| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Technical Developments |
1 From the Departments of Radiology (E.D.P., R.E.J., D.B.W.) and Pathology (M.V.I., C.A.L.) and the UNC-Lineberger Comprehensive Cancer Center (E.D.P.), University of North Carolina School of Medicine, Manning Dr, Chapel Hill, NC 27599-7510; the Department of Physics and Center for Synchrotron Radiation Research and Instrumentation, Illinois Institute of Technology, Chicago (D.C.); the Nuffield Department of Pathology and Bacteriology, University of Oxford, United Kingdom (J.G.); the Department of Physics, North Carolina State University, Raleigh (D.E.S., M.Z.K.); the Brookhaven National Laboratory, Upton, NY (Z.Z.); and the European Synchrotron Radiation Facility, Grenoble, France (W.C.T.). Received February 1, 1999; revision requested March 31; final revision received June 25; accepted July 20. Supported in part by U.S. Army grant DAMD 17-96-1-6143, by the National Synchrotron Light Source at Brookhaven National Laboratory under U.S. Department of Energy contract DE-AC 02-76CH00016, and by the State of Illinois Higher Education Cooperative Agreement. Address reprint requests to E.D.P. (e-mail: etpisano@med.unc.edu).
 |
Abstract |
|---|
 TOP Abstract IntroductionMaterials and MethodsResultsDiscussionAppendixReferences |
|---|
Index terms: Breast radiography, technology, 00.119 • Synchrotron
|
Introduction |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionAppendixReferences |
|---|
Other researchers have applied diffractive optics to imaging problems (4–7) and have observed refraction effects. There is also interest in phase-contrast imaging that makes use of the high transverse coherence of third-generation synchrotron sources. However, these types of measurements are limited to use with materially thin objects or high x-ray imaging energies to obtain phase-contrast images of the object (8,9). The diffraction-enhanced imaging technique is successful with samples of the same thickness as the human breast and, after further development, should be possible without the use of a synchrotron.
For this project, the mammographic technique under development makes use of the high intensity and collimation of synchrotron radiation to create a monoenergetic line scanning system with very little scatter (10). Synchrotron radiation is the electromagnetic radiation emitted as charged particles (electrons or positrons) change direction while passing through electromagnets. The magnets restrain the particles to a circular orbit in large accelerators called "synchrotrons," although most of the modern machines are really storage rings in which the electric currents persist for many hours at constant energy. The particles are highly relativistic, traveling at or near the speed of light, so that the radiation produced has several unique properties. The intensity of the radiation is five to six orders of magnitude higher than that of a laboratory source. The energy spectrum is smooth and continuous from the infrared to the hard x-ray range. The radiation is inherently collimated in the plane of the orbit. The radiation is emitted from the machine through metal windows and delivered to experimental stations through long vacuum vessels. These tubes are called beam lines. The unique radiation properties, when coupled with modern perfect crystal x-ray optics, are used to define very high intensity beams that are highly monochromatic at any selected energy and are highly collimated for medical and other applications.
In addition, an analyzer crystal was used in this study as a scatter rejection optic.
Herein, we describe our first results with imaging of human breast cancer specimens. We have previously reported our results with phantoms (10).
|
Materials and Methods |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionAppendixReferences |
|---|
|
The detector was a photostimulable phosphor image plate typically used for radiologic examinations (models HR5 [high spatial resolution] and ST5 [standard spatial resolution]; Fuji Medical Systems USA, Stamford, Conn). The image recorded on the plate was digitized, stored, and displayed with a reader and workstation (model AC3; Fuji Medical Systems) or with a reader system (model BAS2000; Fuji Medical Systems). The image plates were read with a 2,560 x 2,048 matrix, which resulted in an image with 100 µm per pixel (0.1 x 0.1 mm).
The diffraction angle of the analyzer crystal could be finely tuned by using a stepper-motor driven translation stage to push on a long bar attached to an axle to which the crystal was attached (tangent arm). The spatial resolution limit of the tangent arm was 0.1 µradian, which was sufficient for placing the Bragg analyzer crystal at a selected position on its rocking curve.
For each sample, a "normal" radiograph with the monochromatic beam could be obtained by moving the image plate to a location just downstream of the object on the sample scanning stage and imaging the combined object and image plate through the fan beam. Diffraction-enhanced images were then acquired with the analyzer crystal tuned to various positions on the rocking curve by translating the sample and the image plate in opposite directions at the same speed through the fan beam. The change in scanning direction arises from the beam inversion from the analyzer crystal. At a scanning speed of about 10 mm/sec, the surface dose on the sample was a few milligray at 18 keV and 10ths of a milligray at 30 keV. Rocking curves through a specific location within the specimen were obtained by fixing the specimen in the fan beam and performing a series of exposures by incrementally changing the position of the analyzer crystal and the vertical position of the image plate.
Seven formalin-fixed human breast cancer specimens were imaged, three infiltrative lobular and four infiltrating ductal carcinomas. These specimens were approximately 1 cm thick. Each biologic sample was sealed in a plastic bag and compressed between two Lucite plates so that the absorbing thickness was approximately 40 mm. Images were obtained at 18 keV by using the reflection of silicon 333 crystals as monochromator and analyzer. The crystals were in symmetric Bragg mode, meaning that the beam made the same angle with the crystal surface. For each sample, five images were obtained at angular positions of 0 (peak position), ±W/2, and ±W on the rocking curve, where W is the full-width at half maximum point on the rocking curve.
Images of the same specimens were also obtained with use of a digital mammographic unit (SenoScan; Fischer Medical Imaging, Boulder, Colo).
Diffraction-enhanced and digital images of the specimens were subsequently evaluated by one experienced breast imager (E.D.P.). She identified regions of interest within the diffraction-enhanced images that showed apparently increased lesion information, for example, areas of increased surface spiculation. This was done by comparing the digital radiograph and the diffraction-enhanced image side by side and circling areas on the latter. No regions of interest were identified on the standard digital radiographs.
Histologic whole-mount slides of the specimens were made. With the assistance of the radiologist and direct comparison of the whole-mount slides to the diffraction-enhanced images of the specimen, an experienced breast pathologist (J.G.) evaluated the regions of interest in the specimens to determine whether the information apparent on the diffraction-enhanced images correlated with real histopathologic structures.
|
Results |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionAppendixReferences |
|---|
The figures show the improved depiction of spiculations corresponding with tumor extension (Figs 2–4) and with fibrosis (Fig 5).
|
|
|
|
|
|
|
|
|
|
|
|
Discussion |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionAppendixReferences |
|---|
Estimates of the flux from conventional x-ray sources have yielded scanning time estimates of approximately 1,000–10,000 seconds. Clearly, such long scanning times to deliver a diffraction-enhanced image set would be unacceptable. These times are based on commercially available x-ray sources and indicate that the application of diffraction-enhanced imaging with conventional sources will be challenging. However, there are reasons to believe that this will be possible.
First, conventional x-ray imaging makes use of an x-ray beam that arises directly from the target of the x-ray tube. The spectrum may be filtered. In the implementation of diffraction-enhanced imaging, a monochromator must be used to collimate the imaging beam. This monochromating element allows the operational parameters of the tube to be optimized for the creation of the flux at the desired imaging energy. This eliminates concern for other parts of the spectrum that would normally deliver unnecessary dose, for example, the creation of bremsstrahlung radiation. Use of high accelerating voltages becomes attractive since the flux contained in the emission from the source is enhanced at higher accelerating voltages.
Second, the prospect of obtaining diagnostic information from the diffraction image leads to higher optimal imaging energies. Since the tissue absorption is reduced at higher energy, the transmission and thus flux requirements are reduced. However, two competing effects limit the imaging energy. First, the refraction sensitivity is reduced; second, the flux diffracted by the monochromator is reduced. Estimates and measurements indicate that the optimal imaging energy is in the range of 30 keV. An additional benefit of this energy is reduced dose delivery (15 times less than that at 18 keV) and possible reduction in breast compression.
Finally, the advent of digital detectors for mammography may allow more efficient use of exposure. A contributing element may also be the lack of scatter in the acquired image, which may allow a reduction in exposure in the acquisition of useful diagnostic information. Perhaps this method, paired with more efficient detectors, will allow use of higher energy beams and lower patient dose.
The combination of these four factors may allow construction of a nonsynchrotron, conventional source. However, determination of the optimal parameters for a conventional system will take time. The synchrotron, with its wide flexibility in the trial and modeling of source parameters, will play a major role in this endeavor.
These results suggest that there is potential for improved visualization of breast cancer detail with use of the diffraction component of the x-ray beam. This improved detail was achieved without optimization of this technique for breast imaging. To obtain the maximum amount of information on diffraction-enhanced images for breast cancer detection and characterization, careful evaluation of image detail with histologic correlation by experts should be carried out. In this manner, the optimal crystal type, reflection type, imaging energy, and rocking curve location can be determined so that the specifications for a clinically useful (office-based) system can be developed.
Given the increased visibility of lesion features with diffraction-enhanced imaging in all three cases of infiltrating lobular carcinoma that we evaluated, we believe this method might improve detectability of this frequently occult lesion. Infiltrating lobular carcinoma is the second most common type of invasive breast cancer, occurring in approximately 15% of women with infiltrative tumors. In fact, lobular tumors are bilateral in 10% of cases. Frequently, the extent of this type of tumor is underestimated with traditional imaging methods. Perhaps diffraction-enhanced imaging can improve the detection of this lesion by increasing the visibility of subtle spiculations and architectural distortions.
This study is limited in that we have yet to evaluate any specimens that predominantly contain clustered calcifications. We are currently imaging breast samples that contain this type of mammographic lesion. In addition, this study is essentially qualitative, and no images of normal control tissue or benign lesions were evaluated. A clinically realistic analysis of the potential benefits of this technique will require the imaging of a wider set of sample types and ultimately of patients.
In summary, we believe that these preliminary results suggest promise for diffraction-enhanced imaging as applied to breast cancer detection and diagnosis. Further studies are needed to determine the optimal diffraction-enhanced imaging parameters for these clinical tasks.
|
Appendix |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionAppendixReferences |
|---|
X-ray diffraction from perfect crystals, with narrow reflection angular width (on the order of a few microradians) and peak reflectivity of close to unity, provides the tools necessary to prepare and analyze x-ray beams that traverse an object on the microradian scale (12). Such crystals, typically silicon, are routinely used in the semiconductor industry to make integrated circuits and electronic devices. The purity and perfection of these crystals have allowed many advances in x-ray diffraction techniques, in particular at x-ray synchrotron sources.
The condition for x-ray diffraction from a crystal is met only when the incident beam makes the correct angle to the atomic lattice planes in the crystal for a given x-ray energy or wavelength. When this condition is met, the beam diffracts from the planes over a narrow range of incident angles, so-called Bragg diffraction. As the crystal is rotated about an axis parallel to the lattice planes and perpendicular to the incident beam direction, an intensity variation is observed. This is referred to as the "rocking curve." The shape of this curve is roughly triangular, with the peak reflectivity approaching nearly 100%.
In diffraction-enhanced imaging, an imaging beam is prepared by means of diffraction of the polychromatic beam from the synchrotron to create a nearly monoenergetic imaging beam. This beam is then passed through the object being imaged as in conventional radiography. However, a matching crystal is placed between the object and the detector. This crystal is set at or near the peak of Bragg diffraction and is called the analyzer crystal. A schematic representation of a synchrotron radiography and diffraction-enhanced imaging system is shown in Figure 1.
Since the condition for diffraction from this crystal limits the x rays that can be diffracted into the detector, it automatically provides a high degree of scatter rejection, which results in improved image contrast. The range of angles accepted by the analyzer crystal is a few microradians, which thus provides scatter rejection more completely than does the capabilities of conventional antiscatter techniques such as slit collimation and grids.
The rejected scatter falls into a category referred to as "small-angle scattering" (13), or scattering that arises from diffraction from very small, organized structures. This scattering intensity, which would normally appear on the image, is missing and appears as does absorption on the image. We call the improved image contrast obtained by means of scatter rejection of this sort "extinction contrast," which is drawn from a similar term used in optics and x-ray diffraction to describe intensity loss due to diffraction and scattering. Therefore in diffraction-enhanced imaging, the image that represents the absorption of the object by x rays is referred to as the "apparent absorption image," since it has contrast derived from both absorption and scatter rejection, or extinction.
The shape of the rocking curve of the analyzer crystal introduces a sensitivity to refraction that occurs within the object when the analyzer crystal is detuned from the peak position. Density, thickness, or material variations in an object will refract the x rays as they cross through the material. These small angular variations are generally in a range smaller than a microradian. The steep sides of the reflectivity curve convert these subtle angle variations into intensity variations, which thus makes refraction effects visible on an image. In this study, we found that by acquiring an image pair with the analyzer crystal set to diffract on each side of the rocking curve, we can separate refraction effects from combined absorption and extinction effects (3,14).
Thus, the diffraction-enhanced imaging technique introduces two sources of image contrast to radiography: refraction and extinction (3). Each of these image contrast sources may be further developed to apply to medical and biologic imaging.
Glossary of Terms
Analyzer crystal.—A crystal that is placed in the x-ray beam that emerges from the target material and is used to reflect the x-ray beam at the Bragg angle onto the detector.
Bragg angle.—The angle 
at which an electromagnetic beam is incident to and reflected from a crystal plane. The relationship between the wavelength, 
, the distance between the crystal planes, d, and the angle of reflection, , is given with the Bragg equation: = 2d sin .
Diffraction.—A change in direction of an x ray when it is incident on a substance that has a periodic structure at the molecular level.
Extinction.—A term used in x-ray optics to describe the loss of intensity from an x-ray beam owing to the scattering of photons out of the beam at very small angles, primarily because of diffraction effects.
Reflection.—The changing of direction of an electromagnetic beam when it is incident on the surface of a plane. The reflection angle is equal but opposite in direction to the incident angle.
Refraction.—A change in direction of the electromagnetic beam when it passes obliquely across a boundary between two materials in which the velocity of propagation is different. This can also occur in a single medium owing to variations in its physical properties.
Rocking curve.—The intensity of the x-ray beam that is reflected from the analyzer crystal as the analyzer angle is varied above and below the Bragg angle that corresponds to the incident beam. The intensity of the reflected beam varies correspondingly from zero to a maximum (at the Bragg angle) and to zero again.
Synchrotron radiation.—Electromagnetic radiation emitted as the path of a charged particle (an electron or positron) is bent as it passes through a magnetic field.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
|
References |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionAppendixReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  J. Keyrilainen, M. Fernandez, M.-L. Karjalainen-Lindsberg, P. Virkkunen, M. Leidenius, K. von Smitten, P. Sipila, S. Fiedler, H. Suhonen, P. Suortti, et al. Toward High-Contrast Breast CT at Low Radiation Dose Radiology, October 1, 2008; 249(1): 321 - 327. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Muehleman, D. R. Sumner, and Z. Zhong Refraction Effects of Diffraction-Enhanced Radiographic Imaging: A New Look at Bone J Am Podiatr Med Assoc, September 1, 2004; 94(5): 453 - 455. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R A Lewis, C J Hall, A P Hufton, S Evans, R H Menk, F Arfelli, L Rigon, G Tromba, D R Dance, I O Ellis, et al. X-ray refraction effects: application to the imaging of biological tissues Br. J. Radiol., May 1, 2003; 76(905): 301 - 308. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
