Image enhancement method and system for boron concentration imaging in BNCT, equipment and medium

Abstract

The invention discloses an image enhancement method and system for boron concentration imaging in BNCT, equipment and a medium wherein the image enhancement system for boron concentration imaging in BNCT comprises an image acquisition module, an image registration fusion module, a gamma source spatial distribution image acquisition module, a boron concentration inversion module, an image fusion module and an image enhancement module. And finally obtaining an enhanced voxel internal three-dimensional boron concentration distribution diagram. According to the method, the data of the two gamma rays are utilized at the same time, the constructed boron concentration inversion model is combined, meanwhile, accurate correspondence of the gamma data and the organization structure in the three-dimensional space can be achieved, in addition, the trained image enhancement model is utilized for conducting image enhancement on the three-dimensional boron concentration distribution diagram in the voxel, and the image quality is further improved. The technical problem that in the prior art, a boron concentration imaging measurement method in the BNCT process is poor in imaging quality can be solved.

Description

Technical Field

[0001] This invention relates to the field of neutron therapy technology, specifically to image enhancement methods, systems, devices, and media for boron concentration imaging in BNCT. Background Technology

[0002] Boron neutron capture therapy (BNCT) is a highly selective new radiotherapy technique whose therapeutic mechanism is based on... 10 B(n,α) 7 Li nuclear reaction, under thermal neutron irradiation, interacts with tumor cells enriched within... 10 Boolean atoms undergo a capture reaction, releasing high-linear-energy particles with a range limited to the cellular level, thus achieving precise killing of tumor tissue. This highly precise treatment method has an effect on the body. 10 The B distribution places extremely high demands on patients, especially during treatment, as it requires accurate differentiation between tumor and normal tissue. 10 The real-time concentration and spatial distribution of B can easily cause dosage errors, affecting the therapeutic effect and even triggering potential side effects.

[0003] Currently, a series of in vivo imaging measurement methods have emerged in recent years to attempt to achieve non-invasive boron concentration monitoring during BNCT, such as positron emission tomography (PET), single-photon emission tomography (SPECT), and ultra-short echo time magnetic resonance imaging (UTE-MRI) in nuclear magnetic resonance technology. However, these online boron concentration monitoring technologies generally face the following key problems: First, the gamma signal itself has a low yield and is subject to scattering and absorption interference during tissue propagation, resulting in poor image quality; second, the neutron beam itself has energy spectrum and spatial distribution inhomogeneity, causing the gamma yield and boron concentration in different depth regions to no longer have a linear relationship, further leading to poor imaging quality.

[0004] Therefore, it is necessary to improve the in vivo imaging method for boron concentration during BNCT to address the aforementioned technical problems. Summary of the Invention

[0005] The purpose of this invention is to provide an image enhancement method, system, device, and medium for boron concentration imaging in BNCT, in order to solve the technical problem of poor imaging quality in existing boron concentration imaging measurement methods during BNCT.

[0006] To solve the above-mentioned technical problems, the present invention adopts the following technical solution:

[0007] In a first aspect, an image enhancement method for boron concentration imaging in BNCT includes the following steps:

[0008] S1. Acquire MRI and CT images of the target object, and perform registration and fusion to obtain an MRI-CT fused three-dimensional voxel model;

[0009] S1. Acquire spatial distribution images of the first gamma source corresponding to 478keV and 2.224MeV gamma rays, as well as the corresponding first gamma count and second gamma count;

[0010] S3. Construct a boron concentration inversion model;

[0011] S4. Obtain the photon transport response matrix, and input the first gamma count and the second gamma count into the boron concentration inversion model to perform inversion and obtain the first boron concentration distribution data.

[0012] S5. Obtain the tissue mask function in the MRI-CT fused three-dimensional voxel model, and fuse the tissue mask function, the spatial distribution image of the first gamma source, and the first boron concentration distribution data to obtain a three-dimensional boron concentration distribution map within the voxel.

[0013] Furthermore, the image enhancement method for boron concentration imaging in BNCT also includes: S6, training a deep learning model with the three-dimensional boron concentration distribution map within the voxel to obtain a boron concentration image enhancement model, and enhancing the image of the three-dimensional boron concentration distribution map within the voxel to obtain an enhanced three-dimensional boron concentration distribution map within the voxel.

[0014] Furthermore, in the boron concentration inversion module, the method for obtaining the photon transport response matrix is ​​as follows:

[0015] The MRI-CT fused three-dimensional voxel model and the treatment room spatial parameters were input into the Monte Carlo simulation platform;

[0016] The location information of multiple gamma ray detectors is marked in the treatment room space;

[0017] Define neutron beam source term parameters;

[0018] The multi-channel photon transport process under monoenergetic gamma source conditions of 478 keV and 2.224 MeV was simulated respectively, and the photon transport response function of each channel was obtained.

[0019] The photon transport response functions of each channel are coupled to obtain the photon transport response matrix.

[0020] Furthermore, the photon transport response function of the i-th channel is:

[0021] C i =∫∫∫VR i (r)·S γ (r)dr;

[0022]

[0023] In the formula, C i R is the observation count of the i-th gamma-ray detector channel; i (r) represents the gamma source intensity at voxel location r, N i,r N represents the number of photons effectively detected by the i-th channel after a gamma photon is emitted at voxel position r. emitted S represents the total number of gamma photons emitted at voxel position r; γ (r) is the response function of the photon at this position propagating to the i-th gamma-ray detector; V is the voxel volume.

[0024] Furthermore, the boron concentration inversion model is as follows:

[0025]

[0026] In the formula, N B per unit volume 10 B atomic density, N H The density of hydrogen atoms, for 10 The (n,α) reaction cross section of B for 1 The (n,γ) reaction cross section of H, f γ for 10 B(n,α) 7 The Li reaction produces a branching ratio of 478 keV gamma; ∈478 represents the detector's detection efficiency for 478 keV gamma; ∈2.224 represents the detector's detection efficiency for 2.224 MeV gamma, N 478 N 2.224 These are the first gamma count and the second gamma count, respectively.

[0027] Furthermore, the boron concentration image enhancement model is trained using the following method:

[0028] S1. Obtain a set of boron concentration distribution images, the set of images including the obtained reference boron concentration distribution map and the corresponding degraded boron concentration distribution map;

[0029] S2. Preprocess the images in the boron concentration distribution image set, and divide them into training set and test set according to a preset ratio;

[0030] S3. Initialize the generator and discriminator of the GAN, and set the optimizer; the generator is used to extract and enhance features from the degraded boron concentration distribution map, and the discriminator is used to determine the difference between the generated image and the reference boron concentration distribution map.

[0031] S4. The model is trained using the training set, the total loss is defined, and the model is evaluated using the test set. After multiple rounds of iterative training, the boron concentration image enhancement model is obtained.

[0032] Secondly, an image enhancement system for boron concentration imaging in BNCT, utilizing the image enhancement method for boron concentration imaging in BNCT as described above, is characterized by comprising:

[0033] The image acquisition module is used to acquire MRI and CT images of the target object;

[0034] The image registration and fusion module is used to register and fuse the acquired MRI images with CT images to obtain an MRI-CT fused three-dimensional voxel model.

[0035] The gamma source spatial distribution image acquisition module is used to acquire the first gamma source spatial distribution image and the corresponding first gamma count and second gamma count for gamma rays of 478 keV and 2.224 MeV, respectively.

[0036] The boron concentration inversion module is used to input the photon transport response matrix, the first gamma count, and the second gamma count into the boron concentration inversion model to perform inversion and obtain the first boron concentration distribution data.

[0037] The image fusion module is used to obtain the tissue mask function in the MRI-CT fused three-dimensional voxel model, and fuse the tissue mask function, the spatial distribution image of the first gamma source, and the first boron concentration distribution data to obtain a three-dimensional boron concentration distribution map within the voxel.

[0038] Furthermore, the image enhancement system for boron concentration imaging in BNCT also includes: an image enhancement module, which trains a deep learning model with the three-dimensional boron concentration distribution map within the voxel to obtain a boron concentration image enhancement model, and enhances the image of the three-dimensional boron concentration distribution map within the voxel to obtain an enhanced three-dimensional boron concentration distribution map within the voxel.

[0039] Thirdly, the present invention also provides an apparatus, characterized in that it includes a processor and a memory, the processor being coupled to the memory; the memory is used to store a computer program, and the processor is used to execute the computer program stored in the memory to implement the image enhancement method for boron concentration imaging in BNCT as described above.

[0040] Fourthly, the present invention also provides a medium for storing a computer program, characterized in that the computer program is executed by a processor to implement the image enhancement method for boron concentration imaging in BNCT as described above.

[0041] The present invention has the following beneficial effects: The image enhancement method and system for boron concentration imaging in BNCT disclosed in this invention utilizes data from two types of gamma rays and constructs a concentration inversion model using the ratio of 478keV to 2.224MeV dual-energy gamma rays. This avoids reliance on empirical coefficients and biological models, and directly establishes measurement indicators based on nuclear reaction yields, providing a clear physical source and quantitative interpretation capability. Furthermore, it obtains a photon transport response matrix, ensuring accurate correspondence between gamma data and tissue structure in three-dimensional space, achieving millimeter-level spatial resolution concentration distribution map reconstruction, and significantly improving spatial positioning accuracy. Attached Figure Description

[0042] To make the objectives, technical solutions, and advantages of the invention clearer, the invention will now be described in further detail with reference to the accompanying drawings, wherein:

[0043] Figure 1 This is a schematic diagram of the image enhancement method for boron concentration imaging in BNCT according to the present invention.

[0044] Figure 2 This is a schematic diagram of the image enhancement system for boron concentration imaging in BNCT according to the present invention. Detailed Implementation

[0045] The technical solutions of some embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments disclosed in the present invention, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments disclosed in the present invention are within the scope of protection of the present invention. It should be noted that in the drawings with reference numerals, the same reference numerals and letters represent similar parts. Once a part is defined in a drawing, it will not be defined and explained again in subsequent drawings.

[0046] This invention can be applied to the real-time monitoring of boron concentration distribution in vivo during BNCT, and solves the technical problem of poor imaging quality in existing boron concentration imaging measurement methods during BNCT.

[0047] The image enhancement method, system, equipment, and medium for boron concentration imaging in BNCT disclosed in this invention have the following technical advantages:

[0048] 1. The image enhancement method and system for boron concentration imaging in BNCT disclosed in this invention utilizes data from two types of gamma rays and constructs a concentration inversion model using the ratio of 478keV to 2.224MeV dual-energy gamma rays. This avoids reliance on empirical coefficients and biological models and directly establishes measurement indicators based on nuclear reaction yield, thus having a clear physical source and quantitative interpretation capability.

[0049] 2. The image enhancement method and system for boron concentration imaging in BNCT disclosed in this invention also realizes the acquisition of photon transport response matrix based on Monte Carlo simulation, ensuring the accurate correspondence between gamma data and tissue structure in three-dimensional space, realizing the reconstruction of concentration distribution map with millimeter-level spatial resolution, and significantly improving spatial positioning accuracy.

[0050] 3. The image enhancement method and system for boron concentration imaging in BNCT disclosed in this invention also utilizes a trained image enhancement model to enhance the image of the three-dimensional boron concentration distribution map within voxels, thereby further improving image quality;

[0051] 4. The image enhancement method and system for boron concentration imaging in BNCT disclosed in this invention combine non-invasiveness, physical accuracy, spatial precision, and clinical applicability, providing a brand-new dose monitoring method and intelligent control support for BNCT treatment, and has outstanding scientific research value and translational prospects.

[0052] To further illustrate the system and method of the present invention, the following embodiments are disclosed.

[0053] In some embodiments, an image enhancement method for boron concentration imaging in BNCT is provided; please refer to [link to relevant documentation]. Figure 1 This includes the following steps:

[0054] S1. Acquire MRI and CT images of the target object, and perform registration and fusion to obtain an MRI-CT fused three-dimensional voxel model;

[0055] S1. Acquire spatial distribution images of the first gamma source corresponding to 478keV and 2.224MeV gamma rays, as well as the corresponding first gamma count and second gamma count;

[0056] S3. Construct a boron concentration inversion model;

[0057] S4. Obtain the photon transport response matrix, and input the first gamma count and the second gamma count into the boron concentration inversion model to perform inversion and obtain the first boron concentration distribution data.

[0058] S5. Obtain the tissue mask function in the MRI-CT fused three-dimensional voxel model, and fuse the tissue mask function, the spatial distribution image of the first gamma source, and the first boron concentration distribution data to obtain a three-dimensional boron concentration distribution map within the voxel.

[0059] This embodiment utilizes data from two types of gamma rays simultaneously and constructs a concentration inversion model using the ratio of 478 keV to 2.224 MeV dual-energy gamma rays. This avoids reliance on empirical coefficients and biological models, and directly establishes measurement indicators based on nuclear reaction yields, providing a clear physical source and quantitative interpretation capability. Furthermore, it obtains a photon transport response matrix, ensuring an accurate correspondence between gamma data and tissue structure in three-dimensional space, achieving millimeter-level spatial resolution concentration distribution map reconstruction, and significantly improving spatial positioning accuracy.

[0060] In some embodiments, MRI and CT images of the target object are acquired, registered, and fused. The process is as follows:

[0061] MRI image I MRI Compared with CT images I CT (r) After registration, the following spatial coordinate mapping relationship is satisfied:

[0062] I CT (r)≈I MRI (T(r)) (1)

[0063] In equation (1), r is the position of the voxel in the MRI coordinate system; T(r) is the affine or non-rigid transformation function during the registration process; I MRI This represents the grayscale intensity of CT and MRI images.

[0064] After registration, regions such as tumors, brain tissue, soft tissue, blood vessels, and cavities are segmented, and a three-dimensional spatial voxel model consistent with the gamma detector coordinate system is generated. A tissue mask function M is defined. k Let (r)∈{0,1}, indicating whether position r belongs to the k-th type of tissue (such as tumor, white matter, etc.). Then the tissue attribute mapping is:

[0065] μ(r)=∑ k M k (r)·μ k (2)

[0066] ρ(r)=∑ k M k (r)·ρ k (3)

[0067] In equations (2) and (3), r is the position of the voxel in the MRI coordinate system, and each r corresponds to a spatial sampling point after MRI-CT registration; k is the index number of the tissue type, and different k correspond to different tissue categories, such as tumor, white matter, gray matter, soft tissue, bone tissue, blood, cavity, etc.; M kρ(r) is the tissue mask function, indicating whether position r belongs to the k-th tissue type; ρ(r) is the mass density at position r or an equivalent parameter used to calculate the hydrogen atom number density. Specifically, in boron concentration inversion, ρ(r) can be combined with the hydrogen mass fraction to calculate N. H (r)(hydrogen atom number density); ρ k The mass density of the k-th type of tissue, or defined in the design as a parameter that can be directly mapped to the hydrogen atom number density; ∑ k To sum over all organization types, the mask Mk(r) is summed with the corresponding physical property μ. k or ρ k Multiply and then add them together to obtain the equivalent physical properties of the voxel.

[0068] In some embodiments, spatial distribution images of the first gamma source corresponding to 478 keV and 2.224 MeV gamma rays, as well as the corresponding first and second gamma counts, are acquired, respectively, using a high-purity germanium (HPGe) detector or a LaBr3 (Ce) scintillator array. The detector array is deployed at multiple locations around the main beam channel of the treatment room, with some areas using a fixed multi-angle arrangement.

[0069] In some embodiments, the method for obtaining the photon transport response matrix in the boron concentration inversion module is as follows:

[0070] The MRI-CT fused three-dimensional voxel model and the treatment room spatial parameters were input into the Monte Carlo simulation platform;

[0071] The location information of multiple gamma ray detectors is marked in the treatment room space;

[0072] Define neutron beam source term parameters;

[0073] The multi-channel photon transport process under monoenergetic gamma source conditions of 478 keV and 2.224 MeV was simulated respectively, and the photon transport response function of each channel was obtained.

[0074] The photon transport response functions of each channel are coupled to obtain the photon transport response matrix.

[0075] In this technical solution, a corresponding photon response function library needs to be constructed based on the actual BNCT treatment device before implementation. This ensures the spatial accuracy and quantitative accuracy of boron concentration inversion. The process relies on Monte Carlo platforms such as Geant4 or MCNP. In the three-dimensional patient model and treatment environment simulation, the real parameters such as the energy spectrum structure, flux distribution, and angular distribution of the accelerator neutron source are reproduced. At the same time, combined with the geometric model of the treatment beam collimator, multilayer water or graphite modulator and detector arrangement, a γ response matrix covering the entire treatment field of view is generated.

[0076] Specifically, the photon transport response function of the i-th channel is:

[0077] C i =∫∫∫VR i (r)·S γ (r)dr (4)

[0078]

[0079] In equations (4) and (5), C i R is the observation count of the i-th gamma-ray detector channel; i (r) represents the gamma source intensity at voxel location r, N i,r N represents the number of photons effectively detected by the i-th channel after a gamma photon is emitted at voxel position r. emitted S represents the total number of gamma photons emitted at voxel position r; γ (r) is the response function of the photon at this position propagating to the i-th gamma-ray detector; V is the voxel volume.

[0080] In some embodiments, the boron concentration inversion model is:

[0081]

[0082] In equations (6) and (7), N B per unit volume 10 B atomic density, N H The density of hydrogen atoms, for 10 The (n,α) reaction cross section of B for 1 The (n,γ) reaction cross section of H, f γ for 10 B(n,α) 7 The Li reaction produces a branching ratio of 478 keV gamma; ∈478 represents the detector's detection efficiency for 478 keV gamma; ∈2.224 represents the detector's detection efficiency for 2.224 MeV gamma, N 478 N 2.224 These are the first gamma count and the second gamma count, respectively. Specifically, the... f γ The parameters ∈478 and ∈2.224 were obtained from the constructed photon response function library.

[0083] Specifically, the derivation process of the boron concentration inversion model is as follows:

[0084] 1) Relationship between reaction flux and gamma intensity ratio:

[0085] Let the intravoxel thermal neutron flux be Φ. n unit volume10 The atomic density of B is N B The hydrogen atom density is N H The gamma emissivity per unit volume is:

[0086] 478keV γ source strength:

[0087] S 478 =Φn·NB·σB (n,α) ·f γ (8)

[0088] 2.224keV γ source strength:

[0089] S 2.224 =Φn·NB·σB (n,γ) (9)

[0090] In equations (8) and (9), for 10 The (n,α) reaction cross section of B for 1 The (n,γ) reaction cross section of H, f γ for 10 B(n,α) 7 The Li reaction produces a branching ratio of 478 keV γ.

[0091] 2) Detection ratio and concentration inversion model:

[0092]

[0093] Therefore, it is possible to extract voxels. 10 B concentration (molar ratio):

[0094]

[0095] If N in the organization is known H (N can be determined through CT density conversion or tissue classification) H (This allows for further obtaining the unit mass or unit volume) 10 B mass fraction W B : In the formula, M B for 10 The mass fraction of B.

[0096] The above-mentioned gamma ratio inversion formula forms the theoretical basis of this invention, and its core advantages are reflected in the following aspects:

[0097] First, with the help of 10 B(n,α) 7 The characteristic 478 keV gamma line produced by the Li reaction and 1H(n,γ) 2The intensity ratio between the 2.224 MeV gamma lines produced by the D reaction can be effectively eliminated by this method to reduce the influence on the absolute neutron flux Φ. n This characteristic significantly improves the stability and robustness of the measurement system under complex field strength conditions, and reduces the accuracy requirements for neutron field calibration.

[0098] Secondly, this model constructs a proportional relationship by using the relative counts of two energy windows in the gamma spectrum, enabling it to achieve [the desired result] without relying on invasive sampling. 10 The quantitative inversion of B concentration has high physical traceability and algorithmic versatility.

[0099] Furthermore, the local boron concentration calculated in the model can be directly mapped to medical image registration space (such as UTE-MRI or CT anatomical voxels), thereby achieving 3D tissue-scale reconstruction of boron spatial distribution. This characteristic provides strong data support for real-time feedback and dose control during BNCT treatment, and helps to advance the establishment of an integrated closed-loop control system of "targeting-feedback-regulation".

[0100] Specifically, the tissue mask function in the MRI-CT fused three-dimensional voxel model is obtained, and the tissue mask function, the spatial distribution image of the first gamma source, and the first boron concentration distribution data are fused to obtain a three-dimensional boron concentration distribution map within the voxel. This is achieved through the following process:

[0101] Solving the linear inverse problem: C measured =R·S γ +ε; This yields the voxel-level gamma source distribution Sγ, where ε is the noise term.

[0102] Subsequently, based on the inversion relationship between gamma source intensity and boron concentration:

[0103] S γ =Φ n ·N B ·σ B ·V (12)

[0104] In equation (12), Φ n For voxel endothermic neutron flux, N B For voxel intrathermal neutron flux voxel intra 10 B atomic number density, σ B for 10 B(n,α) 7 Li reaction cross section, V is voxel volume.

[0105] The two characteristic γ generation terms at voxel r are:

[0106] S 478 (r)=κ 478 ·Φ n(r)·σ B ·N B (r)·V (13)

[0107] S 2.224 (r)=κ 2.224 ·Φ n (r)·σ H ·N H (r)·V (14)

[0108] In equations (13) and (14), N B per unit volume 10 B atomic density, N H For the density of hydrogen atoms, σ B σ H For the corresponding microscopic cross section, κ 478 and κ 2.224 The normalization factors (κ) are for 478 keV and 2.224 MeV, respectively.

[0109] Specifically, the normalization factor κ is a comprehensive proportionality coefficient that converts the theoretical gamma generation rate into the detection count rate. It includes terms such as gamma branching ratio, geometric efficiency, detection efficiency, and medium attenuation, and is determined through Monte Carlo simulation and experimental calibration.

[0110] Based on this, the dual-energy ratio is constructed as follows:

[0111]

[0112] in This is a calibrable constant; the remaining characters have the same meaning as described above.

[0113] Voxel-level 10B concentration can be obtained from: Find it. If N H Given the data (obtained from CT density / tissue classification), the mass fraction W can be further calculated. B (r).

[0114] N B (r) converted to intravoxel boron concentration

[0115]

[0116] In equation (16), N A It is Avogadro's constant, N B (r) is 10 B atomic density, M B for 10 The mass fraction of B.

[0117] The dual-energy ratio method used in this embodiment constructs a ratio from the counts of two energy windows, 478 keV and 2.224 MeV, and participates in the inversion, automatically canceling Φ. n Common factors such as voxel volume V reduce the impact on absolute neutron flux Φ n Dependence.

[0118] In some embodiments, the image enhancement method for boron concentration imaging in BNCT further includes: S6, training a deep learning model with the voxel three-dimensional boron concentration distribution map to obtain a boron concentration image enhancement model, and enhancing the voxel three-dimensional boron concentration distribution map to obtain an enhanced voxel three-dimensional boron concentration distribution map.

[0119] This embodiment utilizes a deep learning model to enhance the image of the three-dimensional boron concentration distribution map within voxels, further improving image quality. Specifically, the deep learning model employs either GAN or Transformer.

[0120] In some embodiments, the deep learning is preferably GAN.

[0121] In some embodiments, the boron concentration image enhancement model is trained using the following method:

[0122] S1. Obtain a set of boron concentration distribution images, the set of images including the obtained reference boron concentration distribution map and the corresponding degraded boron concentration distribution map;

[0123] S2. Preprocess the images in the boron concentration distribution image set, and divide them into training set and test set according to a preset ratio;

[0124] S3. Initialize the generator and discriminator of the GAN, and set the optimizer; the generator is used to extract and enhance features from the degraded boron concentration distribution map, and the discriminator is used to determine the difference between the generated image and the reference boron concentration distribution map.

[0125] S4. The model is trained using the training set, the total loss is defined, and the model is evaluated using the test set. After multiple rounds of iterative training, the boron concentration image enhancement model is obtained.

[0126] Specifically, in step S2, the preprocessing of the images in the boron concentration distribution image set includes standardizing the boron concentration distribution images to ensure spatial consistency and statistical stability of the input data. Specifically, standardizing the boron concentration distribution images includes normalization and size unification processing. Specifically, the three-dimensional boron concentration distribution map within the voxel is normalized to unify the grayscale value range to the [0,1] interval; simultaneously, the image is uniformly adjusted to 256×256 or 512×512 pixels to meet the input dimension requirements of GAN.

[0127] Specifically, the preset ratio is: the ratio of training set to test set is 8:2, that is, 80% of the samples are used for model training and 20% of the samples are used for testing and verification; in other embodiments, a ratio of 7:3 or 9:1 can also be used to verify the robustness and stability of the model under different sample sizes.

[0128] Specifically, the degraded boron concentration distribution map is obtained using a SPECT system or by artificially adding noise.

[0129] Specifically, the total loss is:

[0130] L total =αL phys +βL MAE +γL SSIM +δL adv +ηL cycle-perc (17)

[0131] In equation (17), L phys For physical consistency loss; L MAE For the mean absolute error loss; L adv To combat losses; L SSIM For structural similarity loss; L cycle-perc The loss is the cyclic perceptual consistency loss; α, β, γ, δ, and η are the weight coefficients of each item, and the preferred values ​​are: α = 1.0, β = 1.0, γ = 0.2, δ = 0.5, and η = 0.2.

[0132] In some embodiments, please refer to Figure 2 This embodiment provides an image enhancement system for boron concentration imaging in BNCT, comprising:

[0133] The image acquisition module is used to acquire MRI and CT images of the target object;

[0134] The image registration and fusion module is used to register and fuse the acquired MRI images with CT images to obtain an MRI-CT fused three-dimensional voxel model.

[0135] The gamma source spatial distribution image acquisition module is used to acquire the first gamma source spatial distribution image and the corresponding first gamma count and second gamma count for gamma rays of 478 keV and 2.224 MeV, respectively.

[0136] The boron concentration inversion module is used to input the photon transport response matrix, the first gamma count, and the second gamma count into the boron concentration inversion model to perform inversion and obtain the first boron concentration distribution data.

[0137] The image fusion module is used to obtain the tissue mask function in the MRI-CT fused three-dimensional voxel model, and fuse the tissue mask function, the spatial distribution image of the first gamma source, and the first boron concentration distribution data to obtain a three-dimensional boron concentration distribution map within the voxel.

[0138] The image enhancement system for boron concentration imaging in BNCT disclosed in this embodiment utilizes data from two types of gamma rays and constructs a concentration inversion model using the ratio of 478 keV to 2.224 MeV dual-energy gamma rays. This avoids reliance on empirical coefficients and biological models, and directly establishes measurement indicators based on nuclear reaction yields, providing a clear physical source and quantitative interpretation capability. Furthermore, it ensures the precise correspondence between gamma data and tissue structure in three-dimensional space through the photon transport response matrix, achieving millimeter-level spatial resolution concentration distribution map reconstruction and significantly improving spatial positioning accuracy.

[0139] The solution objective of this embodiment is at the voxel level. 10 B Concentration distribution C B (r) can also be converted to mass fraction W B (r). The process is as follows: based on the multi-channel probe counting vector C measured Using the system response matrix, first solve the linear inverse problem to obtain the voxel-level gamma source intensity S. γ Then, using nuclear reaction relations, S γ Inversion C B (r).

[0140] In some embodiments, the image enhancement system for boron concentration imaging in BNCT further includes an image enhancement module, used to train a deep learning model with the intravoxel three-dimensional boron concentration distribution map to obtain a boron concentration image enhancement model, and to enhance the intravoxel three-dimensional boron concentration distribution map to obtain an enhanced intravoxel three-dimensional boron concentration distribution map.

[0141] In some embodiments, the image enhancement system for boron concentration imaging in BNCT further includes a visualization and ROI statistics module, which is used to visualize the enhanced three-dimensional boron concentration distribution map within voxels output by the image enhancement module, and to calculate and output statistical indicators of boron concentration corresponding to any ROI region.

[0142] Specifically, the Region of Interest (ROI) statistics module can analyze user-specified tumor areas or normal tissue areas Ω. i Automatic calculation:

[0143]

[0144] The results are output in numerical tabular form, which can be used for comparative analysis between treatment courses and model validation.

[0145] Specifically, based on the reconstruction of CB Based on the neutron flux distribution Φn(r), the boron dosage factor can be further calculated:

[0146] BDF(r) = C B (r)·Φ n (r)·σ B ·E a Li (19)

[0147] In equation (19), E a Li is 10 The B(n,α)7Li reaction releases energy.

[0148] In some embodiments, the statistical indicators of boron concentration include one or a combination of the average, maximum, minimum, and standard deviation.

[0149] In some embodiments, a device is also provided, characterized in that it includes a processor and a memory, the processor being coupled to the memory;

[0150] The memory is used to store computer programs, and the processor is used to execute the computer programs stored in the memory to implement the image enhancement method for boron concentration imaging in BNCT as described above.

[0151] In some embodiments, a medium is also provided for storing a computer program, characterized in that the computer program is executed by a processor to implement the image enhancement method for boron concentration imaging in BNCT as described above.

[0152] In the description disclosed in this invention, it should be understood that the terms "center", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to has a specific orientation, or is constructed and operated in a specific orientation. Therefore, they should not be construed as limiting the disclosure of this invention.

[0153] Unless the context otherwise requires, throughout the specification and claims, the term "comprising" is interpreted as open-ended and encompassing, meaning "including, but not limited to." In the description, terms such as "one embodiment," "some embodiments," "exemplary embodiment," "exemplary," or "some examples" are intended to indicate that a particular feature, structure, material, or characteristic associated with that embodiment or example is included in at least one embodiment or example disclosed in the invention. The illustrative representations of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics mentioned may be included in any suitable manner in any one or more embodiments or examples.

[0154] The terms "first" and "second" are used merely to distinguish different descriptive objects and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated; that is, they do not limit the position, order, priority, quantity, or content of the described objects. Therefore, a feature specified as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of the embodiments disclosed in this invention, unless otherwise stated, "a plurality of" means two or more.

[0155] In describing some embodiments, the term "connection" and its derivative expressions may be used. For example, the term "connection" may be used in describing some embodiments to indicate that two or more components have direct physical or electrical contact with each other. The embodiments disclosed herein are not necessarily limited to the content of this document.

[0156] "At least one of A, B, and C" has the same meaning as "at least one of A, B, or C," both including the following:

[0157] Combinations of A, B, and C: A only, B only, C only, combinations of A and B, combinations of A and C, combinations of B and C, and combinations of A, B, and C. In the description of the embodiments of this application, unless otherwise stated, " / " means "or", for example, A / B can mean A or B; "A and / or B" includes the following three combinations: A only, B only, and combinations of A and B.

[0158] The use of “configured as” in this article implies an open and inclusive language that does not exclude the applicability to or configuration of devices to perform additional tasks or steps.

[0159] In addition, the use of “based on” implies openness and inclusivity, because processes, steps, calculations or other actions “based on” one or more of the stated conditions or values ​​may in practice be based on additional conditions or values ​​beyond those stated.

[0160] As used herein, “about” and “approximately” include the values ​​stated and the average values ​​within an acceptable range of deviation from a particular value, wherein the acceptable range of deviation is determined by a person skilled in the art taking into account the measurement under discussion and the error associated with the measurement of the particular quantity (i.e., the limitations of the measurement system).

[0161] This document describes exemplary embodiments with reference to cross-sectional views and / or plan views, which are idealized exemplary drawings. In the drawings, the thickness of layers and regions is enlarged for clarity. Therefore, variations in shape relative to the drawings are contemplated due to, for example, manufacturing techniques and / or tolerances. Thus, exemplary embodiments should not be construed as limited to the shapes of the regions shown herein, but rather include shape deviations due to, for example, manufacturing processes. For example, etched regions shown as rectangular would typically have curved features. Therefore, the regions shown in the drawings are schematic in nature, and their shapes are not intended to show the actual shapes of the regions of the device, nor are they intended to limit the scope of the exemplary embodiments.

[0162] The above description is merely a specific embodiment of the present invention, but the scope of protection disclosed in the present invention is not limited thereto. Any variations or substitutions conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection disclosed in the present invention. Therefore, the scope of protection disclosed in the present invention should be determined by the scope of the claims.

Claims

1. An image enhancement method for BNCT boron concentration imaging, characterized by, The method comprises the following steps: S1, acquiring MRI images and CT images of a target object, and performing registration and fusion to obtain an MRI-CT fusion three-dimensional voxel model; S1, respectively acquiring a first gamma source spatial distribution image corresponding to 478keV and 2.224MeV gamma rays and corresponding first and second gamma counts; S3, constructing a boron concentration inversion model; S4, obtaining a photon transport response matrix, and inputting the boron concentration inversion model with the first and second gamma counts to perform inversion to obtain first boron concentration distribution data; S5, obtaining a tissue mask function in the MRI-CT fusion three-dimensional voxel model, and fusing the tissue mask function, the first gamma source spatial distribution image, and the first boron concentration distribution data to obtain a three-dimensional boron concentration distribution map in a voxel.

2. The image enhancement system for boron concentration imaging in BNCT according to claim 1, further comprising: S6, training the three-dimensional boron concentration distribution map in a voxel in a deep learning model to obtain a boron concentration image enhancement model, and performing image enhancement on the three-dimensional boron concentration distribution map in a voxel to obtain an enhanced three-dimensional boron concentration distribution map in a voxel.

3. The image enhancement system for boron concentration imaging in BNCT according to claim 1, wherein: In the boron concentration inversion module, the photon transport response matrix is obtained by: Inputting the MRI-CT fusion three-dimensional voxel model and the treatment room space parameters into a Monte Carlo simulation platform; Labeling position information of a plurality of gamma ray detectors in the treatment room space; Defining a neutron beam source term parameter; Simulating multi-channel photon transport processes under conditions of 478keV and 2.224MeV single-energy gamma sources respectively to obtain respective channel photon transport response functions; Coupling the respective channel photon transport response functions to obtain a photon transport response matrix.

4. The image enhancement system for boron concentration imaging in BNCT according to claim 3, wherein: The photon transport response function of the i th channel is: C i = ∫∫∫VR i (r) · S γ (r) dr; In the formula, C i is the observed counts for the i-th gamma-ray detector channel; R i (r) is the gamma source strength at voxel position r, N i,r is the number of photons emitted at voxel position r that are effectively detected by the i-th channel after the photon has traveled to the i-th gamma ray detector; V is the voxel volume. emitted is the total number of gamma photons emitted at voxel position r; S γ (r) is the response function for a photon at that position to travel to the i-th gamma ray detector; V is the voxel volume.

5. The image enhancement system for boron concentration imaging in BNCT according to claim 1, wherein: The boron concentration inversion model is: where N B is the atomic density of B 10 is the atomic density of B H is the atomic density of H is the atomic density of B 10 is the (n, a) reaction cross section of B is the atomic density of B 1 is the (n, y) reaction cross section of H γ is the atomic density of B 10 is the (n, a) reaction cross section of B 7 is the branching ratio of Li reaction producing 478 keV gamma; is the detection efficiency of the detector for 478 keV gamma; is the detection efficiency of the detector for 2.224 MeV gamma, N 478 , N 2.224 are the first and second gamma counts, respectively.

6. The image enhancement system for boron concentration imaging in BNCT according to claim 2, wherein: The boron concentration image enhancement model is obtained by training in the following manner: S1, obtaining a boron concentration distribution image set, wherein the image set includes a reference boron concentration distribution image obtained and a corresponding degraded boron concentration distribution image; S2, preprocessing images in the boron concentration distribution image set and dividing a training set and a test set according to a preset ratio; S3, initializing a generator and a discriminator of a GAN, and setting an optimizer; the generator is used for feature extraction and enhancement of the degraded boron concentration distribution image, and the discriminator is used for judging differences between generated images and the reference boron concentration distribution image; S4, performing model training using the training set, defining a total loss, and performing model evaluation using the test set, and obtaining a boron concentration image enhancement model after multiple rounds of iterative training.

7. An image enhancement system for boron concentration imaging in BNCT using the image enhancement method for boron concentration imaging in BNCT according to any one of claims 1 to 6, characterized by The method comprises: An image acquisition module is configured to acquire MRI images and CT images of a target object; An image registration and fusion module is configured to register and fuse the acquired MRI images and CT images to obtain an MRI-CT fusion three-dimensional voxel model; A gamma source spatial distribution image acquisition module is configured to acquire a first gamma source spatial distribution image corresponding to 478 keV and 2.224 MeV gamma rays, and corresponding first and second gamma counts, respectively; A boron concentration inversion module is configured to input a photon transport response matrix, the first gamma count and the second gamma count into a boron concentration inversion model, perform inversion, and obtain first boron concentration distribution data; An image fusion module is configured to obtain a tissue mask function in the MRI-CT fusion three-dimensional voxel model, and fuse the tissue mask function, the first gamma source spatial distribution image and the first boron concentration distribution data to obtain a three-dimensional boron concentration distribution map in a voxel.

8. The image enhancement system for boron concentration imaging in BNCT according to claim 7, further comprising an image enhancement module configured to train a deep learning model using the three-dimensional boron concentration distribution map in a voxel to obtain a boron concentration image enhancement model, perform image enhancement on the three-dimensional boron concentration distribution map in a voxel to obtain an enhanced three-dimensional boron concentration distribution map in a voxel. A processor, a memory, the processor being coupled to the memory; 9. An apparatus, comprising: The memory is configured to store a computer program, and the processor is configured to execute the computer program stored in the memory to implement the image enhancement method for boron concentration imaging in BNCT according to any one of claims 1 to 6. The computer program is executed by the processor to implement the image enhancement method for boron concentration imaging in BNCT according to any one of claims 1 to 6.

10. A medium for storing a computer program, characterized by ​

Images