Mammary gland background essence enhancement automatic quantification method, system and equipment based on magnetic resonance imaging and medium

Abstract

The invention discloses an automatic quantification method, system and device for breast background essence enhancement based on magnetic resonance imaging and a medium. The method comprises the following steps: preprocessing and pre-segmenting MRI multi-stage image data to be segmented to obtain a pure mammary gland region mask; intercepting an unenhanced stage image by using a pure mammary gland region mask to obtain a pure mammary gland region in an unenhanced stage, finely tuning the nnUNet pre-training model, and segmenting a mammary gland mask; obtaining tumor position information according to the breast tumor binary mask information marked in the MRI multi-stage image data to be segmented; according to the pure mammary gland area mask, cutting the image in the enhanced peak period and the image in the non-enhanced period after registration to respectively obtain image areas with the same size as the mammary gland mask, and performing array operation by using the mammary gland mask and the obtained image areas; and calculating quantitative parameters of the BPE. According to the method, false positive regions generated in segmentation results such as normal organs and artifacts are effectively eliminated.

Description

Technical Field

[0001] The present invention belongs to the technical field of computer vision, and in particular relates to a method, system, device and medium for automatic quantification of breast background substance enhancement based on magnetic resonance imaging. Background Art

[0002] Magnetic resonance imaging (MRI) is one of the most commonly used non-invasive diagnostic and treatment tools in clinical practice. DCE-MRI is an important imaging mode for evaluating breast tumors. It highlights the morphology of blood vessels and capsules in breast tissue by injecting contrast agents, and can provide morphological characteristics and functional properties of breast tumors. In addition to using DCE-MRI to discover suspected lesions, it can also be used to evaluate the vascular distribution of normal breast tissue. This enhancement of normal breast tissue in DCE-MRI is called breast background parenchymal enhancement (BPE).

[0003] BPE is qualitatively divided into four categories in the fifth edition of the BI-RADS standard: minimal, mild, moderate, and marked. BPE changes dynamically, and BPE varies in different time periods and different women. Many studies have shown that there is a strong correlation between BPE levels and breast cancer risk and treatment outcomes. Previously, the quantification of BPE was mainly based on manual classification based on BI-RADS, which was highly subjective. As BPE evaluation is increasingly used as an evaluation of breast cancer screening and the efficacy of neoadjuvant chemotherapy, the quantitative calculation of BPE is becoming increasingly important. The key step in the quantitative analysis of BPE is the segmentation of breast fibroglandular tissue (FGT). DCE-MRI scans contain multi-phase images at multiple time points. The tumor and vascular areas gradually become highlighted and difficult to distinguish. Therefore, the images before enhancement are used as training data. The powerful medical image segmentation model nnUNet extracts the features of the training data, obtains accurate and reliable FGT masks, and is used to accurately calculate BPE.

[0004] The above prior art has the following disadvantages: (1) Currently in clinical practice, BPE is usually assessed clinically and subjectively, and there may be wide subjective differences, so there is no standard method to quantify BPE; (2) MRI images do not have standardized measurement units like HU values ​​for CT images. The grayscale values ​​of scans using different models and different parameters in different centers vary significantly. There is no standardized and complete breast DCE-MRI preprocessing process. (3) In MRI images, image artifacts will be generated due to uneven magnetic field strength and scanning environment. The grayscale value distribution of some artifacts is also similar to that of the gland area. However, the existing technology has not effectively eliminated the false positive areas generated in the segmentation results, such as normal organs and artifacts. Summary of the invention

[0005] The main purpose of the present invention is to overcome the shortcomings and deficiencies of the prior art and to provide a method, system, device and medium for automatic quantification of breast background parenchymal enhancement based on magnetic resonance imaging.

[0006] In order to achieve the above object, the present invention adopts the following technical solutions: One aspect of the present invention provides a method for automatically quantifying breast background parenchymal enhancement based on magnetic resonance imaging, comprising the following steps: Preprocess the multi-phase MRI image data to be segmented, including resampling, transfer, N4 correction, registration, noise reduction and labeling; Based on the nnUNet pre-trained model and the corrosion operation, the pre-processed MRI unenhanced image data to be segmented is pre-segmented to obtain a pure breast area mask with the skin and chest wall removed. Use the pure breast region mask to intercept the unenhanced image to obtain the pure breast region in the unenhanced period and fine-tune the nnUNet pre-trained model to segment the breast gland mask; Obtaining tumor location information based on breast tumor binary mask information annotated in the multi-phase MRI image data to be segmented; According to the pure breast area mask, the enhanced peak period image S1 and the unenhanced period image S0 after registration are cropped to obtain image areas of the same size as the breast gland mask, and then the breast gland mask is used to perform array operations on the image areas of the same size as the breast gland mask to obtain the corresponding S sub1 , S sub0 ; Using tumor location information and S sub1 , S sub0 The quantitative parameters of BPE were calculated, including glandular volume, breast volume, fat volume, glandular ratio, and quantitative parameters of BPE for the left and right sides and the healthy side at different thresholds.

[0007] As a preferred technical solution, the resampling and transfer are specifically as follows: The trilinear interpolation algorithm is used for the segmented MRI multi-phase image data, and the nearest neighbor interpolation algorithm is used for the tumor mask to resample the image to a resolution of (1,1,1), and the original DICOM image is converted into a NII image.

[0008] As a preferred technical solution, the N4 correction is specifically as follows: According to the low-frequency information of the image, the bias field of the non-enhanced period image and the image with the most significant enhancement after registration is estimated and corrected to complete the noise reduction.

[0009] As a preferred technical solution, the registration is specifically as follows: The deedsBCV image registration algorithm is used to enhance the peak phase image in the corrected multi-phase image. Non-enhanced images Registration.

[0010] As a preferred technical solution, the denoising is specifically as follows: Non-local adaptive denoising, Gaussian filtering, median filtering or wavelet transform are used to reduce the noise of the unenhanced period images and the enhanced peak period images after registration.

[0011] As a preferred technical solution, the pre-segmentation of the pure glandular area of ​​the breast based on the nnUNet pre-training model and the corrosion operation is specifically as follows: Based on the nnUNet pre-trained model, the model is fine-tuned using local data to obtain the breast area segmentation results; The erosion operation in OpenCV morphological operation is used to further organize the stripping segmentation to obtain the pure gland environment pre-segmentation as follows: ; in, represents the image after corrosion, represents the pixel position in the image, S represents the structural element, Represents the position of the structural element in the original image The corresponding pixel value, min in the formula means taking the minimum pixel value in the structural element; According to the characteristic that skin tissue is mostly presented as the linear edge of the breast part in the image, the breast area mask result A obtained in the previous step is raw Corrosion is performed to obtain pure mammary gland environment area A pure The edge of the original image minus the erosion image is the skin tissue part A. skin : A skin =A raw -A pure .

[0012] As a preferred technical solution, the quantitative parameters of the BPE are calculated by the following formula: ; PE represents the enhancement rate at each pixel position; ; BPE represents the average enhancement rate or average degree of glandular enhancement, and Volume represents the volume of the gland involved in the calculation.

[0013] Another aspect of the present invention further provides a magnetic resonance imaging-based breast background parenchyma enhancement automatic quantification system, which is applied to the above-mentioned magnetic resonance imaging-based breast background parenchyma enhancement automatic quantification method, including a preprocessing module, a breast region pre-segmentation module, a glandular segmentation module and a BPE quantification calculation module; The preprocessing module is used to preprocess the MRI multi-phase image data to be segmented, including resampling, transfer, N4 correction, registration, noise reduction and labeling; The breast region pre-segmentation module is used to pre-segment the pre-processed MRI unenhanced period image data to be segmented based on the nnUNet pre-training model and the corrosion operation to obtain a pure breast region mask with the skin and chest wall removed; The gland segmentation module is used to use the pure breast area mask to intercept the unenhanced period image to obtain the pure breast area of ​​the unenhanced period and fine-tune the nnUNet pre-training model to segment the breast gland mask; according to the pure breast area mask, the enhanced peak period image S1 and the unenhanced period image S0 after registration are cropped to obtain image areas of the same size as the breast gland mask, and then the breast gland mask is used to perform array operations on the image areas of the same size as the breast gland mask to obtain the corresponding S sub1 , S sub0 ; The BPE quantitative calculation module is used to obtain tumor location information according to the breast tumor binary mask information marked in the multi-phase MRI image data to be segmented; using the tumor location information and S sub1 , S sub0 The quantitative parameters of BPE were calculated, including glandular volume, breast volume, fat volume, glandular ratio, and quantitative parameters of BPE for the left and right sides and the healthy side at different thresholds.

[0014] Another aspect of the present invention also provides an automatic quantification device for breast background parenchymal enhancement based on magnetic resonance imaging, the automatic quantification device for breast background parenchymal enhancement based on magnetic resonance imaging comprising: a memory and at least one processor, the memory storing instructions, the memory and the at least one processor being interconnected via lines; the at least one processor calls the instructions in the memory so that the automatic quantification device for breast background parenchymal enhancement based on magnetic resonance imaging executes the above-mentioned automatic quantification method for breast background parenchymal enhancement based on magnetic resonance imaging.

[0015] Another aspect of the present invention provides a storage medium storing a program, which, when executed by a processor, implements the above-mentioned magnetic resonance imaging-based automatic quantification method for breast background parenchyma enhancement.

[0016] Compared with the prior art, the present invention has the following advantages and beneficial effects: (1) Compared with the prior art, the present invention designs a set of image preprocessing processes and breast segmentation of pure glandular areas for glandular segmentation based on the characteristics of MRI images.

[0017] (2) In clinical practice, BPE is usually clinically qualitative and subjectively assessed, and there may be wide subjective differences. There is currently no standard method for quantitative BPE classification. Therefore, the present invention first designs a series of DCE-MR image preprocessing and breast region pre-segmentation methods, and then obtains pure breast regions without skin and chest wall through morphological operations; to eliminate motion artifacts generated during image acquisition and false positives generated by irrelevant regions (e.g., chest cavity, skin, etc.). Then, the most advanced medical image segmentation model nnUNet is used to train the breast gland segmentation model to obtain a reliable breast gland mask; finally, the above-mentioned gland mask is used to obtain the pure gland region of the image, and the relevant measurement indicators of BPE are automatically calculated. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 It is a flow chart of an automatic quantification method of breast background parenchyma enhancement based on magnetic resonance imaging according to an embodiment of the present invention; Figure 2 is a schematic diagram of the process of image preprocessing in an embodiment of the present invention; Figure 3 It is a structural schematic diagram of an automatic quantification system for breast background parenchyma enhancement based on magnetic resonance imaging according to an embodiment of the present invention; Figure 4 It is a schematic diagram of the structure of the storage medium of an embodiment of the present invention. DETAILED DESCRIPTION

[0019] In order to enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without making creative work are within the scope of protection of the present application.

[0020] Example: like Figure 1 As shown, this embodiment provides a method for automatic quantification of breast background parenchyma enhancement based on magnetic resonance imaging, comprising the following steps: Step 1: Preprocess the multi-phase MRI image data to be segmented, including resampling, transfer, N4 correction, registration, denoising and labeling; the process is as follows Figure 2 As shown; (1) Use the trilinear interpolation algorithm for the segmented MRI multi-phase image data and the nearest neighbor interpolation algorithm for the tumor mask to resample the image to a resolution of (1,1,1), and use the Python language SimpleITK package to convert the original DICOM image into a NII image. In addition, linear interpolation, Spline interpolation and other interpolation methods can also be used to resample the segmented MRI multi-phase image data and tumor mask.

[0021] (2) During the scanning period, due to the influence of the patient's respiratory cycle and self-motion, the images of different phases will have different degrees of offset. First, the deedsBCV image registration algorithm is used to register the most significant phase image in the original multi-phase MRI image. Non-enhanced images Registration is used to make the pixel-to-pixel difference calculation of the two images more accurate. In addition, the registration methods provided by python libraries such as ANTspy and SimpleITK can also be used for medical image registration.

[0022] (3) Use N4 correction and Gaussian denoising to perform denoising on the non-enhanced period image and the image with the most significant enhancement period after registration. The N4 correction estimates and corrects the bias field of the non-enhanced period image and the image with the most significant enhancement period after registration based on the low-frequency information of the image to complete the denoising. In addition, the N4BiasFieldCorrectionImageFilter class of the SimpleITK library in Python can be used to perform the N4 correction operation to obtain the corrected image. The N4 correction method is also provided in the antspyx library of Python and can be used. In addition to Gaussian denoising, non-local adaptive denoising, Gaussian filtering, median filtering or wavelet transform can also be used to perform denoising on the non-enhanced period image and the image with the most significant enhancement period after registration.

[0023] Step 2: Pre-segment the MRI unenhanced image based on the nnUNet pre-trained model and corrosion operation to obtain the breast region mask; (1) nnUNet has high flexibility and scalability, and can be adaptively adjusted according to different tasks and data. It has been applied in multiple modal medical image segmentation tasks and has achieved reliable and efficient segmentation results. The present invention is based on the nnUNet pre-trained model and uses local data to fine-tune the model to obtain a more reliable breast area segmentation result. However, the segmentation result includes the entire breast part. Compared with the related tasks of the mammary gland, the skin and other areas are still redundant information.

[0024] (2) The present invention further performs tissue stripping and segmentation by using the erosion operation in OpenCV morphological operations to obtain a pure breast region without the skin and chest wall. The general formula of the erosion operation is as follows: ; in, represents the image after corrosion, represents the pixel position in the image, S represents the structural element, Represents the position of the structural element in the original image The corresponding pixel value. The min in the formula means taking the minimum pixel value in the structural element. This achieves the effect of reducing the image or erasing a small area in the image.

[0025] According to the characteristic that skin tissue often appears as the linear edge of the breast part in the image, the reliable breast area mask result A obtained in the previous step is raw Corrosion is performed to obtain pure mammary gland environment area A pure , and use the original image to subtract the erosion image to get the edge, which is the skin tissue part: A skin =A raw -A pure .

[0026] Step 3: Gland segmentation and BPE quantification calculation; (1) Gland segmentation: Use the pure breast area mask to intercept the unenhanced image to obtain the pure breast area in the unenhanced period and fine-tune the nnUNet pre-trained model to segment the breast gland mask.

[0027] (2) BPE automatic quantization calculation: 1. Obtain tumor location information based on the binary mask information of breast tumors marked in the multi-phase MRI image data to be segmented and save it as a csv table.

[0028] 2. According to the pure breast area mask, the enhanced peak period image S1 and the unenhanced period image S0 after registration are cropped to obtain image areas of the same size as the breast gland mask, and then the breast gland mask is used to perform array operations on the image areas of the same size as the breast gland mask to obtain the corresponding S sub1 , S sub0 ; 3. Using tumor location information and S sub1 , S sub0 Calculate the quantitative parameters of BPE, including glandular volume, breast volume, fat volume, glandular ratio, and quantitative parameters of BPE for the left and right sides and the healthy side under different thresholds. The main formula is as follows: ; PE represents the enhancement rate at each pixel location.

[0029] ; BPE represents the average enhancement rate or average degree of glandular enhancement, and Volume represents the volume of the gland involved in the calculation.

[0030] The above results were saved for BPE-related analysis of breast DCE-MRI.

[0031] like Figure 3 As shown, in another embodiment of the present application, a magnetic resonance imaging-based breast background parenchymal enhancement automatic quantification system is provided, the system comprising a preprocessing module, a breast region pre-segmentation module, a glandular segmentation module and a BPE quantification calculation module; The preprocessing module is used to preprocess the MRI multi-phase image data to be segmented, including resampling, transfer, N4 correction, registration, noise reduction and labeling; The breast region pre-segmentation module is used to pre-segment the pre-processed MRI multi-phase image data to be segmented based on the nnUNet pre-training model and the corrosion operation to obtain a pure breast region mask with the skin and chest wall removed; The gland segmentation module is used to use the pure breast area mask to intercept the unenhanced period image to obtain the pure breast area of ​​the unenhanced period and fine-tune the nnUNet pre-training model to segment the breast gland mask; according to the pure breast area mask, the enhanced peak period image S1 and the unenhanced period image S0 after registration are cropped to obtain image areas of the same size as the breast gland mask, and then the breast gland mask is used to perform array operations on the image areas of the same size as the breast gland mask to obtain the corresponding S sub1 , S sub0 ; The BPE quantitative calculation module is used to obtain tumor location information according to the breast tumor binary mask information marked in the multi-phase MRI image data to be segmented; using the tumor location information and S sub1 , S sub0 The quantitative parameters of BPE were calculated, including glandular volume, breast volume, fat volume, glandular ratio, and quantitative parameters of BPE for the left and right sides and the healthy side at different thresholds.

[0032] It should be noted here that the system provided in the above embodiment is only illustrated by the division of the above functional modules. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure is divided into different functional modules to complete all or part of the functions described above. The system is an automatic quantification method for breast background parenchymal enhancement based on magnetic resonance imaging applied to the above embodiment.

[0033] In another embodiment of the present application, there is also provided an automatic quantification device for breast background parenchymal enhancement based on magnetic resonance imaging, the automatic quantification device for breast background parenchymal enhancement based on magnetic resonance imaging comprising: a memory and at least one processor, the memory storing instructions, the memory and the at least one processor being interconnected via lines; the at least one processor calling the instructions in the memory so that the automatic quantification device for breast background parenchymal enhancement based on magnetic resonance imaging executes the automatic quantification method for breast background parenchymal enhancement based on magnetic resonance imaging as described in the above-mentioned embodiment.

[0034] like Figure 4 As shown, in another embodiment of the present application, a storage medium is further provided, storing a program, and when the program is executed by a processor, a method for automatically quantifying breast background parenchymal enhancement based on magnetic resonance imaging is implemented, specifically: Preprocess the multi-phase MRI image data to be segmented, including resampling, transfer, N4 correction, registration, noise reduction and labeling; Based on the nnUNet pre-trained model and corrosion operation, the pre-processed MRI multi-phase image data to be segmented is pre-segmented to obtain a pure breast area mask with the skin and chest wall removed; Using the pure breast region mask, the breast gland mask is segmented using the nnUNet pre-trained model fine-tuned with labeled data; Obtaining tumor location information based on breast tumor binary mask information annotated in the multi-phase MRI image data to be segmented; According to the pure breast area mask, the enhanced peak period image S1 and the unenhanced period image S0 after registration are cropped to obtain image areas of the same size as the breast gland mask, and then the breast gland mask is used to perform array operations on the image areas of the same size as the breast gland mask to obtain the corresponding S sub1 , S sub0 ; Using tumor location information and S sub1 , S sub0 The quantitative parameters of BPE were calculated, including glandular volume, breast volume, fat volume, glandular ratio, and quantitative parameters of BPE for the left and right sides and the healthy side at different thresholds.

[0035] It should be understood that the various parts of the present application can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0036] The above embodiments are preferred implementation modes of the present invention, but the implementation modes of the present invention are not limited to the above embodiments. Any other changes, modifications, substitutions, combinations, and simplifications that do not deviate from the spirit and principles of the present invention should be equivalent replacement methods and are included in the protection scope of the present invention.

Claims

1. A method for automatic quantification of breast background parenchymal enhancement based on magnetic resonance imaging, characterized in that: The steps include: Preprocess the multi-phase MRI image data to be segmented, including resampling, transfer, N4 correction, registration, noise reduction and labeling; Based on the nnUNet pre-trained model and the corrosion operation, the pre-processed MRI unenhanced image data to be segmented is pre-segmented to obtain a pure breast area mask with the skin and chest wall removed. Use the pure breast region mask to intercept the unenhanced image to obtain the pure breast region in the unenhanced period and fine-tune the nnUNet pre-trained model to segment the breast gland mask; Obtaining tumor location information based on breast tumor binary mask information annotated in the multi-phase MRI image data to be segmented; According to the pure breast area mask, the enhanced peak period image S1 and the unenhanced period image S0 after registration are cropped to obtain image areas of the same size as the breast gland mask, and then the breast gland mask is used to perform array operations on the image areas of the same size as the breast gland mask to obtain the corresponding S sub1 , S sub0 ; Using tumor location information and S sub1 , S sub0 The quantitative parameters of BPE were calculated, including glandular volume, breast volume, fat volume, glandular ratio, and quantitative parameters of BPE for the left and right sides and the healthy side at different thresholds.

2. The method for automatic quantification of breast background parenchyma enhancement based on magnetic resonance imaging according to claim 1, characterized in that: The resampling and transfer are specifically as follows: The trilinear interpolation algorithm is used for the segmented MRI multi-phase image data, and the nearest neighbor interpolation algorithm is used for the tumor mask to resample the image to a resolution of (1,1,1), and the original DICOM image is converted into a NII image.

3. The method for automatic quantification of breast background parenchyma enhancement based on magnetic resonance imaging according to claim 1, characterized in that: The N4 correction is specifically: According to the low-frequency information of the image, the bias field of the non-enhanced period image and the image with the most significant enhancement after registration is estimated and corrected to complete the noise reduction.

4. The method for automatic quantification of breast background parenchyma enhancement based on magnetic resonance imaging according to claim 1, characterized in that: The registration is specifically as follows: The deedsBCV image registration algorithm is used to enhance the peak phase image in the corrected multi-phase image. Non-enhanced images Registration.

5. The method for automatic quantification of breast background parenchyma enhancement based on magnetic resonance imaging according to claim 1, characterized in that: The denoising is specifically as follows: Non-local adaptive denoising, Gaussian filtering, median filtering or wavelet transform are used to reduce the noise of the unenhanced period images and the enhanced peak period images after registration.

6. The method for automatic quantification of breast background parenchyma enhancement based on magnetic resonance imaging according to claim 1, characterized in that: The pre-segmentation of the pure glandular area of ​​the breast based on the nnUNet pre-training model and the corrosion operation is specifically as follows: Based on the nnUNet pre-trained model, the model is fine-tuned using local data to obtain the breast area segmentation results; The erosion operation in OpenCV morphological operation is used to further organize the stripping segmentation to obtain the pure gland environment pre-segmentation as follows: ； in, represents the image after corrosion, represents the pixel position in the image, S represents the structural element, Represents the position of the structural element in the original image The corresponding pixel value, min in the formula means taking the minimum pixel value in the structural element; According to the characteristic that skin tissue is mostly presented as the linear edge of the breast part in the image, the breast area mask result A obtained in the previous step is raw Corrosion is performed to obtain pure mammary gland environment area A pure The edge of the original image minus the erosion image is the skin tissue part A. skin : A skin =A raw -A pure 。 7. The method for automatic quantification of breast background parenchyma enhancement based on magnetic resonance imaging according to claim 1, characterized in that: The quantitative parameters of BPE are calculated by the following formula: ； PE represents the enhancement rate at each pixel position; ； BPE represents the average enhancement rate or average degree of glandular enhancement, and Volume represents the volume of the gland involved in the calculation.

8. An automatic quantification system for breast background parenchyma enhancement based on magnetic resonance imaging, characterized in that: An automatic quantification method for breast background parenchymal enhancement based on magnetic resonance imaging, applied to any one of claims 1-7, comprising a preprocessing module, a breast region pre-segmentation module, a glandular segmentation module and a BPE quantification calculation module; The preprocessing module is used to preprocess the MRI multi-phase image data to be segmented, including resampling, transfer, N4 correction, registration, noise reduction and labeling; The breast region pre-segmentation module is used to pre-segment the pre-processed MRI unenhanced period image data to be segmented based on the nnUNet pre-training model and the corrosion operation to obtain a pure breast region mask with the skin and chest wall removed; The gland segmentation module is used to use the pure breast region mask to obtain image regions of the same size as the breast gland mask for the registered enhanced peak period image S1 and the unenhanced period image S0, and then use the breast gland mask to perform array operations on the image regions of the same size as the breast gland mask to obtain the corresponding S sub1 , S sub0 ; The BPE quantitative calculation module is used to obtain tumor location information according to the breast tumor binary mask information marked in the multi-phase MRI image data to be segmented; using the tumor location information and S sub1 , S sub0 The quantitative parameters of BPE were calculated, including glandular volume, breast volume, fat volume, glandular ratio, and quantitative parameters of BPE for the left and right sides and the healthy side at different thresholds.

9. An automatic quantification device for breast background parenchyma enhancement based on magnetic resonance imaging, characterized in that: The automatic quantification device for breast background parenchymal enhancement based on magnetic resonance imaging comprises: a memory and at least one processor, the memory storing instructions, the memory and the at least one processor being interconnected via lines; the at least one processor calling the instructions in the memory so that the automatic quantification device for breast background parenchymal enhancement based on magnetic resonance imaging executes a method for automatic quantification of breast background parenchymal enhancement based on magnetic resonance imaging as described in any one of claims 1 to 7.

10. A storage medium storing a program, characterized in that: When the program is executed by a processor, the method for automatic quantification of breast background parenchyma enhancement based on magnetic resonance imaging as described in any one of claims 1 to 8 is implemented.

Images