WO2016077865A1

WO2016077865A1 - Single/limited angle gamma/x-ray tomography

Info

Publication number: WO2016077865A1
Application number: PCT/AU2015/000695
Authority: WO
Inventors: Jeremy Michael Cooney BROWN; Marcus John Kitchen
Original assignee: Monash University
Priority date: 2014-11-21
Filing date: 2015-11-17
Publication date: 2016-05-26

Abstract

A tomography system (30) comprises one or more first radiation detectors (42) and one or more large area sources of gamma rays or x-rays (32). In use, the one or more first radiation detectors (42) detect photons transmitted through at least part of an object positioned in the path of the respective beam of radiation and output first data describing the positions on the detection face of the respective first radiation detector (42) at which transmitted photons are received. The system also comprise at least one scatter camera (38). Each scatter camera (38) is positioned laterally of each beam of radiation to receive photons scattered from the object.

Description

Title

SINGLE/LIMITED ANGLE GAMMA/X-RAY TOMOGRAPHY Field of Invention

The present invention relates to the field of computed gamma/x-ray tomography. Embodiments of the invention provide both a system and method for gamma/x-ray tomographic imaging and gamma/x-ray tomographic image reconstruction from either a single or limited number of angular gamma/x-ray projections.

Background

An X-ray Computed Tomography (X-ray CT) scanner acquires between a few hundred and a few thousand x-ray images of an object/patient at multiple angles to reconstruct its/their three- dimensional structure via digital image processing. The dose given to the object/patient during a clinical X-ray CT scan typically ranges between 1 mSv and 20 mSv. However, in the most extreme cases the level of dose can approach values of approximately 100 mSv. As such, X- ray CT spatial resolution is kept as low as possible to minimize the dose as it is inversely proportional to the fourth power of the spatial resolution.

Several epidemiological studies were recently completed that assessed the potential risk of developing cancer due to the exposure of ionizing radiation from X-ray CT scans. Whilst it is difficult to calculate the exact likelihood of developing cancer from a single source, due to genetic predisposition, environmental factors, etc., one such study has reported that the risk is roughly 1/2,400. Another has suggested that the risk could be as low as -1/1 1 ,000 for a head X-ray CT, and as high as -1/270 cancer cases for coronary angiography X-ray CT. Given the tens of millions of X-ray CT scans administered globally each year, this amounts to the potential for tens of thousands of new cancer victims each year; an estimated 29,000 cases per year in the US alone. These reports validate the current practice of limiting X-ray CT to only essential cases. The basic design of scientific, pre-clinical and clinical X-ray CT imaging systems have remained relatively unchanged since the release of the first system, the EMI-scanner, in the early 1970's. These systems are comprised of 3 basic elements: an x-ray source, object / patient stage, and a position resolving x-ray detector. The X-ray source is orientated such that the emitted field of x-rays will travel through the object located on the object / patient stage and then strike the front of the position resolving x-ray detector. Each exposure in this geometry will generate a

^"shadowgraph", an image created from the differences in x-ray absorption of the materials within the object. By taking a number of these shadowgraphs at different rotations around the centre of the object / patient stage, moving the x-ray source and position resolving in unison, sufficient data can be measured to apply digital imaging algorithms to reconstruct a three- dimensional estimate of the object. Over the last 40 years the majority of X-ray CT system developments have been in an attempt to increase the spatial resolution of the reconstructed images/volumes, remove blurring effects in reconstructed images/volumes due to object movement, minimise patient dose, and enable time resolved volumetric imaging. Whilst the development of ultra-bright x-ray sources, advanced detection technology, novel system rotation gantries and novel image reconstruction algorithms has made significant headway in meeting these aims, traditional X-ray CT imaging is starting to approach its physical limit due to the level of information that can be gained through transmission data alone. This limitation is due to the fact that, in the energy range of clinical X- ray CT (viz. 50 to 120 keV) the dominant physical interaction processes of x-rays with biological materials are Rayleigh and Compton scattering. In some cases almost 90 % of all x-rays will scatter in the object in all directions, with those that reach the position resolving x-ray detector degrading the quality of the collected shadowgraph. Attempts have been made to use this scattered information to reconstruct a three-dimensional electron density estimate of an object. In 1959, P. G. Lane employed a pencil beam of 1 MeV gamma-rays to raster scan a number of specimens of interest with a collimated Anger camera placed at 90° to the beam's direction and confirmed that it was possible to accurately measure the electron density of these samples. The basic geometry of this first Compton Scatter Tomography (CST) imaging system has served as the template for the majority of scatter based X-ray CT imaging systems that have been developed to date. Four primary CST modalities exist: point by point, line by line, plane by plane and energy dispersive. Point by point, line by line and plane by plane CST systems utilise a collimated γ- / x-ray source which is raster-scanned or swept over the object of interest. The scattered photons are detected by a collimated energy resolving radiation detector and, in some cases, the attenuated beam of photons is detected by a position resolving radiation detector forming a partial shadowgraph. In the cases of the line by line and plane by plane systems the collimated scatter detectors are both energy and position resolving. The energy dispersive CST modality utilises a wide area collimated point source and energy resolving radiation detector that are both directed at the object of interest. The angular separation of the source and energy resolving radiation detector is varied to measure the required data for imaging reconstruction. However, all four of these CST modalities are limited by the long scanning processed required to obtain sufficient data for image reconstruction, the need of energy resolving radiation detectors, and an ineffective approach in combining both scatter and transmission datasets. Summary of Invention

Embodiments of the present invention employ an imaging modality that is able to reconstruct three-dimensional electron density estimates, i.e. volumetric imaging, of an object through the use of a single or limited number of wide area γ- / x-ray field projections. This is achieved through the measurement and combination of both scattered and transmitted data types in the electron density reconstruction process. The present inventors have identified a system geometry for large area γ- / x-ray field imaging enabling single and limited angle X-ray computed tomography; a collimator design for scatter tomography with, in advantageous embodiments, position resolving radiation detectors; and a novel image reconstruction approach that enables the combination of both scatter and transmission datasets. In some embodiments, this novel imaging modality provides improved three-dimensional imaging and enables true four-dimensional imaging (time resolved volumetric imaging). In some

embodiments, this novel imaging modality, allows the use of radiation doses up to thousands of times lower than today's scientific, pre-clinical and clinical CT systems.

The first aspect of the present invention details a system geometry for large area γ- / x-ray field imaging enabling single and limited angle X-ray computed tomography. This system is composed of three basic elements: a large area γ- / x-ray field source, a position resolving radiation detector (or a position and energy resolving radiation detector if desired) positioned to detect photons transmitted through the object, and either a single or multiple collimated position, and if desired energy, resolving radiation detectors for detecting scattered photons. In some embodiments, such as in medical imaging applications, an object stage may be provided on which an object such as a patient can be positioned whilst exposed to radiation.

In an embodiment of the first aspect there is provided a tomography system comprising:

one or more first radiation detectors, each configured to resolve the positions of photons incident on a detection face of the first radiation detector;

one or more large area sources of gamma rays or x-rays, each configured to emit a beam of radiation towards at least one first radiation detector, whereby, in use, the one or more first radiation detectors detect photons transmitted through at least part of an object positioned in the path of the respective beam of radiation and output first data describing the positions on the detection face of the respective first radiation detector at which transmitted photons are received; and

at least one scatter camera, each scatter camera being positioned laterally of each beam of radiation to receive photons scattered from the object, each scatter camera comprising a collimator for filtering the photons scattered from the object and a second radiation detector configured to resolve the positions of photons incident on a detection face of the second radiation detector that pass through the collimator and output second data describing the positions on the detection face of the second detector at which scattered photons are received.

In an embodiment, at least one of the first radiation detector and the second radiation detector is further configured to detect energy of incident photons.

In an embodiment, the first and second radiation detectors are only configured to resolve the position of incident photons.

In an embodiment, there are a plurality of scatter cameras, each positioned laterally of the beam of radiation to receive photons scattered from the object. 7. A tomography system as claimed in claim 6, comprising at least one pair of scatter cameras positioned opposing one another with their respective detection faces parallel to one another.

The transmission and single or multiple collimated radiation detectors measure two different types of shadowgraphs, obtained from transmission and scattered photons, respectively. These two data types contain complementary information about the electron density within the object and, through the application of the novel image reconstruction approach outlined, can be combined to generate true four-dimensional datasets, multiple frames per second volumetric imaging, or a volumetric estimate from very low dose radiation exposures.

In some embodiments, and unlike previous CST systems, a collimator is used in front of each scatter detector with the resulting advantage that the radiation detector only needs to be configured to resolve the positions of incident photons. An advantage of this unique collimator design, in combination with the system geometry described above and image reconstruction approach outlined below, is that it removes the need for both energy and position resolving radiation detectors. At present, position only resolving detectors have a small fraction of detection dead time, have a higher spatial resolution and are able to record shadowgraphs at multiple frames per second. These factors increase the maximum frame rate of true four- dimensional volumetric imaging and further reduce the level of radiation exposure required to reconstruct a volumetric estimate of an object of interest. In an embodiment, the collimator filters the scattered photons to exclude photons incident on the scatter camera at greater than an acceptable scatter angle from reaching the detection surface. In an embodiment, the collimator comprises a grid of collimating cells.

In an embodiment, the number of cells of the collimator corresponds to a number of detection elements on the detection surface of the second detector. In an embodiment, a collimator depth (/) and septal thicknesses between cells in x and y (t(_X,_y)) are optimised as a function of maximum acceptable scatter angle variation (ΔΘ), minimum active detector pixel size {d_<Xiy)) and γ- / X-ray energy attenuation length (μ) in accordance with the following relationships: *.y ~ (d_x>y/tan (Δθ))-(3/_μ;) ; and > ^dx,y

~ tan(A0) ^'

The second aspect of the present invention details a novel image reconstruction approach which is able to combine both scattered and transmission data types. Through the use of an iterative reconstruction framework, the estimates of the electron density from transmission and scattered photons can be combined to calculate both true four-dimensional and three- dimensional volumetric estimates. The key to this approach is the use of detailed physics models in the back and forward projection processes, which enable the replication of self- attenuation effects of the scattered photons within an object of interest. Up until now the issue of self-attenuation of the scattered radiation was thought to be a limiting effect of CST; however, the present inventors have realised that these observed effects, seen in the scattered shadowgraphs, can be exploited to yield additional information about the electron density of an object of interest.

In an embodiment of the second aspect, there is provided an electronic method of generating a volumetric estimate of an object comprising:

(a) forming a first backprojection from first data indicative of the position of photons transmitted through the object and forming a second backprojection from second data indicative of the position of photons scattered from the object; (b) forming a volumetric estimate of the object by combining the first and second backprojections according to a ratio derived from first and second estimations of the electron density of the object obtained from the first and second data respectively;

(c) generating updated first data and updated second data by comparing the first data and second data with data derived by a forward projection of the volumetric estimate;

(d) repeating steps (a) to (c) with the updated first and second data until a current volumetric estimate formed at step (b) meets a convergence condition with respect to a previous volumetric estimate; and

(e) outputting the current volumetric estimate of the object upon the convergence condition being met.

In an embodiment, the method comprises determining the ratio of estimations of the electron density based on the relationships: p^Ne ⁼ % ^{; and}

°s,e

Another embodiment provides computer program code for implementing the above method.

Another embodiment provides a tomography system comprising:

one or more large area sources of γ rays or X-rays, each configured to emit a beam of radiation towards the at least one first radiation detector, whereby, in use, the one or more first radiation detectors detect photons transmitted through at least part of an object positioned in the path of the respective beam of radiation and outputs first data describing the positions on the detection face of the respective first radiation detector at which transmitted photons are received; and

at least one scatter camera, each scatter camera being positioned laterally of the beam of radiation to receive photons scattered from the object, each scatter camera comprising a second radiation detector configured to resolve at least the positions of photons incident on a detection face of the second radiation detector, and output second data describing the positions on the detection face of the second detector at which scattered photons are received; and an image reconstruction module configured to process the first data in conjunction with the second data to generate a volumetric estimate of the object by: (a) forming a first backprojection from the first data and forming a second backprojection from the second data;

(b) forming a volumetric estimate of the object by combining the first and second backprojections according to a ratio derived from first and second estimations of the electron density of the object obtained from the first and second data respectively;

(c) generating updated first data and updated second data by comparing the first data and second data with data derived from a forward projection of the volumetric estimate; and

(d) repeating steps (a) to (c) with the updated first and second data until a current volumetric estimate formed at step (b) meets a convergence condition with respect to a previous volumetric estimate.

Brief Description of Drawings In order that the invention may be more clearly ascertained, embodiments will now be described by way of example, with reference to the accompanying drawings in which:

Figure 1 is a schematic view of a tomography system according to an embodiment of the present invention;

Figures 2A and 2B are side- and front-views of a pre-clinical single / limited angle γ- / X-ray tomography system according to an embodiment of the present invention;

Figures 3A and 3B are side- and front-views of a clinical single / limited angle γ- / X-ray tomography system according to an embodiment of the present invention;

Figures 4A and 4B are cross section and front-views of a collimator according to an

embodiment of the present invention, for use in the collimated radiation detectors of the tomography systems of figures 1 , 2A, 2B, 3A and 3B;

Figure 5 is a flow chart of an embodiment;

Figure 6A is a schematic illustration of a sample used in a simulated example of the invention; Figure 6B is a standard x-ray of the sample of Figure 6A;

Figures 6C and 6D were formed by simulating radiation scattered by the sample; and Figures 7A to 7D show examples of slices of the 3D sample object volume reconstructed using filtered back projection, based solely on scattered photons.

Detailed Description

Single/limited angle γ- / X-ray tomography is an imaging modality applicable to the functional and structural volumetric imaging of any object. Embodiments of the invention combine data from transmitted and scattered photons to obtain at least one volumetric estimate of an object. The use of data from both transmitted and scattered photons has the advantage that a complete reconstruction of the object can be generated using much less radiation than conventional tomographic methods that use only the transmitted photons. In some

embodiments a plurality of images are obtained in order to advantageously allow the object to be observed as moving in three-dimensional space over time. When γ- / X-rays penetrate through matter, some of them are absorbed whilst others pass through the object without interaction. This selective absorption as measured by a transmission detector produces image contrast. However, up to 90% of the non-absorbed photons can be scattered in random directions via processes including Rayleigh (elastic) and Compton

(inelastic) scattering. Most of those scattered photons do not reach the transmission detector. In some embodiments of the present invention, the transmitted X-rays are captured by a detector to form a single two-dimensional (2D) projection image (shadowgraph) of the object. Located laterally of the object position, are one or more scatter cameras having detectors that capture these scattered X-rays. In some embodiments, a collimator is applied to the face of each detector to impose a spatial relationship between the object and pixels of the detector.

Figure 1 is a schematic view of a single/limited angle γ- / X-ray tomography system 10 according to an embodiment of the present invention. Tomography system 10 includes a large area γ- / X-ray field source 12, an object stage 14 (towards which γ- / X-ray field source 12 directs a γ- / X-ray field 15), a shielding ring (not shown) surrounding object stage 14, and a plurality of static radiation detectors (termed 'scatter cameras') 18 orientated with their active sensors directed towards object stage 14. Tomography system 10 also includes a γ- / X-ray field backstop (not shown) on the distal side of object stage 14 (relative to source 12) that includes a transmission detector 20 oriented towards γ- / X-ray field source 12. Transmission detector 20 is a position resolving radiation detector (or, optionally, a position and energy resolving radiation detector). Transmission detector 20 comprises, for example, of a 100 mm x 100 mm x 1 mm thick silicon photon counting hybrid pixel detector with a 50 μιη pixel size, and is arranged to detect γ-rays and X-rays transmitted through an object (such as a patient) on object stage 14, that is, radiation that has passed through that object, largely without interacting with the object. The output of transmission detector 20 is used by system 10 to form a transmission shadowgraph. Persons skilled in the art will appreciate that multiple transmission detectors could be used depending on the orientation of the γ- / X-ray source/sources. It will be appreciated that in some embodiments an object stage is not necessary, for example, in industrial applications, objects to be scanned could pass through the tomography system on a conveyor, or the object could be free standing and the tomography system could be moved into place around the object. However, in medical applications an object stage is advantageous as it makes it easier for the patient to stay still while the tomography system is operating.

Scatter cameras 18 comprise collimated position resolving radiation detectors (or optionally collimated position and energy resolving radiation detectors), and are provided to collect photons scattered from the object on object stage 14. Each scatter camera 18 comprises, for example, a 100 mm x 100 mm x 1 mm thick silicon photon counting hybrid pixel detector with a 50 μιη pixel size. Tomography system 10 is shown with two scatter cameras 18, but may include more or fewer scatter cameras. Desirably the scatter camera or cameras 18 subtend as large an angle as practical about object stage 14 in order to gather as much information about the scattered photons as practical, which will commonly be most convenient when plural, individual scatter cameras are employed. In particular, the use of multiple scatter cameras allows the detector bank to gain multiple views of an object. In some embodiments, this allows a complete three-dimensional tomographic reconstruction of the object to be created without moving the scatter cameras relative to the object. In other embodiments, the scatter cameras may be moved to improve the quality of the reconstruction of the image of the object. However, it will be appreciated that in many applications multiple scatter cameras will be used and accordingly movement of the scatter detectors will be through a smaller angle than in conventional systems because movement will only be necessary to capture information not captured by another scatter camera. For example, in an embodiment where four cameras are employed angular movement will typically be through less than 90°. Obtaining images either from one angle or a limited range of angles has the advantage of reducing the radiation dose imparted by the scanner.

In an embodiment, scatter cameras 18 are arranged around object stage 14 with their active sensors directed towards object stage 14, though without encroaching on the γ- / X-ray field from γ- / X-ray field source 12 or the field of view of transmission detector 20. Scatter cameras 18 are arranged in respective pairs orientated towards object stage 14 such that the scatter cameras of each pair face each other with detection surfaces 19 mutually parallel. This mutual pairing yields additional information, on a voxel by voxel basis, due to co-linear measurement of scattered radiation self-absorption within the object.

Tomography system 10 includes a control system 22 that controls the operation of system 10, including in particular source 12 and detectors 18 and 20, and which receives the output of detectors 18 and 20. Control and data connections are shown schematically at 24a-e.

Tomography system 10 includes an image reconstruction module in the form of image reconstruction server 26, which receives image data from control system 22, reconstructs images of the subject (as described below) from that data, and outputs a Digital Imaging and Communications in Medicine (DICOM) file (or other appropriate image format) comprising the images. This file can be viewed using any conventional computer provided with the appropriate viewing software.

Figures 2A and 2B are side- and front-views of a specific implementation of tomography system 10, comprising a pre-clinical single/limited angle γ- / X-ray tomography system 30 according to an embodiment of the present invention. Tomography system 30 includes a large area γ- / X- ray field source 32, an object stage 34 (towards which γ- / X-ray field source 32 directs a y- I X- ray field 35), and a shielding ring 36 surrounding object stage 34. The shielding ring 36 includes a plurality of scatter cameras 38 orientated with their active sensors directed towards object stage 34. Tomography system 30 also includes a γ- / X-ray field backstop 40 that includes a transmission detector 42 oriented towards γ- / X-ray field source 32.

The γ- / X-ray field source 32 of this embodiment is a microfocus liquid metal jet X-ray source, with a target of indium alloy and an approximate flux of 10¹¹ photons /(s mm²). Persons skilled in the art will appreciate that other X-ray and gamma ray sources could be used, for example, an Amercium 241 based γ-ray source. Further multiple sources could be used in a single system. Where multiple sources are used they could all be of the same type or composed of a mixture of X-ray and gamma ray sources. For example an array of Americium 241 based γ-ray sources, an array of microfocus liquid metal jet X-ray sources, or a combination of both, could be orientated around the patient stage. Other possible sources include fixed or rotating anode X-ray generators, or γ-ray sources Cobalt 57, Cobalt 60 and Caesium 137.

Transmission detector 42, like that of system 10, is a position resolving radiation detector (or, optionally, a position and energy resolving radiation detector). Transmission detector 42 comprises, for example, a 100 mm x 100 mm x 1 mm thick silicon photon counting hybrid pixel detector with a 50 μιη pixel size, and is arranged to detect γ-rays and X-rays transmitted through an object (such as a patient) on object stage 34, that is, radiation that has passed through that object, largely without interacting with the object. The output of transm^' detector 42 is used by the system 30 to form a transmission shadowgraph.

Scatter cameras 38, like those of system 10, comprise collimated position resolving radiation detectors (or optionally collimated position and energy resolving radiation detectors), and are provided to collect photons scattered from the object on object stage 34. In the illustrated embodiment, each scatter camera 38 comprises, for example, a 100 mm x 100 mm x 1 mm thick silicon photon counting hybrid pixel detector with a 50 μιη pixel size. Tomography system 30 includes eight such scatter cameras 38, but tomography system 30 may include more or fewer scatter cameras. Indeed, in a variant embodiment only one such scatter camera is used, though desirably the scatter camera or cameras subtend as large an angle as practical about object stage 34, which will commonly be most convenient when plural, individual scatter cameras are employed. Scatter cameras 38 are arranged around object stage 34 with their active sensors directed towards object stage 34, without encroaching on the γ- / X-ray field from γ- / X-ray field source 32 or the field of view of transmission detector 42. Scatter cameras 38 are arranged in respective pairs orientated towards object stage 34 such that the scatter cameras of each pair face each other with detection surfaces 39 mutually parallel. Alternatively, since each scatter camera comprises a plurality of detection elements or pixels (as described below), the outputs of respective pairs of opposed detection elements or pixels may be paired during analysis, irrespective of the actual number of scatter cameras 38 into which they are assembled.

The output of scatter cameras 38 is used by the system 30 to form one or more scatter shadowgraphs.

Figures 3A and 3B are side- and front-views of another specific implementation of tomography system 10, comprising a clinical single/limited angle γ- / X-ray tomography system 50.

Tomography system 50 includes an object stage 52 and a shielding ring 54 surrounding an object stage 52. Tomography system 50 further includes, mounted to a rotatable annular mount 56 within shielding ring 54, a large area γ- / X-ray field source 58 directed towards object stage 52 and six scatter cameras 60 with their active sensors also directed towards object stage 52 and with respective pairs of detection surfaces 61 facing each other. While tomography system 50 includes six scatter cameras 60, it will be appreciated that, in variants of this embodiment, as many scatter cameras as practical and desired may be employed. The γ- / X-ray field source 58 of this embodiment is a rotating anode X-ray source, with a target of tungsten and an approximate flux of 10⁹ photons /(s mm²).

Tomography system 50 includes a γ- / X-ray field backstop 62 that includes a transmission detector 64 (in the form of, for example, a 750 mm x 750 mm x 1 mm thick Csl(TI) flat panel detector with a 200 μιη effective pixel size), also located on annular mount 56, with

transmission detector 64 directed towards γ- / X-ray field source 58. Annular mount 56 hence rotates source 58, scatter cameras 60, backstop 62 and transmission detector 64 in unison. Scatter cameras 60 each comprise, for example, a 500 mm x 500 mm x 2 mm thick Cadmium Zinc Telluride photon counting hybrid pixel detector with 250 μιη pixel size.

Clinical tomography system 50 is able to operate in both a single shot and limited angle tomography modes.

Referring to figures 4A and 4B, each scatter camera 18, 38, 60 in anembodiment includes a grid pattern collimator 70, as shown schematically in cross section (figure 4A, shown with detection surface 72 comprising a plurality of detection elements 74) and from the front (figure 4B). Each cell 76 corresponds to a respective detection element or pixel 74. In Figure 4, the illustrated exemplary collimator 70 comprises a 10 x 10 grid of cells 76 but it will be appreciated that this is for illustrative purposes to show that the collimator has a distance between walls d(x,y) (here corresponding to active detector pixel size (d(x,y)), collimator depth (/), and a septal thicknesses in x and y {t(x,y)) leading to a scatter angle variation (ΔΘ). In the example described below in connection with Figures 6 and 7, collimator 70 comprises a 2000 x 2000 grid of square section collimating cells 76, each with d_{<Xi y)} = 40 x 40 μιη and a depth / of 5 mm. The wall or 'septal' thicknesses between cells 76 is _Xiy) = 10 μιη. These dimensions can be varied, especially according to the considerations discussed below. It will also be appreciated that while matching the number of cells of the collimator to the number of detection elements maximises the resolution of the detector, having fewer cells than the number of detection elements improves sensitivity due to the increased angle of acceptance through the collimator. Accordingly, in other embodiments the collimator has fewer cells than there are detection elements, for example, one cell for each 2 x 2 array of detection elements.

Collimator 70, in combination with the image reconstruction approach described below, removes the need for energy resolution in detection elements 74. If these detection elements 74 are in the form of position only resolving radiation detectors (i.e. not energy resolving) it enables the use of high flux, high spatial resolution radiation detectors, such as photon counting hybrid pixel detectors, which possess spatial resolutions of tens of microns and individual pixel maximum photon counting rates of 100,000 counts per second.

Removal of the dependence on detector energy resolution is achieved by having a collimator depth (/) and septal thicknesses in x and y ( _x,yj) optimised as a function of maximum acceptable scatter angle variation (ΔΘ), minimum active detector pixel size {d_<Xiy)) and γ- / -ray energy attenuation length (μ): χ·ν ~ (i tan (ΔΘ))-(3 ) Equation 1

Equation 2 tan(A0)

In Eqns. 1 and 3, the values of ΔΘ and _{(Χ )} are traded-off to find the optimal balance between spatial resolution and sensitivity for the single/limited angle X-ray tomographic imaging systems intended application.

In other embodiments, energy resolving radiation detectors can be employed with the advantage that additional information can be obtained about the materials that are present but typically with the disadvantage of a slower counting rate.

Imaging Reconstruction

In these embodiments, transmission and scattered datasets are combined through the use of an iterative reconstruction method. In these embodiments, this is in the form of Maximum

Likelihood Expectation Maximisation (ML-EM). In its simplest form, ML-EM can be expressed as:

Equation 3

This approach is implemented in the present embodiments with the use of a back-projection (BP) approach that has been developed to combine the two complementary transmission and scatter datasets. Conventional techniques are employed in the calculation of Systems Matrices and modelling of the forward projection (P) process on a specific system-by-system basis. One back-projection approach suitable for modification in accordance with the present approach is the Siddon's line backprojection method. Both transmission and scatter datasets are back-projected using Siddon's Line of Response (LoR) approach in which, for the present embodiments, the effective electron density within the region of the projected γ- / X-ray field is estimated. The following derivation takes the results of Jackson et al. (Physics Reports, Volume 70, Issue 3, 169-223, April 1981 ) and applies it to the situation of the electron density estimation from standard transmission and scattered datasets. For a single element material:

- = θχρ(-μί) Equation 4

'0 where I / 1₀ is the measured normalised beam intensity, μ is the energy dependent attenuation coefficient and f is the thickness of the material. Here: μ 1 Ν^σ

Equation 5 p P A where p is the material density, N_A is Avogadro's number, A is the relative atomic mass of the element and σ is the atomic cross-section. Since σ = Ζσ_β and μ = nZo_e Eqn. 5 becomes:

- tion 6 p = N ^A _A ua

where o_e is the cross-section of an electron, nZ is the number of electrons per unit volume and N_g is the number of electrons per unit mass.

In a multiple element amorphous material:

Equation 7

where:

Equation 8

∑f aj Aj and a, is the fractional composition of element /^' in the material. Then:

N_g =∑i N' = N_A Equation 9 Eqn. 6 becomes:

7 Equation 10

where:

Equation 11

Rearranging Eqn. 10:

V- ⁼ {p^Ng)∑i i ^a< Equation 12 where (pA/_g) is the electron density of the material.

Through the substitution of the cross-section a with the attenuation cross-section and scattering cross-sections, a_a = σ + σ + σ and σ₃ = σ + σ , respectively, and σ_θ = σ/Ζ, simplification yields the following two expressions that are able to convert attenuation coefficients to electron densities of a material: pN₍ Equation 13

Equation 14 s,e where a_{s e} is the combined coherent scattering (CS) and incoherent scattering (ICS) cross- section for a single electron and <f^E is the photoelectric absorption cross-section. This combined scattering cross-section is dependent on the orientation of each scatter camera 18, 40 with respect to the incident γ- / X-ray field and angular acceptance of the collimator opening ΔΘ.

A description of an exemplary iterative image reconstruction method 500 for the above electron density mapping approach follows. For a given x-ray/gamma-ray exposure of arbitrarily length, two datasets, scatter data (SD) 501 and transmission data (TD) 502, are recorded. These datasets are piped into the update scatter data estimate (USDE) function 51 1 and update transmission data estimate (UTDE) function 512 which compare 550 the forward projection 545 of the volumetric estimate of the systems active volume with the current value of the volumetric estimate. In the first iteration the current value of the volumetric estimate is a map of the system sensitivity, and the USDE function 51 1 and UTDE function 512 are divided by the current normalised estimate of the systems active volume. In higher order iterations the current value of the volumetric estimate consists of the current electron density estimate of the systems active volume. The outputs of USDE function 51 1 and UTDE function 512 are then backprojected 520 into individual volumetric estimates and combined into a single dataset through the use of a system specific weighting function 525. This weighting function combines the two temporary volumetric estimates through the fractional ratio of each dataset electron density maps corresponding cross-sectional estimate with respect to the sum of both cross-sectional estimates (derived from Equations 13 and 14). The result is a single electron density volumetric estimate that is then used to update the current estimate of the systems active volume 530.

At this point the volumetric estimate is assessed for convergence 535, i.e. has the current volumetric estimate after update changed less than a specific defined measure with respect to the previous estimate. If the volumetric estimate has converged, the iterative loop is broken and a final volumetric estimate of the systems active volume 570 is returned. In the case of where convergence has not been reached, the iterative process continues to the start of the loop with the current estimate forward projected 545 and compared 550 with the last iterations respective scatter and transmission datasets. This process will continue until either convergence or a user defined number of iterations is reached.

Further aspects of the method will be apparent from the above description of the system. It will be appreciated that at least part of the method will be implemented electronically, for example, digitally by a processor executing program code. In this respect, in the above description certain steps are described as being carried out by a processor, it will be appreciated that such steps will often require a number of sub-steps to be carried out for the steps to be implemented electronically, for example due to hardware or programming limitations. For example, to carry out a step such as comparing, a processor may need to compute several values and compare those values.

As indicated above, the method may be embodied in program code. The program code could be supplied in a number of ways, for example on a tangible computer readable storage medium, such as a disc or a memory device or as a data signal (for example, by transmitting it from a server). Further different parts of the program code can be executed by different devices, for example in a client server relationship. Persons skilled in the art, will appreciate that program code provides a series of instructions executable by the processor. Simulation

Single shot X-ray tomography according to an embodiment of the invention was simulated in order to show the quantity of X-rays that would be detected at 90° to the main X-ray beam upon scattering from an object. The Monte Carlo radiation transport toolkit Geant4 (trade mark) was employed. Geant4 enables accurate simulation of the transport of X-rays as they travel through a user defined volume. At each point along the X-ray path, the likelihood of the numerous possible physical interactions is calculated. This includes the probability of X-rays being photo- electrically absorbed, being redirected elastically via Rayleigh scattering, or redirected inelastically by Compton scattering. For photons within the diagnostic x-ray energy regime, Geant4 utilises the interaction cross-sections for the given materials from the Evaluated Nuclear Data Library of the OECD Nuclear Energy Agency. A simple phantom sample was simulated, the sample comprising a cylinder of lung tissue 10 mm in diameter and 10 mm in height. Inside this were two smaller cylinders running parallel to the long axis, the first comprising a bone equivalent plastic 3 mm in diameter and 8 mm in height, the second an air-filled cavity 1 .5 mm in diameter and 10 mm in height. Figure 6A is a schematic illustration of this sample 100, showing the lung tissue 102, air-filled cavity 104 and bone equivalent plastic 106. The simulation included a tomography system comparable to the embodiment shown in Figures 2A and 2B, although with eight scatter cameras, each comprising silicon based photon-counting detectors with a depth of 500 μιη and a pixel size of 250 μιη. These scatter cameras 100 were arranged in an equal angular spacing of 45° to obtain 8 different projections. The scatter cameras collected photons scattered by sample 100 at 90° to the main beam. A checkerboard tungsten collimator, with each checker face measuring 250 μιη x 250 μιη x 12.5 mm thick, was simulated as aligned flush to the surface of each scatter camera.

An incident beam of 30 keV X-rays was simulated, illuminating the entire sample. The beam intensity was set to match that produced in Hutch 3B at the Imaging and Medical Beamline

(IMBL) at the Australian Synchrotron. The exposure was set to 20 ms (assuming 10¹² photons / (s mm²)) to show the potential for performing dynamic imaging at this spatial resolution. This simulation took approximately 410 thousand CPU hours to run on the MASSIVE supercomputer cluster housed at the Australian Synchrotron and Monash University.

The results of this Monte Carlo based simulation are shown in Figures 6B to 6D. Figure 6B shows the standard X-ray image 1 10 that would be formed by the transmission detector, so may be regarded as a standard radiograph. Figures 6C and 6D were formed by using radiation scattered by the sample and subsequently incident on the collimated photon-counting detectors. The cross-hatching evident in figures 6C and 6D is caused by the cross-hatched collimator design used for this simulation. These results show that scattered X-rays can be used to form a clear image of a sample when using collimated, photon counting detectors.

As may be seen from these results, the direct (i.e. transmission) detector produces an attenuation map (shadowgraph) of the sample object in which all three cylinders 102, 104, 106 can be clearly seen. Most impressively, the scatter cameras (cf. Figures 6C and 6D) show different projections of the sample object, all of which clearly show the large cylinder 102 of lung tissue. Standard, non-collimated detectors would not yield any such contrast and would appear essentially featureless. The lateral scatter image (Figure 6D) shows the two internal cylinders overlapping in projection, making them difficult to distinguish. In the vertical scatter camera projections (Figure 6C), bone equivalent plastic cylinder 104 can be seen above the image noise. Interestingly, the air-filled cavity 104 is very easily seen in most projections owing to the larger difference in density between it and the lung tissue 102, than between the bone equivalent plastic 106 and lung tissue 102.

All detectors in the simulation operated as noise-free photon counting detectors. Such detectors have only recently become commercially available with sufficiently large areas and frame rates. Noise in the simulations results from limited quantum statistics, that is, the small number of scattered photons reaching the detectors. Conventional tomographic reconstruction (i.e. filtered backprojection) techniques are poorly suited to this type of imaging. Nevertheless, Figures 7A to 7D show examples of slices of the 3D sample object volume reconstructed using simple backprojection, based solely on scattered photons. Figures 7A and 7B show axial and coronal slices, respectively, reconstructed using four projections, while figures 7C and 7D show the same axial and coronal slices, respectively, reconstructed with eight projections.

Although only the scattering data was used for these figures, using data from only four and eight scatter cameras, respectively, Figures 7A to 7D nonetheless show the clear potential for the method, but also that a small number of scatter cameras using standard algorithms may produce images of less than diagnostic quality.

The apparent streaking artefacts in Figures 7A to 7D are expected for backprojection methods using limited numbers of projections. Conventional backprojection methods typically require a few hundred projections to perform a reconstruction of diagnostic quality. Iterative algorithms start with an assumed image, compute projections from this image, and then compare the original projection data and update the image based upon the difference between the calculated and the actual projection. This framework enables, with the appropriate forward projection algorithms, the ability to extract additional information out of the scatter shadowgraphs from the effects of the self-absorption of object scattered radiation.

The above embodiments are described in relation to arrangements suitable for medical imaging. Medical applications include diagnostic, prognostic or predictive applications such as scanning the patient to detect abnormalities such as tumours (benign or malignant), oedema, internal haemorrhaging, tissue necrosis, heart defects, cardiovascular disease (including

atherosclerosis and inflammatory conditions of blood vessels, ischemia, stroke, and

thrombosis), inflammation (e.g. inflammation caused by arthritis, such as rheumatoid arthritis) and bronchopulmonary dysplasia. Other diagnostic applications include obtaining images of complex fractures, blocked blood vessels and obstructions of the bowel. Such technology would also be useful for surgical procedures where implants need to be accurately positioned within the body (e.g. stents). This imaging technology may also be used for imaging patients during radiation therapy to aid in alignment of the x-ray beam with the position of the tumour. In such uses the radiation beam could be used as the radiation source for imaging the three- dimensional structure of the patient. Embodiments of the invention may also be used in fluoroscopy.

Embodiments of the present invention are particularly suited to diagnostics in respect of parts of the body that move during normal operation such as the heart or the lungs where it is desired to capture images of one or more operation cycles of the organ. Imaging such body parts with prior art techniques is difficult due to the need to rotate the source and detector around the object to obtain sufficient projections for volumetric reconstruction. This process typically results in object misalignment from projection to projection and/or motion blur from either the movement of the object or movement within the object and rapid movement of the gantry. To obtain time resolved volumetric estimates in this manner, a number of shadowgraphs are measure at each projection which encompasses the entire operation cycle leading to a significant dose of radiation. As embodiments of the invention require little or no movement to gather an image of the object, time resolved volumetric information can be gathered more quickly than in conventional techniques with a resultant reduction in radiation dosage. Further, as no breath or heart beat is exactly the same, gathering data more quickly reduces the complexity of integrating data from a number of different cycles.

As indicated above, embodiments of the invention may also be used in industrial applications such as flaw detection, failure analysis, material analysis, and security scanning. Again, an advantage of embodiments of the invention is that information can be gathered more quickly than in conventional techniques allowing for more efficient throughput. A further advantage in industrial applications is that the lower dose requirements of the technique can lead to power savings.

Modifications within the scope of the invention may be readily effected by those skilled in the art. It is to be understood, therefore, that this invention is not limited to the particular embodiments described by way of example hereinabove. In the claims that follow and in the preceding description of the invention, except where the context requires otherwise owing to express language or necessary implication, the word "comprise" or variations such as "comprises" or "comprising" is used in an inclusive sense, that is, to specify the presence of the stated features but not to preclude the presence or addition of further features in various embodiments of the invention.

Further, any reference herein to prior art is not intended to imply that such prior art forms or formed a part of the common general knowledge in any country.

Claims

1 . A tomography system comprising:

2. A tomography system as claimed in claim 1 , further comprising an image reconstruction module configured to process the first data in conjunction with the second data to generate an image of the object.

3. A tomography system as claimed in claim 2, wherein the image reconstruction module processes the first data and second data by:

(a) forming a first backprojection from the first data and forming a second

backprojection from the second data;

d) repeating steps (a) to (c) with the updated first and second data until a current volumetric estimate formed at step (b) meets a convergence condition with respect to a previous volumetric estimate.

4. A tomography system as claimed in any one of claims 1 to 3, wherein at least one of the first radiation detector and the second radiation detector is further configured to detect energy of incident photons.

5. A tomography system as claimed in any one of claims 1 to 3, wherein the first and second radiation detectors are only configured to resolve the position of incident photons.

6. A tomography system as claimed in any one of claims 1 to 5, comprising a plurality of scatter cameras, each positioned laterally of the beam of radiation to receive photons scattered from the object.

7. A tomography system as claimed in claim 6, comprising at least one pair of scatter cameras positioned opposing one another with their respective detection faces parallel to one another.

8. A tomography system as claimed in any one of claims 1 to 7, wherein the collimator filters the scattered photons to exclude photons incident on the scatter camera at greater than an acceptable scatter angle from reaching the detection surface.

9. A tomography system as claimed in claim 8, wherein the collimator comprises a grid of collimating cells.

10. A tomography system as claimed in claim 9, wherein the number of cells of the collimator corresponds to a number of detection elements on the detection surface of the second detector.

1 1 . A tomography system as claimed in claim 8 or claim 9, wherein a collimator depth (/) and septal thicknesses between cells in x and y ( _x,yj) are optimised as a function of maximum acceptable scatter angle variation (ΔΘ), minimum active detector pixel size {d_<Xiy)) and γ- / X- ray energy attenuation length (μ) in accordance with the following relationships: x'^y ~ (dx,_y/tan (Δθ))-(3/_μ;) I and > ^dx,y

~ tan(A0) ^'

12. An electronic method of generating a volumetric estimate of an object comprising: (a) forming a first backprojection from first data indicative of the position of photons transmitted through the object and forming a second backprojection from second data indicative of the position of photons scattered from the object;

13. A method as claimed in claim 12, comprising determining the ratio of estimations of the electron density based on the relationships:

°s,e

14. Computer program code which when executed by one or more processors implements the method of claim 12 or claim 13.

15. A tangible computer readable medium comprising the computer program code of claim 14.

16. A tomography system comprising:

one or more large area sources of γ rays or X-rays, each configured to emit a beam of radiation towards the at least one first radiation detector, whereby, in use, the one or more first radiation detectors detect photons transmitted through at least part of an object positioned in the path of the respective beam of radiation and output first data describing the positions on the detection face of the respective first radiation detector at which transmitted photons are received; and at least one scatter camera, each scatter camera being positioned laterally of the beam of radiation to receive photons scattered from the object, each scatter camera comprising a second radiation detector configured to resolve at least the positions of photons incident on a detection face of the second radiation detector, and output second data describing the positions on the detection face of the second detector at which scattered photons are received; and

an image reconstruction module configured to process the first data in conjunction with the second data to generate a volumetric estimate of the object by:

(a) forming a first backprojection from the first data and forming a second backprojection from the second data;

17. A tomography system as claimed in claim 16, wherein at least one of the first radiation detector and the second radiation detector is further configured to detect energy of incident photons.

18. A tomography system as claimed in claim 16 or claim 17, wherein the first and second radiation detectors are only configured to resolve the position of incident photons.

19. A tomography system as claimed in any one of claims 16 to 18, comprising a plurality of scatter cameras, each positioned laterally of the beam of radiation to receive photons scattered from the object.

20. A tomography system as claimed in claim 19, comprising at least one pair of scatter cameras positioned opposing one another with their respective detection faces parallel to one another.