WO2003052397A1

WO2003052397A1 - A phantom

Info

Publication number: WO2003052397A1
Application number: PCT/NZ2002/000282
Authority: WO
Inventors: Geordie Robert Burling-Claridge; Philip Edward Petch; Serguei Timofeevich Zavtrak
Original assignee: AgResearch Ltd
Current assignee: New Zealand Institute for Bioeconomy Science Ltd
Priority date: 2001-12-19
Filing date: 2002-12-19
Publication date: 2003-06-26
Anticipated expiration: 2004-06-19
Also published as: AU2002366300A1

Abstract

A phantom, for use with a composition analysis device configured to estimate the composition of a target material by creating a profile corresponding to a measured characteristic of the target material, wherein the phantom substantially mimics the profile of the target material across a region of interest. The phantom is produced by assessing the profiles of two or more component materials, determining whether the combination of the materials in a predetermined quantity substantially mimics the profile of the target material in a region of interest, and forming a phantom from component materials satisfying these conditions.

Description

A PHANTOM

TECHNICAL FIELD

This invention relates to the selecting of materials used to produce a phantom that is used to calibrate a machine.

In particular, this present invention relates to the selecting of materials used to produce X-ray phantoms for use in conjunction with Dual energy X-ray Absorptiometry (DXA) machines.

BACKGROUND ART

Devices that estimate a target material's composition by measuring that material's characteristic profile require calibration and corrections for measurement drift over time.

These devices have an important role, and can include machines such as X-ray machines, Dual Analysis X-Ray machines (DXA), Nuclear Magnetic Resonance machines (NMR), Magnetic Resonance Imaging (MRI) and Ultrasound devices.

DXA is often used to determine the relative proportions of bone to soft tissue, or soft tissue to fat in meat, be it human or otherwise. Study of the human body using DXA can provide information, for example, about percentage body fat for study of wasting disease and drug administration thereof, or information relating to osteoporosis or Paget's disease, both bone density diseases.

DXA is also used in airports around the world to examine the contents of luggage for restricted substances such as certain foods, metals and explosive materials or the like.

It is essential that these types of devices are calibrated both to prevent drifting of the measurements of a single machine over time and, in the case where multiple machines are used in a study, to ensure consistency among results. The standard methods of calibrating those devices, as listed above, are known to those skilled in the art. As an example, the building of a calibration model for fat in meat is undertaken by correlating the attenuation of the low energy/high energy ratio to the fat reference results. This method is however sensitive to the thickness of the sample and is therefore not useful when studying samples with varying heights or thickness, which is often the case with portions of meat or the like.

PCT/DKOO/00588 describes a method and apparatus for determination of properties of food for feed and focuses on the use of DXA to undertake that analysis. The application discusses the use of a calibration model to define relations between the plurality of values and properties of a medium, being in this case, raw material of food or feed. A limitation of the method described in this application is that the samples that are used for calibration consist of known amounts of raw material. In particular, the calibration samples contain minced pork meat of varying fat content. It is a disadvantage of this application that the method of calibration involves raw material. While the content is known, the material is open to decay or inhomogeniety of the sample.

Phantoms and reference tiles are often used in place of raw material to provide a means for calibration and drift compensation of such devices.

US Patent No. 6,315,447 describes the use of a variable composition phantom for simulating varying degrees of human body fat for dual energy X-ray machine calibration. Instead of using raw material of known composition to calibrate the machine being utilized for analysis, phantoms have been developed. These phantoms are made of a limited number of calibrated plates of materials simulating different body fat percentages. By combining the plates, a range of simulated body fat compositions may be obtained. However, there is no indication of a means of asserting the appropriate constituents for phantoms simulating materials other than body fat. It is also a disadvantage of the patent that the phantom only mimics distinct values, or ratios of meat and fat for two points, corresponding to the high energy and low energy x-ray attenuation values of the detector.

Furthermore, US Patent No. 6,315,447 provides for a modular phantom made from acrylic, polyvinylchloride and white vinyl. It does not however, justify the selection of these materials, or provide a method of selection of materials that could be combined to produce a phantom.

It would be an advantage to have a selection criteria or methodology for producing phantoms or reference tiles that simulates or mimics specific properties of a target material. This target material can include any material of interest, such as meat, cheese, bouillon, explosives, composites, coal or the like. It would also be an advantage to have a phantom that mimics an entire region of interest, rather than only distinct points along that region. This would increase the accuracy of the calibration, and mean that, regardless of the machine used to take the measurement, the results could be directly compared, as the entire profile has been mimicked, not just distinct points that could shift with differing machines.

All references, including any patents or patent applications, cited in this specification are hereby incorporated by reference. No admission is made that any reference constitutes prior art. The discussion of the references states what^' their authors assert, and the applicants reserve the right to challenge the accuracy and pertinency of the cited documents. It will be clearly understood that, although a number of prior art publications are referred to herein, this reference does not constitute an admission that any of these documents forms part of the common general knowledge in the art, in New Zealand or in any other country.

It is an object of the present invention to address the foregoing problems or at least to provide the public with a useful choice. Further aspects and advantages of the present invention will become apparent from the ensuing description that is given by way of example only.

DISCLOSURE OF INVENTION

According to one aspect of the present invention there is provided a phantom, for use with a composition analysis device designed to estimate the composition of a target material by creating a profile corresponding to a measured characteristic of the target material.

characterised in that the phantom mimics substantially the profile of the target material across a region of interest.

According to another aspect of the present invention there is provided a method of manufacturing a phantom substantially as described above

characterised by the steps of

a) assessing the profiles of two or more component materials, and

b) determining whether the combination of the materials in a predetermined quantity mimics substantially the profile of the target material in a region of interest, and

c) once the component materials satisfy the above condition, forming them into a phantom.

In preferred embodiments, the term 'composition analysis device' in accordance with the present invention is a machine that estimates a target material's composition by measuring the said characteristics of the target material, as defined herein. These machines could, for example, include a Magnetic Resonance Imaging (MRI) machine, a Nuclear Magnetic Resonance (NMR) machine, an X-Ray Analyse machine, a Dual X-ray Analysis (DXA) machine or an Ultrasonic device. The specification will now focus on the use of the present invention with DXA machines, as the present invention has proved to be particularly useful for this application, though it will be appreciated that the invention is not solely restricted to same.

The term 'composition' in accordance with the present invention should be understood to include the ratio of one measured component of a target material to another, and to a measurement of the absolute quantity of the constituents of the target material. For example, a DXA machine can be used to examine the ratio of muscle tissue to fat in meat, or bone to muscle tissue or the like. The composition of such a material can play a vital role in determining the market value of an end product, such as Chemical Lean (CL) of manufacturing meat or the like, for supply to fast food chains.

The term 'target material' in accordance with the present invention should be understood to mean any material whose composition is desired to be determined by a device as described previously. The target material can include meat of any kind, cheese, coal, explosives, luggage or the like. However, these are listed by way of example only, and should not be seen to be limiting.

The term 'profile' in accordance with the present invention should be understood to mean the characteristic response generated by a material when it is examined by a composition analysis device, as described above. The profile is an information data set which may be compiled by iterative, periodic, or sequential measurements, which are capable of representation in any convenient graphical form. However these are listed by way of example only and are not meant to be limiting.

In preferred embodiments, the composition analysis device for measuring a profile is a DXA machine, and as such, the generated profile is an absorption coefficient profile.

The term 'characteristic' is intended to refer to any physical feature of the target material that is desired to be measured. For example, the characteristic may be the response of the target to electromagnetic radiation, such as the absorption, transmission or reflection of X-rays, optical light, gamma rays, infrared, ultraviolet, radio and the like.

In other embodiments the characteristic may be a response to ultrasound, thermal radiation and so forth.

In a preferred embodiment, the measured characteristic is the absorption of X-rays.

The term 'mimics' should be understood to mean that the phantom profile substantially corresponds (either directly or via a transformation, scaling, multiplying factor or equivalent) to the shape of the measured characteristic of the target profile over the region of interest but optionally having a different amplitude or order of magnitude.

To date, phantoms are produced to mimic the response of a target material at individual points on the corresponding target profile, and not to the whole of the significant parts thereof. In DXA analysis for example, the individual points that are selected correspond with both a high energy and low energy region at points of maximum detector sensitivity.

By mimicking at least a significant portion of the target material profile, measurements of profiles generated by varying machines can be directly compared. Previously, the high energy and low energy points that are measured by DXA machines are generated by each machine, and are therefore machine dependant. By mimicking at least a significant portion of the target material profile, any drift of the machine's output away from the optimised high and low energy regions will not cause a loss of accuracy in measurements, as the machine has been calibrated across a region, and not two selected points. This is a distinct advantage when multiple machines are being utilised across multiple lines in a factory or the like.

The term 'region of interest' in accordance with the present invention should be understood to mean the portion of the profile, be it an absorption profile or otherwise, that the user is interested in with respect to the target, or that is definable by a measurable occurrence or criteria related to the analysis of material composition. In one embodiment for example, the region of interest is defined by the maximum bandwidth of detectors in said material analysis device used to quantify said measured characteristic.

Alternatively, the region of interest may be a specific energy range corresponding to the maximum sensitivity of the said detectors for example.

The region of interest for this type of device can be defined by mathematically combining the x- ray generators spectral production with the detector response curve, to allow appropriate x-ray beam energy cutoffs to be selected. It should be appreciated however that a precise definition of the upper and lower boundaries of the region of interest is not critical to the process described, and the region of interest need not be defined particularly accurately with respect to the x-ray generator/detector pairing.

Ideally, the upper and lower boundaries of the region of interest should be relatively wide. This will ensure that the final fitted curves (target material and phantom material) cover at least the region of the spectrum that corresponds to the detector/generator x-ray production and detection.

It should appreciated however, that the region of interest can also encompass a much narrower portion of the region of the spectrum, provided that the calibration that is achieved is adequate to satisfy the requirements of the user.

In preferred embodiments, it should be appreciated that a phantom can be generated by selecting at least a first material, which has a lower x-ray absorption profile that the target material profile and combining it with at least a second material which has a higher x-ray absorption profile that the target material profile. The combination of these profiles can therefore provide an absorption profile that mimics the absorption profile of the target material across the absorption region of interest. The term 'first material' and 'second material' in accordance with the present invention should be understood to mean materials that are capable of being manufactured into a phantom and each exhibit a profile distinct from the other.

It should be appreciated that the phantom can be made up of more than just the first and second material; phantoms made up of three or more materials could work equally as well.

It should also be appreciated that the final structural makeup of the phantom can vary. The components selected to make up the phantom can be combined as distinct portions, layered in a sandwich structure or the like. The combination of components can also be mixed or blended together in such a way to provide a substantially homogenous phantom. This combining of materials can be undertaken by any convenient means such as pulverising the components and recombining them into a desired final form, or even the melting and reconstituting of constituents and casting into a preferred shape.

In preferred embodiments, the target material that is being measured is meat. This meat can include manufacturing meat, which is commonly included in non-specific meat products such as hamburger patties, sausages, etc. Manufacturing meat comprises a large portion of the meat sold nationally and internationally. Currently around 70% of beef sold out of New Zealand is sold as manufacturing meat. Such meat is shipped in plastic bags within cardboard boxes of various dimensions, containing approximately 27kg of product.

Manufacturing meat is often a target material due to the importance of Chemical Lean (CL), or the ratio of fat to muscle, as the monetary value of manufacturing meat is calculated using the CL ratio. If there is too much or too little fat in the target material, penalties can be incurred, or the meat rejected for sale. In preferred embodiments, the target material could be human tissue, for example, analysis of human bones and tissue for hip replacement X-rays or the like. The study of the ratio of muscle to fat can provide information on blocked arteries and the like also.

In other embodiments, the target material that is being measured can include explosive materials. DXA machines are employed in airports to scan luggage for weapons, explosives or undesirable or illegal materials. The ability to increase the accuracy of a machine for target materials, such as explosives, would be a distinct advantage in the current climate.

The method selects combinations of materials whose graphical representation of x-ray absorption coefficient closely resembles that of the target material, although in absolute terms the absorbance values of one absorption curve may be a multiple of those of the other by a constant value.

The method can be used to select a family of composites to produce a phantom whose x-ray absorption curves match those of variations of the target material. For example, for boxed meat, the method would select a range of composite materials that are phantoms for 100% CL to 30% CL.

There are several advantages in selecting stable, non-organic phantom materials rather than a phantom composed of the actual target material. The target material may, for example:

• have an irregular composition (e.g. unregulated lumps of composite material)

• be an irregular shape (e.g. the human body)

• present a safety hazard for workers (e.g. radioactive or poisonous materials)

• be chemically reactive (e.g. tendency to oxidise or absorb water )

• be perishable (e.g. prone to microbial attack/spoilage) In principle, the x-ray absorption curves for both the target material and components of the candidate phantom materials should be defined at all points as continuous curves within the region of interest. However, in practice such curves consist of discrete points that can be at some distance from each other.

The absorption curves for either the target material or components of the phantom material can be defined by direct measurement (using an x-ray spectrophotometer for example) or from published data.

If parts of a curve are not defined for all points within the region of interest, an interpolation and/or extrapolation procedure, such as regression fitting, cubic spline, or bezier curve fitting, may used as is convenient for the available data and processing capabilities.

X-ray beams are attenuated, i.e, the photon flux is reduced, as they pass through a material. This attenuation is described by de Beers Law. For any given mono-energetic x-ray beam of intensity, I₀ , the intensity, / , after exiting material with attenuation coefficient, μ₁ relative to material thickness, and thickness, t_x , is

( T /\

In 0/ = βχ 1 "1 v ^/z .

Take for example, where there are two section of material, with the x-ray beam passing through both. For any individual mono-energetic beam, the relationship above applies for passage through the second material. However, the relationship between different mono-energetic beams would be altered by the differences in attenuation coefficients for those beams.

In practice, x-ray detectors integrate over a number of beam energies. Since any material has different attenuation for different x-ray frequency, the total effect of several layers of material is not merely an amalgamation of individual layers. Each layer sees an effective x-ray beam with different frequency distribution, due to the differential absorption within the preceding material. This effect is termed "beam hardening" and for thick material, this effect can be quite large.

It is essential for the phantom to mimic the target material as closely as possible. A well-selected phantom will not only mimic the absorption profile of the target material, but also the secondary beam hardening effects as well.

For a material composed of several components, de Beer's law may be generalised as:

In ^fl 0/ j \

+ μ₂m₂ + ... (1)

where:

I₀ is the intensity of a mono-energetic (i.e. single frequency) x-ray beam with no material in the beam path.

I is the intensity of a mono-energetic (i.e. single frequency) x-ray beam with some mass

of a composite material in the beam path.

μ₁,ju₂ are the attenuation coefficients relative to mass of the composite material (1,2, ...) comprising the material.

m₁,m₂ are the masses in the beam path of the composite material (1, 2, ...) comprising the material.

Equation 1 can be re-expressed in the form:

where:

m_τ = the total mass within the beam. Equation 2 can be simplified to A i /\

In = μ_em_τ (3)

\ J

where the effective absorption coefficient, μ_e , is the mass-biased mean of the absorption coefficients of the constituent material:

μ_e = μ₁ — + μ₂ — + ... (4) m_τ m_τ

The systematic method described is based on the matching of a target absorption curve to a phantom composite effective x-ray absorption curve. An x-ray absorption curve is simply a collection of all the individual absorption coefficients for a range of mono-energetic x-ray beams.

Note that an x-ray device consists of an x-ray beam supply, which emits a range of x-ray frequencies, and one or more x-ray detectors. No detector can respond perfectly to all x-ray frequencies. Therefore, to model the x-ray environment, one must mathematically combine the x- ray generator's spectral production with the detector response curve over the region of interest.

In the current art, potential phantom material combinations are matched with the target material by measuring one or more discrete absorption coefficients on the actual x-ray device in question.

This procedure correctly accounts for device-specific differences arising from the x-ray generator and the detector response. However, if the absorption curve (rather than one or more discrete absorption coefficients) of the phantom material differs significantly from that of the target material, while agreeing at a few points, then a small change in x-ray environment could seriously degrade the agreement in absorption between the phantom and the target material. This is a distinct disadvantage.

Possible reasons for changes to the x-ray environment include:

• drift in the spectrum of x-rays generated, due to - electrical supply fluctuation,

- changes within the x-ray generator due to, for example, warming of the tube, causing subsequent changes in x-ray spectrum;

• changes in detector geometry;

• drift of the x-ray beam relative to detector location;

• direct changes in detector response due to;

- changes in electrical component operational parameters due to heating, electricity supply, etc,

- changes in detector composition (particularly water absorption or desorption).

• variations in scintillator afterglow

Changes in the x-ray environment may change the actual region of interest, or, more usually, may result in changes in relative spectral levels within the region of interest.

The use of phantom material can take such changes into account, allowing absorption effects due to these changes to be corrected from the phantom absorption data, but only if the absorption curve of the phantom material resembles that of the target material.

In principle, therefore, one would like to precisely match the x-ray absorption curve of the target material and that of the composite phantom material over the region of interest. Practically, however, only the actual target material itself will ever display the precise attenuation curve of the target material. However, the method described here will allow a systematic choice of appropriate materials to best match the target attenuation curve over the defined region of interest. In preferred embodiments, two different materials are chosen for their particular X-ray absorption properties. These materials are then combined in a certain ratio, such that the resultant absorption spectrum mimics that of the target material. This has the advantage of reducing the margin of error of any reading and providing a more accurate analysis of the target material.

By combining a material of both a low and high X-ray absorption response profile, both the minima and maxima values are covered. These two materials are then combined in a certain ratio, such that the resultant combination provides the X-ray absorption response profile that matches most closely the target material response.

Any variation of the expected response of the target material can then be attributed to a change in the ratio of the materials of interest, for example, fat to muscle tissue, and a value obtained.

The advantages conferred by mimicking as closely as possible, the absorption response profile of the target material are substantial. Firstly, the size of any error associated with each reading is reduced. Error reduction occurs because the traditional method of producing a phantom is to combine materials that provide an absolute maximum and absolute minimum value. The result of such a minimum/maximum approach provides for points on a line that are the greatest distance apart. A line can then be interpolated between those two points and an intermediate value obtained. The errors associated with such a method are greater than providing a phantom that mimics as closely as possible, the target material.

Another advantage conferred by the use of an accurate phantom is that the majority of traditional phantoms were actually true samples of the target materials. As such, they were often organic, and therefore open to degradation or decay over time. The inorganic phantoms that are produced by the current described method are;

• Easy to duplicate • Easy to manufacture

• Easy to characterize

• Physically robust and;

• Easy to store and transport.

BRIEF DESCRIPTION OF DRAWINGS

Further aspects of the present invention will become apparent from the following description which is given by way of example only and with reference to the accompanying drawings in which:

Figure 1 illustrates the absorption curves of PNC, Plexiglas and meat;

Figure 2 illustrates absorption curves of PNC-Plexiglas phantom and meat;

Figure 3 illustrates the absorption curves of carbon, glass and meat

Fi ure 4 illustrates absorption curves of Carbon-Glass phantom and meat.

BEST MODES FOR CARRYING OUT THE INVENTION

The following description illustrates the process of selecting appropriate materials to produce a phantom that mimics substantially the profile of a target material.

Although the phantom may be used in a variety of applications, the embodiment illustrated relates to the use of a phantom with a Dual Energy X-ray Absorptiometry (DXA) device for the analysis of meat to establish the Chemical Lean (CL) of manufactured meat.

It will be understood that this is for exemplary purposes only and the present invention may be utilised, to mimic other target materials using other material analysis devices. a) Methodology to choose a suitable composite for use with a DXA machine:

1. Calculate an effective absorption coefficient from the mass-biased mean of the known absorption coefficients of the phantom's individual components.

2. Calculate the average of the ratio of the effective absorption curve to the target's absorption curve over the region of interest (using, for example regression techniques)

3. Calculate a ratio-corrected effective absorption curve for the composite phantom material.

4. Calculate the difference between the ratio-corrected phantom absorption curve and the target absorption curve over the region of interest.

5. Repeat steps 1-4 for a number of composites of different relative mass between the composite materials.

6. Select the composite phantom with the best-fit curve for the target material over the region of interest.

7. Decide on the suitability of this potential phantom, considering

a. Curve distance between the target and phantom attenuation curves within the region of interest

b. Target and phantom properties (e.g. thickness, handling issues, operator safety, etc)

b) Curve distance calculation:

Curve distance can be calculated in any convenient manner suited to the data. Normally, a simple root-mean-square (RMS) method is sufficient, but any valid curve distance or multivariate distance method can be used (e.g. Mahalanobis distance, Linear Regression fit, or the like)

c) Quality of fit between target material and phantom:

The criteria that indicate an acceptable fit between the target material and phantom material curves depend somewhat on the measurement circumstances and materials used. The degree of match needs to be closer for systems with fine tolerance on measurement results. The limits to define acceptability of fit may need to be broader in extreme cases; for example, if target absorption curves are difficult to match, or if there are other constraints (e.g., the maximum physical size of the phantom is restricted, or a harsh environment limits the choice of phantom material).

However, a proposed composite would typically not be considered to be suitable if, for example in a Linear Regression fit:

• the remaining regressed slope (after ratio adjustment) was not within 0.9 to 1.1;

• the remaining bias was greater than 10% of the average absorption coefficient over the region of interest; or

• the fit parameter (R²) was less than 0.9.

EXAMPLE 1

Region of interest: 40 keV to 120 keN In this example, the aim is to develop a phantom composite material set to mimic meat, which is itself a composite of muscle (lean) and fat. If the target material is meat containing 30% fat (and is therefore 70% lean), the meat composite absorption curve for this case is shown in Figure (1).

It should be appreciated that a composite of aluminium and polyethylene has been used previously as a calibration phantom for DXA machines. When the above detailed method is applied to a mixture of aluminium and polyethylene to produce a phantom for real meat, the phantom produces an effective Chemical Lean above 300. The Chemical Lean should ideally fall between zero and 100.

1. Phantom material made of PNC and Plexiglass (Perspex)

The absorption curves of PVC and Plexiglass and Meat are shown in Figure 1.

A suitable choice of mass ratio for a PNC and Plexiglass phantom composite material might be approximately 40/60 in mass. The absorption curve for this composite material is shown in Figure 2 and is compared with that for meat (70% CL) after ratio correction. The degree of fit between these curves (R²) is 99.28%.

2. Glass and carbon

It should be appreciated that the phantom that is detailed below is, in itself, novel. A glass and carbon combination has never before been produced for use as a phantom.

The absorption curves of carbon (in graphite form) and Bora-Float glass are shown in Figure 3.

A suitable choice of mass ratio for carbon and glass might be 50/50 in mass. The absorption curve for this composite material is shown in Figure 4 and is compared with that for meat (70% CL) after ratio correction. The degree of fit between these curves is (R²) 96.96%. Note that these curves coincide at only two points, but their shape is very similar. The fit criteria suggest that the x-ray absorption curve of a composite based on these materials would react similarly to that of 70% CL meat should there be changes in the x-ray environment.

2a) Glass and Carbon - a worked example (matching the attenuation curves for muscle (CL=100) and a carbon-glass phantom)

The current method relies on matching the whole attenuation curve for phantoms and meat in some energy band of X-ray quanta. The outcome of matching the whole attenuation curve is that there is less dependence on sensitivities of the detectors.

Below is an example of the procedure used to match the curves. In this example, the attenuation curves for muscle (CL=100) are matched with carbon-glass phantoms. The energy band is 40 < E . < 120 keN. With accuracy R² =0.999, attenuation curves for glass (p_g = 2.33 g/cm³), carbon

(p_c = 1.7 g/cm³) and muscle (lean) ( _L = 1.05 g/cm³), taken from published data sets, may be written as follows:

Glass: _/^ I p_g =exp( 0.0037*x^Λ6-0.0918*x^Λ5 + 0.8277*x -3.1491*x^Λ3 + 3.7912*x^Λ2 + 1.3176*x + 0.2057)

Carbon: _Ju_c / _c = exp(-0.0003*x^Λ6 + 0.0111*x^Λ5-0.1499*x^Λ4 + 0.9109*x^Λ3-2.1262*x^Λ2-1.3295*x + 7.5561)

Muscle:

μ_c I p_c = exp(0.0009*x^Λ6-0.0299*x^Λ5+0.2653*x^Λ4-0.8583*x^Λ3+1.0432*x^Λ2-3.1774*x+8.2464)

Here;

x = ln(E) where E is measured In keN .

For a muscle layer h_L (cm) thick, the attenuation coefficient is (μ_L I p_L )• p_L - h_L . For the (glass/carbon) phantom, the attenuation coefficient is [μ_g I p_g p_g - h_g + (μ_c I p_c )• p _c ■ h_c , where h_s (h_c) is the thickness of glass (carbon).

Ideally, these attenuation coefficients should be equivalent over the frequency range, however this does not hold in practice and the function

(E) = μ_g.a + μ_c.b - μ_L

determines the difference between the phantom and muscle attenuation coefficients. Here

a = h_g / h_L,

b = h_c /h_L

To match the attenuation curves, we must find a minimum of integral

J = J_na² + 2J_l2ab + J₂₂b² - 2J_wa -2J₂₀b + J_Q0

where

(E)dE

^min

^min ~ 40 . E_max — 120 The minimum is determined by equations dJ I da = 9/ / db = 0 that results in a linear system of equations for a and b:

J_na + J_l2b = J 10

J₁ 12α + J_∞ 22b' = J 20

This means that a 100 mm thick box of muscle (CL=100) may be theoretically replaced by a phantom of 14.6 mm glass and 50.5 mm carbon thickness.

3. Example of Phantom set required for calibration

Two types (A and B) of phantoms are need for calibration.

Number of phantoms of each type: 3

Total number of phantoms: 6

Calibration of the detectors may be achieved by using combinations of phantoms (see Table 1): 3 A, 3B, 2A+B, 2B+A, 2A, 2B, A and B to cover all the spectrum of chemical lean and meat mass, as shown in Table 1.

Aspects of the present invention have been described by way of example only and it should be appreciated that modifications and additions may be made thereto without departing from the scope thereof.

Claims

CLAIMS:

1. A phantom, for use with a composition analysis device configured to estimate the composition of a target material by creating a profile corresponding to a measured characteristic of the target material characterised in that the phantom substantially mimics the profile of the target material across a region of interest.

2. The phantom as claimed in claim 1, wherein said composition analysis device include Magnetic Resonance Imaging (MRI) machines, Nuclear Magnetic Resonance (NMR) machines, a Dual X-ray Analysis (DXA) machine and/or an Ultrasonic devices.

3. The phantom as claimed in claim 1 or claim 2, wherein said composition includes both the ratio of one measured constituent of a target material to another, and to a measurement of an absolute quantity of the constituents of the target material.

4. The phantom as claimed in any one of the preceding claims, wherein said profile is an information data set compiled by iterative, periodic, and/or sequential measurements.

5. The phantom as claimed in any one of the preceding claims, wherein said profile is capable of representation in a graphical form.

6. The phantom as claimed in any one of the preceding claims, wherein the composition analysis device for measuring a profile is a DXA machine.

7. The phantom as claimed in any one of the preceding claims, wherein said profile is an absorption coefficient profile.

8. The phantom as claimed in any one of the preceding claims, wherein the said measured characteristic is the absorption of x-rays with respect to x-ray energy.

9. The phantom as claimed in any one of the preceding claims, wherein said phantom profile shape directly mimics the target profile of the measured characteristic of the target profile over the region of interest.

10. The phantom as claimed in any one of claims 1-8, wherein said phantom profile shape indirectly mimics the measured characteristic of the target profile over the region of interest and is transformed into a direct mimicking of said target profile via application of a transformation, scaling, multiplying factor or other equivalent mathematical manipulation.

11. The phantom as claimed in any one of the preceding claims, wherein the region of interest is defined by the maximum bandwidth of detectors in said material analysis device used to quantify said measured characteristic.

12. The phantom as claimed in any one of claims 1-10, wherein the region of interest is an energy range corresponding to the maximum sensitivity of the said detectors.

13. The phantom as claimed in any one of claims 1-10, wherein the region of interest is defined mathematically by combining an x-ray generator's spectral production with a detector response curve.

14. The phantom as claimed in any one of the preceding claims, wherein the phantom is generated by selecting at least a first material, which has a lower x-ray absorption profile that the target material profile and combining it with at least a second material which has a higher x-ray absorption profile that the target material profile.

15. The phantom as claimed in claim 14, wherein said first material and second material are materials capable of being manufactured into a phantom and each exhibit a profile distinct from the other.

16. The phantom as claimed in any one of the preceding claims, wherein the phantom is formed from three or more constituents.

17. The phantom as claimed in any one of the preceding claims, wherein the constituents of said phantom are formed as distinct portions, layered in a sandwich structure.

18. The phantom as claimed in any one of claims 1-16, wherein the combination of components are mixed or blended together to provide a substantially homogenous phantom.

19. The phantom as claimed in any one of claims l-16,and 18 wherein said combining of materials is achieved by either pulverising the components and recombining them into a desired final form, or melting and reconstituting of constituents and casting into a preferred shape.

20. The phantom as claimed in any one of the preceding claims, formed from carbon and glass with mass ratio of approximately 50/50.

21. The phantom as claimed in any one of the preceding claims, formed from 14.6 mm glass and 50.5 mm carbon thickness as a mimic of a 100 mm thick box of muscle with a Chemical Lean of 100%.

22. A method of producing a phantom as claimed in any of the preceding claims, characterised by the steps of

- assessing the profiles of two or more component materials, and

- determining whether the combination of the materials in a predetermined quantity substantially mimics the profile of the target material in a region of interest, and

- forming a phantom from component materials satisfying the above conditions.

23. The method as claimed in claim 22, further characterised by the selection of said component materials from those whose x-ray absorption coefficient graphical representation closely resembles that of the target material.

24. The method as claimed in claim 23, wherein said graphical representations of a component material closely resembles that of the target material by application of a scaling factor.

25. The method as claimed in claim 24, wherein the absorption curves for either the target material or components of the phantom material is defined by direct measurement.

26. The method as claimed in claim 25, wherein said direct measurement is by an x-ray spectrophotometer.

27. The method as claimed in claim 24, wherein the absorption curves for either the target material or components of the phantom material are defined from published data.

28. The method as claimed in any one of claims 22-27, wherein to determine whether the combination of the materials in a predetermined quantity substantially mimics the profile of the target material within the region of interest, an interpolation, extrapolation and/or curve fitting procedure is used.

29. The method as claimed in claim 28, wherein said interpolation, extrapolation and/or curve fitting procedures include regression fitting, cubic spline, and/or Bezier procedures.

30. The method as claimed in any one of the preceding claims, wherein said step of assessing the profiles of two or more component materials to choose a suitable composite phantom for use with material analysis machine includes the further steps of:

- calculating an effective absorption coefficient from a mass-biased mean of known absorption coefficients of the phantom's individual components; - calculating the average of the ratio of an effective absorption curve to the target's absorption curve over the region of interest;

- calculating a ratio-corrected effective absorption curve for the composite phantom material;

- calculating the difference between the ratio-corrected phantom absorption curve and the target absorption curve over the region of interest.

31. The method as claimed in claim 30, further including the steps of

- repeating the steps of claim 30 for a number of composites of different relative mass between the composite materials;

- selecting the composite phantom with the best-fit curve for the target material over the region of interest.

32. The method as claimed in claim 30 or claim 31, further characterised by the steps in which;

a decision is made on the suitability of the potential phantom by considering at least one of;

a) the curve distance between the target and phantom attenuation curves within the region of interest;

b) target and phantom properties ; and/or

c) curve distance calculation:

33. The method as claimed in claim 32, wherein said curve distance is calculated in any known manner suited to the data.

34. The method as claimed in claim 33, wherein said curve distance is calculated by a method including, a simple root-mean-square (RMS) method, Mahalanobis distance, Linear Regression fit, and/or any other valid curve distance or multivariate distance procedures.

35. The method as claimed in any one of claim 30-34 wherein a proposed composite would not be considered suitable according to a Linear Regression fit if:

- the remaining regressed slope (after ratio adjustment) was not within 0.9 to 1.1; or

- the remaining bias was greater than 10% of the average absorption coefficient over the region of interest; or

- the fit parameter (R ) was less than 0.9.

36. A sample material analysis device substantially as hereinbefore described with reference to, and as shown in the accompanying drawings.

37. A method substantially as hereinbefore described with reference to, and as shown in the accompanying drawings.