Patent Application
20200345322
The following relates generally to the medical imaging arts, positron emission tomography (PET) imaging arts, and related arts.
Positron emission tomography (PET) scanners typically include a cylindrical bore-type housing supporting several PET detector rings for detecting 511 keV gamma rays. These PET scanners have a field of view (FOV) with fixed axial and radial dimensions. Commercial PET scanners have been developed with increasingly large bore diameters, so as to accommodate larger patients. However, such designs increase the cost as the number of detectors increases with bore diameter. In the axial direction, the usual solution is to employ multi-stage imaging, in which the patient is stepped through the bore and imaged at several positions that overlap in the axial direction. These individual PET images are then “stitched” together at the axial overlaps to form a whole-body image (or other image with extended axial extent). This solution has disadvantages including potential for error at the stitched overlap regions, and increased imaging session time required to acquire the multiple images along the axial direction. It is also not possible to perform certain continuous acquisition dynamic studies with a single bed position.
To increase the axial FOV (AFOV) without a concomitant increase in the number of PET detector modules (and hence increased cost), it is known to provide gaps between adjacent PET detector rings. The axial FOV of the system is approximately the sum of the axial dimension of the rings and the gaps between the rings. In another approach, the detectors can be sparsely populated around the circumference of each detector ring, so that each ring has a reduced number of PET detector modules so that more rings can be added to increase the axial FOV. Zhang et al., “PET System With Crystal or Detector Unit Spacing”, WO 2015/019312 A1 discloses embodiments of “sparse” designs, including embodiments in which the spacing(s) between adjacent detector rings can be adjusted for specific imaging tasks.
In a variant approach (see Gagnon et al., “Positron emission tomography system with hybrid detection geometries and sampling”, U.S. Pat. No. 8,558,181), an adjustable axial FOV is provided. Detector bars are arranged parallel to the axial axis of the bore of the PET scanner and populated along a circle surrounding the patient. The bars can shift in the axial direction relative to each other by the desired amount to achieve the desired axial FOV while keeping a central axial region into which all the bars extend, providing full detector coverage for this central axial region. The regions/organs of interest will be aligned with the central axial region to optimize the imaging for such regions/organs.
In further previous approaches (see Gagnon et al., “Modular multi-geometry PET system”, U.S. Pat. No. 8,378,305), a dual detector PET system includes two detector sets to image different portions of a patient and an adjustable detector ring having one set of detectors that can move in and out radially to form different size of transaxial rings to image patient of different sizes, while the other set of detectors can acquire data simultaneously if desired.
The following discloses new and improved systems and methods.
In one disclosed aspect, a PET imaging device includes a plurality of PET detector modules; and a robotic gantry operatively connected to the PET detector modules. The robotic gantry is configured to control a position of each PET detector module along at least two of an axial axis, a radial axis, and a tangential axis of the corresponding PET detector module.
In another disclosed aspect, a PET imaging device includes a plurality of PET detector modules; and a robotic gantry operatively connected to the PET detector modules. The robotic gantry is configured to control a position of each PET detector module along each of an axial axis, a radial axis, and a tangential axis of the corresponding radiation detector.
In another disclosed aspect, a PET imaging device includes a plurality of PET detector modules and a plurality of radiation shields disposed in gaps between neighboring PET detector modules. A robotic gantry is configured to control a position of each radiation detector along at least two of an axial axis, a radial axis, and a tangential axis of the corresponding radiation detector. The robotic gantry is operatively connected to the radiation shields to selectively extend or retract individual radiation shields. A plurality of racks is connected to the robotic gantry and upon which the PET detector modules are mounted, each rack being oriented parallel with the axial direction of the bore and each PET detector module robotically movable in the axial direction along the rack supporting the PET detector module.
One advantage resides in providing a positron emission tomography (PET) imaging device with radiation detectors or detector modules being individually controllable in multiple directions (e.g. axially and/or radially and/or tangentially) to configure the PET scanner for a particular patient and/or task.
Another advantage resides in providing an imaging device with movable radiation detectors to increase or decrease an axial field of view of the imaging device with reduced loss of data coverage in the increased FOV configuration by way of oscillating the detector modules axially and/or tangentially to provide oversampling.
Another advantage resides in providing an imaging device with an increased axial field of view and a reduced number of detectors.
Another advantage resides in providing an imaging device with movable detectors that conform to a subject geometry of a patient.
A given embodiment may provide none, one, two, more, or all of the foregoing advantages, and/or may provide other advantages as will become apparent to one of ordinary skill in the art upon reading and understanding the present disclosure.
The disclosure may take form in various components and arrangements of components, and in various steps and arrangements of steps. The drawings are only for purposes of illustrating the preferred embodiments and are not to be construed as limiting the disclosure.
The following proposes a configurable PET scanner in which the PET detector positions can be optimized for a particular imaging task. In one embodiment, the PET detector modules are mounted on racks, which are tracks that allow the PET detector modules to be moved longitudinally (i.e. axially, i.e. in the z-direction) along the length of the rack. Further, each PET module is mounted to the rack via a robotic telescoping arm to provide for movement in the radial direction. Still further, each rack may be moved along the tangential (i.e. angular) direction. With these three degrees of robotic freedom, a wide range of PET scanner configurations can be achieved. For example, a larger axial FOV can be achieved by allowing for larger gaps between rings of PET modules via movement of the PET modules in the axial direction. In another approach, sets of PET modules in different angular intervals can be relatively offset to provide increased axial FOV.
The flexibility also allows for non-uniform PET module positioning, for example in a cardiac scan the density of rings near to the heart location can be increased (up to having no gap between adjacent rings) relative to the peripheral rings. In this regard, it is contemplated to mount radiation shields between adjacent PET modules on independent robotic telescoping arms, thus allowing these shields to be withdrawn from between central rings to maximize ring density in the central region.
In some embodiments, the PET modules can be moved during a PET imaging data acquisition. For example, if a larger axial FOV is achieved by way of spacing apart the neighboring PET rings, those rings could be moved during the acquisition so that there are no axial gaps in the final collected data set. A similar concept is “oversampling”, in which detectors are moved back and forth during the acquisition to increase axial detector resolution; similar angular back and forth movement could be used to increase detector resolution in the tangential direction.
In some embodiments, different PET modules may comprise different detector types, e.g. a mix of TOF-PET modules and non-TOF-PET modules, and the configurability of the PET scanner leveraged to optimally position the mixture of PET module types. In a similar fashion, PET modules with larger density of dead pixels could be compensated by moving other PET modules with fewer dead pixels close by to compensate for the dead pixels.
In some embodiments, in any PET scanner configuration for which the acquisition setup is less than ideal, e.g. with gaps between rings or so forth, it may be possible to acquire phantom data for the PET scanner configured both with such non-ideality and in a more ideal configuration (e.g. without gaps), and deep learning used to train a transform to adjust an image acquired using the non-ideal configuration to more closely mimic the ideal configuration.
In other embodiments, the disclosed PET system may require different or additional configuration robotics as compared with the illustrative rack arrangement. For example, in the case of a breast examination, detectors could additionally be provided with a tilt robotic adjustment, and two PET modules located between the breasts can be tilted to face the opposing breasts, thereby providing PET counts in those directions. Advantageously, with such an arrangement it may be possible to image both breasts simultaneously, whereas current PET breast imagers use a single cup and image one breast at a time.
In addition to appropriate robotic manipulators, the robotic controller tracks the current location (and angulation, in the case of tilting PET modules) of each PET detector module in order to accurately record the line of response (LOR) spatial trajectories. In one approach, a detector is defined to have a nominal position (z, r, θ) where z is the default axial position, r is the default radial position, and θ is the default tangential (i.e. angular) position. This is updated in a particular PET module to a value (z+Δz, r+Δr, θ+Δθ) where Δz is the axial shift of the PET module along the rack, Δr is the radial shift of the PET module, and Δθ is the tangential (angular) shift of the supporting rack. The LOR is then conventionally defined given the positions of the two involved detectors in three-dimensional space. Additionally, the sensitivity matrix used in PET image reconstruction may need to be adjusted, especially when the PET detector modules are configured to a non-uniform arrangement which may, for example, increase sensitivity near the center of the scanner versus the axial periphery by having a higher density of detector rings positioned at/near scanner center.
The choice of PET scanner configuration for a particular imaging task can be variously chosen. In the simplest approach, the configuration is chosen manually, e.g. adding annular gaps between neighboring PET rings (or axial offsets between angularly neighboring racks) sized to achieve a desired axial FOV, setting radial positions of the PET modules to a minimum practical radial position for a patient of a certain girth, or so forth. In a more complex approach, it is contemplated to use an inverse optimization algorithm similar to inverse planning optimization employed in Intensity-modulated radiation therapy (IMRT) planning. In inverse IMRT planning, the radiation sources configuration is chosen, the resulting dose (or fluence) distribution in the target modeled taking into account radiation absorption using an attenuation map, and the sources configuration updated to improve matching between the modeled dose or fluence distribution and dose optimization goals. In similar fashion, the PET modules configuration can be chosen, the resulting counts distribution in the target modeled taking into account radiation absorption using an attenuation map, and the PET modules configuration updated to improve matching between the modeled counts distribution and dose optimization goals.
In imaging tasks employing multiple bed positions, the disclosed PET scanner configuration could in general be different for each bed position.
The disclosed system contains multiple racks distributed along the frame of the gantry that surrounds the patient (and bed or pallet on which the patient lies). For a simple implementation of this idea, the racks are parallel to each other and they are all parallel to the axial axis of the gantry. The racks are sufficiently long for the maximal AFOV to be achieved. In other examples, each rack can have multiple segmented pieces, and the segments can have offset in the plane perpendicular to the racks; racks or segments of racks can be in different orientations relative to each other and they don't have to be parallel to each other, etc.
The disclosed detectors can be designed as plug-in components with associated assembly peripherals. For example, the detectors can be plugged into the racks and their positions on the racks can be independently controlled and is programmable by the system. The system can include a mechanism to allow the detector to move closer to or retract from an object of interest either by moving the racks or rack segments or by moving the PET detector modules. The system can also allow the detector to reorient to point to an object of interest. The PET detector modules includes a mechanism to move and reorient the detector as controlled by the system. In some embodiments, the PET detector modules can include a collision detection and prevention mechanism. For example, one or more pressure sensors (not shown) can be disposed at corners of the PET detector modules. One or more electronic processors can analyze pressure signals obtained by the pressure sensors to determine if a collision between PET detector modules (or between one of the PET detectors and a patient to be imaged) is occurring. The processors can then control the PET detector modules to move away from each other avoid a collision.
For example, if each rack includes five detectors, and the detectors are pushed together to align all the detectors on all of the racks, the system configuration is the same as a conventional PET system with about 16 cm AFOV (assuming the detector assembly has dimension of 3.2 cm×3.2 cm).
In some examples, since the detectors can be independently controlled and their locations on the racks are known, the neighboring PET detector modules 18 can be positioned so that there is a pre-determined gap along the rack (along an axial direction).
In other examples, all the detectors on one rack can be moved together, but the detectors on different racks can be moved by different amounts.
The previous two examples can be implemented in the disclosed PET system to increase the AFOV of image acquisition. These two example embodiments can be implemented to further increase the AFOV of the disclosed PET system.
Since the pre-determined gaps between the detectors on the racks is programmable, the disclosed PET system has the flexibility of adjusting the gap as desired. For imaging of small organs/objects, the gap can be set to zero so that maximal sensitivity can be obtained using the same number of detector areas.
In some examples, the number of detectors on each rack can be different. For example, one third of the racks of the gantry can have, for example, seven detectors (˜22.4 cm), and the remaining portion of the racks can have, for example, four detectors each (˜12.8 cm). The racks with seven detectors can provide an effective AFOV of 22.4 cm. The detectors on the four-detector racks can be shifted according to the intended application for optimal performance. The total number of detector area is the same as a system with five detectors on each rack (or a system with five fully populated rings), the effective AFOV of which, however, is increased from 16.4 cm to 22.4 cm.
The configurable design of the disclosed PET system can allow the manufacturing of low-cost, high-performing, and scalable systems. For example, if each rack has three detectors, it is equivalent to a three-ring system with AFOV of 9.6 cm. Using the disclosed configurable design, the effective AFOV can be extended to that of five virtual rings, i.e., 16 cm.
The three-ring system described above can be easily upgraded to a four-ring, a five-ring system, etc. by adding one or two detectors to each rack. This flexibility will significantly help customers with different upgrading needs.
The flexibility of extending the AFOV allows for a lot dynamic studies to be performed. In contrast, such studies cannot be done on the conventional systems with the same detector area because the effective AFOV is too small.
The position of the detectors can be controlled individually and optimized collectively for the intended applications using an optimization program (similar to that in radiation therapy in which the multi-leaf collimators openings, delivery length at each angle, etc. are optimized).
Since the detectors are designed as swappable plug-in components, they can be shared among multiple systems as needed. For example, one or more detectors can (temporarily) be removed from an available three-ring system and use them on another three-ring system to achieve maximum effectiveness of a six-ring system. If the other system is a combined PET/CT system, removal of the PET detector modules 18 from the system does not impact the performance of the CT part of the system.
The detectors can be turned or rotated to point to organs of interest and or move to and from the organs of interest to improve the sensitivity, reduce background activity impact, and improve the quality of the acquired data.
When detectors on neighboring racks are shifted relatively in a predefined pattern, axial oversampling of the disclosed PET system can be achieved.
The disclosed detector configuration can change during the scan via an optimization program. For example, for a multi-frame whole-body scan using step-and-shoot bed motion, the detectors can be positioned differently at the head, head/neck frames, torso, and lower body acquisition frame. This potentially reduces the total acquisition time and improve the clinical workflow and patient throughput.
The disclosed racks are not necessarily implemented as racks in the system design if other alternatives are desired. For example, a two-dimensional (2D) surface surrounding the patient can be designed as the base for detectors to be mounted/plugged. The position of the detectors on a 2D surface can be programmable. The detectors can also include the mechanism to reorient or move toward or away from the patient.
In some examples, the disclosed PET system can be transformed into a mammography PET scanner, forming the rings of PET detector modules 18 around patients breasts, with few extra detectors positioned to the sides and to the back of the patient for extra projection views to allow for complete tomographic data.
In other examples, the disclosed PET system can be transformed into brain PET scanner by forming the rings around patient head.
In further examples, the disclosed PET system can also be reconfigured into pre-clinical small animal scanner.
In some example embodiments, an angle of a detector can be optimized by positioning the detectors closer to the patient's body. To do so, the thickness of the detector crystals can be reduced in order to minimize the effect of depth-of-interaction (DOI). This potentially reduces the cost of goods for the production while the effective sensitivity of the PET camera would still be large due to optimized solid angle.
In other example embodiments, the detectors can have different configurations, or can be positioned closer to the patient with a smaller crystal size, for optimization optimized for higher spatial resolution.
With reference to
A robotic gantry 20 is operatively connected to the plurality of PET detector modules 18. The robotic gantry 20 configured to control a position of each PET detector module 18 along an axial axis z and/or a radial axis r (
As shown in
With continuing reference to
As seen in
As labeled in
In some embodiments, as shown in
Referring back to
The at least one electronic processor 12, 14, 40 is operatively connected with one or more non-transitory storage media 46 (such as a magnetic disk, RAID, or other magnetic storage medium; a solid state drive, flash drive, electronically erasable read-only memory (EEROM) or other electronic memory; an optical disk or other optical storage; various combinations thereof; or so forth) which stores instructions which are readable and executable by the at least one electronic processor 12, 14, 40 to perform operations disclosed herein such as performing a detector configuration update method or process 100, 200 (see
With reference to
With reference to
At 204, the at least one electronic processor 40 is programmed to acquire imaging data with the configuration of the PET detector modules 18. To do so, the at least one electronic processor 40 is programmed to operate the robotic gantry 20 to arrange the plurality of PET detector modules 18 in the determined detector configuration. With the plurality of PET detector modules 18 arranged in the determined detector configuration, the at least one electronic processor 12, 14, 40 is programmed to control the robotic gantry 20 to acquire imaging data using the PET detector modules and reconstruct the imaging data to generate a reconstructed image. In some examples, the at least one electronic processor 12, 14, 40 is further programmed to, during the acquisition of imaging data, operate the robotic gantry 20 to oscillate the PET detector modules 18 in at least one of the axial direction and the tangential directions, so as to perform oversampling. This can be useful if the configuration spaces the detector modules apart with gaps between the detector modules in order to cover a larger FOV—the oversampling can reduce the impact of the gaps on the completeness of the acquired imaging data set. In other examples, the PET detector modules 18 are disposed in a predefined range along at least one of the two directions (z, r, and θ). For oversampling operations, the PET detector modules 18 can be controlled by the robotic gantry 20 to move the PET detector modules to a different location, either continuously or in multiple steps for data acquisition. When a scan is done, PET detector modules 18 can resume an original position.
The illustrative robotic gantry 20 of
As another example, a robotic gantry providing the axial (z) and tangential (θ) degrees of freedom, but not the radial (r) degree of freedom, can be useful in accommodating patients of different heights and also employing fewer PET detector modules by providing for gaps between adjacent detector modules along the circumferential direction.
As another example, to implement a breast examination with conformal placement of PET detector modules around both left and right breasts, the robotics for positioning the PET detector modules could optionally be provided with a tilt robotic adjustment (not shown). With this additional robotic degree of freedom, two PET modules can be placed between the breasts, with one tilted to face the left breast and the other tilted to face the right breast, thereby providing PET counts in those directions. Advantageously, with such an arrangement imaging data can be collected for both breasts simultaneously.
In addition to appropriate robotic manipulators such as those described with reference to
In some embodiments, the at least one electronic processor 40 is programmed to repeat the determination of the detector configuration, the operating of the robotic gantry 20 to arrange the plurality of PET detector modules 18 in the determined detector configuration, and the acquisition of imaging data for a plurality of bed positions to perform multi-station imaging.
At 206, the at least one electronic processor 40 is programmed to model a counts distribution of the acquired imaging data using an attenuation map and a dose distribution.
At 208, the at least one electronic processor 40 is programmed to update the configuration of the PET detector modules 18 with the counts distribution and the dose distribution.
The PET detector modules 18 are configurable in many suitable desired configurations. For example, the PET detector modules 18 can be configured as tiles, and designed as plug-in components. The PET detector modules 18 can be plugged into the racks 24 to face the patient, and can also be move to and from the patient and reorient to the regions of interest for optimized imaging.
The imaging system 10 can include optimization software to compute the optimal position/orientation of each PET detector module 18 according to the imaging task. For example, for a system with PET detector modules 18 equivalent to a conventional five-ring system with AFOV of 16.4 cm, if the imaging task needs an effective AFOV greater than 16.4 cm, the system can program AFOV extension and move the PET detector modules 18 accordingly to achieve the desired AFOV.
In another example,
The above-described examples can be optimized based on patient size, imaging protocol, CT information, etc. to optimize the position, orientation, etc. of each PET detector module 18. The PET detector modules 18 are positioned based on the programmed optimized position/orientation prior to or during the scans. In a first example, for a PET/CT system, a CT surview image is used to define the patient dimension and location in the imaging space. The optimization program can determine which PET detector modules 18 can be moved closer to the patient to form a ring or partial ring of smaller radius to surround the patient for optimal imaging. In a second example, a system is configured as a conventional PET/CT system with AFOV of 16.4 cm for cardiac scans. CT image (e.g., surview) shows that the patient heart has a dimension of 15 cm in the axial direction. Imaging using conventional configuration with AFOV of 16.4 will lead to significantly higher noise level near the ends of the AFOV, also correction of scatter is challenging. With the optimization program, the system can be configured to have an AFOV of 19.6 cm to allow high quality cardiac scan in one frame. Such optimization can be realized by moving the PET detector modules 18 to form a desired geometry, or by introducing gaps between PET detector modules 18 on the racks while the gap size and pattern are obtained from the optimization program. In a second example, for a conventional PET/CT system with an AFOV of 16.4 cm for cardiac scans, a CT image (e.g., surview image) shows that the patient heart has a dimension of 15 cm in the axial direction. Imaging using the conventional configuration with AFOV of 16.4 leads to significantly higher noise level near the ends of the AFOV. With the optimization program, the system can be configured to have an AFOV of 19.6 cm to allow high quality cardiac scan in one frame. Such optimization can be realized by moving the PET detector modules 18 to form a desired geometry, or by introducing gaps between radiation detectors on the racks while the gap size and pattern are obtained from the optimization program. In addition, the optimization program can include an optimization program described in co-pending Application Ser. No. 62/586,229, filed Nov. 15, 2017, which is incorporated herein by reference in its entirety.
The plug-and-play configuration of the PET detector modules 18 allows for easy upgrade and maintenance, e.g., from conventional three-ring system to a five-ring system by adding two detectors on each rack in a plug-and-play model. This allows for detector sharing between scanners, maximization performance/cost for sites with multiple systems, and minimizing costs for maintenance etc.
The dynamic configuration of the PET detector modules PET detector modules 18 allows for a change to compensate for the speed ramping-up and ramping-down sensitivity change during a continue-couch-motion scan. This configuration can change during a whole-body scan to allocate better sensitivity for regions of interest, e.g. tumor area. In addition, the dynamic configuration can allow for a change during a whole-body scan to allocate less sensitivity (e.g., enlarged crystal axial distance) to less important region in the image to allow fast scans, (e.g. a leg area without a tumor). This potentially reduces the total acquisition and improve the clinical workflow and patient throughput.
The disclosure has been described with reference to the preferred embodiments. Modifications and alterations may occur to others upon reading and understanding the preceding detailed description. It is intended that the invention be construed as including all such modifications and alterations insofar as they come within the scope of the appended claims or the equivalents thereof.
| Filing Document | Filing Date | Country | Kind |
|---|---|---|---|
| PCT/EP2018/083148 | 11/30/2018 | WO | 00 |
| Number | Date | Country | |
|---|---|---|---|
| 62593296 | Dec 2017 | US |
