processing and recognition technology, and in particular to a method for processing and recognizing CT medical images for sarcopenia. Background Technology

[0002] Sarcopenia is a systemic syndrome characterized by decreased skeletal muscle mass, reduced muscle strength, and declining physiological function. It is highly prevalent in the elderly, patients with chronic diseases, and cancer patients. Long-term, continuous monitoring of muscle tissue is necessary to assess disease progression, treatment effectiveness, and rehabilitation. CT imaging, due to its ability to precisely quantify muscle area, volume, and density, has become the preferred imaging method for dynamic assessment of sarcopenia.

[0003] Traditional clinical diagnosis of sarcopenia routinely relies on single static analyses, which cannot achieve continuous tracking and quantitative comparison of the same patient across multiple time points. Manual measurements and semi-automated software depend on manual sketching, data recording, and offline calculations, resulting in cumbersome processes, poor repeatability, and significant individual errors, making it difficult to objectively reflect subtle changes in muscles. Commercial body composition analysis systems only support single-scan data processing and lack the ability to link multiple time-series data, calculate changes, and visualize data, thus failing to meet the needs of sarcopenia disease progression monitoring.

[0004] The existing patent technologies have not completely solved the above problems. For example, Chinese patent document CN119151967A only achieves the comparison of three-dimensional lung models and does not involve muscle and body composition analysis; Chinese patent document CN118429632A focuses on bone segmentation and cannot distinguish multiple tissues such as muscle and fat and complete quantitative evaluation; Chinese patent document CN118505615A uses a deep learning model for body composition segmentation, but only completes static analysis and does not support longitudinal comparison of data at different time points, judgment of the significance of changes, and visualization.

[0005] Therefore, there is a need for a CT image processing method and system that can automatically perform quantitative analysis, follow-up tracking, and intuitive display of muscle tissue in patients with sarcopenia, in order to meet the needs of clinical assessment, efficacy monitoring, and prognosis. Summary of the Invention

[0006] This invention aims to provide a method and system for analyzing muscle tissue in sarcopenic patients based on CT images, so as to achieve fully automated processing, multi-tissue quantification, longitudinal change comparison, significance judgment and visualization report output for CT images of the same patient at multiple time points, solving the problems of existing technologies such as inability to track longitudinally, inaccurate quantification, reliance on manual comparison and unintuitive results.

[0007] Therefore, the first objective of this invention is to provide a method for analyzing muscle tissue in sarcopenic patients based on CT images, which performs image segmentation, component analysis, and longitudinal comparison on CT images at different time points, and outputs a report.

[0008] The second objective of this invention is to provide a muscle tissue analysis system for sarcopenia patients based on CT images.

[0009] A third objective of this invention is to provide a computer device.

[0010] A fourth object of the present invention is to provide a computer-readable storage medium.

[0011] To achieve the above objectives, the first aspect of this invention proposes a method for analyzing muscle tissue in sarcopenic patients based on CT images. The method includes: acquiring abdominal CT images of the same subject at baseline and at least one follow-up examination, and performing image preprocessing operations such as normalization, denoising, and format standardization; employing a deep learning segmentation model based on the UNet architecture, combined with an HViTER feature extraction module, to perform pixel-level segmentation on the preprocessed CT images, distinguishing and locating regions of muscle tissue, subcutaneous adipose tissue, and visceral adipose tissue in the CT images, and outputting the pixel count, physical quantization results, and density characterization results for each tissue; calculating the area and / or volume changes and HU value changes of each tissue based on the CT image output results of the same subject at different time points, performing longitudinal quantification analysis and comparison based on the timeline, generating difference maps and / or heatmaps based on pixel-level HU differences, and visually marking regions with increased, decreased, and no significant changes using different colors, and outputting the visualization marking results.

[0012] In one embodiment of the present invention, the physical quantification result includes real physical quantities of area and / or volume, which are obtained by converting the number of pixels based on the number of pixels of the corresponding tissue and in combination with the pixel spatial resolution parameter.

[0013] In one embodiment of the present invention, the density characterization result is obtained from the gray-level statistics within the segmented region, including: acquiring the gray-level values ​​at each coordinate position in the medical image and calculating the average gray-level density; calculating the standard deviation of tissue density of the corresponding part based on the gray-level values ​​at different coordinate positions; and determining the corresponding tissue based on the gray-level density and tissue density standard deviation calculated in the first two steps.

[0014] In one embodiment of the present invention, the area and / or volume change data includes the amount of area and / or volume change and the rate of area and / or volume change, and the HU value change data includes the amount of HU value change and the rate of HU value change. The clinical significance of sarcopenia is determined based on the change data and a set threshold.

[0015] In one embodiment of the present invention, the longitudinal analysis and comparison timeline includes the baseline examination time point and at least one follow-up examination time point. The change data at different time points are compared with a set threshold to determine the significance of changes in sarcopenia characteristics.

[0016] In one embodiment of the present invention, the longitudinal quantitative analysis and comparison includes: establishing an image comparison group based on preprocessed and segmented baseline examination and at least one follow-up examination CT images; generating a difference map and / or heat map for each image group based on the pixel-level differences between each image; highlighting the change area to generate the visualization marker.

[0017] In one embodiment of the present invention, based on the visual markings, the muscle, subcutaneous fat, and visceral fat tissue regions are modified to generate diagnostic opinions.

[0018] To achieve the above objectives, a second aspect of the present invention proposes a muscle tissue analysis system for sarcopenia patients based on CT images. The system includes: an image processing module for acquiring CT images of the same subject at different time points, performing image preprocessing and image segmentation; a component statistics module for acquiring image data of different body tissues based on the segmented CT images, and outputting pixel counts, physical quantification results, and density characterization results for muscles, subcutaneous fat, and visceral fat; and an analysis and comparison module for performing longitudinal analysis and comparison of medical images of the same subject at different time points, and includes at least a change calculation unit, a clinical significance judgment unit, a change area visualization unit, and a report generation unit.

[0019] In one embodiment of the present invention, the component statistics module includes: a pixel statistics module, which counts pixels of different categories in the segmented medical image to obtain the number and / or proportion of pixels for each body component; an area measurement module, which converts the number of pixels into a real physical quantity based on the number of pixels and in combination with pixel spatial resolution parameters, wherein the real physical quantity is area and / or volume; and a density evaluation module, which calculates gray-scale statistics within the segmented region to characterize tissue density characteristics.

[0020] In one embodiment of the present invention, the analysis and comparison module includes: a change calculation unit, which calculates at least the area / volume change and its rate of change, and the grayscale statistical change and its rate of change based on the data output by the component statistics module; a clinical significance judgment unit, which compares the area / volume change and / or grayscale statistical change with a preset threshold and outputs the change significance result and corresponding clinical prompt information; a change area visualization unit, which generates a difference map and / or heatmap based on the pixel-level differences between the baseline and follow-up images to highlight the change area, wherein the pixel-level differences include at least HU differences and / or segmentation category differences, and visualizes areas that have increased, decreased, or have no significant changes; and a report generation unit, which generates an analysis and comparison report based on the change calculation result, the clinical significance judgment result, and the change area visualization result.

[0021] To achieve the above objectives, a third aspect of the present invention provides a computer device including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the method described in the first aspect.

[0022] To achieve the above objectives, a fourth aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon, characterized in that the computer program, when executed by a processor, implements the steps of the method described in the first aspect.

[0023] The beneficial effects of this invention are:

[0024] This invention precisely selects specific indicators for sarcopenia, assessing the tissue regions and changes in three core components: muscle, subcutaneous fat, and visceral fat. It quantifies the assessment results using multiple dimensions, including pixel count, actual physical area or volume, grayscale density, and standard deviation. The assessment is comprehensive, objective, and repeatable, meeting the needs of precise clinical evaluation. It can intuitively identify key pathological changes such as muscle loss and fat infiltration, providing a reliable basis for early intervention. Furthermore, this invention visualizes areas of muscle loss, fat gain, and stable regions using difference maps and heatmaps. The changed areas are clearly and intuitively presented, facilitating rapid interpretation by physicians and significantly improving clinical interpretation efficiency.

[0025] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0026] Figure 1 This is a flowchart of a method provided in Embodiment 1 of the present invention.

[0027] Figure 2 This is a segmented CT image from Embodiment 1 of the present invention.

[0028] Figure 3 This is a heatmap of the amount and rate of change provided in Embodiment 1 of the present invention.

[0029] Figure 4 This is a difference map of the amount and rate of change provided in Embodiment 1 of the present invention.

[0030] Figure 5 This is a schematic diagram of the system structure provided in Embodiment 2 of the present invention.

[0031] Figure 6 This is a structural block diagram of the computer device provided in Embodiment 3 of the present invention. Detailed Implementation

[0032] To enable those skilled in the art to better understand and implement the technical solutions of the present invention, the present invention will be described in detail, clearly, and completely below with reference to the accompanying drawings and embodiments. Obviously, the described embodiments are merely some embodiments of the present invention, and not all embodiments. Furthermore, based on the embodiments of the present invention, any modifications, equivalent substitutions, improvements, etc., made by those skilled in the art to all other embodiments obtained without creative effort should be included within the protection scope of the present invention.

[0033] Sarcopenia, also known as muscle wasting syndrome, is a systemic degenerative disease characterized by the continuous loss of skeletal muscle mass, decreased muscle strength, and reduced muscle function. Its pathogenesis is related to multiple factors, including age-related hormonal changes, chronic low-grade inflammation, malnutrition, lack of exercise, and neuromuscular degeneration. It has a very high incidence in the elderly, patients with chronic kidney disease, chronic obstructive pulmonary disease, heart failure, and malignant tumors. Sarcopenia can be classified into four levels according to its course and severity: normal, pre-sarcopenia, sarcopenia, and severe sarcopenia. Its core pathological changes are manifested in skeletal muscle atrophy, reduced muscle fiber volume, increased intramuscular fat infiltration, decreased muscle density, blurred muscle boundaries, and a reduction in muscle functional units. These changes directly lead to decreased motor ability, weakened balance, and an increased risk of falls and fractures. They also significantly reduce patients' tolerance to surgery and radiotherapy / chemotherapy, making them important risk factors affecting long-term prognosis and quality of life. In clinical practice, the diagnosis of sarcopenia relies heavily on quantitative assessment of skeletal muscle. Among them, the transverse CT image of the third lumbar vertebra is the standard imaging level for sarcopenia assessment because it can fully display the core muscle groups such as the psoas major, erector spinae, transverse abdominis, and rectus abdominis. The assessment indicators mainly include skeletal muscle cross-sectional area, muscle density, and skeletal muscle index. These indicators all need to be accurately measured on the basis of clear, stable, and high-fidelity images.

[0034] In CT images, distinguishing between muscle and adipose tissue pixels is crucial for accurate quantification and longitudinal comparison, and is also a recognized technical challenge in the industry. Normal muscle tissue in CT images exhibits concentrated and uniform pixel grayscale values, with HU values ​​typically between 30 and 60. Pixel textures are continuous and dense, with clear and regular boundaries. Adipose tissue, on the other hand, shows significantly lower grayscale values, with HU values ​​mostly between -190 and -30. Pixel distribution is loose, and textures are uniform but lack structural features. In images of sarcopenia patients, muscle tissue often exhibits atrophy, decreased density, and fatty infiltration, resulting in numerous low-grayscale pixels within the muscle region that closely resemble adipose tissue pixels. This leads to blurred boundaries, overlapping grayscale values, and mixed textures, making conventional thresholding and edge detection prone to misclassification and omissions. Longitudinal quantification and comparison of CT images of sarcopenia patients at different time points is a recognized high-difficulty scenario in medical image analysis. Compared to single static segmentation, the technical challenges are exponentially increased, and current technologies struggle to achieve stable results. Furthermore, the same patient may exhibit differences in respiratory amplitude, body position, and abdominal fullness between baseline and follow-up examinations. In addition, sarcopenia progresses slowly, resulting in small, gradual, and diffuse changes in muscle area, volume, and density, often less than 5% in the short term—subclinical, subtle changes. Conventional image analysis software can only perform coarse measurements, unable to distinguish between genuine tissue changes and spurious changes caused by scanning noise and segmentation errors. It struggles to extract truly clinically significant changes from background interference, leading to false positives and false negatives. Moreover, sarcopenia progression is not a single tissue change but rather involves simultaneous and interdependent muscle loss, subcutaneous fat remodeling, and visceral fat increase, with uneven degrees of change across different regions. Current technologies can only measure individual tissues separately and compare offline, failing to achieve pixel-level synchronous calculation of changes in muscle, subcutaneous fat, and visceral fat, regional differential labeling, and full-area visualization. This is a key reason why current technologies cannot achieve stable, accurate, and reproducible longitudinal analysis of sarcopenia.

[0035] Example 1

[0036] This embodiment provides a CT medical image processing and recognition method for sarcopenia, as shown in the attached figure. Figure 1-5 The method includes steps S101 to S103.

[0037] Step S101: Obtain CT images of the same object at different time points, and perform image preprocessing and image segmentation.

[0038] In one embodiment of the present invention, the image preprocessing includes normalization and formatting, that is, converting the CT image acquired from the CT image into a preset format, such as an internal processing format like PNG, and setting a uniform size to ensure consistency in subsequent processing.

[0039] In one embodiment of the present invention, image segmentation can be performed using some methods in the prior art, such as the image segmentation method disclosed in Chinese invention patent document CN 113327235 A, or by referring to image segmentation methods in the same field. The purpose is to make the image easier for subsequent tissue component analysis. However, it should be noted that the segmentation in this embodiment is based on the analysis of components such as muscle tissue, adipose tissue, and fat infiltration. Therefore, it is different from the segmentation methods for other body components (such as internal organs).

[0040] In one embodiment of the present invention, image segmentation is performed using the following method: a preprocessed 512×512 CT image is fed into a UNet architecture; the HViTER module in the encoder extracts features at multiple levels; the first-layer HViTER module (64 channels) extracts shallow edge and texture features; the second-layer HViTER module (128 channels) extracts mid-layer tissue morphology features; the feature representation capability is enhanced by residual feature fusion and global feature fusion; the decoder upsamples progressively through transposed convolution and combines the feature maps of skip connections; finally, a 3-channel probability map (muscle, subcutaneous fat, visceral fat) is output.

[0041] The result after segmentation is as follows Figure 2 As shown, the abdominal CT image was segmented into muscle tissue (marked in red), subcutaneous fat (marked in purple), and visceral fat (marked in yellow). The segmentation results had smooth boundaries and conformed to anatomical characteristics, with Dice coefficients of 0.958 (muscle), 0.946 (subcutaneous fat), and 0.937 (visceral fat).

[0042] Step S102: Based on the segmented CT images, acquire image data of different body tissues and output the pixel count, physical quantification results, and density characterization results of muscles, subcutaneous fat, and visceral fat.

[0043] In step 102, firstly, pixel statistics are performed on the segmented CT image, and each pixel in the segmented medical image is assigned a different category label. Then, different colors are used to mark different category regions. After that, all pixels in the segmented medical image are traversed, and the number of pixels corresponding to each category is counted. Finally, the pixel ratio of different categories is counted.

[0044] Taking a specific statistical process as an example, firstly, human tissue and background are defined based on the HU value range. A threshold (-250HU to +1500HU) is applied to filter all human tissue pixels. Then, based on the anatomical knowledge in the existing technology, the specific HU value range is reconfirmed. Muscle tissue is defined as -29HU to +150HU, subcutaneous fat as -190HU to -30HU, and visceral fat as -175HU to -25HU. Misclassified pixels that do not conform to the above specific tissue HU value range are corrected.

[0045] A multi-channel labeled map (each channel corresponding to a tissue component) was generated using a segmentation model. Each channel was binarized (pixel values ​​> 0.5 were classified as belonging to that category), and the total number of pixels and their coordinate distribution for each channel were counted. Then, priority rules were applied to process overlapping areas. Based on the counted pixels, the percentage of each tissue pixel in the total human tissue pixels, the ratio of muscle to total fat (subcutaneous + visceral), and the ratio of subcutaneous fat to visceral fat were calculated. The results are shown in the table below.

[0046]

[0047] As can be seen from the table above, the key ratio indicators for this result are: muscle / total fat ratio: 0.78; subcutaneous fat / visceral fat ratio: 1.59; muscle mass index (SMI): 54.2 cm² / m² (assuming the patient's height is 1.68m).

[0048] The image and results correspondence shows that in the color-coded segmentation results: red area: muscle tissue (24,682 pixels); purple area: subcutaneous fat (19,428 pixels); yellow area: visceral fat (12,176 pixels).

[0049] by Figure 2 Taking the segmentation results in the image as an example, the distribution characteristics of different tissues can be clearly observed: muscle tissues such as the psoas major and erector spinae muscles are located around the spine; subcutaneous fat forms a continuous ring band located on the outer side of the abdominal wall; visceral fat is mainly distributed around the abdominal organs.

[0050] The clinical significance of the segmentation can be analyzed based on the above image segmentation results: the patient's SMI value is 54.2 cm² / m², which is higher than the diagnostic threshold for sarcopenia (male <55 cm² / m², female <39 cm² / m²); the visceral fat / subcutaneous fat ratio is 0.63, which is lower than the high-risk threshold for metabolic syndrome (>1.0); the mean muscle HU value is 56.8, which is higher than the threshold for muscle mass loss (<30).

[0051] Step S103 involves performing a longitudinal analysis and comparison of the changes in CT images of the same object at different time points, and visualizing the areas of change. Specifically, this includes:

[0052] Calculate the change in volume: For a given tissue c (e.g., skeletal muscle), calculate the change in volume at two different time points (baseline and follow-up). and rate of change :

[0053]

[0054]

[0055] in: It is the tissue volume at baseline (initial scan). It refers to the tissue volume at the time of follow-up.

[0056] Clinical interpretation:

[0057] Will and preset threshold Comparison,

[0058] like If abnormal growth is detected, the system can issue a warning (such as edema or tumor proliferation).

[0059] like In such cases, the system can issue a warning about tissue atrophy (such as muscle loss or fat reduction).

[0060] Calculating the change in HU value: The HU value in medical imaging (CT) reflects tissue density. For example, a higher HU value in skeletal muscle indicates greater density and better muscle quality. Therefore, this embodiment calculates:

[0061] Change in HU average value:

[0062]

[0063] HU change rate:

[0064]

[0065] in, The HU value indicates the value at the time of follow-up (re-examination). This indicates the HU value at the baseline (initial scan). This indicates the amount of change in the HU value.

[0066] Clinical interpretation:

[0067] Muscle HU decline The system can then issue alerts regarding muscle atrophy and fat infiltration.

[0068] An increase in fat levels can trigger systemic alerts that may be related to inflammation.

[0069] The clinical significance of the change is determined by a preset threshold, for example:

[0070] Muscle atrophy:

[0071] Decreased bone density:

[0072] Abnormal fat changes: .

[0073] To visually represent the changes in tissues such as muscle, subcutaneous fat, and visceral fat, this embodiment uses heatmaps and difference maps for visualization.

[0074] Heatmaps, such as Figure 3 As shown, the goal of a heatmap is to visually represent the changes in volume and density of different tissues.

[0075] The heatmap is implemented as follows:

[0076] Red indicates a decrease: This means that the volume or HU value of the area has decreased (such as muscle atrophy).

[0077] Blue indicates an increase: This means that the volume or HU value of the area has increased (such as tumor growth).

[0078] Green represents no significant change: indicating that the change in this area is within the normal range.

[0079] For example Figure 3 As shown in the heatmap, the visceral fat area is displayed in red, indicating a decrease in visceral fat; the muscle area is displayed in red, indicating muscle atrophy and a decrease in visceral fat density (HU); and the subcutaneous fat area is displayed in blue, indicating subcutaneous fat hyperplasia.

[0080] Difference Map, for example Figure 4 As shown. The goal of a difference map is to generate a difference map by calculating the pixel-level differences between two scans, highlighting areas of change.

[0081] The specific calculation method is as follows:

[0082]

[0083] in, This represents the pixel value during follow-up scans. The pixel value represents the baseline. This is the difference between the two.

[0084] like If the volume or density of the region increases (marked in blue), then the region's volume or density increases.

[0085] like If the volume or density of the region decreases (marked in red), then the region's volume or density decreases.

[0086] like If no significant changes are observed in the region, then no significant changes will be observed.

[0087] The difference map is visualized by overlaying it onto the original CT image, highlighting the areas of change. An adjustable transparency difference map is generated to allow clinicians to observe subtle changes.

[0088] Example 2

[0089] This embodiment provides a CT medical image processing and recognition system for sarcopenia, such as... Figure 6 As shown, the system includes:

[0090] Image processing module 201 acquires CT images of the same object at different time points, and performs image preprocessing and image segmentation;

[0091] The component statistics module 202 acquires image data of different body tissues based on the segmented CT images and outputs the number of pixels, physical quantification results, and density characterization results of muscles, subcutaneous fat, and visceral fat.

[0092] The analysis and comparison module 203 performs longitudinal analysis and comparison of medical images of the same object at different time points, and includes at least a change calculation unit, a clinical significance judgment unit, a change area visualization unit, and a report generation unit.

[0093] In some implementations, the component statistics module 202 further includes:

[0094] The pixel statistics module 204 counts the pixels of different categories in the segmented medical image to obtain the number and / or proportion of pixels for each body component.

[0095] The area measurement module 205 converts the number of pixels into a real physical quantity based on the number of pixels and combined with the pixel spatial resolution parameter. The real physical quantity is area and / or volume.

[0096] The density assessment module 206 calculates gray-scale statistics within the segmented region to characterize tissue density properties.

[0097] In some implementations, the analysis and comparison module 203 further includes:

[0098] The change calculation unit 207 calculates at least the area / volume change and its rate of change, as well as the grayscale statistical change and its rate of change, based on the data output by the component statistics module.

[0099] The clinical significance judgment unit 208 compares the changes in area / volume and / or grayscale statistics with preset thresholds, and outputs the results of the changes in significance and corresponding clinical prompts.

[0100] The change region visualization unit 209 generates a difference map and / or heat map based on the pixel-level differences between the baseline and follow-up images to highlight the change region, wherein the pixel-level differences include at least HU differences and / or segmentation category differences, and visualizes and marks the regions that have increased, decreased, and have no obvious changes.

[0101] The report generation unit 210 generates an analysis and comparison report based on the calculation results of the change amount, the clinical significance judgment results, and the visualization results of the change area.

[0102] Example 3

[0103] This embodiment provides a computer device 300, the structure of which is as follows: Figure 6 As shown, the device includes a memory 301 and a processor 302. The memory 301 stores a computer program 303, and the processor 302 executes the computer program 301 to implement the processing and identification method described in Embodiment 1. The computer device described in this embodiment may also include a communication interface and other external devices.

[0104] The aforementioned computer equipment can be a server, desktop computer, laptop computer, tablet computer, smartphone, industrial control computer, embedded device, or dedicated smart terminal. Its hardware configuration can be flexibly adjusted according to the performance requirements, power consumption requirements, and size requirements of the actual application scenario, and it has good versatility and scalability when executing the methods of the embodiments of the present invention.

[0105] In some specific implementations, the processor may be one or more combinations of a central processing unit (CPU), a graphics processing unit (GPU), a neural network processor (NPU), a digital signal processor (DSP), or an embedded microcontroller. During instruction execution, the processor performs data processing, logical judgments, model calculations, flow control, and other operations to implement the steps described in this application.

[0106] Embodiments of this specification also provide a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method described in Embodiment 1.

[0107] Communication interfaces are used to enable data communication between computer devices and external devices or networks, and can include wired and wireless communication interfaces. Wired communication interfaces may include Ethernet interfaces, USB interfaces, serial ports, parallel ports, HDMI interfaces, SFP interfaces, etc., for high-speed and stable data transmission via cables. Wireless communication interfaces may include Wi-Fi modules, Bluetooth modules, ZigBee modules, mobile cellular communication modules, etc., for enabling short-range or long-range data interaction in scenarios without physical connections, thereby receiving external input data or outputting processing results.

[0108] Other external devices can be various input / output devices connected via peripheral interfaces to meet human-computer interaction or external data acquisition needs. For example, input devices may include keyboards, mice, touchscreens, scanners, image acquisition devices, sensors, and voice acquisition devices, used to input operating instructions, raw data, image information, audio information, or environmental parameters into the computer. Output devices may include monitors, indicator lights, speakers, printers, and projectors, used to display processing results, indicate operating status, or output final files.

[0109] Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for analyzing muscle tissue in sarcopenic patients based on CT images, characterized in that, The method includes: Abdominal CT images of the same subject at baseline examination and at least one follow-up examination were acquired, and image preprocessing operations such as normalization, noise reduction and format standardization were performed. A deep learning segmentation model based on the UNet architecture, combined with the HViTER feature extraction module, is used to perform pixel-level segmentation on CT images after image preprocessing. This distinguishes and locates the regions of muscle tissue, subcutaneous adipose tissue and visceral adipose tissue in CT images, and outputs the number of pixels, physical quantization results and density characterization results of each tissue. Based on the CT image output results of the same subject at different time points, the area and / or volume change data and HU value change data of each tissue are calculated. Longitudinal quantitative analysis and comparison are performed with the timeline as the benchmark. Based on the pixel-level HU difference, difference maps and / or heat maps are generated respectively. Regions with increase, decrease and no obvious change are marked with different colors and the visualization marking results are output.

2. The method for analyzing muscle tissue in sarcopenic patients based on CT images according to claim 1, characterized in that, The physical quantification results include real physical quantities of area and / or volume, which are obtained by converting the number of pixels based on the number of pixels of the corresponding tissue and combined with the pixel spatial resolution parameter.

3. The method for analyzing muscle tissue in sarcopenic patients based on CT images according to claim 1, characterized in that, The density characterization results are obtained from the gray-level statistics within the segmented region, including: Obtain the gray values ​​at each coordinate position in the medical image and calculate the average gray density; Calculate the standard deviation of tissue density for the corresponding area based on the gray values ​​at different coordinate positions; The corresponding tissue is determined based on the gray density and tissue density standard deviation calculated in the first two steps.

4. The method for analyzing muscle tissue in sarcopenic patients based on CT images according to claim 1, characterized in that, The area and / or volume change data includes the amount of area and / or volume change and the rate of area and / or volume change. The HU value change data includes the amount of HU value change and the rate of HU value change. The clinical significance of sarcopenia is determined based on the change data and a set threshold.

5. The method for analyzing muscle tissue in sarcopenic patients based on CT images according to claim 1, characterized in that, The longitudinal analysis and comparison timeline includes the baseline examination time point and at least one follow-up examination time point. The change data at different time points are compared with the set threshold to determine the significance of changes in sarcopenia characteristics.

6. The method for analyzing muscle tissue in sarcopenic patients based on CT images according to claim 1, characterized in that, The longitudinal quantitative analysis and comparison includes: establishing image comparison groups based on preprocessed and segmented baseline examination and at least one follow-up examination CT images; generating difference maps and / or heat maps for each image group based on pixel-level differences between each image; highlighting the areas of change; and generating the visualization markers.

7. The method for analyzing muscle tissue in sarcopenic patients based on CT images according to claim 6, characterized in that, Based on the visual markers, the areas of muscle, subcutaneous fat, and visceral fat tissue are refined to generate diagnostic opinions.

8. A muscle tissue analysis system for sarcopenia patients based on CT images, characterized in that, The system includes: The image processing module acquires CT images of the same patient at different time points, and performs image preprocessing and image segmentation. The component statistics module, based on the segmented CT images, acquires image data of different body tissues and outputs the number of pixels, physical quantification results, and density characterization results for muscles, subcutaneous fat, and visceral fat. The analysis and comparison module performs longitudinal analysis and comparison of medical images of the same object at different time points, and includes at least a change calculation unit, a clinical significance judgment unit, a change area visualization unit, and a report generation unit.

9. A muscle tissue analysis system for sarcopenia patients based on CT images according to claim 8, characterized in that, The component statistics module includes: The pixel statistics module counts pixels of different categories in the segmented medical image to obtain the number and / or proportion of pixels for each body component. The area measurement module converts the number of pixels into a real physical quantity based on the number of pixels and combined with the pixel spatial resolution parameter. The real physical quantity is area and / or volume. The density assessment module calculates gray-scale statistics within segmented regions to characterize tissue density properties.

10. A muscle tissue analysis system for sarcopenia patients based on CT images according to claim 8, characterized in that, The analysis and comparison module includes: The change calculation unit calculates at least the area / volume change and its rate of change, as well as the grayscale statistical change and its rate of change, based on the data output by the component statistics module. The clinical significance judgment unit compares the changes in area / volume and / or grayscale statistics with preset thresholds, and outputs the results of the changes in significance and corresponding clinical prompts. The change region visualization unit generates a difference map and / or heat map based on the pixel-level differences between the baseline and follow-up images to highlight the change regions, wherein the pixel-level differences include at least HU differences and / or segmentation category differences, and visualizes and marks regions that have increased, decreased, or have no significant changes. The report generation unit generates an analysis and comparison report based on the calculation results of the change amount, the clinical significance judgment results, and the visualization results of the change area.

11. 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 7.

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

Images