Processing a pipeline for immediate particle image reconstruction
阅读说明:本技术 处理管道以用于立即进行粒子图像重建 (Processing a pipeline for immediate particle image reconstruction ) 是由 D·F·德容格 E·A·德容格 K·达芬 N·卡罗尼克斯 C·E·奥多涅斯 J·威南斯 于 2018-12-27 设计创作,主要内容包括:提供了用于生成医学图像的计算机实施的系统和方法。在一些实施例中,通过以下操作来生成医学图像:确定针对第一跟踪探测器相对于粒子射束系统的定位和对准;确定从粒子射束系统生成的射束的方向;还确定来自撞击在第一跟踪探测器上的探测到的粒子的第一粒子的第一位置;根据撞击在剩余射程探测器上的探测到的粒子来确定第一粒子的第一剩余射程;该系统基于第一粒子的位置、对准、第一位置以及第一剩余射程来重建针对第一粒子的路径;该系统基于针对第一粒子的重建路径来生成结果得到的医学图像。(A computer-implemented system and method for generating medical images is provided. In some embodiments, the medical image is generated by: determining a position and an alignment for the first tracking detector relative to the particle beam system; determining a direction of a beam generated from a particle beam system; also determining a first position of a first particle from the detected particles impinging on the first tracking detector; determining a first remaining range of the first particle based on the detected particles impinging on the remaining range detector; the system reconstructs a path for the first particle based on the position, alignment, first position, and first remaining range of the first particle; the system generates a resulting medical image based on the reconstructed path for the first particle.)
1. A computer-implemented method for generating a medical image, the method comprising:
determining a position and an alignment for the first tracking detector relative to the particle beam system;
determining a first position and a first initial kinetic energy of a first particle generated from the beam system from the detected particles impinging on the first tracking detector;
determining a first direction of the first particle at the first tracking detector;
determining a first remaining range of the first particle from the detected particles impinging on a remaining range detector;
reconstructing a path for the first particle based on the positioning, the alignment, the first position, the first direction, and the first initial kinetic energy; and is
Generating the medical image based on a reconstructed path for the first particle and the first remaining range and the first initial kinetic energy of the first particle.
2. The method of claim 1, wherein generating the medical image comprises: iteratively adjusting a first value for a first voxel associated with the medical image and a second value for a second voxel associated with the medical image, wherein the adjusting is based on the first remaining range.
3. The method of claim 2, wherein the first value and the second value are relative stop power values.
4. The method of claim 2, wherein the first voxel is projected along a direction of a particle beam.
5. The method of claim 1, further comprising: selecting an initial image approximation for the medical image, wherein generating the medical image is further based on the initial image approximation.
6. The method of claim 5, wherein the initial image approximation result is based on one or more of: (i) data based on at least one of the first tracking probe, the beam system, or remaining range finder; (ii) a CT image; or (iii) images from other modalities.
7. The method of claim 2, further comprising: a variable chord length for one or more touching voxels is determined based on linear interpolation between voxel boundaries using a step size larger than the voxels.
8. The method of claim 1, wherein determining the positioning and the alignment comprises: directing beams at different positions around an isocenter of the particle beam system.
9. The method of claim 2, further comprising: a variable chord length for one or more contacted voxels is determined based on a subdivision step of the voxel boundaries into segments.
10. The method of claim 1, further comprising: correlating a time at which the first particle impinges on the first tracking detector with an expected turn of the particle beam system at that time in an accelerator plan.
11. The method of claim 1, further comprising: the particles are iteratively eliminated based on Water Equivalent Path Length (WEPL) measurements and/or measurements of lateral deflection between tracking planes.
12. The method of claim 1, further comprising: particles with appropriate initial energies are selected for a given region of the image.
13. The method of claim 1, wherein the path comprises a straight line interpolation between a tracking detector strike and an estimated particle entry point and a particle exit point on a geometric housing of an object to be imaged.
14. The method of claim 13, wherein for the case where no exit direction is measured, the particle exit point from the geometric housing is estimated as the point at which the exit direction points to the location of the measurement on the tracking plane.
15. The method of claim 1, wherein the reconstructing may include: multiple data sets taken in different lateral positions of the patient are combined into a single voxel reference frame by applying an offset to the measurement positions in the detector reference frame.
16. One or more non-transitory computer-readable storage media comprising a plurality of instructions that, in response to being executed, cause a computing device to:
determining a position and an alignment for the first tracking detector relative to the particle beam system;
determining a first position and a first initial kinetic energy of a first particle generated from the beam system from the detected particles impinging on the first tracking detector;
determining a first direction of the first particle at the first tracking detector;
determining a first remaining range of the first particle from the detected particles impinging on a remaining range detector;
reconstructing a path for the first particle based on the positioning, the alignment, the first position, the first direction, and the first initial kinetic energy; and is
Generating the medical image based on a reconstructed path for the first particle and the first remaining range and the first initial kinetic energy of the first particle.
17. The one or more non-transitory computer-readable storage media of claim 16, wherein generating the medical image comprises: iteratively adjusting a first value for a first voxel associated with the medical image and a second value for a second voxel associated with the medical image, wherein adjusting is based on the first residual range.
18. The one or more non-transitory computer-readable storage media of claim 17, wherein the first value and the second value are relative stop power values.
19. The one or more non-transitory computer-readable storage media of claim 17, wherein the first voxel is projected along a direction of a particle beam.
20. The one or more non-transitory computer-readable storage media of claim 16, further comprising: instructions for selecting an initial image approximation for the medical image, wherein generating the medical image is further based on the initial image approximation.
21. The one or more non-transitory computer-readable storage media of claim 20, wherein the initial image approximation result is based on one or more of: (i) data based on at least one of the first tracking probe, the beam system, or remaining range finder; (ii) a CT image; or (iii) images from other modalities.
22. The one or more non-transitory computer-readable storage media of claim 17, further comprising: instructions for determining a variable chord length for one or more touching voxels based on linear interpolation between voxel boundaries using a step size larger than a voxel.
23. The one or more non-transitory computer-readable storage media of claim 16, wherein determining the positioning and the alignment comprises: directing beams at different positions around an isocenter of the particle beam system.
24. The one or more non-transitory computer-readable storage media of claim 17, further comprising: a variable chord length for one or more contacted voxels is determined based on a subdivision step of the voxel boundaries into segments.
25. The one or more non-transitory computer-readable storage media of claim 16, further comprising: correlating a time at which the first particle impinges on the first tracking detector with an expected turn of the particle beam system at that time in an accelerator plan.
26. The one or more non-transitory computer-readable storage media of claim 16, further comprising: the particles are iteratively eliminated based on Water Equivalent Path Length (WEPL) measurements and/or measurements of lateral deflection between tracking planes.
27. The one or more non-transitory computer-readable storage media of claim 16, further comprising: instructions for selecting particles having an appropriate initial energy for a given region of the image.
28. The one or more non-transitory computer-readable storage media of claim 16, wherein the path comprises a straight line interpolation between a tracking detector strike and an estimated particle entry point and a particle exit point on a geometric housing of an object to be imaged.
29. The one or more non-transitory computer-readable storage media of claim 28, wherein for the case that no exit direction is measured, the particle exit point from the geometric shell is estimated as a point at which the exit direction points to a location of the measurement on the tracking plane.
30. The one or more non-transitory computer-readable storage media of claim 16, wherein the reconstructing may include: multiple data sets taken in different lateral positions of the patient are combined into a single voxel reference frame by applying an offset to the measurement positions in the detector reference frame.
31. A computing system for generating a medical image, the computing system comprising:
one or more processors;
a memory having stored thereon a plurality of instructions that, when executed by the one or more processors, cause the computing system to:
determining a position and an alignment for the first tracking detector relative to the particle beam system;
determining a first position and a first initial kinetic energy of a first particle generated from the beam system from the detected particles impinging on the first tracking detector;
determining a first direction of the first particle at the first tracking detector;
determining a first remaining range of the first particle from the detected particles impinging on a remaining range detector;
reconstructing a path for the first particle based on the positioning, the alignment, the first position, the first direction, and the first initial kinetic energy; and is
Generating the medical image based on a reconstructed path for the first particle and the first remaining range and the first initial kinetic energy of the first particle.
32. The computing system of claim 31, wherein generating the medical image comprises: iteratively adjusting a first value for a first voxel associated with the medical image and a second value for a second voxel associated with the medical image, wherein adjusting is based on the first residual range.
33. The computing system of claim 32, wherein the first value and the second value are relative stop power values.
34. The computing system of claim 32, wherein the first voxel is projected along a direction of a particle beam.
35. The computing system of claim 31, further comprising: instructions for selecting an initial image approximation for the medical image, wherein generating the medical image is further based on the initial image approximation.
36. The computing system of claim 35, wherein the initial image approximation result is based on one or more of: (i) data based on at least one of the first tracking probe, the beam system, or remaining range finder; (ii) a CT image; or (iii) images from other modalities.
37. The computing system of claim 32, further comprising: instructions for determining a variable chord length for one or more touching voxels based on linear interpolation between voxel boundaries using a step size larger than a voxel.
38. The computing system of claim 31, wherein determining the positioning and the alignment comprises: directing beams at different positions around an isocenter of the particle beam system.
39. The computing system of claim 32, further comprising: instructions for determining a variable chord length for one or more contacted voxels based on a subdivision step of the voxel boundaries into segments.
40. The computing system of claim 31, further comprising: instructions for correlating a time at which the first particle impinges on the first tracking detector with an expected turn of the particle beam system at the time in an accelerator plan.
41. The computing system of claim 31, further comprising: instructions for iteratively eliminating particles based on Water Equivalent Path Length (WEPL) measurements and/or measurements of lateral deflection between tracking planes.
42. The computing system of claim 31, further comprising: instructions for selecting particles having an appropriate initial energy for a given region of the image.
43. The computing system of claim 31, wherein the path comprises a straight line interpolation between a tracking detector strike and an estimated particle entry point and a particle exit point on a geometric housing of an object to be imaged.
44. The computing system of claim 43, wherein, for a case where no exit direction is measured, the particle exit point from the geometric housing is estimated as a point at which the exit direction points to a location of measurement on a tracking plane.
45. The computing system of claim 31, wherein the reconstructing may include: multiple data sets taken in different lateral positions of the patient are combined into a single voxel reference frame by applying an offset to the measurement positions in the detector reference frame.
Technical Field
The present disclosure relates to medical imaging systems; in particular, the present disclosure relates to techniques for generating medical images from proton/particle radiography.
Background
Over 50% of the 160 million americans diagnosed with cancer annually require radiation therapy. There is a conservative estimate of 137000 new cancer patients in the united states per year that could benefit from proton therapy, far exceeding current therapeutic capabilities. Proton radiation therapy may protect large amounts of normal tissue from low to moderate radiation doses and avoid organs at risk. This can reduce late effects and improve quality of life, and is particularly important for patients with high cure rates and long survival times. Recent policy statements published by the american society for radiation therapy oncology (ASTRO) have cited scientific evidence to demonstrate that proton beam therapy is particularly useful in many pediatric patients, particularly those with brain tumors, and in certain adult cancer patients requiring high doses in close proximity to critical structures. Clinical trials supported by NCI are recruiting patients from proton therapy facilities in the united states of america for more common sites of cancer disease (e.g., breast, prostate, and lung), and additional studies are currently being conducted on all three sites of cancer disease. At present, there are 73 ion treatment facilities (62 proton treatment facilities, 11 carbon ion treatment facilities) running worldwide (25 proton treatment facilities in the united states), of which there are 42 treatment facilities at the construction stage (40 proton treatment facilities, 1 carbon ion treatment facilities, 1 proton and carbon ion treatment facilities) (10 proton treatment facilities in the united states) and 22 treatment facilities at the planning stage (21 proton treatment facilities, 1 carbon ion treatment facilities) (4 proton treatment facilities in the united states).
The bragg peak phenomenon (a sharp dose peak occurring at the end of the particle range) enables particles such as protons and other ions (e.g. deuterium, helium or carbon) to be accurately radiotreated with a tumor target, while healthy tissue will receive a much smaller dose compared to X-ray treatment systems. However, proton radiation therapy requires precise patient alignment and also requires adjustment of the initial proton energy so that the maximum dose corresponding to the bragg peak is deposited in the intended tissue. To adjust the range of the proton beam so that the maximum dose corresponding to the bragg peak is deposited in the intended tissue, the treatment plan may require a three-dimensional map of the specific patient in terms of relative stopping power (energy loss of the proton beam in the material versus the proton beam in water).
Disclosure of Invention
According to one aspect, the present disclosure provides a computer-implemented method for generating a medical image. The method comprises the following steps: determining a position and an alignment for the first tracking detector relative to the particle beam system; determining a first location of a first particle generated from the beam system from the detected particles impinging on the first tracking detector; a first direction of the first particle at the first tracking detector is determined. The method further comprises the steps of: determining a first remaining range of the first particle from the detected particles impinging on a remaining range detector; reconstructing a path for the first particle based on the positioning, the alignment, the first position, and the first direction; generating the medical image based on a reconstructed path for the first particle and the first remaining range of the first particle.
According to another aspect, the present disclosure provides one or more non-transitory computer-readable storage media comprising a plurality of instructions that, in response to being executed, cause a computing device to: determining a position and an alignment for the first tracking detector relative to the particle beam system; determining a first location of a first particle generated from the beam system from the detected particles impinging on the first tracking detector; determining a first direction of the first particle at the first tracking detector; determining a first remaining range of the first particle from the detected particles impinging on a remaining range detector; reconstructing a path for the first particle based on the positioning, the alignment, the first position, and the first direction; and generating the medical image based on the reconstructed path for the first particle and the first remaining range of the first particle.
According to a further aspect, the present disclosure provides a computing system for generating a medical image. The computing system includes: one or more processors; and a memory having stored thereon a plurality of instructions that, when executed by the one or more processors, cause the computing system to: determining a position and an alignment for the first tracking detector relative to the particle beam system; determining a first location of a first particle generated from the beam system from the detected particles impinging on the first tracking detector; determining a first direction of the first particle at the first tracking detector; determining a first remaining range of the first particle from the detected particles impinging on a remaining range detector; reconstructing a path for the first particle based on the positioning, the alignment, the first position, and the first direction; and generating the medical image based on the reconstructed path for the first particle and the first remaining range of the first particle.
Drawings
The disclosure will be described hereinafter with reference to the accompanying drawings, given as non-limiting examples only, in which:
fig. 1 illustrates a proton radiography system according to the present disclosure.
Fig. 2 shows the system of fig. 1 in more detail in block diagram form.
Fig. 3 shows a plurality of views of a tracking detector of the system of fig. 1.
FIG. 4 is a flow chart of an exemplary process for generating a medical image.
Fig. 5 shows a beam focus and a plurality of voxels.
FIG. 6 illustrates an example medical image generated using an embodiment of the present technique.
Fig. 7 shows the use of two positions to scan the entire patient in the case of a large patient.
Corresponding reference characters indicate corresponding parts throughout the several views. The components in the figures are not necessarily to scale, emphasis instead being placed upon illustrating the principles of the invention. The exemplifications set out herein illustrate embodiments of the invention, and such exemplifications are not to be construed as limiting the scope of the invention in any manner.
Detailed Description
While the concepts of the present disclosure are susceptible to various modifications and alternative forms, specific exemplary embodiments thereof have been shown by way of example in the drawings and will herein be described in detail. It should be understood, however, that there is no intent to limit the concepts of the present disclosure to the particular forms disclosed, but on the contrary, the intention is to cover all modifications, equivalents, and alternatives falling within the spirit and scope of the present disclosure.
Proton (and other particle) radiotherapy requires a three-dimensional map of the patient in terms of Relative Stopping Power (RSP) (energy loss of the beam in the material versus energy loss of the beam in the water) for treatment planning. This information is used to adjust the range of the ion beam. In some cases, the maps are obtained from an X-ray CT scan. Such a map introduces uncertainty regarding the conversion of X-ray absorption (hounsfield) units to RSP, and heterogeneity within the patient adds additional uncertainty. In addition, metal implants or other high density materials can cause shadowing artifacts and streaking.
The exemplary techniques described below reduce alignment uncertainty, reduce range uncertainty, and enable more complex treatments with more ion directions, and deliver higher doses to the tumor per treatment. In addition to the benefits of directly measuring proton stopping power, proton imaging deposits much less dose to the patient than comparable resolution X-ray images. The use of proton beams in imaging and treatment can simplify patient setup and quality assurance procedures, reduce alignment uncertainty, and reduce range uncertainty.
The proton trajectory deviates from a straight line due to multiple coulomb scattering, and thus it is more complicated to form an image with optimal spatial resolution than to form an image with optimal spatial resolution using X-ray radiography.
Some embodiments of the techniques described below avoid the need for extensive analysis and do not require calibration of the raw data and conversion of the raw data to proton paths and residual range. Some embodiments allow for the generation of images for verification immediately prior to treatment immediately following data acquisition, which was not previously possible.
Fig. 1 depicts a
The
Note that
FIG. 2 depicts the
By using different initial energies at different lateral positions of the tracking detector and the object being imaged, the remaining range of the ions can be kept low and the depth of the remaining range detector can be kept small. For example, regardless of the thickness or density of the object along a particular path, the initial energy can be selected to maintain the remaining range across the field to be imaged between 0 and 10 cm. The larger the range of possible initial energy, the smaller the remaining range finder is likely to be.
In some examples for the architecture of tracking
In some examples, the fibers 208 are oriented perpendicular to the fibers 210. If the light detector 212 indicates that protons are passing through one of the fibers 208 and one of the fibers 210 and the computing system 214 knows the locations of the two fibers, the computing system 214 can determine the X-Y coordinates of the
As the protons of
In some examples, the fibers 218 are oriented perpendicular to the fibers 220. If light detector 222 indicates that a proton passed through one of fibers 218 and one of fibers 220 and computing system 214 knows the locations of the two fibers, computing system 214 can determine the X-Y coordinates of tracking
Although an exemplary architecture of a tracking probe has been described, other architectures are possible. For example, if the fibers are sufficiently stiff, the fibers may be bonded together to avoid the use of a substrate. As another example, with respect to tracking
As the protons of the
The protons enter the
The use of multiple photon detectors also provides the potential to obtain additional position data for the position of protons leaving the
Fig. 3 depicts two cross-sections of tracking detectors that may be used to implement tracking
As depicted in
The fibers of adjacent layers can be bundled together so they are connected to a single photodetector channel. For example, referring to fig. 3,
Multiple fibers or duplexes of fibers may be organized in logical strips. For example, if
To further reduce the number of channels required for the light detectors, similarly positioned fibers or duplexes of fibers in strips positioned on the sides of the
In different variations of this technique, the two types of bundling described above (i.e., bundling adjacent fibers of different layers and bundling fibers or fiber duplexes of different strips) can be used together or separately.
Although an exemplary architecture of a tracking detector has been described in fig. 3, the same bundle architecture may be used for other configurations of tracking detectors. For example, with respect to the tracking probe of FIG. 3, if the fibers are sufficiently stiff, the fiber layers 308, 310, 312, and 314 may be bonded together to avoid the use of a substrate. As another example, fiber layers 308 and 310 and
Computing system 214 (fig. 2) may implement embodiments of the techniques to generate images from data collected from a proton radiography system. In an example embodiment, computing system 214 implements pipeline processing techniques to quickly generate images from particle measurements.
Fig. 4 depicts a flow diagram of a
In
For example, a test pattern consisting of a collection of discrete proton spots may be used to align tracking
The upstream and downstream tracking planes (tracking
One alignment algorithm is the product of the projection geometry theorem in which any four points on the projection plane (any four points, including infinity points, none of which are collinear) can be transformed to any other four points using a projective transformation. The projective transformation preserves all straight lines and includes scaling, rotation, and translation. The 2D projective transformation includes projecting a 3D rotation onto 2D. In the 2D projection plane, the points are represented by a homogenous triplet x y w, which is unique to the scaling factor. To create a proxel from cartesian points, it is sufficient to set w to 1. The conversion from the proxels to the cartesian points is achieved by dividing by w (for non-infinity points).
With the positions of the four measurement points in the plane and the four desired positions, a 3 x 3 matrix representing the projective transformation between the two sets can be solved directly. In essence, the transformation will take all points in the measured coordinate system and transform them to the desired coordinate system in a manner consistent with a combination of Euclidean transformations. This approach can be sensitive to location noise. In fact, a collection of many corresponding point pairs is used to robustly derive a single projective transformation by non-linear optimization, a process known in the computer vision community as bundle adjustment.
In one embodiment, twenty-five well-separated spots are emitted for alignment purposes, each of which has thousands of protons. An average spot position is calculated for each spot on the upstream and downstream detector planes. These points are measurement points. By using an accelerator plan for alignment and knowledge of the detector and beam geometry, the position can be calculated for each theoretical spot center on the upstream and downstream detector planes. These positions are referred to as ideal point positions. For each of the detector planes, a projective transformation from the measured coordinate system to the ideal coordinate system is calculated.
A projective transformation is used in the image reconstruction step to obtain the intersection of each proton trace with the detector plane at an unknown position and orientation, and to convert that intersection into the corresponding coordinates of the detector plane at its ideal position. For each detector plane, each coordinate transformation involves fourteen floating point operations for each proton. The alignment correction step is one of the preprocessing steps.
With this procedure, an image can be directly reconstructed in the isocenter coordinate system of the beam scanning system without relying on external alignment and other quality assurance steps. This enables radiography for patient alignment examination as well as examination of the integrated proton range.
A variant of this procedure can be used to check the tracking detector coordinate system. For example, if the tracking detector channels (e.g., the channels connected to the photodetectors 212 and 222) are comprised of scintillating fibers or other types of strips, the tracking detector coordinates may be based on a survey of the strips laid down during the build process. Stepping the pencil beam in the x-direction and y-direction while examining the average change in the reconstructed detector coordinates can provide an accurate correction for the investigation.
In
In one example, the time of the event is related to the expected turning of the ion beam at that time in the accelerator plan. The occurrence of the first event can establish a start time for both systems. The relative clock rates of the two systems can be calibrated with special runs using bursts of events at regular time intervals. The pattern of the impact tracking channel with respect to time can be monitored to update the relative clock rates of the accelerator and detector systems. For example, the beam can be turned off for short time intervals to enable the relative timing to be checked.
In another example, different solutions will have different relative channel strikes in the upstream and downstream tracking detectors. Solutions can be established on an ion-by-ion basis, or can be found by averaging multiple ions nearby in time.
As another example, range detectors may have positional sensitivity and can be used ion by ion, or to find the correct solution by averaging multiple ions in the vicinity in time. The method may depend on the number and arrangement of photon detectors in the remaining range detectors.
In
In 408, the computing system 214 converts the impacts in the remaining
In
In
In some cases, the previous CT scan was acquired using the same patient table as the current ion tomography or radiography. If not, the couch may be removed from the previous image and replaced with the couch for the ion image. However, the couch is designed to be of low quality, and the iterative algorithm converges even if the initial approximation of the couch is not perfect.
The housing may be used for an initial approximation. If the hull is accurately known, the hull can also be used as a boundary to define which voxels participate in the iterative algorithm. Voxels outside the boundary can be defined as air or table material with fixed values, or the table can be included inside the boundary. This will reduce the computation time and improve convergence since a large volume is not sampled for density variations.
In the case of tomography, the preliminary image can be efficiently obtained from the filtered back projection obtained using the straight-line approximation result for the proton path. This would produce an image similar to the final image, but with poor spatial resolution and could be used as an initial approximation. If a previous CT scan is available, it can be compared to the image to verify patient placement.
In the case of radiography, a preliminary image may be formed by extrapolating ions from an upstream detector to the isocenter, binning the ions into pixels, and averaging the WEPL of the ions. If a previous CT scan is available, a prospective preliminary image can be derived and compared to the actual preliminary image. This can be used to verify patient placement. If a previous CT scan is not available, an approximate shell can be constructed from: the preliminary image, in conjunction with WEPL, assumes water density to obtain a length along the beam path, and isocenter, or some other assumption based on an examination of the patient's placement. Further refinement can be made based on the amount of multiple scattering between the upstream and downstream planes. In the case of radiography, for example, as shown in fig. 5, by using protons from a single direction, the final step for obtaining a 2D image requires summing the voxels along columns in the direction of the proton ray, thereby collapsing the intermediate 3D map into the final 2D image.
An optional alternative to an approximate housing is to use a separate housing for each proton, with the length along the beam path matching the WEPL of that proton. This length may assume the density of the water or, if the geometric length is known from previous images, may be fixed in length, but the assumed material density may be adjusted. The shape may be a box sufficient to accommodate any reasonable proton path.
The procedure typically involves careful placement and rotation of the patient in the beam system to correspond to a previous CT scan. Comparison of the preliminary proton image with a previous CT scan can verify patient placement and rotation. The lateral shift looks like a simple lateral offset. When the proton beam diverges, the shift along the beam direction appears to be a magnification difference. If the patient rotates, the images may not correspond. A desired image assuming many different rotations can be prepared to find the best match. At the initial approximation, any significant shift, offset or rotation should be taken into account (otherwise the patient should be repositioned).
In
In one example, one algorithm can avoid such additional image noise sources while still reducing computation time compared to other algorithms that use per voxel chord length. Given the magnitude of the lateral multiple scatter, a series of straight line segments, each segment being several voxels in length, can be used without significantly degrading accuracy. The length of the segments can be kept constant throughout the imaging volume or can be scaled for expected multiple scatter. This enables longer step sizes to be used for the MLP algorithm, saving computation time. The interpolation between steps can be linear in (x, y, z) coordinates, or can be linear in (theta, theta2, z) coordinates used in projection radiography. Using these straight line segments, the voxel boundaries can be directly detected and assigned per voxel chord length. In the case of radiography, further simplification can be obtained by using a step size corresponding to an integer number of voxels in z, so that each segment ends on the z-boundary of the voxel.
In another example of a method for calculating a variable chord length for each contacted voxel:
the method does not involve the calculation of the intersection of line segments with the voxel sides. The method relies on the following stepping strategy throughout the reconstructed volume: each step is exactly one pixel along the beam direction; each step along the beam direction starts and ends at the center of two adjacent voxels. These conditions ensure that the step size will always span at least two voxels. It is not necessary to check whether the step size ends in the same voxel or in a different voxel. If the step size happens to span two voxels, the chord degree assigned to each voxel will increase by half the distance between the two end points. If the line joining the two end points spans more than two voxels, a semi-exact calculation is performed here. The line segment is divided into a plurality of smaller sub-segments. In one embodiment, we use twenty segments for every one millimeter step. Then testing each sub-section to which the voxel belongs; the chord length in the voxel then increases with the cumulative length of the sub-segments in the voxel. If a sub-segment spans two voxels, the chord lengths of both voxels increase by half the length of the sub-segment. Although the computation time is increased, the chord length approximation results improve as the number of sub-segments increases.
In some embodiments, the MLP algorithm relies on selecting an end point for the particle path through the hull that is the most likely intersection of the trajectory with the hull. If the direction of the particles upstream and downstream of the housing is known, direct extrapolation of the path from the tracking detector to the housing may be involved. In the case of only one tracking detector downstream of the housing, providing a position measurement but not a direction measurement, some iterations may be required to find the most likely intersection of the particle path with the housing. MLP provides the expected direction for any assumed position of the particle. The downstream intersection of the path with the hull can be selected as the point on the hull where the MLP points to the downstream tracking measurement.
At
The range finder has an upper and lower limit of the remaining range that it can reliably measure, unlike its physical boundaries. Not all ions detected in the range detector fall within this range. As a result of the range straggling, the ions reach the range finder with a distribution of remaining ranges, the mean of which is used to determine the WEPL passing through the patient. In order not to deviate from the mean, it is important that the entire distribution should be within the measurable limits of the detector. A more restrictive set of limits can then be defined for the mean remaining range. These limits are used to establish a minimum measurable WEPL and a maximum measurable WEPL for a given initial energy. When multiple energies are used, it is necessary to determine the region of the patient where each energy is active, i.e. where the measured WEPL falls within limits.
X-ray CT data can be used to estimate the WEPL through the patient and determine the regions where each energy should be used. However, there will be significant overlap due to the spread of pencil beams and the margin added for uncertainty. In the case where only part of the remaining range distribution is measured, the overlap area may be a problem. In regions where WEPL is too high for energy, some events at the upper end of the range profile may still reach the range finder. In regions where the WEPL is too low for energy, some events at the lower end of the distribution may be within the upper limit of the remaining range measurements. Including these events creates a biased WEPL profile in the region where the energy overlaps, so the software must identify and remove these events.
In
To save time, one option is to first determine the position on the binning plane using only MLP in order to apply these cuts. The complete MLP process including voxels and chord lengths can then be applied only to the remaining events.
At
For radiography, it is not necessary to view the entire lateral size of the patient, but this may be desirable. For tomography, a complete 3D reconstruction requires ions that pass through all lateral positions and angles of the patient (a 180 degree range is sufficient). This can be achieved by making two patient crossings in the frame of the beam system and laterally shifting the patient between the two crossings, as shown in fig. 7. For fixed beam and rotating patients, this is a simple lateral displacement of the rotating chair. In the case of a rotating gantry, the patient moves in a small semicircle within the frame of the treatment room. Patient movement must be considered as a shift in ion coordinates before image reconstruction can take place.
Fig. 7 depicts the use of two positions to scan the entire patient in the case of a large patient (the same technique can be extended to cover more than two positions). Each curved arrow indicates the size of the patient rotating about a different axis. In a scanning beam frame in which the patient rotates, two rotations in which the axis of rotation is displaced laterally are simply run. In a frame with a scanning beam system mounted on the gantry and rotating with the gantry, the patient table must be programmed to move in a small half circle as the gantry moves around the patient, however, the picture appears the same as seen in a frame in which the scanning beam system is stationary.
The methods, systems, and devices discussed above are examples. Various configurations may be omitted, replaced, or various method steps or flow or system components may be added as appropriate. For example, in alternative configurations, the methods may be performed in an order different than that described, and/or stages may be added, omitted, and/or combined. Also, features described with respect to certain configurations may be combined in various other configurations. Different aspects and elements of the configurations may be combined in a similar manner. Moreover, technology is evolving, and thus, many elements are examples and do not limit the scope of the disclosure or claims.
This description provides example configurations only and does not limit the scope, applicability, or configuration of the claims. Rather, the previous description of the configurations will provide those skilled in the art with an enabling description for implementing the described techniques. Various changes may be made in the function and arrangement of elements without departing from the spirit or scope of the disclosure.
Also, a configuration may be described as a process, which is depicted as a flowchart or a block diagram. Although each process may describe the operations as a sequential process, many of the operations can be performed in parallel or concurrently. In addition, the order of the operations may be rearranged. The process may have additional steps not included in the figures. Furthermore, examples of the methods may be implemented by hardware, software, firmware, middleware, microcode, hardware description languages, or any combination thereof. When implemented in software, firmware, middleware or microcode, the program code or code segments to perform the tasks may be stored in a non-transitory computer-readable medium such as a non-transitory storage medium. In some examples, one or more processors perform the tasks described above.
Furthermore, the example embodiments described herein may be implemented as logical operations in a computing device in a networked computing system environment. The logical operations may be implemented as: (i) a sequence of computer implemented instructions, steps or program modules running on a computing device; (ii) interconnected logic units or hardware modules that operate within the computing device.
Examples of the invention
Illustrative examples of the methods and systems disclosed herein are provided below. Embodiments of the method and system may include any one or more of the examples described below, and any combination thereof.
Example 1 is a computer-implemented method for generating a medical image. The method comprises the following steps: determining a position and an alignment for the first tracking detector relative to the particle beam system; determining a first location of a first particle generated from the beam system from the detected particles impinging on the first tracking detector; a first direction of the first particle at the first tracking detector is determined. The method comprises the following steps: determining a first remaining range of the first particle from the detected particles impinging on a remaining range detector; reconstructing a path for the first particle based on the positioning, the alignment, the first position, and the first direction; generating the medical image based on a reconstructed path for the first particle and the first remaining range of the first particle.
In example 2, the subject matter of example 1 is further configured such that generating the medical image comprises: iteratively adjusting a first value for a first voxel associated with the medical image and a second value for a second voxel associated with the medical image, wherein the adjusting is based on the first remaining range.
In example 3, the subject matter of example 2 is further configured such that the first value and the second value are relative stop power values.
In example 4, the subject matter of example 2 is further configured such that the first voxel is projected along a direction of the particle beam.
In example 5, the subject matter of example 1 is further configured to include selecting an initial image approximation result for the medical image, wherein generating the medical image is further based on the initial image approximation result.
In example 6, the subject matter of example 5 is further configured such that the initial image approximation result is based on one or more of: (i) data based on at least one of the first tracking probe, the beam system, or remaining range finder; (ii) a CT image; or (iii) images from other modalities.
In example 7, the subject matter of example 2 is further configured to include: a variable chord length for one or more touching voxels is determined based on linear interpolation between voxel boundaries using a step size larger than the voxels.
In example 8, the subject matter of example 1 is further configured such that determining the positioning and the alignment comprises: directing beams at different positions around an isocenter of the particle beam system.
In example 9, the subject matter of example 2 is further configured to include: a variable chord length for one or more contacted voxels is determined based on a subdivision step of the voxel boundaries into segments.
In example 10, the subject matter of example 1 is further configured to include: correlating a time at which the first particle impinges on the first tracking detector with an expected turn of the particle beam system at that time in an accelerator plan.
In example 11, the subject matter of example 1 is further configured to include: the particles are iteratively eliminated based on Water Equivalent Path Length (WEPL) measurements.
In example 12, the subject matter of example 1 is further configured to include: particles with appropriate initial energies are selected for a given region of the image.
Example 13 is one or more non-transitory computer-readable storage media comprising a plurality of instructions that, in response to being executed, cause a computing device to: determining a position and an alignment for the first tracking detector relative to the particle beam system; determining a first location of a first particle generated from the beam system from the detected particles impinging on the first tracking detector; determining a first direction of the first particle at the first tracking detector; determining a first remaining range of the first particle from the detected particles impinging on a remaining range detector; reconstructing a path for the first particle based on the positioning, the alignment, the first position, and the first direction; and generating the medical image based on the reconstructed path for the first particle and the first remaining range of the first particle.
In example 14, the subject matter of example 13 is further configured such that generating the medical image comprises: iteratively adjusting a first value for a first voxel associated with the medical image and a second value for a second voxel associated with the medical image, wherein adjusting is based on the first residual range.
In example 15, the subject matter of example 14 is further configured such that the first value and the second value are relative stop power values.
In example 16, the subject matter of example 14 is further configured such that the first voxel is projected along a direction of the particle beam.
In example 17, the subject matter of example 13 is further configured to include: instructions for selecting an initial image approximation for the medical image, wherein generating the medical image is further based on the initial image approximation.
In example 18, the subject matter of example 17 is further configured such that the initial image approximation result is based on one or more of: (i) data based on at least one of the first tracking probe, the beam system, or remaining range finder; (ii) a CT image; or (iii) images from other modalities.
In example 19, the subject matter of example 14 is further configured to include: instructions for determining a variable chord length for one or more touching voxels based on linear interpolation between voxel boundaries using a step size larger than a voxel.
In example 20, the subject matter of example 13 is further configured to determine the positioning and the aligning comprises: directing beams at different positions around an isocenter of the particle beam system.
In example 21, the subject matter of example 14 is further configured to include: a variable chord length for one or more contacted voxels is determined based on a subdivision step of the voxel boundaries into segments.
In example 22, the subject matter of example 13 is further configured to include: correlating a time at which the first particle impinges on the first tracking detector with an expected turn of the particle beam system at that time in an accelerator plan.
In example 23, the subject matter of example 13 is further configured to include: the particles are iteratively eliminated based on Water Equivalent Path Length (WEPL) measurements.
In example 24, the subject matter of example 13 is further configured to include: instructions for selecting particles having an appropriate initial energy for a given region of the image.
Example 25 is a computing system for generating a medical image, the computing system comprising: one or more processors; and a memory having stored thereon a plurality of instructions that, when executed by the one or more processors, cause the computing system to: determining a position and an alignment for the first tracking detector relative to the particle beam system; determining a first location of a first particle generated from the beam system from the detected particles impinging on the first tracking detector; determining a first direction of the first particle at the first tracking detector; determining a first remaining range of the first particle from the detected particles impinging on a remaining range detector; reconstructing a path for the first particle based on the positioning, the alignment, the first position, and the first direction; and generating the medical image based on the reconstructed path for the first particle and the first remaining range of the first particle.
In example 26, the subject matter of example 25 is further configured such that generating the medical image comprises: iteratively adjusting a first value for a first voxel associated with the medical image and a second value for a second voxel associated with the medical image, wherein adjusting is based on the first residual range.
In example 27, the subject matter of example 26 is further configured such that the first value and the second value are relative stop power values.
In example 28, the subject matter of example 26 is further configured such that the first voxel is projected along a direction of the particle beam.
In example 29, the subject matter of example 25 is further configured to include: instructions for selecting an initial image approximation for the medical image, wherein generating the medical image is further based on the initial image approximation.
In example 30, the subject matter of example 29 is further configured such that the initial image approximation result is based on one or more of: (i) data based on at least one of the first tracking probe, the beam system, or remaining range finder; (ii) a CT image; or (iii) images from other modalities.
In example 31, the subject matter of example 26 is further configured to include: instructions for determining a variable chord length for one or more touching voxels based on linear interpolation between voxel boundaries using a step size larger than a voxel.
In example 32, the subject matter of example 25 is further configured such that determining the positioning and the alignment comprises: directing beams at different positions around an isocenter of the particle beam system.
In example 33, the subject matter of example 26 is further configured to include: instructions for determining a variable chord length for one or more contacted voxels based on a subdivision step of the voxel boundaries into segments.
In example 34, the subject matter of example 25 is further configured to include: instructions for correlating a time at which the first particle impinges on the first tracking detector with an expected turn of the particle beam system at the time in an accelerator plan.
In example 35, the subject matter of example 25 is further configured to include: instructions for iteratively eliminating particles based on a Water Equivalent Path Length (WEPL) measurement.
In example 36, the subject matter of example 25 is further configured to include: instructions for selecting particles having an appropriate initial energy for a given region of the image.
Although the subject matter has been described in language specific to structural features and/or methodological acts, it is to be understood that the subject matter defined in the claims is not necessarily limited to the specific features or acts described above. Rather, the specific features and acts described above are disclosed as example forms of implementing the claims.
- 上一篇:一种医用注射器针头装配设备
- 下一篇:超声造影成像方法、超声成像装置和存储介质