WO2012040611A1 - Anthropomorphic, x-ray and dynamic contrast-enhanced magnetic resonance imaging phantom for quantitative evaluation of breast imaging techniques - Google Patents

Anthropomorphic, x-ray and dynamic contrast-enhanced magnetic resonance imaging phantom for quantitative evaluation of breast imaging techniques Download PDF

Info

Publication number
WO2012040611A1
WO2012040611A1 PCT/US2011/053053 US2011053053W WO2012040611A1 WO 2012040611 A1 WO2012040611 A1 WO 2012040611A1 US 2011053053 W US2011053053 W US 2011053053W WO 2012040611 A1 WO2012040611 A1 WO 2012040611A1
Authority
WO
WIPO (PCT)
Prior art keywords
phantom
lesion
breast
images
tissue mimicking
Prior art date
Application number
PCT/US2011/053053
Other languages
French (fr)
Inventor
Melanie Freed
Aldo Badano
Original Assignee
The United States Of America, As Represented By The Secretary Department Of Health & Human Services
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Priority to US38592910P priority Critical
Priority to US61/385,929 priority
Priority to US201061424495P priority
Priority to US61/424,495 priority
Application filed by The United States Of America, As Represented By The Secretary Department Of Health & Human Services filed Critical The United States Of America, As Represented By The Secretary Department Of Health & Human Services
Publication of WO2012040611A1 publication Critical patent/WO2012040611A1/en

Links

Classifications

    • AHUMAN NECESSITIES
    • A61MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61BDIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B6/00Apparatus for radiation diagnosis, e.g. combined with radiation therapy equipment
    • A61B6/58Testing, adjusting or calibrating devices for radiation diagnosis
    • A61B6/582Calibration
    • A61B6/583Calibration using calibration phantoms
    • GPHYSICS
    • G09EDUCATION; CRYPTOGRAPHY; DISPLAY; ADVERTISING; SEALS
    • G09BEDUCATIONAL OR DEMONSTRATION APPLIANCES; APPLIANCES FOR TEACHING, OR COMMUNICATING WITH, THE BLIND, DEAF OR MUTE; MODELS; PLANETARIA; GLOBES; MAPS; DIAGRAMS
    • G09B23/00Models for scientific, medical, or mathematical purposes, e.g. full-sized devices for demonstration purposes
    • G09B23/28Models for scientific, medical, or mathematical purposes, e.g. full-sized devices for demonstration purposes for medicine
    • G09B23/286Models for scientific, medical, or mathematical purposes, e.g. full-sized devices for demonstration purposes for medicine for scanning or photography techniques, e.g. X-rays, ultrasonics
    • GPHYSICS
    • G09EDUCATION; CRYPTOGRAPHY; DISPLAY; ADVERTISING; SEALS
    • G09BEDUCATIONAL OR DEMONSTRATION APPLIANCES; APPLIANCES FOR TEACHING, OR COMMUNICATING WITH, THE BLIND, DEAF OR MUTE; MODELS; PLANETARIA; GLOBES; MAPS; DIAGRAMS
    • G09B23/00Models for scientific, medical, or mathematical purposes, e.g. full-sized devices for demonstration purposes
    • G09B23/28Models for scientific, medical, or mathematical purposes, e.g. full-sized devices for demonstration purposes for medicine
    • G09B23/30Anatomical models
    • G09B23/34Anatomical models with removable parts
    • AHUMAN NECESSITIES
    • A61MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61BDIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B2560/00Constructional details of operational features of apparatus; Accessories for medical measuring apparatus
    • A61B2560/02Operational features
    • A61B2560/0223Operational features of calibration, e.g. protocols for calibrating sensors
    • A61B2560/0228Operational features of calibration, e.g. protocols for calibrating sensors using calibration standards

Abstract

Anthropomorphic breast phantoms comprise a combination of adipose tissue mimicking components and fibroglandular tissue mimicking components. Typically, x-ray attenuation coefficients or magnetic resonance relaxation times Ύι and T2 are selected that are sufficiently similar to actual patient tissues. The mimicking components are distributed within the phantom such that images of the phantom contain features similar to those of patient tissues. A breast phantom can be based on a lard/egg white combination that is shaped to approximate a human breast, or a compressed human breast as prepared for mammography. The phantoms can include lesion chambers that permit the introduction of contrast agents to simulate benign or malignant lesions, and contrast agent concentration can be time varied to produce washout curves.

Description

ANTHROPOMORPHIC, X-RAY AND DYNAMIC CONTRAST-ENHANCED MAGNETIC RESONANCE IMAGING PHANTOM FOR QUANTITATIVE EVALUATION OF BREAST IMAGING TECHNIQUES CROSS REFERENCE TO RELATED APPLICATIONS

This application is claims the benefit of U.S. Provisional Patent Application 61/385,929, filed September 23, 2010 and U.S. Provisional Patent Application 61/424,495, filed December 17, 2010, both of which are incorporated herein by reference in their entireties.

BACKGROUND

Advanced imaging techniques have been developed for diagnosis and treatment of many significant medical conditions. For example, sophisticated x-ray and magnetic resonance imaging (MRI) techniques have been applied to early detection of breast cancer. These techniques often provide detailed images that can permit precise location of malignant tissues as well as the overall extent of any malignancy. With the views provided by such imaging techniques, clinicians can establish treatments that are precisely targeted to the condition of each patient. X- ray mammography in combination with palpation has become standard for breast cancer screening. Unfortunately, x-ray mammography can miss many cancers, especially for patients with dense breasts.

Sophisticated MRI methods such as dynamic contrast-enhanced (DCE) MRI having high sensitivity have also been developed. These methods take advantage of specialized breast coils and Gd-based contrast agents to overcome the overlap of intrinsic T\ and T2 relaxation times. While DCE-MRI has high sensitivity, this high sensitivity is associated with a low specificity so that DCE-MRI tends to produce increased numbers of false positives. The increase in false positives may be correlated with a possible increase in mastectomy rates. Other techniques have been applied or are under development for the detection of breast cancer. Both breast tomosynthesis and dedicated breast CT have been demonstrated to have a potential for high sensitivity, but these techniques also may require increased radiation dose levels. Although these advanced imaging techniques can provide detailed images of patient tissues, comparison of the images generated by these methods for disease diagnosis or treatment can be subjective. Indeed, even subjecting a patient to a common imaging protocol in different hospitals or clinics can produce different images of similar tissues. Such differences can be associated with the use of differing imaging systems, differing set-ups of the same imaging systems, or other factors. Objective comparison is difficult so that it is correspondingly difficult to accurately compare images obtained at different clinics or with different imaging systems in a common clinic. In addition, due to individual system variations, it can be difficult to compare the many available techniques to identify one that is most appropriate for any particular patient. Inter-comparison and calibration of differing imaging methods could enhance the usefulness of established imaging techniques, permit reliable inter-comparisons, and aid in the development of additional imaging protocols.

Several breast MR phantoms have already been developed that can aid in the evaluation and development of imaging methods. See, for example, Mazzara et al., "Use of a modified polysaccharide gel in developing a realistic breast phantom for MRI," Magn. Reson. Imaging 14: 639-648 (1996), and Liney et al., "A simple and realistic tissue equivalent breast phantom for MRI," J. Magn. Reson. Imaging

10:968-971 (1999). In addition, the American College of Radiology (ACR) phantom for accreditation (http://www.acr.org/accreditation/mri.aspx) is available for the quantitative evaluation of image quality parameters such as resolution, signal to noise ratio, and the presence of artifacts. However, images based on these phantoms do not have the appearance of actual patient tissues.

SUMMARY

Disclosed herein are phantoms and methods of making and using such phantoms for both x-ray and magnetic resonance imaging. In some examples, magnetic resonance imaging phantoms comprise an adipose tissue mimicking component and a fibroglandular tissue mimicking component distributed in the adipose tissue mimicking component, wherein the combined components are shaped to mimic a body part. In some examples, the phantoms include a container having a shape that mimics the body part, wherein the combined components are retained in the container. In additional examples, the container has a shape that mimics the compressed body part. In further examples, the body part is human breast. In some embodiments, the adipose and fibroglandular tissue mimicking components are lard and egg whites, respectively. In other examples, the fibroglandular tissue mimicking component is distributed in the adipose component such that a normalized stationary covariance has a full width at half maximum of less than about 3 mm, 5 mm, or 10 mm. In further representative examples, the phantoms comprise one or more lesion chambers situated within the combined adipose and fibroglandular tissue mimicking components, the lesion chambers defining respective lesion volumes. In other examples, a plurality of tubes is fluidically coupled to the lesion volumes, and the lesion volumes contain a magnetic resonance contrast agent. In still further examples, the lesion chambers include a plurality of lobulations.

Methods of making a phantom comprise providing an at least partially liquefied adipose tissue mimicking component and distributing a fibroglandular tissue mimicking component within the at least partially liquefied adipose tissue mimicking component. Typically the adipose tissue mimicking component is fully liquefied. The combined adipose tissue mimicking component and the

fibroglandular tissue mimicking component are cooled so as to solidify the combination. In further examples, the adipose and fibroglandular tissue mimicking components are lard and egg whites, respectively, and the combination is shaped so as to mimic a body part. In some examples, the body part is a human breast, and the combination is shaped so as to be at least partially spherical. In additional examples, a lesion chamber is provided within the combination.

Methods comprise providing a breast phantom that includes a lesion chamber having at least one fluid inlet and fluidically coupling a contrast agent to the lesion chamber. The contrast agent is controlled so that contrast agent concentration in the lesion chamber is time varying so as to correspond to a human tumor washout curve. As used herein, a washout curve refers to time a varying image signal intensity that is a function of time from application of a contrast agent. A washout curve generally refers to signal intensity decline after an initial rise in signal intensity. In some examples, the contrast agent is controlled by varying a mixture of a contrast agent containing fluid with a tissue mimicking fluid. In typical examples, the contrast agent includes gadolinium.

These and other features and aspects of the disclosed technology are set forth below with reference to the accompanying drawings.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a sectional view of a representative container for forming a breast imaging phantom.

FIG. 2 is a block diagram illustrating a method of producing a phantom.

FIG. 3 is a graph of Ύι and T2 relaxation times for materials used in the disclosed phantoms along with human data from selected references.

FIGS. 4A-4B are graphs illustrating fractional changes in T\ and T2 relaxation times, respectively for lard and egg white (as incorporated into a phantom) as a function of time since phantom production.

FIGS. 5A-5J are representative patient and phantom images. FIG. 5A is T weighted, fat suppressed patient image. FIG. 5B is a Ti-weighted patient image without fat suppression. FIG. 5C is a T2 spectral adiabatic recovery (SPAIR) fat suppressed patient image. FIG. 5D is a Ti-weighted, fat suppressed phantom image. FIG. 5E is a Ti-weighted phantom image without fat suppression. FIG. 5F is a T weighted phantom image without fat suppression. FIG. 5G is a Ti-weighted, fat suppressed phantom image. FIG. 5H is a short T\ inversion recovery (STIR) phantom image with fat suppression. FIG. 51 is a T2-weighted phantom image without fat suppression. FIG. 5J is a T2-weighted, fat suppressed phantom image. All are shown with a 10 mm scale bar.

FIG. 6 includes representative images of regions of interest from fat suppressed, Tt -weighted MRI images based on phantoms and patients. A top row contains patient image data and a bottom row contains phantom data.

FIGS. 7A-7B show patient and phantom overall stationary covariance matrices, respectively.

FIGS. 8A-8B are graphs of stationary covariance as a function of position in anterior-posterior and right- left directions. FIG. 9 is a graph of RMS variation in a stationary covariance matrix estimate due to instrumentation noise as a function of a number of 35 pixel by 35 voxel regions of interest (ROIs) for offset distances in a range of 4.0 mm to 20.0 mm.

FIG. 10 illustrates a simulated lesion having a lobulated spherical lesion chamber with fluid inlet/outlet tubes.

FIGS. 11A-11B illustrate alternative geometries for lesion chambers.

FIGS. 12A-12B are photographs of additional lesion surrogates that include spherical lesion chambers with and without lobulations.

FIGS. 13A-13B are fat-suppressed, TVweighted, gradient echo magnetic resonance images produced with simulated lesions such as those shown in FIGS. 12A-12B, respectively.

FIG. 14 is a schematic diagram of a tissue-mimicking phantom with a lesion chamber configured to produce time varying contrast agent concentrations for forming images corresponding to patient washout curves.

FIGS. 15A-15B illustrate time varying contrast agent distributions in collinear and non-collinear lesion chambers in different flow planes.

FIGS. 16A-16B are graphs of average fractional contrast fluid concentration as a function of time in benign and malignant lesions, respectively.

FIG. 17 is a photograph of a tissue-mimicking phantom having a compressed shape such as associated with mammography.

FIG. 18 illustrates results of a segmentation algorithm on a central slice of the phantom for inclusion in x-ray scatter simulations. For each voxel, the fraction of material that is air, the jar, lard, and egg whites is indicated.

FIG. 19A is a photograph of a compressed phantom.

FIG. 19B is an example x-ray image of a phantom.

FIG. 19C is an example patient mammogram.

FIGS. 20A-20B provide comparisons of x-ray mass attenuation coefficients for breast adipose tissue and adipose-mimicking phantom material, and for breast glandular tissue and glandular-mimicking phantom material, respectively.

FIG. 21 contains images of example patient and phantom ROIs. Upper images are patient images, lower images are phantom images. All ROIs represent areas 3.5 cm by 3.5 cm in object space. FIGS. 22A-22B depict overall stationary covariance matrices for patient and phantom data sets.

FIGS. 23A-23B are graphs illustrating patient and phantom overall stationary covariance matrices (as shown in FIGS. 22A-22B) in the anterior-posterior and superior- inferior directions.

FIGS. 24A-24B illustrate the influence of scatter on a stationary covariance matrix.

FIG. 25 illustrates Monte Carlo simulations to estimate the amount of scatter produced by an example phantom.

FIGS 26A-26B illustrate validation of the Monte Carlo simulated scatter results for a heterogeneous phantom. FIG. 26A is an image of a phantom with tungsten discs labeled 1-5 in place. FIG. 26B is a comparison of the estimated scatter-to-primary ratio for each disc location. DETAILED DESCRIPTION

As used in this application and in the claims, the singular forms "a," "an," and "the" include the plural forms unless the context clearly dictates otherwise. Additionally, the term "includes" means "comprises." Further, the term "coupled" does not exclude the presence of intermediate elements between the coupled items.

The systems, apparatus, and methods described herein should not be construed as limiting in any way. Instead, the present disclosure is directed toward all novel and non-obvious features and aspects of the various disclosed

embodiments, alone and in various combinations and sub-combinations with one another. The disclosed systems, methods, and apparatus are not limited to any specific aspect or feature or combinations thereof, nor do the disclosed systems, methods, and apparatus require that any one or more specific advantages be present or problems be solved. Any theories of operation are to facilitate explanation, but the disclosed systems, methods, and apparatus are not limited to such theories of operation.

Although the operations of some of the disclosed methods are described in a particular, sequential order for convenient presentation, it should be understood that this manner of description encompasses rearrangement, unless a particular ordering is required by specific language set forth below. For example, operations described sequentially may in some cases be rearranged or performed concurrently. Moreover, for the sake of simplicity, the attached figures may not show the various ways in which the disclosed systems, methods, and apparatus can be used in conjunction with other systems, methods, and apparatus. Additionally, the description sometimes uses terms like "produce" and "provide" to describe the disclosed methods. These terms are high-level abstractions of the actual operations that are performed. The actual operations that correspond to these terms will vary depending on the particular implementation and are readily discernible by one of ordinary skill in the art.

Phantoms that exhibit suitable features for use in the inter-comparison of imaging methods, and the calibration and assessment of particular methods and imaging systems are described below. The disclosed phantoms and associated methods are suitable for both magnetic resonance and x-ray based evaluation methods. Because many such methods are known, for convenience, they are not described further herein. Generally, such phantoms can be subjected to the same or similar imaging protocols as in patient exposures, and images obtained of the phantoms used for comparison and assessment. Such phantoms are configured to mimic actual patient tissue response in some or all image methods under evaluation. For example, breast imaging can be based on x-ray exposure or magnetic resonance (MR), so that phantoms for comparison of these imaging methods preferably provide suitable response to x-rays and magnetic fields. In addition, a preferred phantom can include portions that have structure similar to that of normal patient tissues as well as portions similar to lesions or other disease. Thus, phantom Ύι and T2 relaxation times for MRI and x-ray attenuation for x-ray based methods can be selected to provide contrast similar to that available in actual tissues. For example, a breast phantom preferably mimics the complex structure between fibroglandular and adipose tissues present in the human breast since this provides a confounding signal that can obscure or mimic lesions and complicate diagnoses. Lesions included in a phantom can mimic properties that clinicians use to diagnose breast cancer such as border shape and dynamic washout curve shape for MRI. Phantom tissue and lesion properties can be known so that estimates derived from images can be related to actual phantom structure.

Anthropomorphic MR Phantoms

In the examples described below, phantoms are configured to provide T\ and

T2 times that are similar to those of adipose and glandular human breast tissues. In addition, phantom structures are provided to mimic normal anatomical structures and lesions. The phantoms also permit active suppression of MR signals from adipose-simulating phantom portions and can be configured to include simulated lesions.

A fatty or fat like material is selected to mimic adipose tissue. Human white adipose tissue consists mostly of lipids in the form of triglycerides. Myristic, palmitic, palmitoleic, stearic, oleic, and linoleic fatty acids make up more than 90% of the triglyceride component. Lard is composed entirely of lipids, about 98% of which are in the form of triglycerides, wherein myristic, palmitic, palmitoleic, stearic, oleic, and linoleic acids comprise over 90% of total lipid content. See, for example, B.D. Flickinger and N. Matsuo, "Diacyl glycerols," in Baileys Industrial Oil and Fat Products: Volume 3, "Edible oil and fat products: specialty oils and oil products," F. Shahidi ed. (John Wiley & Sons, Inc., Hoboken, New Jersey, 6th ed, 2005), p. 38, and USDA National Nutrient Database for Standard Reference,

Release 22, U.S. Department of Agriculture, Agricultural Research Service (2009). Thus, lard is a convenient material to mimic adipose tissue as it has a similar composition, is readily available, requires no special handling, and is stable.

In addition to adipose tissue, the human female breast is also made up of ducts, lobules, and the associated fibrous stromal compartment that make up the so- called glandular portion of the breast. Both ducts and lobules consist of epithelial and myoepithelial cells surrounded by a basement membrane that is made up of primarily laminin and type IV collagen proteins. The main extracellular matrix component of the interstitial stromal compartment is type I collagen protein. See, for example, Keely et al., "The spatial and temporal expression of the α2β1 integrin and its ligands, collagen I, collagen IV, and laminin, suggest important roles in mouse mammary morphogenesis," Differentiation 59: 1-13 (1995). Therefore, a large part of the glandular tissue comprises proteins. Thus, hen's egg whites were chosen to mimic glandular tissue since egg whites have a similar composition, and have unique functional properties that aid in structure formation. Other types of eggs could be used.

A significant fraction of glandular breast tissue is made up of water. One study has shown that normal patient glandular breast tissue has a water content of 41-76% by weight. See, for example, Campbell and Land, "Dielectric properties of female human breast tissue measured in vitro at 3.2 GHz," Phys. Med. Biol. 37: 193- 210 (1992). Therefore, a glandular- tissue-mimicking phantom material should be high in both water and protein content. Egg whites provide a good first order match to human glandular tissue makeup. Egg whites are made up of mostly water (87.6%) and proteins (10.9%). See, for example, USDA National Nutrient Database for Standard Reference, Release 22, cited above. The major proteins are ovalbumin, ovotransferrin, ovomucoid, ovoglobulin G2, ovoglobulin G3, and lysozyme, which make up about 91.9% of the total protein content of egg 573 whites. See, for example, Burley and Vadehra, The avian egg: chemistry and biology, 1st ed. (John Wiley & Sons, Inc., New York, 1989), p. 72, Table 4.3. While the types of protein in egg whites and human breast tissue differ, the elemental composition of different proteins is almost indistinguishable. See, for example, E.G. Young, "Occurrence, classification, preparation and analysis of proteins," in Comprehensive

Biochemistry: Proteins (Part 1), Florkin and Stotz eds. (Elsevier Publishing

Company, Amsterdam, 1st ed, 1963), p. 28. In addition, all proteins are made up of amino acids, which all bind H in similar ways. Therefore, the MRI signal of the proteins in egg whites is similar to that of the proteins found in human breast tissue.

In addition to having a suitable composition, and being readily available, egg whites also have several functional properties that make them suitable for forming structures with no intervening septum. They coagulate irreversibly after heating and do not dissolve into the surrounding material. While the shelf-life of egg whites is markedly less than refined lard, egg whites can be heated, stored in an air-tight container, and mixed with a preservative to improve shelf-life. The true shelf-life of both phantom components can be measured as described below. Phantoms can be conveniently shaped so as to simulate or approximate actual breast shapes. With reference a sectional view shown in FIG. 1, a phantom preparation container 100 includes a hemispherical section 102, a cylindrical section 104, and a lid 108. The hemispherical section 102 has a radius selected to approximate human breast shape and volume, and in a typical example, defines a volume 106 of about 425 ml. The lid 108 includes ports 110, 112 for introducing or removing material from the container 100. The lid 108 can be secured to the cylindrical section 104 with screws 114, and the ports 110 can be sealed with screws 116. The numbers of screws used to secure the lid can be selected as need, and two, three or more ports can be provided as convenient.

Different volumes and shapes can be selected as well. For example, hemi- elliptical, conical, and other shapes can be used, and larger or smaller volumes can be selected. In addition, shapes can be defined by different curvatures in different directions so that container cross sections can be oval, elliptical, or other shapes. If preferred, a container can be provided that matches any particular patient by molding a container shape based on a particular patient shape and size. Container shapes can be selected based on anticipated or actual breast distortions associated with breast contact during imaging. The container need not define a smooth curve but can be formed as a series of segments as well, if more convenient.

With reference to FIG. 2, method 200 of constructing a phantom includes selecting an adipose-mimicking component (such as lard) and a glandular tissue mimicking component (such as egg white) at 201. At 202, a phantom composition is selected based on relative proportions of adipose and glandular tissue mimicking components. Phantom composition can be selected to conform to average breast composition, or tailored to a particular patient. In one example, an adipose tissue mimicking component and a glandular tissue mimicking component were selected and combined so that the glandular mimicking component was about 24% by volume of the mixture. Actual patient breast densities can range from about 2% to about 72%, and higher densities can be achieved by increasing the relative proportion of the adipose mimic, and lower densities by increasing the relative proportion of the adipose mimic. As noted above, convenient materials are lard and egg whites. At 204, the mimicking components are prepared for introduction into a phantom container. For example, if egg white is selected to mimic glandular tissue, a preservative such as 0.2% weight per volume water soluble DOWICIL

preservative (active ingredient (l-(3-chloroallyl)-3,5,7-triaza-l-azoniaadamantane chloride in sodium bicarbonate) can be added to the egg whites. In other examples, other antimicrobial materials can be selected. If a solid adipose mimic such as lard is selected, the adipose mimic is melted. The adipose mimic (melted) and the glandular mimic (with preservative) are mixed and heated and stirred at 206.

Heating and stirring parameters can be selected to produce a preferred distribution of the fibro glandular mimic in the phantom. Air bubbles can be removed by placing the mixture in a vacuum at 208. The mixture is then place in a suitably shaped container at 210, and then cooled and rotated at 212. One or more rotations during cooling can redistribute the glandular mimic in the adipose mimic, and produce an outermost layer of adipose mimic. Example: Lard/Egg White Breast DCE-MRI Phantom

Egg whites were prepared by separating egg whites from yolks, and placing the separated egg whites into a 600 ml glass Griffin beaker until an egg white volume of 100 ml was achieved. The egg whites were not mixed to homogenize them but could be mixed to produce smaller structures, if desired. A preservative (DOWICIL 75) was added in a ratio of 0.2% DOWICIL by volume, i.e., for 100 ml of egg whites (100 ml) *(0.2g/ 100ml) = 0.2 g DOWICIL 75 was added. The preservative was placed in the Griffin beaker and mixed very gently with a glass stirring rod so as to avoid breaking the egg white structure.

A water bath was prepared by placing a container of water on a stirring hot plate and heating the water bath without stirring until the water bubbled slightly, but did not boil. A stirring rod was placed in a 1000 ml glass Griffin beaker, and about 700 ml of lard was then added. The beaker and lard were then placed in the hot water bath and stirred continuously by hand with the glass stirring rod to start melting the lard. After the lard was almost liquid, the water bath was removed from the hot plate, and the lard beaker was placed directly on the hot plate heating surface. The hot plate surface temperature was about 270°C and stirring was at 125 RPM. Lard temperature was monitored with a thermometer, and bubbles were skimmed from the lard surface when noted. After the lard reached a temperature of between about 100°C and 110°C, the stirring rate was maintained at 125 RPM or increased to 350 RPM and the egg white mixture was added rapidly to the lard. Different rotation rates were selected to vary phantom internal structure. The lard/egg white mixture was cooked and stirred for 30 seconds, and then removed from heat.

A portion of the egg white/lard mixture was then poured into a phantom container such as illustrated in FIG. 1, leaving space between the lard/egg white mixture so that the lid of the container could be added later. Additional liquefied lard could be added later through fill ports. The phantom container and its contents were allowed to cool at room temperature for about 10 min. The phantom container was then placed in a desiccator and subjected to vacuum for about 20 minutes. For the last 10 minutes of vacuum exposure, the desiccator and phantom were periodically agitated slightly to displace any trapped bubbles.

While lard was cooling under vacuum, inner holes of hex plugs for the container were filled with lard using a syringe and allowed to solidify. The container lid flange and screw holes in the lid were coated with a thin layer of a silicone epoxy, and the lid was attached with screw to the top of the jar. The remaining volume of the phantom container was gently filled with additional liquid lard through the fill ports using a funnel so as to fill but avoiding introducing bubbles. The phantom container was tilted to remove any remaining small bubbles. When filled so that the lard reached both the hex port plugs, filling was halted and the funnel was removed.

The threads of the fill port plugs were wrapped with TEFLON tape, and the bottom of the flange of one of the hex plugs was coated with silicone epoxy. This plug was slowly threaded into a fill port so that the lard level remained flush with the top of the other fill port as the lard cooled. After about 5-10 min., the plug was completely threaded into the selected fill port. The bottom of the flange of the remaining fill plug was coated with the silicone epoxy, and the plug was slowly screwed into the second fill port over a period of between about 5 to 10 minutes until fully inserted. Once both plugs were securely in place, the phantom container was shaken to break up egg pieces, and then placed curved side down to cool. About four hours after the egg whites and lard were mixed, the phantom container was flipped over once, and then placed with the curved side up to cool overnight.

Ti and T2 relaxation times were measured using inversion recovery and spin echo sequences. For lard, relaxation times were measured using a lard filled tube and assuming mono-exponential signal behavior. Relaxation values for egg white material were obtained using phantom measurements as these times are known to vary with preparation technique. These measurements were repeated for three different phantoms made with different construction parameters (such as lard heating temperatures and mixing rotational velocities). FIG. 3 is a graph illustrating Ti and T2 relaxation times for the materials used in these representative phantoms (points indicated with squares) along with human data from selected references (points shown with circles, triangles, diamonds, and lines). Error bars are one standard error. Lines are plotted for data from Graham et al., "Magnetic resonance properties of ex vivo breast tissue at 1.5 T," Magn. Reson. Med. 38:669-677 (1997) and indicate contours that include a calculated 12.5% and 7.5% probability of their measured tissues. The Ti and T2 relaxation times of the phantom materials fall within two standard errors of the human data for both the adipose- and

fibroglandular- mimicking materials. The phantom Ti relaxation times are a better match to human data than the phantom T2 relaxation times and are the primary determinants of image contrast for DCE-MRI studies. As shown in FIG. 3, Ti relaxation times are about 200 ms and 1100 ms, and T2 relaxation times are about 38 ms and 70 ms for lard and egg whites, respectively.

FIGS. 4A-4B are graphs of fractional change in Ti and T2 relaxation times, respectively for lard and egg white (as incorporated into a phantom) as a function of time since phantom production. Data points are normalized by Ti and T2 relaxation times on the phantom production date. As shown in FIGS. 4A-4B, Ti and T2 relaxation times of lard and egg are stable to within about 8% and 15%, respectively, over a period of 9 months. Error bars are the standard deviation over all image volume elements (voxels) included in the computation.

Phantom structure can be compared with patient data based on covariance matrices. A covariance matrix measures how pixel values in an image vary with respect to values of other pixels within a population of phantom or patient images. A full covariance matrix of an image g having M pixels (M =N ) for an image portion (region of interest or ROI) of N by N pixels can be expressed as an M x M matrix with elements given by

wherein the overbar indicates an average and * indicates complex conjugation. If gi and gj are statistically independent and i = j, then Ky = 0. If i = j, then Kjj is equal to the variance of gj. A satisfactory estimate of Ky can be based on an average over many instances,

Because only a limited number of images are available relative to the number of elements that must be computed for a full covariance matrix, the covariance matrix can be averaged over all positions within an image ROI, therefore assuming wide sense stationarity within the ROI and reducing the variance of the estimate of K at the expense of position-dependent information. Such an averaged covariance matrix is referred to as a stationary covariance matrix and it represents average direction-dependent correlation strength over all positions in the ROI and is a second order estimate of image texture. The stationary covariance matrix has 2N - 1 x 2N - 1 elements and can be expressed as:

stationary / \

^ pq ~ \ iO+P+qN) /ie [(m+nN )eS ] >

wherein p and q are relative offset values or offset indices (both are in a range [-(N - 1), (N - 1)]) in the x and y directions, respectively. Ks ionary is the element of ^stationary assocjatecj wim an average covariance over all pixel pairs in the ROI that are separated by p pixels in the x direction and q pixels in the y direction.

Alternatively, the covariance element Kswnary describes a correlation between any pixel in the ROI and its neighbor p pixels to the right and q pixels up. The element ^stationary ^ ^ ^ center 0f mg matnx. T e two-dimensional integer indices m and n both run from 0 to N - 1 over the two-dimensional ROI in the x and y directions, respectively. The one-dimensional index i is simply a one-dimensional version of the indices m and n that runs over every pixel in the ROI (G [0, N - 1]) and is equal to m + nN. The average over the elements of the full covariance matrix only includes elements for which a (p, q) offset ROI pixel exists, denoted by the set S. Therefore, the number of samples that contribute to the calculation of the pqth element K^s twnary 0f the stationary covariance matrix varies with the exact p, q indices and is given by (N - \p\)(N - \q\). Thus, the stationary covariance elements describing correlations with more distant pixels pairs (i.e., large p and q values) have fewer samples. Both the full and stationary covariance matrices are symmetric, so that Kij = Kji and K ationary = K tionary .

In calculation of the stationary covariance matrix for a 3x3 ROI (N = 3), m, n are integers from 0 to 2, i is an integer ranging from 0 to 8, and p and q range from -2 to 2. For the (p, q) = (1, 2) element of the stationary covariance matrix, the average is computed only for index values i = 0 or 1 because other i index values correspond to ROI pixels that do not have counterparts that are 1 pixel to the right and 2 pixels up. The resultant stationary covariance matrix element is equal to stationary = R^ + R^ md [& based Qn twQ samples similarly, the (p, q) = (1,-1) element of the stationary covariance matrix is equal to

KS tio„ary = R^ + R^ + R^ + R^ and is based on four samples.

For each phantom or patient data set, a largest square ROI was selected by hand which still contained only breast tissue. ROI size varied with breast size for patient data (between 35x35 and 150x150 voxels), but was constant for phantom data (70x70 voxels). This ROI was applied to a set of MR image slices (between 26 and 92 slices per patient depending on breast size and 61 slices per phantom) of the left breast where enough breast tissue was present to fill the chosen ROI. Multiple slices were included in the stationary covariance matrix calculation by concatenating all ROIs from all slices into a single g vector. This provides a single, in-plane, stationary covariance matrix for each patient or phantom, which can be referred to as a patient- or phantom-specific stationary covariance matrix. These matrices can be normalized by average voxel variance (the central pixel of the stationary covariance matrix) to highlight the relative correlation fall-off with spatial position.

An overall stationary covariance matrix for the entire patient or phantom population was then calculated by first converting all patient- specific stationary covariance matrices to the same spatial scale (0.625 mm/voxel) using a cubic convolution interpolation with an interpolation parameter of -0.5 if necessary. See Park and Schowengerdt, "Image reconstruction by parametric cubic convolution," Comput. Vision Graph. 23:258-272 (1983). The difference between the original and interpolated covariance matrices was negligible. Finally, all the resultant matrices were averaged. Error bars on the overall stationary covariance matrix were estimated by calculating the standard deviation of the patient-specific stationary covariance matrices values at each offset position.

To understand whether the difference among patient- specific stationary covariance matrices was due to instrumentation error or anatomical variations, simulated ROIs were created with only Rician noise. In all cases, the noise variance was set to one since the final covariance is normalized by its maximum value, which is equivalent to the average pixel variance. Simulated ROIs were chosen to have a conservative size of 35x35 voxels, which is equal to the smallest ROI used for covariance calculations on the patient data. Sets of between 5 and 95 simulated ROIs were created to bracket the range of the number of slices selected in the patient data. The root mean squared (RMS) variation in the simulated, Rician-noise only, patient-specific stationary covariance estimate was then calculated for 5 different offsets 4-20 mm) by averaging over all offsets whose absolute values

Figure imgf000017_0001
were within 4 mm of the specified offset. The RMS variation was calculated as a function of the number of ROIs used in the covariance estimation (corresponding to the number of patient slices). This process was repeated on eleven independent realizations to improve the RMS estimates and compare with the variation in the patient-specific stationary covariance matrices. Coded patient data were taken from the National Cancer Institute's (NCI) Clinical Genetics Branch's Breast Imaging Study data archive in which high-risk patients were imaged using various MR imaging protocols and scanner types. Patients were included if they were between 25 and 56 years of age when the images were produced and considered at high genetic risk of developing breast cancer.

Seventy-seven patients with MRI data collected on a 1.5T Philips machine with a 7-channel, dedicated breast coil and with the same pre-contrast imaging sequence were selected for additional analysis. Thirteen were excluded because of the presence of an implant or diagnosis of breast cancer before or during the course of the NCI breast imaging study. Sixty-four patients remained for the final analysis. Pre-contrast, TVweighted, gradient-echo, fat-suppressed images were available for each patient with an in-plane resolution ranging from 0.586 to 0.664 mm and a slice thickness ranging from 1.9 to 2.3 mm. Twenty phantoms were fabricated and imaged using the same scanner type, breast coil, and imaging sequence for comparison with patient data.

FIGS. 5A-5J are example patient and phantom images acquired with clinical systems. TV and T2-weighted images of the phantom are shown, acquired using standard clinical breast protocols at two different institutions and with two different clinical scanners (1.5 T Philips and 1.5 T General Electric scanners with dedicated breast coils). FIG. 5A is Ti-weighted, fat suppressed patient image, FIG. 5B is a IV weighted patient image without fat suppression, FIG. 5C is a T2 spectral adiabatic recovery (SPAIR) fat suppressed patient image. FIG. 5D is a TVweighted, fat suppressed phantom image. FIG. 5E is a Ti-weighted phantom image without fat suppression. FIG. 5F is a Ti-weighted phantom image without fat suppression, FIG. 5G is a Ti-weighted, fat suppressed phantom image. FIG. 5H is a short Ύι inversion recovery (STIR) phantom image with fat suppression. FIG. 51 is a T2-weighted phantom image without fat suppression. FIG. 5J is a T2-weighted, fat suppressed phantom image. All are shown with a 10 mm scale bar. As is apparent from these images, the disclosed phantoms mimic breast shapes and internal tissue structures that are much improved over currently available phantoms.

FIG. 6 includes representative images of regions of interest based on phantoms and patients. A top row contains patient image data and a bottom row contains fat suppressed Ti-weighted phantom data. A visual comparison between the patient and phantom ROI images indicates that the phantom has a randomly appearing structure that resembles the complicated patient data image structure. Patient data appears to have a directional preference in the anterior-posterior direction, whereas phantom data is more isotropic. Furthermore, fat suppression in the phantom images appears to be slightly improved as compared to that of patient data. This may be due to the fact that there is no torso attached to the phantom, resulting in improved shimming of the phantom, or that the spectral shape of the fat signature in the phantom may be less complicated than that of patients. There may also be more homogeneity of the material types within a voxel in the phantom than in patient data, resulting in the appearance of better fat suppression.

FIGS. 7A-8B are provided for comparison of covariance matrices associated with patient and phantom data. FIGS. 7A-7B contains images of the patient and phantom overall stationary covariance matrices, respectively. FIGS. 8A-8B are graphs of stationary covariance as a function of position in anterior-posterior and right-left directions. As shown in FIGS. 8A-8B, a covariance length of the patient and phantom images is similar along the anterior-posterior direction. In a right-left direction, the images differ by about two standard error bars, with the phantom image data having a larger correlation length than the patient data. In general, the phantom data is more isotropic than that of the patient, which tends to have structures that elongate along the anterior-posterior direction. Interestingly, the error bars of the patient and phantom data are similar, indicating a similar range of variability among the two populations.

FIG. 9 shows the results of simulations run to estimate the amount of instrumentation error included in the error bars in FIGS. 8A-8B. RMS variations in the stationary covariance matrix estimates are plotted as a function of the number of ROIs used in the estimate. The ROIs were simulated images created with only Rician noise. Results for five different offset distances are shown. For 26 to 92 ROIs, the magnitude of the RMS instrumentation error in the stationary covariance ranges from 0.003 to 0.020 depending on the offset distance and the number of ROIs. The corresponding error bars in FIGS. 8A-8B range from 0.053 to 0.096 for the same set of offset distances. Because the RMS instrumentation errors are less than the error bars in FIG. 8A-8B, the error bars represent mostly anatomical variations.

Simulated Lesions

A phantom that mimics the distribution of adipose and fibro glandular tissue in the human breast and has similar T\ and T2 relaxation times can also be configured to include features that mimic lesions that are indicative of disease. Referring to FIG. 10, a simulated lesion 1000 is situated in a phantom container 1004 and includes a lesion chamber 1006 that is fluidically coupled to one or more tubes 1008, 1009 that permit filling and/or evacuation of the lesion chamber 1006. The chamber 1006 is shown as having an elliptical cross section, but lesion chambers can be formed of spheres, cylinders, ellipsoids, ovoids, cubes, or other regular or irregular shapes defined by curved, planar, or other surfaces. The lesion chamber 1006 also includes lobulations 1010, 1012 that define additional volumes that are coupled to a lesion chamber volume.

Lesion molds can be conveniently fabricated using stereolithography, and lesion chamber wall thickness can be as little as at least 0.60 mm, or smaller, as subject to fabrication limitations of stereolithography or other fabrication techniques. Inlet/outlet tubes can be situated to produce a preferred time-varying distribution of contrast agent in the lesion chambers. Benign lesions tend to have more regular borders than malignant lesions, so that chamber volumes with lobulations tend to produce images more similar to actual malignancies.

FIGS. 1 lA-1 IB illustrate alternative geometries for lesion chambers. FIG. 11A illustrates a spherical chamber 1102 to which flow tubes 1104, 1106 are secured to provide fluid flow into and out of the chamber 1102. These tubes are aligned on a common axis 1105 so that the lesion chamber 1102 can be referred to as a collinear configuration. FIG. 1 IB illustrates a spherical chamber 1112 to which flow tubes 1114, 1116, 1118 are coupled. Tubes 1114, 1118 are at an angle of 15 degrees with respect to the tube 1116 which is aligned along a radius of the chamber 1112. These tubes are aligned along intersecting axes, and this configuration can be referred to as an intersecting configuration. In the examples of FIGS. 1 lA-1 IB, tube inner diameter is 2 mm, and the lesion chambers have inner diameters of about 10 mm.

FIGS. 12A-12B are photographs of additional lesion surrogates that include spherical lesion chambers. The example of FIG. 12A also includes additional lobulations that extend outwardly from the spherical lesion chamber. Fat- suppressed, Ti-weighted, gradient echo magnetic resonance images produced with simulated lesions such as those shown in FIGS. 12A-12B are shown in FIGS. 13A- 13B, respectively. These images were obtained with the lesion chambers filled with an MR contrast agent such as gadolinium doped water. A convenient contrast agent solution comprises 40% glycerol by volume, 60% deionized water by volume, and 150 m Gd-DTPA. In other examples, a GdCl3 solutions can be used at, for example, a concentration of 4.5 mM. With any flowing contrast agent solution, the effects of flow and solution microenvironments on MRI signal intensity and MR relaxation times may be used to produce suitable phantom images. Contrast agent

concentration can be selected as desired, and concentration can be function of time as shown below.

Dynamic Lesions

As shown in FIG. 14, a surrogate lesion 1402 is situated in a container 1404 that holds a mixture 1406 of an adipose mimic and a fibroglandular mimic as described above. Inlet tubes 1408, 1410 are coupled to one or more fluid pumps 1412 and an outlet tube 1414 is coupled to a waste container 1416. A tissue mimicking fluid 1418 and a contrast agent 1420 are coupled to the fluid pumps 1412, and these fluids are directed to the surrogate lesion 1402 by pump outlets 1430, 1432, respectively. The fluids are combined in a bifurcated tube 1434 and the mixture directed to the surrogate lesion 1402 through one or both of the tubes 1408, 1410. A controller 1436 can increase, decrease, or otherwise regulate flow rates associated with the tissue mimicking fluid 1418 and the contrast agent 1420 so as to produce a time varying contrast agent concentration that is directed to the surrogate lesion 1402. Total flow rate can be held constant.

A convenient tissue-mimicking fluid consists of Ni-DTPA, deionized water, and glycerol. Ni-DTPA can be used to match the Ti value of the fluid to that of the glandular-mimicking component of the phantom and can be produced following a method described in Tofts et al., "Ni-DTPA doped agarose gel - a phantom material for Gd-DTPA enhancement measurements," MRI 11:125-133 (1993). Measured Ti values of different concentrations of Ni-DTPA in water can be linearly interpolated to find the concentration that corresponds to a Ti value of a glandular-mimicking component of a phantom (Ti = 1192+49 ms). The resulting concentration of Ni- DTPA was 5.3 mM. Glycerol was diluted in water in a ratio of 40:60 as specified by the manufacturer to provide lubrication for the moving parts in the fluid pump. The contrast agent solution can consist of the tissue mimicking fluid with additional Gd- DTPA added. Two different lesions surrogates were used to model time varying image data. One surrogate was a 10 mm spherical chamber with 2 mm inside diameter, collinear inlet/outlet tubes. A second surrogate had a 10 mm spherical chamber with three 2 mm inside diameter tubes, separated by 15 degrees. These configurations are illustrated in FIGS. 1 lA-1 IB.

Simulation results are shown in FIGS. 15A-15B for a first plane containing an input flow axis and a second plane orthogonal to the first plane, respectively. (These planes are noted as Plane A and Plane B is FIGS. 1 lA-1 IB.) Total flow rate for both designs was 1.0 ml/s and initially the surrogate lesions were assumed to be water filled. At time t = 0, the inlet concentration of contrast agent solution was instantaneously increased to 100%. Referring to FIGS. 15A-15B, it is apparent that the contrast solution distributes more evenly in the intersecting design than the collinear design. By configuring inlet outlet ports, differing time varying distributions of contrast agent can be obtained. FIGS. 16A-16B illustrate the production of washout curves corresponding to benign and malignant lesions in humans, respectively. The curves correspond to design contrast agent

concentrations and data points correspond to measured concentrations based on x- ray data. Anthropomorphic X-Ray Phantoms

A suitable phantom can be produced by combing an adipose-mimic and a fibroglandular mimicking material. Materials and construction can be similar to those described above. Because mammography is performed with a compressed breast, an alternative container to that of FIG. 1 can be used that provides a compressed shape of thickness of about 4.5 cm as shown in FIG. 18. Other construction details can be the same as or similar to those described above. In one example, a phantom was produced with glandular tissue comprising about 29.5% by volume of the phantom. This volumetric density is similar to that of asymptomatic women with high genetic risk factors in one study. X-ray images are based on differences in x-ray attenuation coefficients. For egg whites, typical mass extinction coefficients are 0.426/cm at 27 keV and 0.215 at 60 keV. For lard, typical mass extinction coefficients are 0.410 g/cm at 27 keV and 0.207 g/cm at 60 keV. Phantom tissue structure can be compared with patient data using stationary covariance matrices as described above. Imaging with or without an anti-scatter grid can alter the stationary covariance as well. Stationary covariance matrices were calculated on left and right craniocaudal (CC) patient and phantom mammography images. In order to combine the results for both right and left CC images, the left CC images were flipped about their vertical axis so that the chest wall was on the same side of the image. For patient images, the largest square ROI in the constant thickness region of the breast was selected using the procedure described in Burgess, "Mammographic structure: data preparation and spatial statistics analysis," Proc. SPIE 3661:0277-786X (1999). For the phantom images, the known geometry of the phantom jar was used to select the largest square ROI in the constant thickness region of the phantom. To create an overall stationary covariance matrix for the entire patient or phantom population, the individual stationary covariance matrices were normalized by their average pixel variance and averaged together. Error bars on the overall stationary covariance matrix were estimated by calculating the standard deviation of the patient-or phantom- specific stationary covariance matrix values at each offset position.

For x-ray imaging, mimicking component composition can be selected based on elemental composition of a selected tissue. The table below lists composition data based on published data (Hammerstein et al. "Absorbed radiation dose in mammography," Radiology 130:485-491 (1979) and Poletti et al., "X-ray scattering from human breast tissues and breast-equivalent materials," Phys. Med. Biol. 47:47- 63 (2002)). Values for lard and egg whites are also included for comparison. As is apparent from the table, the only appreciable deficiency in the lard/egg white combination is with respect to nitrogen concentration in lard.

Element Lard Breast Adipose Egg Whites Breast Glandular

Hammerstein Poletti et Hammerstein Poletti (1979) al. [137] (1979) (2002)

C 76.05 61.9 76.5 + 1.1 5.59 18.4 18.4 + 0.9

(49.1-69.1) (10.8-30.5)

H 12.25 11.2 12.4 + 0.1 10.69 10.2 9.3 + 0.5

O 11.69 25.1 10.7 + 1.3 81.64 67.7 67.9 + 2.0

(18.9-35.7) (55.2-75.9)

N 0.007 1.7 0.4 + 0.05 1.56 3.2 4.4 + 0.6 FIGS. 18-26B illustrate a mammography phantom and related images and other data. FIG. 18 illustrates results of a segmentation algorithm on a central slice of the phantom for inclusion in x-ray scatter simulations. For each voxel, the fraction of material that is air, the jar, lard, and egg whites is indicated. FIG. 19A is a photograph of compressed phantom, and FIGS. 19B-19C are a example x-ray images of a phantom and a patient, respectively. FIGS. 20A-20B are comparisons of x-ray mass attenuation coefficients for breast adipose tissue and adipose- mimicking phantom material, and for breast glandular tissue and glandular- mimicking phantom material, respectively. A horizontal line represents equal breast tissue and phantom values. Data and theoretical values are based on Hammerstein et al. "Absorbed radiation dose in mammography," Radiology 130:485-491 (1979), Johns and Yaffe,"X-ray characterization of normal and neoplastic breast tissues," Phys. Med. Biol. 32:675-695 (1987), Al-Bahri and Spyrou, "Photon linear attenuation coefficients and water content of normal and pathological breast tissues," Appl. Radiat. Isot. 47:777-784 (1996), Poletti et al., "X-ray scattering from human breast tissues and breast-equivalent materials," Phys. Med. Biol. 47:47-63 (2002), Tomal et al., "Experimental determination of linear attenuation coefficient of normal, benign and malignant breast tissues," Radiation Measurements 45: 1055- 1059 (2010), and Rao and Gregg. "Attenuation of monoenergetic gamma rays in tissues," Am. J. Roent. 123:631-637 (1975).

FIG. 21 contains images of example patient and phantom ROIs. Upper images are patient images, lower images are phantom images. All ROIs represent areas 3.5 cm by 3.5 cm in object space.

FIGS. 22A-22B depict overall stationary covariance matrices for patient and phantom data sets. The matrices are scaled to have the same intensity at their peak. The phantom and patient overall stationary covariance matrices have similar structure sizes. The patient data set has larger long- scale correlations in the anterior- posterior direction than the phantom data set. FIGS. 23A-23B are graphs illustrating patient and phantom overall stationary covariance matrices (as shown in FIGS. 22A- 22B) in the anterior-posterior and superior-inferior directions. The patient and phantom overall stationary covariance matrices are the same to within their error bars. In the anterior-posterior direction, the patient data appears to have longer correlations on average than the patient data. Error bars are the standard deviation of the individual patient (N=80) and phantom-specific (N=40) stationary covariance matrices at each distance.

FIGS. 24A-24B illustrate the influence of scatter on stationary covariance matrix. Stationary covariance matrices were calculated from simulated images of a single phantom with scatter and with no scatter. Sections of the resultant stationary covariance matrices are shown in the anterior-posterior and superior-inferior directions. Removal of scatter decreases correlations by 15% on average and as much as 33% in the anterior-posterior direction and 6% on average and as much as 10% in the superior- inferior direction. Since the clinical data was acquired with an anti-scatter grid, but the experimental data was not, this effect may account for some of the differences in stationary covariance matrices between the phantom and patient data.

FIG. 25 illustrates Monte Carlo simulations to estimate the amount of scatter produced by the phantom. FIG. 25 includes simulated images of (a) phantom with primary and all scattered x-rays, (b) phantom with primary x-rays only, (c) phantom with scattered x-rays only, (d) SPR of phantom, (e) SPR of a homogeneous version of the phantom. For (e), all voxels of the phantom were converted in the simulation to 30% egg by volume to create a homogeneous phantom.

FIGS 26A-26B illustrate validation of the Monte Carlo simulated scatter results for the heterogeneous phantom. Experimental measurements were performed of SPR at five different locations in the phantom using tungsten discs and also simulated with the same geometry. FIG. 26A is an image of a phantom with discs labeled 1-5 in place. FIG. 26B is a comparison of the estimated scatter-to-primary ratio for each disc location. Error bars are two standard errors.

Thus, the disclosed phantom can be used for a variety of inter-comparisons based on MR or x-ray imaging. Phantom composition can be readily varied, and distribution of material within the phantoms can be varied with a simple mixing procedure. Additional internal structure can be provided in phantoms and surrogate lesions to provide additional details for imaging, or to produce various washout curves. For example, mold surfaces for either phantoms or lesions can have rough or shaped surfaces that include one or more additional protrusions or depressions in a regular, periodic, or random arrangement, or selected to mimic a particular specimen feature. Phantoms or surrogate lesions can also be provided with porous or other materials. For example, a phantom can include plastic beads or sponges so as to provide additional structure in a phantom or surrogate lesion, as well as to adjust washout curves. Washout curves produced with surrogate lesions can also be adjusted by selecting suitable inlet/outlet configurations. In addition, a single phantom can serve as both an x-ray and MR phantom. Thus, the above disclosure includes several representative examples that illustrate some aspects of the disclosed technology. These examples are not to be taken as limiting the scope of the disclosure. In addition, the disclosed phantom's suitability for MR and x-ray imaging is described, but such phantoms can also be used with PET and single photo positron electron computed tomography (SPECT), or other types of imaging. We claim all that is encompassed by the appended claims.

Claims

We claim:
1. An imaging phantom, comprising:
an adipose tissue mimicking component; and
a fibro glandular tissue mimicking component distributed in the adipose tissue mimicking component, wherein the components are distributed so as to produce images corresponding to patient images.
2. The phantom of claim 1, further comprising a container having a shape that mimics the body part, wherein the combined components are retained in the container.
3. The phantom of claim 2, wherein the container has a shape that mimics the compressed body part.
4. The phantom of claim 2, wherein the body part is human breast.
5. The phantom of claim 4, wherein the adipose and fibroglandular tissue mimicking components are lard and egg whites, respectively.
6. The phantom of claim 4, wherein at least one of the fibroglandular tissue mimicking component and the adipose tissue mimicking component have at least one magnetic resonance relaxation time that is approximately the same as that of the corresponding tissue.
7. The phantom of claim 4, wherein the images are x-ray images.
8. The phantom of claim 1, wherein the fibroglandular tissue mimicking component is distributed in the adipose component such that a normalized stationary covariance has a full width at half maximum of less than about 3 mm, 5 mm, or
10 mm.
9. The phantom of claim 1, further comprising a lesion chamber situated within the combined adipose and fibroglandular tissue mimicking components, the lesion chamber defining a lesion volume.
10. The phantom of claim 9, further comprising a plurality of tubes fluidically coupled to the lesion volume.
11. The phantom of claim 9, wherein the lesion chamber contains a magnetic resonance contrast agent.
12. The phantom of claim 9, wherein the lesion chamber includes a plurality of lobulations.
13. A method of making a phantom, comprising:
providing an at least partially liquefied adipose tissue mimicking component; distributing a fibroglandular tissue mimicking component within the at least partially liquefied adipose tissue mimicking component; and
cooling the combined adipose tissue mimicking component and the fibroglandular tissue mimicking component so as to solidify the combination.
14. The method of claim 13, wherein the adipose and fibroglandular tissue mimicking components are lard and egg whites, respectively, and further comprising shaping the combination so as to mimic a body part.
15. The method of claim 14, wherein the body part is a breast, and the combination is shaped so as to be at least partially spherical.
16. The method of claim 15, further comprising providing a lesion chamber within the combination.
17. The method of claim 16, wherein the lesion chamber is provided prior to cooling the combination.
18. A method, comprising:
providing a breast phantom that includes a lesion chamber having at least one fluid inlet;
fluidically coupling a contrast agent to the lesion chamber; and
controlling the contrast agent so that contrast agent concentration in the lesion chamber is time varying so as to correspond to a human washout curve.
19. The method of claim 18, further comprising controlling the contrast agent by varying a mixture of a contrast agent containing fluid with a tissue mimicking fluid.
20. The method of claim 19, wherein the contrast agent includes gadolinium.
21. The method of claim 18, further comprising producing a time varying image associated with the contrast agent concentration.
22. The method of claim 21, wherein the image is an x-ray image or a magnetic resonance image.
23. The method of claim 21, wherein both x-ray and magnetic resonance images are produced.
24. A method, comprising:
providing an imaging phantom having a structure corresponding to that of a patient tissue; and
obtaining first and second images of the phantom based on magnetic resonance and x-ray attenuation, respectively.
25. The method of claim 24, wherein the patient tissue is a human breast, and the phantom is shaped to correspond to a breast compressed for mammography.
PCT/US2011/053053 2010-09-23 2011-09-23 Anthropomorphic, x-ray and dynamic contrast-enhanced magnetic resonance imaging phantom for quantitative evaluation of breast imaging techniques WO2012040611A1 (en)

Priority Applications (4)

Application Number Priority Date Filing Date Title
US38592910P true 2010-09-23 2010-09-23
US61/385,929 2010-09-23
US201061424495P true 2010-12-17 2010-12-17
US61/424,495 2010-12-17

Publications (1)

Publication Number Publication Date
WO2012040611A1 true WO2012040611A1 (en) 2012-03-29

Family

ID=44789607

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/US2011/053053 WO2012040611A1 (en) 2010-09-23 2011-09-23 Anthropomorphic, x-ray and dynamic contrast-enhanced magnetic resonance imaging phantom for quantitative evaluation of breast imaging techniques

Country Status (1)

Country Link
WO (1) WO2012040611A1 (en)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2014173979A (en) * 2013-03-08 2014-09-22 Nihon Medi Physics Co Ltd Heart model and manufacturing method thereof
WO2015001372A1 (en) * 2013-07-03 2015-01-08 General Electric Company Method of contrast enhanced breast imaging, and contrast agent reference insert
WO2015165252A1 (en) * 2014-04-30 2015-11-05 苏州瑞派宁科技有限公司 Imaging subject mould manufacturing method, and individualized imaging method
WO2019034791A1 (en) * 2017-08-17 2019-02-21 Dublin Institute Of Technology Tissue mimicking materials

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
GB2148013A (en) * 1983-10-12 1985-05-22 Yokogawa Hokushin Electric Nuclear magnetic resonance imaging
US6675035B1 (en) * 1997-06-20 2004-01-06 Imaging Diagnostic Systems, Inc. Phantom for optical and magnetic resonance imaging quality control
EP1413251A1 (en) * 2002-10-22 2004-04-28 Iso-Science Laboratories, Inc. Multimodal imaging marker
WO2008151202A2 (en) * 2007-06-03 2008-12-11 The Regents Of The University Of California Elastic deformable heart and torso phantom for nuclear imaging of heart with realistic modes of cardiac and respiratory motion

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
GB2148013A (en) * 1983-10-12 1985-05-22 Yokogawa Hokushin Electric Nuclear magnetic resonance imaging
US6675035B1 (en) * 1997-06-20 2004-01-06 Imaging Diagnostic Systems, Inc. Phantom for optical and magnetic resonance imaging quality control
EP1413251A1 (en) * 2002-10-22 2004-04-28 Iso-Science Laboratories, Inc. Multimodal imaging marker
WO2008151202A2 (en) * 2007-06-03 2008-12-11 The Regents Of The University Of California Elastic deformable heart and torso phantom for nuclear imaging of heart with realistic modes of cardiac and respiratory motion

Non-Patent Citations (19)

* Cited by examiner, † Cited by third party
Title
"USDA National Nutrient Database for Standard Reference, Release 22", 2009, AGRICULTURAL RESEARCH SERVICE
AL-BAHRI, SPYROU: "Photon linear attenuation coefficients and water content of normal and pathological breast tissues", APPL. RADIAT. ISOT., vol. 47, 1996, pages 777 - 784, XP004050767, DOI: doi:10.1016/0969-8043(96)00066-8
B.D. FLICKINGER, N. MATSUO: "Baileys Industrial Oil and Fat Products: Volume 3, ''Edible oil and fat products: specialty oils and oil products", vol. 3, 2005, JOHN WILEY & SONS, INC., article "Diacylglycerols", pages: 38
BURGESS: "Mammographic structure: data preparation and spatial statistics analysis", PROC. SPIE, vol. 3661, 1999, pages 0277 - 786X, XP055066485, DOI: doi:10.1117/12.348620
BURLEY, VADEHRA: "The avian egg: chemistry and biology", 1989, JOHN WILEY & SONS, INC., pages: 72
CAMPBELL, LAND: "Dielectric properties of female human breast tissue measured in vitro at 3.2 GHz", PHYS. MED. BIOL., vol. 37, 1992, pages 193 - 210, XP000248714, DOI: doi:10.1088/0031-9155/37/1/014
E.G. YOUNG: "Comprehensive Biochemistry: Proteins (Part 1", 1963, ELSEVIER PUBLISHING COMPANY, article "Occurrence, classification, preparation and analysis of proteins", pages: 28
GRAHAM ET AL.: "Magnetic resonance properties of ex vivo breast tissue at 1.5 T", MAGN. RESON. MED., vol. 38, 1997, pages 669 - 677, XP055051190, DOI: doi:10.1002/mrm.1910380422
HAMMCRSTCIN: "Absorbed radiation dose in mammography", RADIOLOGY, vol. 130, 1979, pages 485 - 491, XP010529138
HAMMERSTEIN ET AL.: "Absorbed radiation dose in mammography", RADIOLOGY, vol. 130, 1979, pages 485 - 491, XP010529138
JOHNS, YAFFE: "X-ray characterization of normal and neoplastic breast tissues", PHYS. MED. BIOL., vol. 32, 1987, pages 675 - 695
KEELY ET AL.: "The spatial and temporal expression of the a2j31 integrin and its ligands, collagen I, collagen IV, and laminin, suggest important roles in mouse mammary morphogenesis", DIFFERENTIATION, vol. 59, 1995, pages 1 - 13, XP026792107
LINEY ET AL.: "A simple and realistic tissue equivalent breast phantom for MRI", J. MAGN. RESON. IMAGING, vol. 10, 1999, pages 968 - 971, XP002664418
MAZZARA ET AL.: "Use of a modified polysaccharide gel in developing a realistic breast phantom for MRI", MAGN. RESON. IMAGING, vol. 14, 1996, pages 639 - 648
PARK, SCHOWENGERDT: "Image reconstruction by parametric cubic convolution", COMPUT. VISION GRAPH., vol. 23, 1983, pages 258 - 272, XP001377496
POLETTI ET AL.: "X-ray scattering from human breast tissues and breast-equivalent materials", PHYS. MED. BIOL., vol. 47, 2002, pages 47 - 63, XP008060355, DOI: doi:10.1088/0031-9155/47/1/304
RAO, GREGG: "Attenuation of monoenergetic gamma rays in tissues", AM. J. ROENT., vol. 123, 1975, pages 631 - 637
TOFTS ET AL.: "Ni-DTPA doped agarose gel - a phantom material for Gd-DTPA enhancement measurements", MRI, vol. 11, 1993, pages 125 - 133, XP026440349
TOMAL ET AL.: "Experimental determination of linear attenuation coefficient of normal, benign and malignant breast tissues", RADIATION MEASUREMENTS, vol. 45, 2010, pages 1055 - 1059, XP027322828

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2014173979A (en) * 2013-03-08 2014-09-22 Nihon Medi Physics Co Ltd Heart model and manufacturing method thereof
WO2015001372A1 (en) * 2013-07-03 2015-01-08 General Electric Company Method of contrast enhanced breast imaging, and contrast agent reference insert
WO2015165252A1 (en) * 2014-04-30 2015-11-05 苏州瑞派宁科技有限公司 Imaging subject mould manufacturing method, and individualized imaging method
US9931097B2 (en) 2014-04-30 2018-04-03 Raycan Technology Co., Ltd. (Su Zhou) Imaging subject mould manufacturing method, and individualized imaging method
WO2019034791A1 (en) * 2017-08-17 2019-02-21 Dublin Institute Of Technology Tissue mimicking materials

Similar Documents

Publication Publication Date Title
van Rijswijk et al. Diffusion‐weighted MRI in the characterization of soft‐tissue tumors
Müller et al. Morphometric analysis of human bone biopsies: a quantitative structural comparison of histological sections and micro-computed tomography
Ntziachristos et al. MRI-guided diffuse optical spectroscopy of malignant and benign breast lesions
Säbel et al. Recent developments in breast imaging
Donath et al. Toward clinical X-ray phase-contrast CT: demonstration of enhanced soft-tissue contrast in human specimen
Wu et al. A comparison of reconstruction algorithms for breast tomosynthesis
US7424142B2 (en) Display and analysis of medical images using calibrated pixel values in units of known properties of reference materials
US5312755A (en) Contrast resolution tissue mimicking phantoms for nuclear magnetic resonance imaging with stable NMR properties
Liu et al. Morphology enabled dipole inversion for quantitative susceptibility mapping using structural consistency between the magnitude image and the susceptibility map
EP1987523B1 (en) Process for imaging
Kalra et al. Strategies for CT radiation dose optimization
Morsbach et al. Reduction of metal artifacts from hip prostheses on CT images of the pelvis: value of iterative reconstructions
Fuchs et al. A standardized boundary element method volume conductor model
Vilanova et al. Diffusion-weighted whole-body MR screening
Poplack et al. Electromagnetic breast imaging: average tissue property values in women with negative clinical findings
Gutzeit et al. Comparison of diffusion-weighted whole body MRI and skeletal scintigraphy for the detection of bone metastases in patients with prostate or breast carcinoma
Yabuuchi et al. Phyllodes tumor of the breast: correlation between MR findings and histologic grade
Park et al. Breast tomosynthesis: present considerations and future applications
Plewes et al. Visualization and quantification of breast cancer biomechanical properties with magnetic resonance elastography
Samani et al. An inverse problem solution for measuring the elastic modulus of intact ex vivo breast tissue tumours
Manduca et al. Spatio-temporal directional filtering for improved inversion of MR elastography images
Li et al. Methodology for generating a 3D computerized breast phantom from empirical data
Klatt et al. In vivo determination of hepatic stiffness using steady-state free precession magnetic resonance elastography
Miéville et al. Iterative reconstruction methods in two different MDCT scanners: physical metrics and 4-alternative forced-choice detectability experiments–a phantom approach
Gilhuijs et al. Computerized analysis of breast lesions in three dimensions using dynamic magnetic‐resonance imaging

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 11768218

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase in:

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 11768218

Country of ref document: EP

Kind code of ref document: A1