Adaptive automatic segmentation in computed tomography

Abstract

Embodiments of the present disclosure relate to adaptive automatic segmentation in computed tomography. A computer-implemented method of segmenting a reconstructed volume of a patient's anatomical region includes determining an anatomical region associated with the reconstructed volume, detecting one or more metal objects associated with the reconstructed volume arranged in an initial 3D metal object mask, determining, for each metal object of the one or more metal objects arranged in the initial 3D metal object mask, a volume associated with the metal object, determining a value of at least one segmentation parameter based on the anatomical region and the volumes associated with the one or more metal objects, and generating a final 3D metal object mask associated with the reconstructed digital volume using the value of the segmentation parameter.

Description

[0001] Cross-references to this application

[0002] This application claims the benefit of U.S. Provisional Application No. 63 / 250,046, filed September 29, 2021. The aforementioned U.S. Provisional Application, including any appendices or annexes thereof, is incorporated herein by reference in its entirety. Technical Field

[0003] This disclosure relates to a computer-implemented method, and more particularly to a computer-implemented method for modifying an X-ray projection image of an object region. Background Technology

[0004] Unless otherwise stated herein, the methods described in this section are not prior art to the claims of this application and are not acknowledged as prior art by virtue of their inclusion in this section.

[0005] Radiation therapy is a localized treatment used on a specific target tissue (planned target volume), such as a cancerous tumor. Ideally, radiation therapy is performed on the planned target volume, which protects surrounding normal tissue from receiving doses exceeding a specified tolerance, thereby minimizing the risk of damage to healthy tissue. Before administering radiation therapy, an imaging system is typically used to provide three-dimensional images of the target tissue and surrounding area. From this imaging, the size and mass of the target tissue can be estimated, the planned target volume determined, and an appropriate treatment plan generated.

[0006] To ensure that the prescribed dose is correctly delivered to the planned target volume (i.e., the target tissue) during radiotherapy, the patient must be correctly positioned relative to the linear accelerator delivering the radiation. Typically, dosimetry and geometric data are checked before and during treatment to ensure correct patient placement and that the administered radiation matches the previously planned treatment. This procedure is known as image-guided radiotherapy (IGRT) and involves using an imaging system to visualize the target tissue immediately before or during the delivery of radiation to the planned target volume. IGRT incorporates imaging coordinates from the treatment plan to ensure the patient is correctly aligned for treatment within the radiotherapy apparatus. Summary of the Invention

[0007] According to various embodiments, the reconstructed volume of the object region is processed to reduce visual artifacts in the reconstructed volume. Specifically, in some embodiments, novel internal restoration, hybrid, and / or metal object extraction techniques are employed to reduce or eliminate visual artifacts appearing in the reconstructed volume of anatomical regions including one or more metal objects (such as reference objects or dental / orthopedic components).

[0008] In some embodiments, an automated process is employed to extract positional information of metallic objects arranged within a scanned anatomical region. In this automated process, patient anatomical structures within the scanned region are identified and classified, and the types of metals present in the scanned region are determined. Based on the patient's anatomical structures and the metal types, a precise threshold is determined to facilitate the generation of a projected metallic object mask for the scanned region, and this mask is designed to reduce visual artifacts in the reconstructed volume.

[0009] In some embodiments, a novel method is employed to repair portions of a two-dimensional (2D) projection of an object region that are obscured by a metallic object. In this method, a harmonic function is used to mathematically constrain the sharpness between the values ​​of the repaired pixels and pixels with known values, both within the same 2D projection (intra-projective consistency) and between adjacent 2D projections (inter-projective consistency).

[0010] In some embodiments, a novel hybridization method is employed to blend image information from multiple types of 2D projections of an object region (e.g., initial projected image, projection metal mask, flattened projection, and / or restored projection). For example, in one such embodiment, regions within an initial (unrestored) projected image of the object region are indicated by a projection metal mask, and these regions are modified with image information from a flattened projection of the object region, generated via a conventional tissue flattening process. In this novel hybridization method, a harmonic function is used to mathematically constrain the abrupt changes between pixel values ​​within the modified region and pixel values ​​outside the modified region, both within the same 2D projection (intra-projection consistency) and between adjacent 2D projections (inter-projection consistency).

[0011] In some embodiments, a novel blending method is employed to recover metallic objects within a reconstructed volume of an object region. In this novel blending method, a non-binary metal mask is used when blending metallic object information with the reconstructed volume generated by removing the metallic objects. Because the non-binary metal mask makes the edge voxels of the metallic region appear smooth in the final reconstructed volume, this novel blending method improves the appearance of the metallic objects when returning to the reconstructed volume.

[0012] The foregoing description of the invention is merely illustrative and is not intended to be limiting in any way. Apart from the illustrative aspects, embodiments, and features described above, other aspects, embodiments, and features will become apparent from the accompanying drawings and the following detailed description. Attached Figure Description

[0013] The foregoing and other features of this disclosure will become more fully apparent from the accompanying drawings, the following description, and the appended claims. These drawings depict only a few embodiments according to this disclosure and are therefore not intended to limit its scope. This disclosure will be described with additional specificity and detail using the accompanying drawings.

[0014] Figure 1 This is a perspective view of a radiotherapy system that can advantageously realize various aspects of the present disclosure.

[0015] Figure 2 The illustrations illustrate various embodiments. Figure 1 The drive frame and stand for the radiotherapy system.

[0016] Figure 3 The illustrations illustrate various embodiments. Figure 1 The drive frame and gantry of the radiotherapy system.

[0017] Figure 4 This schematically illustrates the various embodiments based on the inclusion of... Figure 1 A digital volume constructed from projection images generated by one or more X-ray images in a radiotherapy system.

[0018] Figure 5 The effect of a metallic object on imaging a region of a patient's anatomical structure, according to one embodiment, is schematically illustrated.

[0019] Figure 6 A flowchart illustrating a computer-implemented process for imaging regions of a patient's anatomical structures, according to one or more embodiments, is provided.

[0020] Figure 7 A flowchart illustrating a computer-implemented process for reconstructing a volume of regions for segmenting a patient's anatomical structures, according to one or more embodiments.

[0021] Figure 8 A flowchart illustrating a computer-implemented process for a 2D projection of a region used to repair a patient's anatomical structure, according to one or more embodiments.

[0022] Figure 9 The illustration schematically shows an acquired 2D projection image combined with a 2D projection metal mask to form a combined 2D projection according to various embodiments.

[0023] Figure 10 A 3D matrix according to various embodiments is illustrated schematically.

[0024] Figure 11 The diagram schematically illustrates a portion of adjacent 3D matrices and projected pixels according to various embodiments.

[0025] Figure 12 This is a description of a computing device configured to execute the various embodiments.

[0026] Figure 13This is a block diagram illustrating an embodiment of a computer program product for implementing one or more embodiments. Detailed Implementation

[0027] In the following detailed description, reference is made to the accompanying drawings, which form part of the description. In the drawings, similar symbols generally denote similar parts unless the context otherwise requires. The illustrative embodiments described in the detailed description, drawings, and claims are not intended to be limiting. Other embodiments may be utilized, and other changes may be made, without departing from the spirit or scope of the subject matter presented herein. It will be readily understood that various aspects of this disclosure, as generally described herein and illustrated in the drawings, can be arranged, substituted, combined, and designed in a variety of different configurations, all of which are expressly contemplated and form part of this disclosure.

[0028] introduction

[0029] Image-guided radiotherapy (IGRT) is used to treat tumors in areas of the body that are subject to voluntary movement (such as the lungs) or involuntary movement (such as organs affected by peristalsis, gas movement, muscle contraction, etc.). IGRT involves using an imaging system to visualize the target tissue (also known as the “target volume”) immediately before or during the delivery of radiotherapy. In IGRT, image-based coordinates of the target volume from a previously determined treatment plan are compared with image-based coordinates of the target volume determined immediately before or during the application of the treatment beam. This allows for the detection of changes in surrounding organs at risk and / or movement or deformation of the target volume relative to the radiotherapy system. Therefore, dose limiting for organs at risk can be precisely implemented based on daily position and shape, and the patient’s position and / or the treatment beam can be adjusted to more precisely target the radiation dose to the tumor. For example, in the treatment of pancreatic tumors, organs at risk include the duodenum and stomach. The shape and relative position of these organs at risk relative to the target volume can change significantly daily. Therefore, precise adjustment of the shape and relative position of such organs at risk allows the dose to be increased to the target volume and achieve better treatment outcomes.

[0030] For the reconstruction of patient anatomy around a target volume, computed tomography (CT) or cone-beam computed tomography (CBCT) is frequently used to generate two-dimensional (2D) projected images from which the patient's anatomy is reconstructed. In this image reconstruction, metallic objects within the scanned anatomical region are a significant source of visual artifacts that negatively impact the quality of the image generated from the reconstructed anatomical region. These visual artifacts can be caused by one or more phenomena associated with the presence of metal in the scanned anatomical region, including beam hardening in multicolor CT and CBCT beams, photon starvation of larger metal components, scattering of imaging X-rays affecting the metal components, and motion of the metallic object during CT or CBCT acquisition. These visual artifacts typically degrade the quality of the reconstruction and the ability to accurately detect the current location of the target volume and / or critical structures adjacent to it.

[0031] According to various embodiments, the reconstructed volume of a region of the patient's anatomical structure is processed to reduce metal-related visual artifacts within the reconstructed volume. Specifically, in some embodiments, one or more automated metal extraction processes, a novel method for repairing 2D projected portions, a novel mixing method for blending image information from multiple types of 2D projections, and / or a novel mixing method for recovering metallic objects within the reconstructed volume are employed to reduce visual artifacts within the reconstructed volume.

[0032] System Overview

[0033] Figure 1 This is a perspective view of a radiotherapy system 100 that can advantageously implement various aspects of the present disclosure. The radiotherapy (RT) system 100 is a radiation system configured to detect intraspinal motion in near real-time using X-ray imaging techniques. Therefore, the RT system 100 is configured to provide stereotactic radiosurgery and precise radiotherapy for indicating lesions, tumors, and conditions anywhere in the body for radiotherapy. Thus, the RT system 100 may include a linear accelerator (LINAC) that generates a megavolt (MV) treatment beam of high-energy X-rays, one or more kilovolt (kV) X-ray sources, one or more X-ray imagers, and one or more of an MV electronic port imaging device (EPID) in some embodiments. As an example, the radiotherapy system 100 described herein is configured with a circular gantry. In other embodiments, the radiotherapy system 100 may be configured with a C-shaped gantry capable of infinite rotation via slip rings.

[0034] Generally, the RT system 100 is capable of kV imaging of a target volume immediately before or during the application of an MV treatment beam, thereby enabling the use of X-ray imaging to perform IGRT and / or intensity-modulated radiotherapy (IMRT) procedures. The RT system 100 may include one or more touchscreens 101, a treatment bed motion controller 102, an aperture 103, a pedestal positioning assembly 105, a treatment bed 107 disposed on the pedestal positioning assembly 105, and an image acquisition and treatment control computer 106, all disposed within the treatment room. The RT system 100 also includes a remote console 110 disposed outside the treatment room and capable of treatment delivery and patient monitoring from a remote location. The pedestal positioning assembly 105 is configured to precisely position the treatment bed 107 relative to the aperture 103, and the motion controller 102 includes input devices (such as buttons and / or switches) that allow a user to operate the pedestal positioning assembly 105 to automatically and precisely position the treatment bed 107 relative to the aperture 103 to a predetermined position. The motion controller 102 also enables the user to manually position the treatment bed 107 to a predetermined location.

[0035] Figure 2 The drive frame 200 and gantry 210 of the RT system 100 according to various embodiments are schematically shown. For clarity, in Figure 2 The cover, base positioning assembly 105, treatment bed 107, and other components of the RT system 100 are omitted. The drive frame 200 is a fixed support structure for the components of the RT treatment system 110, including a gantry 210 and a drive system 201 for rotatably moving the gantry 210. The drive frame 200 rests on and / or is fixed to a support surface (such as the floor of an RT treatment facility) outside the RT treatment system 110. The gantry 210 is rotatably connected to the drive frame 200 and is the support structure on which various components of the RT system 100 are mounted, including a linear accelerator (LINAC) 204, an electronic port imaging device (EPID) 205, an imaging X-ray source 206, and an X-ray imager 207. During operation of the RT treatment system 110, the gantry 220 rotates about the aperture 103 when actuated by the drive system 201.

[0036] A drive system 201 rotatably actuates a gantry 210. In some embodiments, the drive system 201 includes a linear motor that is fixed to the drive frame 200 and interacts with a magnetic track (not shown) mounted on the gantry 210. In other embodiments, the drive system 201 includes another suitable drive mechanism for precisely rotating the gantry 210 about the aperture 201. A LINAC 204 generates a high-energy X-ray (or in some embodiments, electrons, protons and / or other heavily charged particles, ultra-high dose rate X-rays (e.g., for FLASH radiotherapy) or a microbeam for microbeam radiotherapy) MV treatment beam 230, and an EPD 205 is configured to acquire X-ray images using the treatment beam 230. An imaging X-ray source 206 is configured to guide a cone beam of X-rays (hereinafter referred to as imaging X-ray 231) through an isocenter 203 of the RT system 100 to an X-ray imager 207, the isocenter 203 typically corresponding to the location of the target volume 209 to be treated. Figure 2 In the illustrated embodiment, the X-ray imager 207 is depicted as a planar device, while in other embodiments, the X-ray imager 207 may have a curved configuration.

[0037] X-ray imager 207 receives imaging X-rays 231 and generates appropriate projection images therefrom. According to some embodiments, such projection images can then be used to construct or update portions of the imaging data of a digital volume corresponding to a three-dimensional (3D) region including the target volume 209. That is, a 3D image of such a 3D region is reconstructed from the projection image. In some embodiments, cone-beam computed tomography (CBCT) and / or digital tomography synthesis (DTS) can be used to process the projection images generated by X-ray imager 207. CBCT is typically used to acquire projection images over relatively long acquisition arcs (e.g., over a rotation of 180° or greater on gantry 210). Therefore, high-quality 3D reconstructions of the imaging volume can be produced. CBCT is typically employed at the start of a radiotherapy session to generate the established 3D reconstruction. For example, CBCT can be employed immediately before the application of the treatment beam 230 to generate a 3D reconstruction confirming that the target volume 209 has not moved or changed shape. Alternatively, or additionally, in some embodiments, partial data reconstruction is performed by the RT system 100 during a portion of the IGRT or IMRT procedure, wherein a portion of the image data is used to generate a 3D reconstruction of the target volume 209. For example, DTS image acquisition can be performed to generate image data of the target volume 209 as the treatment beam 230 is directed to the isocenter 203 while the gantry 210 rotates through the treatment arc. Because DTS image acquisition is performed over a relatively short acquisition arc (e.g., between approximately 10° and 60°), near real-time feedback on the shape and position of the target volume 209 can be provided via DTS imaging during the IGRT procedure.

[0038] exist Figure 2 In the illustrated embodiment, the RT system 100 includes a single X-ray imager and a single corresponding imaging X-ray source. In other embodiments, the RT system 100 may include two or more X-ray imagers, each having a corresponding imaging X-ray source. Figure 3 An example of such an embodiment is shown.

[0039] Figure 3 The drive frame 300 and gantry 310 of the RT system 100 according to various embodiments are schematically shown. Apart from the components of the RT system 100 mounted on the gantry 310 (including the first imaging X-ray source 306, the first X-ray imager 307, the second imaging X-ray source 308, and the second X-ray imager 309), the drive frame 300 and gantry 310 are substantially similar in construction. Figure 2 The drive frame 200 and gantry 200 are included. In such an embodiment, the RT system 100 includes multiple X-ray imagers, which helps to generate projected images (for reconstructing target volumes) over a shorter image acquisition arc. For example, when the RT system 100 includes two X-ray imagers and corresponding X-ray sources, the image acquisition arc for acquiring a projected image of a certain image quality can be approximately half that used to acquire a similar image quality projected image using a single X-ray imager and X-ray source.

[0040] Projected images generated by X-ray imager 207 (or by first X-ray imager 307 and second X-ray imager 309) are used to construct digital volume imaging data of the patient's anatomy within a 3D region including the target volume. Alternatively or additionally, such projected images can be used to update portions of existing imaging data corresponding to the digital volume of the 3D region. (The following is in conjunction with...) Figure 4 An example of such a digital volume is described.

[0041] Figure 4 A digital volume 400 constructed according to various embodiments is illustrated based on projected images generated by one or more X-ray imagers included in the RT system 100. For example, in some embodiments, the projected image may be generated by a single X-ray imager such as X-ray imager 207, while in other embodiments, the projected image may be generated by multiple X-ray imagers such as first X-ray imager 307 and second X-ray imager 309.

[0042] The digital volume 400 comprises multiple voxels 401 (dashed lines) of anatomical image data, where each voxel 401 corresponds to a different location within the digital volume 400. For clarity, in... Figure 4Only a single voxel 401 is shown. The digital volume 400 corresponds to a 3D region including the target volume 410. Figure 4 In the diagram, the digital volume 400 is depicted as an 8×8×8 voxel cube, but in practice, the digital volume 400 generally includes more voxels, for example, more than... Figure 4 The numbers represent orders of magnitude greater than those shown.

[0043] For the purposes of discussion, target volume 410 may refer to the tumor volume (GTV) for a specific treatment, the clinical target volume (CTV), or the planned target volume (PTV). GTV depicts, for example, the location and extent of the tumor that can be seen or imaged; CTV includes the GTV and additional margins of subclinical disease spread that are typically not imageable; PTV is a geometric concept designed to ensure that an appropriate dose of radiotherapy is actually delivered to the CTV without adversely affecting nearby organs of danger. Therefore, PTV is generally larger than CTV, but in some cases, it may be reduced in certain sections to provide a safety margin around organs of danger. PTV is typically determined based on imaging performed prior to treatment and is facilitated by X-ray imaging of digital volume 400 to ensure alignment of PTV with the current position of the patient's anatomy at the time of treatment.

[0044] According to the various embodiments described below, image information associated with each voxel 401 of the digital volume 400 is constructed using projection images generated by one or more X-ray imagers via a CBCT process. For example, this CBCT process can be employed immediately before the treatment beam 230 is delivered to the target volume 410, thereby allowing the position and shape of the target volume 410 to be confirmed before treatment begins. Furthermore, in some embodiments, image information associated with some or all of the voxels 401 of the digital volume 400 is updated using projection images generated by one or more X-ray imagers via a DTS process. For example, this DTS process can be employed after a portion of the planned treatment has begun and before the planned treatment has been completed. In this way, the position and shape of the target volume 410 can be confirmed during treatment.

[0045] CBCT image acquisition with metallic objects

[0046] Figure 5 The image schematically illustrates the effect of a metallic object 501, according to an embodiment, on imaging a region 502 of a patient's anatomy. Region 502 can be any technically feasible part of the patient's anatomy, including the head, chest, abdomen, etc. Figure 5In the illustrated embodiment, CBCT image acquisition is performed on a digital volume 500 via an imaging X-ray source 206 and an X-ray imager 207. The digital volume 500 includes a target volume 209 and extends to the edge surface 503 of region 502. In other embodiments, multiple X-ray sources and X-ray imagers may be employed. Alternatively or additionally, in some embodiments, the digital volume 500 may not include all or any portion of the edge surface 503.

[0047] Figure 5 This illustrates a 2D projection acquired using an imaging X-ray source 206 and an X-ray imager 207 arranged at a specific image acquisition position 551. In practice, CBCT image acquisition is performed at multiple image acquisition positions around region 502 to enable the generation of a digital reconstruction of digital volume 500. Therefore, when image acquisition position 551 is set, the imaging X-ray source 206 and X-ray imager 207 acquire one of a set of multiple CBCT 2D projection images, which are used together to reconstruct region 502. Furthermore, in Figure 5 In this context, the X-ray imager 207 is observed along its edges, thus depicting it as a one-dimensional imaging structure. In practice, the X-ray imager 207 is typically configured to generate 2D projected images at each of multiple image acquisition locations.

[0048] Metal object 501 can be any metallic object that appears in the X-ray image of region 502 (and / or in the digital reconstruction of region 502). For example, in some cases, metallic object 501 is a reference marker, a surgical staple or other medical device, a dental component, an orthopedic component, etc.

[0049] As shown in the figure, when the X-ray imager 207 is at the first image acquisition position 551, the metallic object 501 appears in pixel 561 of the X-ray imager 207. Therefore, in the 2D projection acquired at the first image acquisition position 551, pixel 561 is associated with the metallic object 501. It should be noted that when the 3D volume of region 502 is reconstructed using the 2D projection image acquired by the X-ray imager 207, the metallic object 501 can significantly cause visual artifacts and / or other inconsistencies. For example, the high contrast of the metallic object 501 modulates the X-ray spectrum in a way that is not modeled by typical reconstruction algorithms (which generally assume that all scanned objects have a radiometric density approximating that of water). Therefore, visual artifacts can occur. Furthermore, in the case of a relatively large high-contrast metallic object 501, other information in the occluded portion 520 (cross-shading lines) of region 502 can be masked, where the occluded portion 520 is the part of region 502 imaged as the metallic object 501 by the same pixels of the 2D projection of region 502.

[0050] Reduce visual artifacts caused by metallic objects

[0051] According to the various embodiments described below, visual artifacts (not shown) appearing in the reconstructed volume of region 502 due to the presence of the metallic object 501 are reduced or eliminated during the computer implementation for imaging the object region (such as a region of patient anatomy). The following is in conjunction with... Figure 6 Describe one such embodiment.

[0052] Figure 6 A flowchart illustrates a computer-implemented process 600 for imaging a region of a patient's anatomical structure according to one or more embodiments. The computer-implemented process 600 can be implemented as an imaging process only, or in conjunction with radiotherapy such as IGRT, stereotactic radiosurgery (SRS), etc. Furthermore, the computer-implemented process 600 can be performed on a single arc of rotation of a gantry of a radiotherapy or imaging system, on a portion of an arc of rotation, or on multiple arcs of rotation. The computer-implemented process 600 may include one or more operations, functions, or actions as shown in one or more boxes in blocks 610-697. Although these boxes are shown in a sequential order, they may be performed in parallel and / or in an order different from that described herein. Furthermore, based on desired implementation, various blocks may be combined into fewer blocks, divided into additional blocks, and / or eliminated. Although combined with the X-ray imaging system described herein as part of a radiotherapy system 100 and... Figure 1-5 The computer-implemented process 600 is described, but those skilled in the art will understand that any properly configured X-ray imaging system is within the scope of this embodiment.

[0053] In step 610, a CT or CBCT scan is performed. For example, in some embodiments, the X-ray imaging system of the radiotherapy system 100 acquires a set of acquired 2D projection images 611 of region 502, which include the target volume 209 and the metallic object 501. Thus, region 502 includes at least one metallic object (e.g., metallic object 501).

[0054] In step 620, a first pass of the reconstruction process is performed. For example, in some embodiments, the X-ray imaging system generates an initial reconstruction volume 621 (first reconstruction volume) of region 502 based on the acquired 2D projection image 611. The initial reconstruction volume 621 is a 3D volume dataset of region 502. In some embodiments, the Feldkamp, ​​Davis, and Kress (FDK) reconstruction algorithm is used to generate the initial reconstruction volume 621. In other embodiments, an algebraic reconstruction technique (ART) or other iterative reconstruction technique is used to generate the initial reconstruction volume 621, while in other embodiments, any other suitable reconstruction algorithm is used. Note that the initial reconstruction volume 621 typically includes visual artifacts due to the presence of a metallic object 501 in region 502.

[0055] In step 630, a novel adaptive automatic segmentation process is performed on the metal object arranged within the initial reconstruction volume 621 to generate a 3D representation of the metal object. For example, in some embodiments, the X-ray imaging system performs automatic segmentation of the metal object 501 to generate a 3D representation 631 of the metal object 501 arranged within region 502. Typically, the adaptive automatic segmentation process in step 630 is performed automatically on the initial reconstruction volume 621 and does not require user input to adjust the segmentation parameter values. The 3D representation 631 includes 3D location information of the metal object 501. In some embodiments, during the novel adaptive automatic segmentation process in step 630, patient anatomy structures in region 502 are identified and classified, and the type of metal present in region 502 is determined. Based on the patient anatomy structures and the type or volume of the metal, appropriate segmentation parameter values ​​(such as thresholds and expansion radii) are determined and used to generate the 3D representation 631 of the metal object 501, sometimes referred to as a 3D metal object mask. The following is in conjunction with... Figure 7 One embodiment of a novel adaptive automatic segmentation process is described. Alternatively, in some embodiments, any suitable segmentation algorithm or software application configured to generate 3D location information of the metal object 501 may be employed in step 630, such as an algorithm that relies on user input indicating patient anatomy and / or the type of metal associated with the metal object 501.

[0056] In step 640, a first pass of forward projection is performed. For example, in some embodiments, the X-ray imaging system performs a forward projection process on the 3D representation 631 to generate a set of 2D projected metal masks 641. Each 2D projected metal mask 641 in the set includes positional information indicating pixels occluded by the metal object 501 during the forward projection process in step 640. In step 640, each generated 2D projected metal mask 641 is selected to correspond to a different acquired 2D projected image 611, which is included in the set of acquired 2D projected images 611 acquired in step 610. That is, for each 2D projected metal mask 641 generated in step 640, the forward projection process is performed using the same projection angle used to acquire one of the acquired 2D projected images 611. Therefore, each 2D projected metal mask 641 is matched to a corresponding acquired 2D projected image 611. For example, in some embodiments, each 2D projection metal mask 641 may be combined with the corresponding acquired 2D projection image 611 during the harmonic repair process in step 650.

[0057] In some embodiments, in step 640, an additional thresholding process is applied to each 2D projected metal mask 641. The thresholding process normalizes the pixels of a particular 2D projected metal mask 641 to the interval [0, 1], where 1 corresponds to the contrast structure and 0 corresponds to the remainder of region 502 (and therefore not part of the 2D projected metal mask 641).

[0058] In step 650, a novel harmonic restoration process is performed. For example, in some embodiments, the X-ray imaging system performs a harmonic restoration process on an acquired 2D projection image 611 to generate a set of 2D restored projections 651 for region 502. Specifically, each restored 2D projection 651 is generated by modifying a portion of the acquired 2D projection image 611. For the acquired 2D projection image 611, visual information (e.g., pixel values) associated with the metal object 501 is removed based on positional information included in the corresponding 2D projection metal mask 641. For example, in one such embodiment, the 2D projection metal mask 641 indicates pixels associated with the metal object 501. The pixel values ​​of these pixels are removed from a particular acquired 2D projection image 611 and replaced with lower-contrast pixel values ​​via a restoration process. (The following is in conjunction with...) Figure 8 One embodiment of the harmonic restoration process is described. Alternatively, in some embodiments, any suitable restoration algorithm may be used in step 650 to generate the 2D restoration projection set 651.

[0059] In step 660, a second reconstruction process is performed. For example, in some embodiments, the X-ray imaging system generates a reconstruction volume 661 (second reconstruction volume) of region 502 based on the repaired 2D projection 651 generated in step 650. In some embodiments, ART is used to generate reconstruction volume 661, while in other embodiments, the FDK reconstruction algorithm or other reconstruction algorithms may be used to generate reconstruction volume 661. Reconstruction volume 661 is similar to the initial reconstruction volume 621, except that the metallic object 501 has been removed.

[0060] In step 670, a tissue flattening process is performed. For example, in some embodiments, the X-ray imaging system performs a flattening process on the reconstructed volume 661 to generate a flattened reconstructed volume 671. Generally, the flattened reconstructed volume 671 contains the original information about the bone, while information related to soft tissue is replaced with a single value. Therefore, in subsequent steps, a blending process restores the bone information, while for soft tissue, the blending process serves as a repair performed in the first pass of reconstruction in step 620. In some embodiments, a simple thresholding method is used in step 670 to generate the flattened reconstructed volume 671, and in other embodiments, filtering and / or convolutional neural networks may also be used in step 670 to generate the flattened reconstructed volume 671.

[0061] In step 680, a second forward projection process is performed. For example, in some embodiments, the X-ray imaging system performs a forward projection process on a flattened reconstruction volume 671 to generate a set of flattened 2D projections 681. It should be noted that in each flattened 2D projection 681, pixels visually obstructed by the metal object 501 (i.e., pixels indicated by positional information in a 2D projected metal mask associated with the metal object 501) have pixel values ​​that do not include contributions from the metal object 501. Instead, in each flattened 2D projection 681, the pixel values ​​of pixels visually obstructed by the metal object 501 are based on the flattened reconstruction volume 671, which includes skeletal information of region 502.

[0062] In step 680, each generated flat 2D projection 681 is selected to correspond to a different acquired 2D projection image 611, which is included in the group of acquired 2D projection images 611 acquired in step 610. That is, for each flat 2D projection 681 generated in step 680, the forward projection process is performed using the same projection angle used to acquire one of the acquired 2D projection images 611. Therefore, each flat 2D projection 681 is matched with a corresponding acquired 2D projection image 611 and can subsequently be combined with it, for example, in the mixing process of step 690.

[0063] In step 690, a novel harmonic mixing process is performed. For example, in some embodiments, the X-ray imaging system generates a set of low-artifact 2D projections 691 for region 502 by modifying the acquired 2D projection image 611 based on a 2D projection metal mask 641 and a flattened 2D projection 681. Specifically, positional information from the 2D projection metal mask 641 indicates certain pixels of the acquired 2D projection image 611 that will have image information (e.g., pixel values) replaced by image information from the corresponding pixels of the flattened 2D projection 681. It should be noted that, generally, most pixels of the low-artifact 2D projections 691 have the same pixel values ​​and / or other image information as the corresponding pixels of the acquired 2D projection image 611. However, pixels of the low-artifact 2D projections 691 indicated as being blocked by or associated with the metal object 501 include pixel values ​​and / or other image information different from the corresponding pixels of the acquired 2D projection image 611.

[0064] In some embodiments, a harmonic mixing process is performed to minimize or otherwise reduce visual artifacts caused by replacing image information in groups of pixels in the acquired 2D projected image 611 with pixel information from a flat 2D projection 681. For example, in one embodiment, the acquired 2D projected image 611 forms a set of projections P1...P n The flat 2D projection 681 forms a set of projections F1...F n Furthermore, the 2D projection metal mask 641 forms a set of masks M1...M n In this embodiment, the projection F1...F n With the mask M1...M n The projections P1...P in the represented metallic region n Combination. Specifically, by dividing P by elements. i / F i To perform projection P1...P n This involves repairing the metal region, and after repair, multiplying the new value of the metal region by the value from F. i The corresponding pixel value. During the repair process, the mask boundary pixels (with respect to mask M1...M) n The projections P1...P of the adjacent pixels at the mask edge of the metal region nThe pixel values ​​of the pixels within the mask boundary are used as boundary conditions. Furthermore, during harmonic restoration, the pixel values ​​of the mask edge pixels are constrained to a certain pixel value based on the pixel values ​​of adjacent mask boundary pixels. Specifically, for each mask edge pixel, the pixel value is constrained such that the change in the slope of the pixel value associated with the mask edge pixel is equal to the change in the slope of the pixel value associated with the adjacent mask boundary pixel. In some embodiments, a harmonic function is used to apply this constraint to the pixel values ​​of the mask edge pixels. In some embodiments, the harmonic restoration process for generating the low-artifact 2D projection 691 is similar to the harmonic restoration process for generating the 2D restored projection 651, which will be discussed below. Figure 7 To describe in more detail.

[0065] In step 693, a third reconstruction process is performed. For example, in some embodiments, the X-ray imaging system uses a set of low-artifact 2D projections 691 of region 502 to perform a reconstruction algorithm to generate a low-artifact reconstruction volume 695 (third reconstruction volume) of region 502. Therefore, the X-ray imaging system generates a low-artifact reconstruction volume 695 of region 502 based on the low-artifact 2D projections 691 generated in step 690. In some embodiments, the FDK reconstruction algorithm is used to generate the low-artifact reconstruction volume 695. In other embodiments, the ART algorithm may be used to generate the low-artifact reconstruction volume 695. In other embodiments, the penalized likelihood (PL) reconstruction algorithm is used to generate the low-artifact reconstruction volume 695. Alternatively, any other suitable reconstruction algorithm may be used in step 690. The low-artifact reconstruction volume 695 is similar to the initial reconstruction volume 621, except that the metallic object 501 has been removed and the artifacts caused by the presence of the metallic object 501 have been removed and / or the visual salience has been reduced.

[0066] In step 697, a hybrid metal restoration process is performed. For example, in some embodiments, the X-ray imaging system generates a final reconstructed volume 699 (fourth reconstructed volume) of region 502. In step 697, the X-ray imaging system blends a low-artifact reconstructed volume 695 with image information (e.g., pixel values) from an initial reconstructed volume 621 and a 3D representation 631 from the metal object 501. Thus, the final reconstructed volume 699 is generated by restoring the metal object information to the low-artifact reconstructed volume 695.

[0067] In some embodiments, the hybrid metal restoration process of step 697 includes an operation that creates a smooth transition between metallic and non-metallic portions within the final reconstructed volume 699. In such an embodiment, the edge voxels of the metallic object 501 are not represented as voxels having a value equal to 100% of the radiographic density of the voxels within the metallic object 501. Instead, each edge voxel of the metallic object 501 in the final reconstructed volume 699 has a value based on a combination of image values ​​from the corresponding voxel in the initial reconstructed volume 621, image values ​​from the corresponding voxel in the low-artifact reconstructed volume 695, and information from the corresponding voxel in the 3D representation of the metallic mask 631. Specifically, the hybrid metal restoration process of step 697 includes a blend of three inputs: the initial reconstructed volume 621, the low-artifact reconstructed volume 695, and a non-binary mask W based on the 3D representation of the metallic mask 631.

[0068] The non-binary mask W is based on the metal mask 3D representation 631, but differs in some respects. The metal mask 3D representation 631 is a binary mask that indicates whether a specific voxel location in the low-artifact reconstruction volume 695 should be considered part of the metal object 501. Therefore, the metal mask 3D representation 631 applies a binary decision for each voxel in the low-artifact reconstruction volume 695. For example, when voxel i of mask M is considered part of the metal object 501, M... voxeli =1, and when voxel i of mask M is considered not to be part of metal object 501, W voxeli =0. In contrast, the non-binary mask W differs from the 3D representation of the metallic mask 631 in that, for some or all of the edge voxels of the metallic object 501, the value of the non-binary mask W can vary between 0 and 1.

[0069] In some embodiments, the value V is based on the corresponding voxel value of the initial reconstructed volume 621. vox Determine the value of each voxel of the non-binary mask W. vox In some embodiments, when V vox Greater than the metal threshold t metal For example, in step 630, when using the metal threshold to generate the 3D representation 631 of the metal mask, W vox =1; when V vox Less than the tissue threshold t tissue (e.g., 300HU) W vox =0; and when V vox At the metal threshold t metal and tissue threshold t tissue In between, W vox = w, where w is a value between 0 and 1. In such an embodiment, w can be based on the value of the corresponding voxel of the initial reconstructed volume 621, the metal threshold t metal and tissue threshold ttissue For example, in one such embodiment, when V vox At the metal threshold t metal and tissue threshold t tissue When w = (V) vox -ttissue) / (t metal -t tissue Therefore, in such an embodiment, linear interpolation is applied when determining the value of w. In other embodiments, when V vox At the metal threshold and tissue threshold t tissue When between, it can be based on including the metal threshold t metal and tissue threshold t tissue Any other suitable algorithm, or V vox The volume of a material is used to calculate w, which corresponds to some other measure of the metallic material.

[0070] In some embodiments, the hybrid metal restoration process in step 697 uses a weighted alpha mixture of three inputs to determine the pixel values ​​of the final reconstructed volume 699: the initial reconstructed volume 621, the low-artifact reconstructed volume 695, and the non-binary mask W. Furthermore, in one such embodiment, a maximum operator is applied to the voxel values ​​of the initial reconstructed volume 621 and the low-artifact reconstructed volume 695. In such an embodiment, the voxel value V of the final reconstructed volume 699 is determined based on Equation 1. FINAL :

[0071] V FINAL = (1 – w) * V 695 +w*max{V 695 V 621} (1)

[0072] Where V 695 This is the value of the voxel corresponding to a low-artifact reconstruction volume of 695, V. 621 This is the value of the corresponding voxel for the initial reconstructed volume 621. In such an embodiment, for V... 695 and V 621 The maximum operator is applied to restore the smooth transition between soft tissue and the final reconstructed metal objects in volume 695.

[0073] Adaptive automatic segmentation of metallic objects

[0074] Proper extraction of the projected metal mask (also known as the metal trace) is crucial in CBCT reconstruction to ensure image quality of the final reconstructed volume for areas of the patient's anatomy. When the metal trace is underestimated, metal-related artifacts are not fully suppressed. Conversely, overestimation of the metal trace can lead to over-reconstruction, resulting in redundant suppression of important non-metallic image information in the final reconstructed volume. In such cases, the image derived from the final reconstructed volume may suffer from unnecessarily lost detail, which is highly undesirable in many applications.

[0075] In the adaptive automatic segmentation process of step 630, automatic segmentation is performed on the metal object 501 arranged within the initial reconstruction volume 621 to generate a metal trace of the metal object 501 (e.g., 3D representation 631). Typically, automatic segmentation of metal objects from the reconstruction volume depends on several segmentation parameters, including a radiographic density threshold (referred to herein as the “threshold parameter”) and an expansion diameter (used to improve mask quality at mask boundaries). Because fully automated segmentation processes rely on correct values ​​of these parameters, they often exhibit poor performance. For example, when applied to detecting dental metal, an automatic segmentation process adapted for detecting hip prostheses may fail due to the much smaller size of dental metal objects and the varying structures of the environment surrounding such objects. Therefore, for a successful automatic segmentation process, user input is generally required to indicate the anatomical region associated with the scanned image data and the specific type of metal object arranged within that region.

[0076] According to various embodiments, the adaptive automatic segmentation process (such as the adaptive automatic segmentation process in step 630) is performed automatically on the reconstructed volume and does not require user input to adjust the segmentation parameter values. In such embodiments, regions of patient anatomy associated with the reconstructed volume are identified and classified, and the type of metal present in the reconstructed volume is determined. Based on the regions of patient anatomy and the type of metal, a 3D metal object mask (such as a 3D representation 631 of metal object 501) is determined and generated using precise segmentation parameter values ​​for the reconstructed volume. The following is in conjunction with... Figure 7 Describe one such embodiment.

[0077] Figure 7A flowchart of a computer-implemented process 700 for segmenting a region of a patient's anatomy according to one or more embodiments is shown. The computer-implemented process 700 can be implemented as part of an imaging process alone, or in conjunction with radiotherapy such as IGRT, stereotactic radiosurgery (SRS), etc. The computer-implemented process 700 may include one or more operations, functions, or actions as shown in one or more boxes in boxes 701-720. Although these boxes are shown in a sequential order, they can be performed in parallel and / or in a different order than described herein. Furthermore, based on the desired implementation, various blocks can be combined into fewer blocks, divided into additional blocks, and / or eliminated. Although combined with the X-ray imaging system described herein as part of a radiotherapy system 100 and... Figures 1-6 The computer-implemented process 700 is described, but those skilled in the art will understand that any properly configured X-ray imaging system is within the scope of this embodiment.

[0078] In step 701, the X-ray imaging system of the radiotherapy system 100 receives a reconstructed digital volume (such as an initial reconstructed volume 621).

[0079] In step 702, the X-ray imaging system generates an anatomical mask for reconstructing volume 621. In some embodiments, the anatomical mask is configured to include the location of the initial reconstructed volume 621 based on an anatomical threshold. For example, in some embodiments, the X-ray imaging system performs a thresholding operation on the initial reconstructed volume 621 using an anatomical threshold. In such embodiments, the anatomical threshold is selected such that voxels associated with patient anatomy are included in the anatomical mask, and voxels associated with air or other substances outside the patient anatomy are excluded from the anatomical mask. In some embodiments, the anatomical threshold is associated with a material whose radioactivity is greater than that of air and / or less than that of water.

[0080] In step 703, the X-ray imaging system identifies the anatomical regions associated with the initial reconstructed volume 621. Examples of identifying the anatomical regions include anatomical masks comprising the head and / or neck, pelvis, abdomen, limb portions, etc. The anatomical mask provides the three-dimensional shape and dimensions of the anatomical region. Therefore, any technically feasible method can be used to identify the specific anatomical regions associated with the initial reconstructed volume 621. For example, in some embodiments, a suitably trained machine learning algorithm can be employed to classify the specific anatomical regions associated with the initial reconstructed volume 621. Because the number of different anatomical regions that may be associated with the initial reconstructed volume 621 is limited, in some embodiments, a simplified geometric analysis is employed to classify the specific anatomical regions associated with the initial reconstructed volume 621.

[0081] In embodiments employing geometric analysis, the geometric analysis includes determining whether an anatomical region is a head region or a body region based on an anatomical mask. In such embodiments, the minimum lateral width of the anatomical mask can indicate whether the anatomical region is a head region or a body region. For example, under normal circumstances, the minimum lateral width of a patient's anatomical structure represented by an anatomical mask greater than 25 cm represents a body scan, while the minimum lateral width of a patient's anatomical structure represented by an anatomical mask less than 20 cm represents a head scan. Therefore, in some embodiments, the maximum lateral width (e.g., in voxels) of each slice in an anatomical mask having a non-zero width is measured. The minimum lateral width of these maximum lateral width measurements is then determined as the minimum lateral width of the anatomical mask. This minimum lateral width, compared with a maximum head width threshold and / or a minimum body width threshold, indicates whether the anatomical region is a head region or a body region. Additionally or optionally, in some embodiments, one or more additional geometric analyses may be applied to certain dimensions of the anatomical mask to determine other anatomical regions, such as the upper arm, lower arm, thigh, etc.

[0082] In step 704, the X-ray imaging system generates an initial 3D metallic object mask M for the initial reconstruction volume 621. init For example, in some embodiments, the X-ray imaging system uses an initial metal threshold t. init An image thresholding operation is performed on the initial reconstructed volume 621 to generate the initial 3D metallic object mask M. init In some embodiments, an initial metal threshold t is selected. init This ensures that all metallic objects set within the initial reconstruction volume 621 are detected. For example, in some embodiments, t init It is a value between approximately 1200 HU and 1800 HU. Furthermore, in some embodiments, to ensure the initial 3D metallic object mask M... init The initial 3D metal object mask M only includes metal objects within the patient's anatomical structure. init Multiply by the anatomical mask (which can be a binary mask with a value of 1 or 0). Thus, in such an embodiment, metallic objects outside the patient's anatomy are eliminated by multiplying the pixel values ​​associated with such metallic objects by 0.

[0083] In step 705, the X-ray imaging system determines whether a metallic object (e.g., a dental filling, reference piece, or orthopedic prosthesis) is positioned within the initial reconstructed volume 621. Typically, the X-ray imaging system uses a mask M derived from the initial 3D metallic object. init The voxels represent the metallic objects arranged within the initial reconstruction volume 621. In some embodiments, the X-ray imaging system also detects the metallic objects contained within the initial 3D metallic object mask M. initOne or more connectivity components within the initial reconstruction volume 621 are used to determine the metallic objects arranged within the initial reconstruction volume 621. In such an embodiment, each connectivity component comprises a set of adjacent and / or consecutive voxels connected within the initial reconstruction volume 621. In such an embodiment, any suitable algorithm can be used to determine the initial 3D metallic object mask M. init The connection components within. For example, in some embodiments, the method for determining the connection components is presented in: "Sequential Operations in Digital Picture Processing", A. Rosenfeld and J. Pfaltz. Journal of the ACM. Vol. 13, Issue 4, Oct. 1966, Pg. 471-494.

[0084] In some embodiments, the X-ray imaging system can determine the size (or volume) of each connected component. c and average radiographic density values ​​(such as average HU value) v c In such an embodiment, vector pairs ([s1,v1],…[s...)) can be generated. n ,v n This is to facilitate determining whether significant metallic objects are arranged within the initial reconstructed volume 621. In such an embodiment, this can be based on vector pairs ([s1,v1],…[s...). n ,v n The metal classification of the initial reconstructed volume 621 is determined by neutralizing and / or associating certain information with the vector pair. Examples of such information include the volume of the largest component among one or more connected components, the cumulative volume of a group of one or more connected components larger than a predetermined volume, and the radiographic density of at least one component among one or more connected components larger than a predetermined volume.

[0085] In step 710, the X-ray imaging system determines whether significant metallic objects (such as metallic objects of significant size) are located within the initial reconstruction volume 621. For example, in some embodiments, the X-ray imaging system performs a classification operation in step 710 to determine whether significant metallic objects are located within the initial reconstruction volume 621. Examples of metallic classification that can be applied to the initial reconstruction volume 621 include metal-free anatomical regions, dental regions, regions containing reference points (such as the abdomen), orthopedic regions (such as the pelvic region), etc.

[0086] In the example metal classification operation performed in step 710, when the paired vectors ([s1,v1],…[s…)…)… n ,v n When ]) is empty, the maximum connectivity component in the head anatomical region has a value less than the first threshold (e.g., 50 mm).3 When the volume is less than the second threshold (e.g., 200 mm), and / or when there is no volume less than the second threshold (e.g., 200 mm). 3 When the connected components of the volume have an average radiographic density value higher than a third threshold (e.g., 2000 HU), the X-ray imaging system identifies the initial reconstructed volume 621 as a metal-free anatomical region. Furthermore, in an exemplary metal sorting operation, when at least one connected component in the head anatomical region has at least a first threshold (e.g., 50 mm...), the X-ray imaging system identifies the initial reconstructed volume 621 as a metal-free anatomical region. 3 When the volume of 621 is defined, the X-ray imaging system identifies the initial reconstructed volume 621 as the tooth region. Furthermore, in an exemplary metal sorting operation, when the connectivity components in the body anatomical region have at least a fourth threshold (e.g., 10000 mm), 3 When the cumulative volume of the X-ray imaging system is reached, the initial reconstructed volume 621 is identified as the orthopedic region. Furthermore, in an exemplary metal sorting operation, when at least one small connectivity component (e.g., a connectivity component with a smaller connectivity component than the second connectivity component) is reached, the X-ray imaging system identifies the initial reconstructed volume 621 as the region containing the reference.

[0087] In step 710, if the X-ray imaging system determines that a significant metallic object is located within the initial reconstruction volume 621, the method proceeds to step 711. If the X-ray imaging system determines that no significant metallic object is located within the initial reconstruction volume 621, the method proceeds to step 720.

[0088] In step 711, for each metallic object or connected component positioned within the initial reconstruction volume 621, the X-ray imaging system determines an appropriate value for a threshold parameter. For example, in some embodiments, a tooth threshold t is selected for the connected component associated with a tooth region. dental Such as 4000HU; select a reference threshold t for the connection components associated with the region containing the reference. fiducial For example, 2000HU; select the shaping threshold t for the connection components related to the shaping area. ortho For example, 1600HU.

[0089] In step 712, for each metal object or connected component positioned within the initial reconstruction volume 621, the X-ray imaging system determines an appropriate value for the expansion radius. For example, in some embodiments, one or more tooth expansion radii, such as an in-plane radius of 2 mm and an out-of-plane radius of 2 mm, are selected for connected components associated with a dental region; one or more reference expansion radii, such as an in-plane radius of 2 mm and an out-of-plane radius of 2 mm, are selected for connected components associated with a reference region; and one or more orthopedic expansion radii, such as an in-plane radius of 3 mm and an out-of-plane radius of 10 mm, are selected for connected components associated with an orthopedic region.

[0090] In step 713, the X-ray imaging system uses the values ​​determined in steps 711 and 712 for the threshold parameter and the expansion radius to generate a final 3D metal object mask (e.g., 3D representation 631) for the initial reconstruction volume 621. Therefore, in step 713, the X-ray imaging system uses the values ​​of the threshold parameter and the expansion radius to perform automatic segmentation of the initial reconstruction volume 621.

[0091] In step 714, the X-ray imaging system generates a set of projection metallic masks (e.g., 2D projection metallic mask 641) for the initial reconstructed volume 621. The projection metallic masks generated in step 714 can be used in multiple steps to generate a digital volume with reduced metallic artifacts, including... Figure 6 The harmonic repair in step 650 and the harmonic mixing in step 690.

[0092] In step 720, the X-ray imaging system aborts the current metal artifact reduction process and typically performs reconstruction, filtering, and / or other processes on the initial reconstruction volume 621.

[0093] Projection Harmonic Repair

[0094] In conventional methods, linear one-dimensional interpolation is used to perform the restoration of 2D projected images (such as...). Figure 6 The 2D projection metallic mask 641 is shown. This restoration produces intra-projective discontinuities (i.e., between adjacent rows within the same 2D projection) and inter-projective discontinuities (i.e., between corresponding pixel rows in adjacent projections). Such discontinuities produce significant visual artifacts in the final reconstructed volume, even though image information obscured by the metallic object has been restored. According to various embodiments, the harmonic restoration process improves intra-projective and inter-projective discontinuities in pixel values, which reduces visual artifacts typically caused by restoration. The following is in conjunction with... Figures 8-11 Describe one such embodiment.

[0095] Figure 8 A flowchart illustrating a computer-implemented process 800 for a 2D projection of a region for repairing a patient's anatomical structures, according to one or more embodiments, is shown. The computer-implemented process 800 can be implemented as part of an imaging process alone, or in conjunction with radiotherapy such as IGRT, stereotactic radiosurgery (SRS), etc. The computer-implemented process 800 may include one or more operations, functions, or actions as shown in one or more boxes in blocks 801-806. Although these boxes are shown in a sequential order, they can be performed in parallel and / or in a different order than described herein. Furthermore, based on the desired implementation, various blocks can be combined into fewer blocks, divided into additional blocks, and / or eliminated. Although combined with the X-ray imaging system described herein as part of a radiotherapy system 100 and... Figures 1-6The computer-implemented process 800 is described, but those skilled in the art will understand that any properly configured X-ray imaging system is within the scope of this embodiment.

[0096] In step 801, the X-ray imaging system of the radiotherapy system 100 receives a set of initial 2D projections and a set of 2D projection metal masks for reconstructing a digital volume, such as the acquired 2D projection image 611 and 2D projection metal mask 641 for the initial reconstructed volume 621.

[0097] In step 802, the X-ray imaging system generates a set of combined 2D projections based on the acquired 2D projection image 611 and the 2D projection metal mask 641. In some embodiments, each combined 2D projection is generated based on the corresponding acquired 2D projection image 611 and the corresponding 2D projection metal mask 641. Generally, for a particular combination of 2D projections, the corresponding acquired 2D projection image 611 and the corresponding 2D projection metal mask 641 both represent projections from the same projection angle. One embodiment of the combined 2D projections generated from the acquired 2D projection image 611 and the 2D projection metal mask 641 is described in... Figure 9 As shown in the image.

[0098] Figure 9 A schematic illustration shows the combination of a captured 2D projection image 611 and a 2D projection metal mask 641, according to various embodiments, to form a combined 2D projection 901. As shown, the captured 2D projection image 611 includes a plurality of pixels 902, each pixel having an associated pixel value, and the 2D projection metal mask 641 includes position information 903 indicating pixels occluded by the metal object 501 during the forward projection process in step 640. The combined 2D projection 901 includes the pixels 902 of the captured 2D projection image 611, except for the pixel positions corresponding to the position information 903 of the 2D projection metal mask 641.

[0099] return Figure 8 In step 803, the X-ray imaging system arranges the combined 2D projections 901 into an array in the projection space. For example, in some embodiments, the array of combined 2D projections is arranged as a 3D matrix or stack. Furthermore, the combined 2D projections 901 are sequentially ordered in the 3D matrix. Thus, in such an embodiment, each combined 2D projection 901 associated with a first projection angle generated from the acquired 2D projection image 611 and the 2D projection metal mask 641 is ordered as a first 2D projection 901 in the 3D matrix, each combined 2D projection 901 associated with a second projection angle generated from the acquired 2D projection image 611 and the 2D projection metal mask 641 is ordered as a second 2D projection 901 in the 3D matrix, and so on. Figure 10 An example of a 3D matrix combining 2D projections is shown.

[0100] Figure 10 A 3D matrix 1000 of a combination of 2D projections 901 according to various embodiments is schematically shown. Figure 10 In the illustrated embodiment, array 1000 includes five combined 2D projections 901. In practice, 3D matrix 1000 may include more than five combined 2D projections 901. As a series of 2D images arranged in three dimensions, 3D matrix 1000 forms a 3D array of image data similar to a 2D CT sine wave. That is, when a CT sine wave expands multiple one-dimensional image datasets into two dimensions by arranging multiple one-dimensional datasets in a planar array, 3D matrix 1000 arranges multiple two-dimensional image datasets associated with each of the different 2D projections in a 3D matrix or "stack".

[0101] As shown in the figure, each combined 2D projection 901 includes mask pixels (black) 1010 and image pixels 1020. Each image pixel 1020 has its associated pixel value (not shown). In contrast, each mask pixel 1010 indicates a metallic object (such as...) Figure 5 The location of pixels blocked by a metallic object 501 is determined and thus corrected during the computer-implemented process 800. Image pixels 1020 include mask boundary pixels 1021 (cross-shading lines), which are pixels adjacent to one or more mask pixels 1010. Mask pixels 1010 include mask edge pixels 1011 adjacent to one or more image pixels 1020. Generally, each mask boundary pixel 1021 is adjacent to at least one mask edge pixel 1011 in the projection space. Figure 10 In the example shown, mask boundary pixel 1021 is shown to be adjacent to mask pixel 1010 within the same combination of 2D projections 901. In other examples, mask boundary pixel 1021 may be adjacent to mask pixel 1010 in different combinations of 2D projections 901 within the 3D matrix 1000. Figure 11 An example of such an embodiment is shown.

[0102] Figure 11 A portion 1100 of a 3D matrix 1000 according to various embodiments and adjacent inter-projection pixels are schematically shown. Figure 11In the illustrated embodiment, portion 1100 includes a first combination of 2D projections 1101, a second combination of 2D projections 1102, a third combination of 2D projections 1103, and a fourth combination of 2D projections 1104. As shown, the second combination of 2D projections 1102 includes mask boundary pixels 1021 (cross-shading lines) adjacent (in projection space) to the corresponding mask edge pixels 1011 included in the third combination of 2D projections 1103. Furthermore, the first combination of 2D projections 1101 includes image pixels 1020 adjacent in projection space to the corresponding mask boundary pixels 1021 in the second combination of 2D projections 1102, and the fourth combination of 2D projections 1104 includes mask pixels 1010 adjacent to the corresponding mask edge pixels 1011 included in the third combination of 2D projections 1103.

[0103] return Figure 8 In step 804, the X-ray imaging system generates a linear algebraic system based on array 1000. In some embodiments, the X-ray imaging system generates the linear algebraic system by applying a harmonic function to a pixel-represented domain included in array 1000, which is an array in the projection space of a set of combined 2D projections 901. Specifically, in such an embodiment, array 1000 is a 3D stack of combined 2D projections 901 representing image f. Therefore, in such an embodiment, Equation 2 is applied to the pixels included in array 1000:

[0104]

[0105] In some embodiments, the X-ray imaging system generates a linear algebraic system (e.g., via the finite difference method) by solving the discrete form of Equation 2 over a domain represented by the pixels of the image. In such embodiments, values ​​from a known region are used as boundary conditions (i.e., Figure 10 and 11 The mask boundary pixels (1021) in the image are used to represent the unknown parts of the image (i.e., Figure 10 and 11Equation 2 is solved for the mask pixels 1010 in the combined 2D projections. In such an embodiment, Equation 2 ensures the continuity and smoothness of pixel values ​​at the edges of the unrepaired portions of each combined 2D projection. Furthermore, since Equation 2 is solved for the domain simultaneously represented by the mask pixels 1010 of all combined 2D projections 901, the continuity and smoothness of pixel values ​​between combined 2D projections 901 are also ensured. Thus, intra-projection and inter-projection discontinuities in pixel values ​​are reduced. Specifically, in the harmonic repair process of the computer-implemented process 800, when the mask pixels of the combined 2D projections 901 are repaired with new pixel values, the change in the slope of the pixel values ​​associated with the mask edge pixels is constrained to be equal to the change in the slope of the pixel values ​​associated with the adjacent mask boundary pixels. Note that the mask edge pixels and mask boundary pixels are adjacent in the projection space and therefore can be included in the same combined 2D projections 901 (e.g., Figure 10 (as shown), or each can be included in different (but adjacent) combinations of 2D projections 901 (such as... Figure 11 (As shown).

[0106] In some embodiments, in step 804, a linear algebraic system (linear equation system) of the form [A][x] = [b] is employed, where [A] is the coefficient matrix of the linear algebraic system, [x] is the variable vector of the linear algebraic system, and [b] is the constant vector of the linear algebraic system. In such embodiments, the magnitudes of [a] and [b] are based on the number of mask pixels 1010 to be repaired in array 1000. In such embodiments, the values ​​of [a] and [b] can be combined using a 7-point finite difference method. For example, in one such embodiment, the values ​​of [a] and [b] are generated using a 3D Laplacian kernel of size 3×3×3 with 7 non-zero values. The method of generating a linear algebraic system by applying a harmonic function to the pixel domain is described in detail in: “On surface completion and image inpainting by biharmonic functions: Numerical aspects”, SBDamelin, NSHooang (2018), International Journal of Mathematics and Mathematical Sciences, 2018.

[0107] In step 805, the X-ray imaging system determines the value of the variable vector [x] of the linear algebraic system by solving the linear algebraic system. In some embodiments, the value of the variable vector [x] of the linear algebraic system is calculated using the conjugate gradient method.

[0108] In step 806, the X-ray imaging system generates a set of repaired 2D projections (e.g., 2D repaired projections 651) by modifying the acquired set of 2D projected images 611 using the values ​​of the variable vectors determined in step 805. In step 806, the pixels modified with the values ​​of the variable vectors are indicated by the position information of mask pixels 1010, which may be included in the 2D projection metal mask 641.

[0109] Example computing device

[0110] Figure 12 This is an illustration of a computing device 1200 configured to perform various embodiments of the present disclosure. For example, in some embodiments, the computing device 1200 may be implemented as Figure 1 The image acquisition and treatment control computer 106 and / or remote console 110 are included. The computing device 1200 may be a desktop computer, laptop computer, smartphone, or any other type of computing device suitable for practicing one or more embodiments of this disclosure. In operation, the computing device 1200 is configured to execute instructions associated with computer-implemented processes 600, 700, and / or 800 as described herein. Note that the computing device described herein is illustrative, and any other technically feasible configuration falls within the scope of this disclosure.

[0111] As shown in the figure, the computing device 1200 includes, but is not limited to, an interconnect (bus) 1240 connecting the processing unit 1250, an input / output (I / O) device interface 1260 coupled to the input / output (I / O) device 1280, a memory 1210, a memory 1230, and a network interface 1270. The processing unit 1250 can be any suitable processor implemented as a central processing unit (CPU), a graphics processing unit (GPU), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), any other type of processing unit, or a combination of different processing units, such as a CPU configured to operate in conjunction with a GPU or a digital signal processor (DSP). Generally, the processing unit 1250 can be any technically feasible hardware unit capable of processing data and / or executing software applications, including computer-implemented process 600, computer-implemented process 700, and / or computer-implemented process 800.

[0112] I / O device 1280 may include devices capable of providing input, such as a keyboard, mouse, touchscreen, etc., and devices capable of providing output (such as a display device, etc.). Additionally, I / O device 1280 may include devices capable of receiving input and providing output, such as a touchscreen, Universal Serial Bus (USB) port, etc. I / O device 1280 may be configured to receive various types of input from the end user of computing device 1200 and also to provide various types of output to the end user of computing device 1200, such as displayed digital images or digital video. In some embodiments, one or more of I / O devices 1280 are configured to couple computing device 1200 to a network.

[0113] Memory 1210 may include random access memory (RAM) modules, flash memory cells, or any other type of memory cell or a combination thereof. Processing unit 1250, I / O device interface 1260, and network interface 1270 are configured to read data from and write data to memory 1210. Memory 1210 includes various software programs executable by processor 1250 and application data associated with those software programs, including computer-implemented processes 600, 700, and / or 800.

[0114] Example computer program product

[0115] Figure 13 This is a block diagram of an illustrative embodiment of a computer program product 1300 for implementing a method for segmenting images according to one or more embodiments of the present disclosure. The computer program product 1300 may include a signal carrying medium 1304. The signal carrying medium 1304 may include one or more sets of executable instructions 1302, which, when executed by a processor of, for example, a computing device, can at least provide the above-mentioned... Figures 1-11 The described function.

[0116] In some implementations, signal-bearing medium 1304 may include non-transitory computer-readable medium 1308, such as, but not limited to, hard disk drives, optical discs (CDs), digital video discs (DVDs), digital magnetic tapes, memory, etc. In some implementations, signal-bearing medium 1304 may include recordable medium 1310, such as, but not limited to, memory, read / write (R / W) CDs, R / W DVDs, etc. In some implementations, signal-bearing medium 1304 may include communication medium 1306, such as, but not limited to, digital and / or analog communication media (e.g., fiber optic cables, waveguides, wired communication links, wireless communication links, etc.). Computer program product 1300 may be recorded on non-transitory computer-readable medium 1308 or another similar recordable medium 1310.

[0117] In summary, the embodiments described herein reduce and / or eliminate visual artifacts in the reconstruction of volumes including one or more significant metallic objects. Furthermore, in some cases, the embodiments reveal structures previously obscured by such visual artifacts. Therefore, the embodiments improve the perceived image quality of CBCT-based reconstructions and, in some cases, improve the accuracy of distinguishing tissue types in reconstructed CBCT images. This improvement over the prior art can be used during adaptive planning and / or radiotherapy.

[0118] Various embodiments have been described for illustrative purposes, but are not intended to be exhaustive or limited to the disclosed embodiments. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments.

[0119] Various aspects of this embodiment can be implemented as a system, method, or computer program product. Therefore, aspects of this disclosure can take the form of a completely hardware embodiment, a completely software embodiment (including firmware, resident software, microcode, etc.), or an embodiment combining software and hardware aspects, all of which are collectively referred to herein as a "circuit," "module," or "system." Furthermore, aspects of this disclosure can take the form of a computer program product implemented on one or more computer-readable media having computer-readable program code implemented thereon.

[0120] Any combination of one or more computer-readable media may be used. A computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. A computer-readable storage medium can be, for example, but not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatuses, or devices, or any suitable combination of the foregoing. More specific examples (not an exhaustive list) of computer-readable storage media will include the following: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable optical disc read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing. In the context of this document, a computer-readable storage medium can be any tangible medium capable of containing or storing a program used by or in conjunction with an instruction execution system, apparatus, or device.

[0121] While various aspects and embodiments have been disclosed herein, other aspects and embodiments will be apparent to those skilled in the art. The various aspects and embodiments disclosed herein are for illustrative purposes and are not intended to be limiting; their true scope and spirit are indicated by the following claims.

Claims

1. A computer-implemented method for reconstructing the volume of a segmented region of a patient's anatomical structure, the method comprising: Identify the anatomical regions associated with the reconstructed volume; Detect one or more metal objects associated with the reconstructed volume, arranged in an initial 3D metal object mask; For each of the one or more metal objects arranged in the initial 3D metal object mask, determine the volume associated with the metal object; Based on the anatomical region and the volume associated with the one or more metal objects, determine the value of at least one segmentation parameter; as well as The values ​​of the segmentation parameters are used to generate a final 3D metallic object mask associated with the reconstructed volume.

2. The computer-implemented method according to claim 1, wherein the at least one segmentation parameter includes a threshold parameter or an expansion radius.

3. The computer-implemented method according to claim 1, further comprising: For each of the one or more metal objects arranged in the initial 3D metal object mask, determine the radiographic density associated with each metal object; as well as The value of the at least one threshold parameter is determined based on the radiographic density associated with each metal object.

4. The computer-implemented method of claim 1, wherein determining the anatomical region includes generating an anatomical mask for the reconstructed volume.

5. The computer-implemented method of claim 4, wherein generating the anatomical mask for the reconstructed volume includes including a position in the anatomical mask based on an anatomical threshold.

6. The computer-implemented method of claim 5, wherein the dissection threshold is associated with a material having a radioactivity density greater than that of air.

7. The computer-implemented method of claim 5, wherein the dissection threshold is associated with a material having a radioactivity density less than that of water.

8. The computer-implemented method of claim 1, wherein determining the anatomical region associated with the reconstructed volume includes determining whether the anatomical region is a head region or a body region based on an anatomical mask of the reconstructed volume.

9. The computer-implemented method of claim 8, wherein determining whether the anatomical region is the head region or the body region based on the anatomical mask includes determining a minimum lateral width of the anatomical mask.

10. The computer-implemented method of claim 9, wherein determining the minimum lateral width of the anatomical mask is based on the maximum lateral width of each slice of the anatomical mask having a non-zero width.

11. The computer-implemented method of claim 1, wherein detecting the one or more metal objects arranged in the initial 3D metal object mask and associated with the reconstructed volume comprises: Generate the initial 3D metal object mask for the reconstructed volume; as well as Detect one or more connecting components contained within the initial 3D metallic object mask.

12. The computer-implemented method of claim 11, wherein detecting the one or more metal objects arranged in the initial 3D metal object mask and associated with the reconstructed volume further comprises determining a corresponding volume for each of the one or more connecting members contained within the initial 3D metal object mask.

13. The computer-implemented method of claim 11, wherein detecting the one or more metal objects arranged in the initial 3D metal object mask and associated with the reconstructed volume further comprises determining a metal classification of the reconstructed volume.

14. The computer-implemented method of claim 13, wherein the metal classification of the reconstructed volume is selected from the group consisting of a metal-free anatomical region, a dental region, a region containing a reference, and an orthopedic region.

15. The computer-implemented method of claim 13, wherein the metal classification for determining the reconstructed volume is based on at least one of the following: the volume of the largest component among the one or more connecting components, the cumulative volume of a group of the one or more connecting components that is greater than a predetermined volume, or the radiographic density of at least one connecting component among the one or more connecting components that is greater than a predetermined volume.

16. The computer-implemented method of claim 11, wherein generating the initial 3D metal object mask for the reconstructed volume includes determining a position in the initial 3D metal object mask based on an initial metal threshold.

17. The computer-implemented method according to claim 1, further comprising: A set of 2D projected metal masks is generated by performing a forward projection process on the final 3D metal object mask associated with the reconstructed volume.

18. The computer-implemented method of claim 17, wherein each of the set of 2D projected metal masks includes position information indicating pixels occluded by one or more connecting components contained within the initial 3D metal object mask during the forward projection process.

19. The computer-implemented method according to claim 1, further comprising: A non-binary mask is generated based on the final 3D metal object mask and the reconstructed volume; as well as The final reconstructed volume is generated based on the non-binary mask, the reconstructed volume, and the low-artifact reconstructed volume.

20. The computer-implemented method of claim 19, wherein generating the final reconstructed volume based on the non-binary mask comprises: For each edge voxel of the metallic object in the reconstructed volume, a value is provided based on the image value from the reconstructed volume, the image value from the low-artifact reconstructed volume, and the value from the non-binary mask.

Images