Medical scanner gantry alignment

Abstract

A framework for gantry alignment of a multimodality medical scanner is provided. First image data of a non-radioactive structure is acquired using intrinsic radiation emitted by scintillator crystals of a detector in a first gantry of the multimodality medical scanner. Second image data of the non-radioactive structure is acquired using a second gantry of the multimodality medical scanner, which is used for another modality. Image reconstruction can be performed based on the first and second image data of the non-radioactive structure to generate first and second reconstructed image volumes. A gantry alignment transformation can then be determined to align the first and second reconstructed image volumes.

Description

Technical Field

[0001] The present disclosure relates generally to image data processing and, more particularly, to a framework for gantry alignment of medical scanners. Background Art

[0002] Multimodal imaging plays an important role in accurately identifying diseased and normal tissue. Multimodal imaging provides combined benefits by fusing images acquired by different modalities. For example, the complementarity between anatomical (e.g., computed tomography (CT) and magnetic resonance (MR) imaging) and molecular (e.g., positron emission tomography (PET)) imaging modalities has led to the widespread use of PET / CT and PET / MR imaging.

[0003] Multimodality scanners require a process to measure the spatial displacement between images produced by the different modalities (e.g., PET to CT displacement, or PET to MR displacement). This process is often referred to as "gantry alignment." The standard gantry alignment process uses a radioactive source or a hot phantom that facilitates the acquisition of PET emission data to form PET images. The radioactive source (e.g., a point source, a line source) is typically positioned in a specific arrangement and imaged using, for example, both PET and CT in a PET / CT scanner, or both PET and MR in a PET / MR scanner.

[0004] For PET / CT scanners, the radioactive source material typically attenuates X-rays sufficiently to produce a CT image. For PET / MR scanners, the radioactive source material is typically invisible in MR imaging sequences, so traditionally, the source is surrounded by an MR-visible material (such as oil), which produces an MR image. The MR-invisible radioactive source material creates a void in the MR image, which is used to identify the source's location.

[0005] However, this gantry alignment process, which relies on an external radioactive positron source, is typically time-consuming because the source must be held and carefully placed in the center of the scanner's gantry. Additionally, humans must repeatedly handle the source, which poses health and safety risks due to its radioactivity. Summary of the Invention

[0006] Described herein is a framework for gantry alignment of a multimodality medical scanner. First image data of a non-radioactive structure is acquired using intrinsic radiation emitted by a scintillator crystal of a detector in a first gantry of the multimodality medical scanner. Second image data of the non-radioactive structure is acquired using a second gantry of the multimodality medical scanner, which is used for another modality. Image reconstruction can be performed based on the first and second image data of the non-radioactive structure to generate first and second reconstructed image volumes. A gantry alignment transformation can then be determined to align the first and second reconstructed image volumes. BRIEF DESCRIPTION OF THE DRAWINGS

[0007] A more complete appreciation of the present disclosure and its many attendant aspects will be readily obtained by reference to the following detailed description when considered in conjunction with the accompanying drawings as they become better understood.

[0008] Figure 1 A block diagram illustrating an exemplary system is shown;

[0009] Figure 2 An exemplary method of stage alignment is shown;

[0010] Figure 3 An exemplary non-radioactive structure is shown;

[0011] Figure 4 Another exemplary non-radioactive structure is shown;

[0012] Figure 5 An exemplary local projection image is shown;

[0013] Figure 6a An exemplary stage alignment transformation is shown; and

[0014] Figure 6b The residual error is shown. DETAILED DESCRIPTION

[0015] In the following description, many specific details are set forth, such as examples of specific components, devices, methods, etc., in order to provide a comprehensive understanding of the implementation of the present framework. However, it will be apparent to those skilled in the art that these specific details do not need to be adopted to practice the implementation of the present framework. In other instances, well-known materials or methods are not described in detail to avoid unnecessarily obscuring the implementation of the present framework. Although the present framework is susceptible to various modifications and alternative forms, specific embodiments thereof are shown by way of example in the accompanying drawings and will be described in detail herein. However, it should be understood that the present invention is not intended to be limited to the specific form disclosed, but on the contrary, it is intended to cover all modifications, equivalents, and alternatives that fall within the spirit and scope of the present invention. In addition, for ease of understanding, certain method steps are depicted as separate steps; however, these separately depicted steps should not be interpreted as necessarily dependent on order in their execution.

[0016] Unless otherwise stated, as will be apparent from the following discussion, it will be appreciated that terms such as "segmentation," "generation," "registration," "determination," "alignment," "positioning," "processing," "computing," "selection," "estimation," "detection," "tracking," and the like may refer to actions and processes of a computer system or similar electronic computing device that manipulates data represented as physical (e.g., electronic) quantities within computer system registers and memories and transforms that data into other data similarly represented as physical quantities within computer system memories or registers or other such information storage, transmission, or display devices. The embodiments of the methods described herein may be implemented using computer software. If written in a programming language conforming to a recognized standard, sequences of instructions designed to implement the methods may be compiled for execution on a variety of hardware platforms and for interfacing with a variety of operating systems. Furthermore, implementations of the present framework are not described with reference to any particular programming language. It will be appreciated that a variety of programming languages ​​may be used.

[0017] A framework for gantry alignment is presented herein. According to one aspect, a non-radioactive structure (or phantom) is used for gantry alignment in multimodality scanners (e.g., PET / CT and PET / MR). Intrinsic radiation from scintillator crystals in the multimodality scanner is used to create transmission images by measuring the transmission of photons through the non-radioactive structure. In some implementations, a neural network can be used to determine the gantry offset from the transmission image and / or to denoise short-duration scintillator crystal transmission images.

[0018] Advantageously, no radioactive source is required, thereby minimizing the customer's ongoing costs for purchasing and replacing sources. This framework is more efficient and safer because it eliminates the need to maintain radioactive sources and avoids the health and safety issues associated with radioactivity. It also advantageously minimizes the customer licensing overhead required to store radioactive sources. These and other exemplary advantages and features are described in greater detail in the following description.

[0019] Figure 1 1 is a block diagram illustrating an exemplary system 100. System 100 includes a computer system 101 for implementing the framework described herein. In some implementations, computer system 101 operates as a standalone device. In other implementations, computer system 101 can be connected (e.g., using a network) to other machines, such as a multi-modality medical scanner 130 and a workstation 134. In a networked deployment, computer system 101 can operate as a server (e.g., in a server-client user network environment), as a client user machine in a server-client user network environment, or as a peer machine in a peer-to-peer (or distributed) network environment.

[0020] In one implementation, computer system 101 includes a processor device or central processing unit (CPU) 104 coupled to one or more non-transitory computer-readable media 106 (e.g., computer storage or memory devices), a display device 108 (e.g., a monitor), and various input devices 109 (e.g., a mouse, touchpad, or keyboard) via an input-output interface 121. Computer system 101 may further include support circuits such as cache memory, a power supply, clock circuits, and a communication bus. Various other peripheral devices (such as additional data storage devices and printing devices) may also be connected to computer system 101.

[0021] The present technology can be implemented in various forms of hardware, software, firmware, a dedicated processor, or a combination thereof, or as part of microinstruction code, or as part of an application or software product, or a combination thereof, which is executed via an operating system. In some implementations, the technology described herein is implemented as computer-readable program code tangibly embodied in one or more non-transitory computer-readable media 106. In particular, the technology can be implemented by a processing module 117. Non-transitory computer-readable media 106 can include random access memory (RAM), read-only memory (ROM), magnetic floppy disks, flash memory, and other types of memory, or a combination thereof. The computer-readable program code is executed by CPU 104 to process data provided by, for example, database 119 and / or multi-modality medical scanner 130. Thus, computer system 101 is a general-purpose computer system that becomes a special-purpose computer system when executing the computer-readable program code. The computer-readable program code is not intended to be limited to any particular programming language or implementation thereof. It will be appreciated that various programming languages ​​and encodings thereof can be used to implement the teachings of the disclosure contained herein. The same or different computer-readable medium 106 may be used to store a database 119 including, but not limited to, image datasets, knowledge bases, individual subject data, medical records, a subject's diagnostic report (or file), or a combination thereof.

[0022] The multimodal medical scanner 130 acquires image data 132 associated with at least one subject. Such image data 132 may be processed and stored in the database 119. The multimodal medical scanner 130 may be a radiology scanner (e.g., a nuclear medicine scanner) and / or appropriate peripheral devices (e.g., a keyboard and display device) for acquiring, collecting, and / or storing such image data 132.

[0023] The multimodality medical scanner 130 can be a hybrid modality designed to acquire image data using at least one modality (e.g., PET) that uses scintillator crystals for detection. For example, the multimodality medical scanner 130 can include a first gantry for PET imaging and a second gantry for CT or MR imaging. The PET imaging gantry can include multiple detectors that include scintillator crystals. The scintillator crystals can be lutetium-based scintillator crystals (such as lutetium silicate (LSO) or lutetium yttrium silicate (LYSO)), containing the radioactive isotope Lu-176. The scintillator (or scintillator crystal) in a PET scanner typically detects one of the gamma photons originating from the annihilation event, while another scintillator crystal detects the other gamma photon. The scintillator crystals are typically part of a detector array that is arranged in a circular or cylindrical configuration around the patient's area. When struck by a gamma photon, each scintillator crystal emits a flash of visible light that is converted into electrons by the photomultiplier tube (PMT) or silicon photomultiplier (SiPM) of the PET scanner for subsequent electrical processing. The intrinsic radiation emitted from the scintillator crystal is used to obtain transmission images of non-radioactive structures. See, for example, Rothfuss H, Panin V, Moor A, Young J, Hong I, Michel C, Hamill J, Casey M, “ LSO background radiation as a transmission source using time of flight ”, Phys Med Biol. 2014 Sep 21;59(18):5483-500 and U.S. Patent Application 20140217294, both of which are incorporated herein by reference in their entirety.

[0024] The workstation 134 may include a computer and appropriate peripheral devices, such as a keyboard and a display device, and may be operated in conjunction with the entire system 100. For example, the workstation 134 may communicate with the multi-modality medical scanner 130 so that medical image data 132 from the multi-modality medical scanner 130 may be presented or displayed at the workstation 134. The workstation 134 may communicate directly with the computer system 101 to display processed data and / or output results 144. The workstation 134 may include a graphical user interface to receive user input via an input device (e.g., a keyboard, mouse, touch screen, voice or video recognition interface, etc.) to manipulate the visualization and / or processing of data.

[0025] It should be further understood that because some of the constituent system components and method steps depicted in the figures can be implemented in software, the actual connections between the system components (or process steps) may vary depending on how the present framework is programmed. Given the teachings provided herein, one of ordinary skill in the relevant art will be able to contemplate these and similar implementations or configurations of the present framework.

[0026] Figure 2 An exemplary method 200 for gantry alignment is shown. It should be understood that the steps of method 200 may be performed in the order shown or in a different order. Additional, different, or fewer steps may also be provided. Furthermore, method 200 may utilize Figure 1 The system 100 may be implemented as described above, a different system, or a combination thereof.

[0027] At 202 , a non-radioactive structure (or phantom) is positioned in the field of view of a multi-modality medical scanner 130 . Figure 3 An exemplary non-radioactive structure 302 is shown. The three-dimensional (3D) non-radioactive structure 302 can be positioned, for example, on a patient table 304 for image acquisition. Alternatively, the non-radioactive structure 302 can be held by a bracket or other support so that it extends into the imaging field of view without being positioned on or above the patient table. In some implementations, the non-radioactive structure 302 comprises an object having a material and a layout that intentionally minimizes the overall attenuation of photons emitted from the scintillator crystal. Highly attenuating objects are preferably placed only in locations that contribute constructively to the process of determining the spatial orientation of the structure 302 in the reconstructed transmission image.

[0028] For MR imaging, both the geometry and the composite material of the phantom structure are selected to account for the homogeneity of the MR magnetic field. Spatial distortion in MR images can be caused by materials with high magnetic susceptibility and by asymmetric geometries. The structure includes objects that substantially attenuate photons from the scintillator crystal and produce MR images (i.e., MR-visible). These objects are interspersed or supported by materials with low attenuation properties and are essentially MR-invisible.

[0029] The exemplary 3D structure 302 minimizes MR image artifacts, minimizes global photon attenuation, and produces sufficient photon attenuation in desired locations and MR signals to provide both transmission and MR images that can be used to determine spatial offsets between gantry positions. The 3D structure 302 includes 3D objects 308 constructed from or filled with a material visible under MR imaging, such as, but not limited to, liquid nickel sulfate (NiSO4). The 3D objects 308 can include, for example, spheres or other non-spherical (e.g., cubic) objects. The 3D objects 308 are placed within a substrate 310 of a low-density material (e.g., polymer foam). The low-density substrate 310 is essentially invisible during standard MR sequences and minimally attenuates photons from the scintillator crystals.

[0030] For CT imaging, structure 302 includes an object 308, which attenuates photons from scintillator crystals and X-rays and is visible under CT. Object 308 is distributed or supported in a material with low attenuation properties, resulting in relatively low Hounsfield Units (HU) in the reconstructed CT image. A non-limiting example of such a material is liquid nickel sulfate (NiSO4). Other materials may also be used. 3D object 308 is placed in a substrate 310 of a low-density material (e.g., polymer foam), which minimizes both the scintillator photons used to create the transmission image and the X-ray photons used to create the CT image.

[0031] Figure 4 Another exemplary non-radioactive structure 402 is shown. In this example, spheres 404 filled with NiSO₄ are connected to rods 406 to form a pyramidal structure. Other types of structures, such as tetrahedral or octahedral structures, are also possible. Rods 406 can be made of polytetrafluoroethylene (PTFE) or any other suitable material. Spheres 404 and rods 406 are visible in transmission and CT images. In MR images, only spheres 404 are visible, while PTFE rods 406 do not appear in the image or create any spatial distortion.

[0032] return Figure 2At 204, first image data of a non-radioactive structure is acquired using intrinsic radiation emitted by a scintillator crystal in the multimodality medical scanner 130. The scintillator crystal may be found in a PET detector in a first gantry of the multimodality medical scanner 130. In some implementations, the scintillator crystal is a lutetium-based scintillator crystal, which is known to have intrinsic radiation originating from the isotope Lu-176, which is present at a concentration of 2.6% in naturally occurring Lutetium. Lu-176 decays through beta decay with cascading gammas having energies of 307, 202, and 88 keV. Depending on the energy of the gamma photons, the first image data (or transmission data) acquired from the Lu-176 decay may be separated into sinograms. The sinograms may be reconstructed using, for example, a maximum likelihood transmission (ML-TR) algorithm, which is then used to reconstruct the transmission image.

[0033] At 206, second image data of the non-radioactive structure is acquired using a second gantry for another modality in the multi-modality medical scanner 130. The second modality may be, for example, CT or MR. The non-radioactive structure preferably does not move between the first and second image data acquisitions. In the case of PET / MR, the MR sequences are performed either simultaneously, partially simultaneously, or sequentially.

[0034] At 208, the processing module 117 performs reconstruction based on the first and second image data to generate first and second reconstructed image volumes, respectively. The first reconstructed image volume can be a volumetric transmission (Tx) image (e.g., an LSO-Tx image) reconstructed from the first image data using a reconstruction method such as a maximum likelihood transmission (ML-TR) image reconstruction technique. The ML-TR image reconstruction technique is an iterative algorithm with quadratic regularization that models the statistics of the transmission data. See, for example, Rothfuss H, Panin V, Moor A, Young J, Hong I, Michel C, Hamill J, Casey M, “ LSO background radiation as a transmission source using time of flight ”, Phys Med Biol. 2014 Sep 21;59(18):5483-500, which is incorporated herein by reference in its entirety. The first reconstructed image volume (such as the LSO-Tx image) replaces the acquired and reconstructed PET emission image of the radioactive source that is typically used in the standard gantry alignment procedure.

[0035] The second reconstructed image volume may be, for example, a volumetric CT or MR image reconstructed from the second image data using a reconstruction technique such as an iterative reconstruction algorithm, filtered backprojection, statistical modeling, or a combination thereof.Other techniques are also useful.

[0036] At 210, the processing module 117 determines a gantry alignment transformation to align the first and second reconstructed image volumes. The gantry alignment transformation describes the mechanical displacement (or spatial offset) between the first and second gantry in the multimodality medical scanner 130. The gantry alignment transformation identifies one or more rotations and / or translations required to align the first and second reconstructed image volumes. For example, the gantry alignment transformation includes three translation values ​​corresponding to three principal orthogonal axes in the image domain (e.g., anterior-posterior, inferior-superior, and left-right) and three rotation values ​​about these orthogonal axes. Applying the gantry alignment transformation to the first reconstructed image volume (e.g., a reconstructed PET image) produces an image that is spatially aligned with the second reconstructed image volume (e.g., a CT image in the case of a PET / CT system, or an MR image in the case of a PET / MR system).

[0037] In some implementations, an analytical method is used to determine the gantry alignment transformation. One example of an analytical method utilizes affine registration to align the first and second image volumes using a cost function based on mutual information. See, for example, P. Viola and W. M. Wells, “Alignment by maximization of mutual information,” Proceedings of IEEE International Conference on Computer Vision , Cambridge, MA, USA, 1995, pp. 16-23, which is incorporated herein by reference in its entirety. Other cost functions are also possible. Another exemplary analytical method explicitly identifies each sphere in the non-radioactive structure in the first and second reconstructed image volumes and determines a transformation that minimizes the distance between the sphere centers. The sphere centers can be identified by calculating the center of mass of each sphere, or by using a priori knowledge of the sphere geometry and positioning the spheres using the Hough transform.

[0038] Figure 5 An exemplary local projection image 502 derived from a reconstructed CT image volume (i.e., the second reconstructed image volume) and an exemplary local projection image 504 derived from a reconstructed LSO-TX image volume (i.e., the first reconstructed image volume) are shown. The local projection images may be xz projection images and xy projection images. A Hough transform may be used to locate spheres 506 and 508 in the respective projection images 502 and 504. A gantry alignment transform may then be determined that minimizes the distance between the center of sphere 506 and the center of sphere 508.

[0039] In other implementations, analytical methods are used in conjunction with deep learning techniques to perform denoising before determining the gantry alignment transformation. Denoising reduces the acquisition time of the first image data (i.e., the transmission image) for non-radioactive structures. To achieve this, a deep learning neural network can be trained to approximate full (or long) duration denoised transmission images (i.e., the first image data) from reduced duration images. The neural network is trained using multiple pairs of full-duration and reduced-duration transmission images, where each reduced-duration transmission image is a subset of the events recorded in the corresponding full-duration transmission image. Alternatively, scanner-acquired CT images aligned using existing gantry alignment methods can be utilized solely during training to associate the LSO TX images with the matching CT dataset and thereby serve as targets for the neural network.

[0040] In still other implementations, deep learning techniques are used to directly determine the gantry alignment transformation without analytical methods. In this method, a transmission image (i.e., first image data) and a CT or MR image (i.e., second image data) are acquired and reconstructed, and a deep learning convolutional neural network is used to directly identify the translation and / or rotation that describes the spatial offset between the two gantries of the multimodality medical scanner 130. The neural network can be trained using multiple examples of matching first and second reconstructed image volumes (e.g., transmission and CT or MR reconstructed image pairs) on a system with a known gantry offset. This known gantry offset can be obtained using radioactive markers (or thermal phantoms), as described in the standard National Electrical Manufacturers Association (NEMA) gantry offset procedure.

[0041] Figure 6a An exemplary gantry alignment transform 602 for a CT-PET medical scanner determined by the present framework is shown. As shown, the gantry alignment transform 602 specifies rotations (ie, roll, yaw, pitch) and translations about the X, Y, and Z axes. Figure 6b The resulting residual error 604 in the gantry alignment transformation is shown. This residual error is determined for each of the five spheres in the non-radioactive structure. It can be observed that the average residual error is 0.00 for all three axes. The mean absolute (abs) errors are 0.95 mm, 0.37 mm, and 0.30 mm for the X, Y, and Z axes, respectively.

[0042] Although the present framework has been described in detail with reference to exemplary embodiments, those skilled in the art will appreciate that various modifications and substitutions may be made to the present framework without departing from the spirit and scope of the present invention as set forth in the appended claims. For example, elements and / or features of different exemplary embodiments may be combined with each other and / or substituted for each other within the scope of the present disclosure and the appended claims.

Claims

1. A stage alignment system comprising: a non-radioactive structure positioned in a field of view of a multimodal medical scanner; a non-transitory memory device for storing computer-readable program code; as well as a processor device in communication with the memory device, the processor device being operable with the computer readable program code to perform steps comprising: performing image reconstruction based on the first and second image data of the non-radioactive structure to generate first and second reconstructed image volumes, wherein the first image data is acquired by using intrinsic radiation emitted by a scintillator crystal of a positron emission tomography (PET) detector in the multimodality medical scanner, wherein the second image data is acquired by using magnetic resonance (MR) or computed tomography (CT) of the multimodality medical scanner, and A gantry alignment transformation is determined that aligns the first and second reconstructed image volumes. 2 . The stage alignment system of claim 1 , wherein the scintillator crystal comprises a lutetium-based scintillator crystal. 3 . The stage alignment system of claim 2 , wherein the lutetium-based scintillator crystal comprises lutetium silicate (LSO) or lutetium yttrium silicate (LYSO).

4. The gantry alignment system of claim 1 , wherein the non-radioactive structure comprises an object constructed of or filled with a material that substantially attenuates photons from the scintillator crystal and generates an MR or CT signal. The stage alignment system of claim 4 , wherein the material comprises nickel sulfate.

6. The gantry alignment system of claim 4, wherein the object is placed in a substrate that is substantially invisible in MR or CT and that minimally attenuates photons from the scintillator crystal. The stage alignment system of claim 6 , wherein the substrate comprises a polymer substrate.

8. The gantry alignment system of claim 4, wherein the object comprises a sphere connected to at least one rod.

9. The gantry alignment system of claim 8, wherein the at least one rod comprises polytetrafluoroethylene.

10. A stage alignment method, comprising: acquiring first image data of a non-radioactive structure by using intrinsic radiation emitted by a scintillator crystal of a detector in a first gantry of a multimodality medical scanner; acquiring second image data of the non-radioactive structure using a second gantry for another modality of the multimodality medical scanner; performing image reconstruction based on the first and second image data of the non-radioactive structure to generate first and second reconstructed image volumes; as well as A gantry alignment transformation is determined that aligns the first and second reconstructed image volumes.

11. The gantry alignment method according to claim 10, wherein acquiring first image data of the non-radioactive structure comprises: First image data is generated using intrinsic radiation emitted by a lutetium-based scintillator crystal in a positron emission tomography (PET) detector.

12. The stage alignment method according to claim 10, wherein acquiring the second image data comprises: Obtain computed tomography (CT) or magnetic resonance (MR) images.

13. The gantry alignment method of claim 10 , wherein performing image reconstruction based on the first and second image data of the non-radioactive structure comprises: A maximum likelihood transmission image reconstruction technique is performed to generate a first reconstructed image volume.

14. The stage alignment method of claim 10, wherein determining the stage alignment transformation comprises: An affine registration is determined to align the first and second reconstructed image volumes.

15. The stage alignment method of claim 10, wherein determining the stage alignment transformation comprises: Spheres are identified in the non-radioactive structure in first and second reconstructed image volumes, and the gantry alignment transformation is determined that minimizes a distance between centers of the spheres.

16. The stage alignment method of claim 15, wherein identifying the sphere comprises: The sphere is positioned using a Hough transform.

17. The gantry alignment method according to claim 10, wherein acquiring first image data of the non-radioactive structure comprises: First image data of reduced duration is acquired, and first image data of full duration is approximated by applying a trained deep learning neural network to the first image data of reduced duration.

18. The stage alignment method of claim 10, wherein determining the stage alignment transformation comprises: A deep learning convolutional neural network is used to identify the stage alignment transformation.

19. The stage alignment method according to claim 18, further comprising: The deep learning convolutional neural network is trained using multiple instances of matching first and second reconstructed image volumes on a system with known gantry offsets.

20. One or more non-transitory computer-readable media embodying instructions executable by a machine to perform operations for stage alignment, the operations comprising: acquiring first image data of a non-radioactive structure by using intrinsic radiation emitted by a scintillator crystal of a detector in a first gantry of a multimodality medical scanner; acquiring second image data of the non-radioactive structure using a second gantry for another modality of the multimodality medical scanner; performing image reconstruction based on the first and second image data of the non-radioactive structure to generate first and second reconstructed image volumes; as well as A gantry alignment transformation is determined that aligns the first and second reconstructed image volumes.

Images