Most applications of x-ray microtomography in medicine have been to the study of bones and teeth. Of these, the majority have been devoted to investigations aimed at characterization of trabecular structure in normal and osteoporotic bone. Such studies are hindered by partial volume effects, both at the level of the trabecular boundaries, and also on a finer scale, from cells and cell processes embedded within the bone tissue. The effect of the latter on the determination of mineral concentration has been modelled by computer simulation. Primarily as a result of these partial volume effects, most studies of trabecular bone have been based on binary images produced by thresholding. If the resolution of laboratory systems were improved, changes in mineral concentration within trabeculae could also be assessed. Although most microtomography uses detector arrays, useful results can be obtained with first generation systems. These are illustrated by quantitative mineral concentration studies of bones and teeth; determination of concentrations of elements with accessible absorption edges derived from absorption spectra recorded at all points in the projections; and demonstration of diffusion of KI into cortical bone. A very exciting use of 3D microtomography is to observe changes in bone structure after specific treatments. Others have used this for in vivo studies, but here is illustrated by preliminary observation of in situ cracking of cortical bone under mechanical load.

Use of synchrotron generated x-ray for micro-CT is particularly powerful for several reasons. These include the high x-ray flux which permits short duration exposure and hence scan durations, the narrow bandwidth of the x-ray energy which permits quantitative CT imaging with high accuracy of the measured attenuation coefficient and the fact that the x-ray photon energy can be adjusted allows element selective imaging. Another advantage is that the radiation is close to parallel so that the tomographic image reconstruction process is facilitated. On the other hand, synchrotron-based micro-CT imaging does have the limitation of a rather small field of view being illuminated. This means that specimens larger than the field of view also create problems for the conventional 'global' tomographic image reconstruction algorithms. Fortunately, recently developed 'local' reconstruction algorithms can, in large measure, overcome this limitation of the synchrotron generated x-ray field.

A monochromatic CT for imaging the human head and neck is being developed at the National Synchrotron Light Source. We compared the performance of this system, multiple energy computed tomography (MECT), with that of a conventional CT (CCT) using phantoms. The advantage in image contrast of MECT, with its beam energy tuned just above the K-edge of contrast element, over CCT carried out at 120 kVp, was approximately equal to 3.2-fold for iodine and approximately equal to 2.2 fold for gadolinium. Image noise was compared by simulations because this comparison requires matching the spatial resolutions of the two systems. Simulations at a 3- rad dose and 3-mm slice height on an 18-cm-diameter acrylic phantom, with MECT operating at 60.5 keV, showed that image noise for MECT was 1.4 HU vs. 1.8 HU for CCT. Simulations in the dual-energy quantitative CT mode showed a two-fold advantage for MECT in image noise, as well as its superior quantification. MECT operated in the planar mode revealed fatty tissue in the body of a rat using xenon K-edge subtraction. Our initial pan for clinical application of the system is to image the composition of carotid artery plaques non-invasively, separating the plaques' main constituents: the fatty, fibrous, and calcified tissues.

The appearance of cancellous bone architecture is different for various skeletal sites and various disease states. In the iliac crest it is more plate-like, whereas in the spine rods dominate. During aging and disease plates are perforated and connecting rods are dissolved. There is a continuous shift from one structural type to the other. So traditional histomorphometric procedures, which are based on a fixed model type, will lead to questionable results. 3D microtomography allows to assess model independent structural parameters so that trabecular thickness, for example, can be determined directly. Not only mean thicknesses are available but also thickness histograms which are helpful to identify pathological states. Other features such as trabecular separation, degree of anisotropy and structural type index can be extracted from the 3D images and allow to characterize cancellous bone and its changes due to aging, disease and treatment. To fully exploit the significance of bone structure on bone strength large scale finite element (FE) analyses are performed. Hence microtomography, performed with a sufficiently fine isotropic spatial resolution, reveals information on the structural features of cancellous bone which were not available so far and which will, hopefully, lead to a better understanding of the pathogenesis of bone diseases and subsequently to improved treatment regimes.

A synchrotron radiation computed microtomography system allowing high resolution 3D imaging of bone samples has been developed at ESRF. The system uses a high resolution 2D detector based on a CCd camera coupled to a fluorescent screen through light optics. The spatial resolution of the device is particularly well adapted to the imaging of bone structure. In view of studying growth, vertebra samples of fetus with differential gestational ages were imaged. The first results show that fetus vertebra is quite different from adult bone both in terms of density and organization.

Thin slice spiral computed tomography techniques open up new avenues in assessing human bone architecture in vivo. A slice images do not yield accurate and reproducible results for structural parameters. The use of standard 2D histomorphometric measures is problematic because even a slice thickness of 1 mm is large compared to a trabecular thickness of less than 200 micrometers . In this study we investigated the use of topological parameters an in particular their dependence on spatial resolution. The topology was derived from a 3D skeleton of a (mu) CT dataset of a real trabecular bovine bone sample. The 3D thinning algorithm from which the skeleton was computed will be descried in detail. Decreasing the spatial resolution of the (mu) CT resulted in the following percent changes: number of branches -46 percent; number of noes -50 percent; avg. branch length +28 percent. These results indicate that it is not possible to determine topological parameters accurately in thin slice CT images. However, diagnostically relevant changes over time may still be quantifiable. In order to better analyze these problems we developed realistic spongiosa models using the rapid prototyping technique of stereolithography. Plastic models of a real trabecular bone network were built. So far these spongiosa models are slightly enlarged compared to the original data. Stereolithographic models of artificial geometries showed that a spatial resolution of 100micrometers and variations of +/- 50 micrometers are technically achievable. (mu) CT imaging of the stereolithographic spongiosa models revealed an excellent agreement between the model and the original dataset. Although the absorption characteristics of plastic and bone are different the CT contrast of the stereolithographic model imaged in air is comparable to the contrast of bone imaged in a marrow matrix. Thus for clinical purposes the plastic models can serve as a standard of trabecular bone which can, for example, be used to compare 2D and 3D structural analysis methods, the impact of spatial resolution, and the influence of segmentation techniques.

Synchrotron microtomography enables us to visualize trabecular bone microarchitecture in 3D with a spatial resolution of a few microns. With recent developments in finite element modeling, it is possible to incorporate these images voxel by voxel into a finite-element model of the elastic properties of the bone. The ability to calculate mechanical behavior from 3D images will help us to understand the relationships between bone structure and properties.

The mechanical behavior of trabecular bone depends on the internal bone structure as well as the load applied. Mechanical stresses and strains influence the modeling process and subsequently the structure and strength of the bone. Although the basic concepts of adaptive bone remodeling are generally accepted, the mathematical laws relating bone remodeling to the stress/strain relations are still under investigation. The aim of this project was to develop an algorithm which allows simulation of the response of the trabecular bone is age-related bone loss and to determine the biomechanical consequences of such a response based on realistic 3D models of the trabecular microstructure. Today, such models can be generated directly using micro-computed tomography ((mu) CT). For the purpose of the study, a compact fan-beam type tomograph was used, also referred to as desktop (mu) CT, providing a nominal isotropic resolution of 14 micrometers . Two groups of seven trabecular bone specimens were measured including specimens from pre- menopausal and post-menopausal women respectively. In order to control bone loss over age, a novel algorithm to simulate bone resorption and adaptive process was developed. The algorithm, also referred to as simulated bone atrophy, generates a set of microstructural models, iteratively derived from the original 3D structure. Simulated bone atrophy was used to 'age-match' the first and the second group incorporating an underlying realistic time-frame for the simulation. Using quantitative bone morphometry and 3D animation tools, the changes in bone density and bone architecture could be monitored in the progress of age- related bone loss over a total observation time of 28 years. The structures at the end-point of the simulations were then compared qualitatively and quantitatively to the structures of the post-menopausal group directly assessed by (mu) CT. The results suggest the possibility of transforming 'normal' to osteopenic' bone on a microstructural level resulting in realistic bone models similar in appearance and structural properties when compared to the post-menopausal group. In the future, the assessment of the biomechanical competence of microstructural bone in the progress of adaptive bone remodeling might result in an improved prospective prediction of individual bone strength as an indicator for fracture risk in patients predisposed to osteoporosis.

The Exxon microtomography facility at the National Synchrotron Light Source (NSLS) located at Brookhaven National Laboratory has constantly evolved since its first operation in 1986. This evolution has been driven by improvement in imaging technology and by research applications particularly the study of fluids in man-made structures and random porous media. The present facility provides tunable x-ray energies between 6 and 35 keV and acquires 1024 cubes of 14 bit x-ray image data with a resolution of a few microns in about one hour. Image data is reconstructed using direct Fourier inversion. We present a brief overview of the system configuration and how the application of differential absorption contrast microtomography across the iodine K edge to the characterization of multi-phase fluids in sandstones. We also present a simple method for computing the nuclear magnetic resonance T₂ relaxation spectrum directly from the CT data. Recently the system has been upgraded to radiographically image at high sped multi-phase fluid flow in small channels such as in petroleum reservoir rocks and in microelectromechancial systems. Expansion of the high speed capability to tomographic imaging of specimens in 3D in near real time is currently in progress. The upgrades improve both image acquisition and reconstruction speed by employing multi-port CCD and parallel processing technologies. We will describe the design capabilities of the high speed system and will show results from the high speed imaging of fluid flow in small tubes.

To maximize the power of synchrotron radiation as an analytic tool, much is dependent on the performance of the detector. A review is presented of state of the art integrating and photon counting detectors using CCDs, currently the subject of much activity and driven by recent developments in very large area arrays. The synergies with other scientific applications are strong. In x-ray astronomy, a number of new missions will be launched this decade which will utilize the latest developments in CCD detector technology for the focal plane sensor. These will provide simultaneous imaging and spectroscopy of astronomical sources, carried out with unprecedented sensitivity. Reviewed will be the key operating parameters of the CCD, such as the energy resolution, detection efficiency and speed together with major aspects of its practical realization in imaging applications.

At BAM 3D-computerized tomography (3D-CT) using x-ray cone beam and area detectors is established as a standard method for materials testing and development. Up to now main applications concerned fiber reinformed plastics and ceramics, density distribution in ceramics, powder metallurgical parts and archaeological objects. Spatial and density resolution depends on the object and on the combination x-ray source - detector system. The maximum spatial resolution is 5 micrometers using a transmission target and 12 micrometers using a standard micro focus tube together with an image intensifier as detector. The main problem of image intensifiers applied to 3D-CT is the rather bad contrast ratio of about 20:1. An object dependent correction for the light scattering in the image intensifier in combination with bam hardening correction is performed at BAM. This contribution will point out the advantages and disadvantages of different detector systems and results will be shown on test samples and selected investigation from our ongoing work.

The principle and experimental l realization of x-ray phase- contrast in compute assisted microtomography ((mu) CT) at the micrometer resolution level is described. The camera used is a modification of a setup previously developed by us for attenuation-contrast (mu) CT using synchrotron x-rays. Phase detection is accomplished by employing the x-ray interferometer. By using x-ray phase contrast it is possible to image structural details in low-z biological tissues much better than with absorption contrast. The advantage of phase over attenuation contrast is not limited to light element or to low x-ray energies. Examples of applying phase contrast (mu) CT to the structural investigation of rat trigeminal nerve are given.

We have shown so far that 3D structures in biological sot tissues such as cancer can be revealed by phase-contrast x- ray computed tomography using an x-ray interferometer. As a next step, we aim at applications of this technique to in vivo observation, including radiographic applications. For this purpose, the size of view field is desired to be more than a few centimeters. Therefore, a larger x-ray interferometer should be used with x-rays of higher energy. We have evaluated the optimal x-ray energy from an aspect of does as a function of sample size. Moreover, desired spatial resolution to an image sensor is discussed as functions of x-ray energy and sample size, basing on a requirement in the analysis of interference fringes.

A new nondestructive method for structural imaging is proposed. The method is based on direct measurements of phase and amplitude changes in a 2D x-ray image. A standing wave is formed between two separated crystals allowing high- resolution imaging of the complex refractive index. A comprehensive analysis of the amplitude-phase contrast is possible because of the precisely controlled variation of the phases between the reference and reflected beams from a crystalline mirror.

3D computed tomographic images with micrometer resolution were made in phase-contrast mode with high energy x-rays at a third generation synchrotron source. The phase-contrast technique enables one to obtain information not only about the amplitude of the wave field behind the object and thus about the absorption, but also about the refractive index distribution inside the sample. Increasing the x-ray energy from the soft x-ray region up to 10-60 keV simplifies the experimental setup and opens the possibility to study organic samples at room-temperature and under normal pressure conditions. The projection data is recorded with a fast, high-resolution x-ray camera consisting of a 5 micrometers thin YAG scintillator crystal, a visible light microscope optics and a slow scan 1k X 1k CCD camera. The spatial resolution of phase-contrast microtomography is currently limited by the resolution of the x-ray detector to about 1-2 micrometers . First applications in biology and geophysics are shown.

Hard x-ray tomographic images with high spatial resolution were obtained at he European Synchrotron Radiation Facility in Grenoble. They clearly display damage initiation within materials submitted to monotonic tensile test. Strain- induced cracks with opening of less than 1 micrometer are detected, as are slight differences in composition between constituent phases. These observations are based on the Fresnel diffraction fringes associated with local jumps in the optical phase of x-rays transmitted through the sample. The high lateral coherence of third generation synchrotron radiation sources is essential for the application of this instrumentally simple technique.

Fluorescent x-ray computed tomography (FXCT) is being developed to detect non-radioactive contrast materials in living specimens. The FXCT systems consists of a silicon channel cut monochromator, an x-ray slit and a collimator for detection, a scanning table for the target organ and an x-ray detector for fluorescent x-ray and transmission x-ray. To reduce Compton scattering overlapped on the K(alpha) line, incident monochromatic x-ray was set at 37 keV. At 37 keV Monte Carlo simulation showed almost complete separation between Compton scattering and the K(alpha) line. Actual experiments revealed small contamination of Compton scattering on the K(alpha) line. A clear FXCT image of a phantom was obtained. Using this system the minimal detectable dose of iodine was 30 ng in a volume of 1 mm³, and a linear relationship was demonstrated between photon counts of fluorescent x-rays and the concentration of iodine contrast material. The use of high incident x-ray energy allows an increase in the signal to noise ratio by reducing the Compton scattering on the K(alpha) line.

We present a novel approach to perform Rayleigh-to-Comptom (RC) measurements using a bent analyzer crystal in Laue geometry. The ratio of the elastically to the inelastically scattered photons gives directly access to the effective atomic number of the same by eliminating several side effects falsifying the scattered spectrum, eg. absorption in the sample. The use of the analyzer crystal, instead of an energy dispersive detector, enables us to measure in forward scattering geometry, where the method is optimized in respect to the contrast-to-noise ratio for low Z materials. The first generation computed tomography approach enables us to reconstruct the RC-value independently from the scattering geometry on a square image grid. We give this method the acronym RC-CT. The identity of values in the RC method and the RC-CT method will be derived. he conical bent crystal is focused on the source line where the incident monochromatic beam traverses the sample. The analyzer crystal reflects the elastic line of the scatter spectrum into a scintillation detector. The inelastic part of the spectrum passes the crystal and is recorded in a second scintillation detector. By rocking the analyzer, the whole energy distribution of the scatter spectrum can be obtained with an energy resolution dominated by the rocking curve width of the bent crystal. We will discuss the requirements and the constraints on the geometrical parameter of the experiment. From this, it is concluded that the method requires high flux sources like synchrotron. Employing the contrast-to-noise ratio introduced by Harding et al, the wavelength dispersive approach to the RC method is evaluated for a wide range of elements. Finally, we show a first reconstruction of a bone sample and discus s possible applications in medical and material science.

Demands on image quality in general and on low-contrast resolution in particular ware very high for medical uses of x-ray computed tomography (CT). Therefore in conventional single-slice CT, perfect planar geometry was strictly adhered to where consistent data over 360 degrees can be acquired. This implied scanning the object slice by slice. Fast volume imaging had been a request in medical imaging for a long time, but results of various approaches like cone-beam CT remained unsatisfactory. Spiral CT, the first non-planar scan mode which gained general acceptance, was first introduced in 1989, 3D volume imaging of 3 to 100 cm body sections within 10 to 60 s by spiral CT is now routine in medical imaging. We will review the spiral scan principle and the different approaches to image reconstruction which are presently implemented. In particular, the type of z- interpolation algorithm can be used to include noise, z-axis resolution and artifact behavior. Aspects of image quality are discussed in detail. There are only subtle differences between conventional single slice CT and spiral volume scanning. Thereby the old paradigm that high quality CT scanning demands perfect planar geometry has become obsolete. The medical use of spiral CT is meanwhile undisputed; but a review of typical applications indicates then need for further improvements in the speed of volume data acquisition, in particular the need for more efficient use of the available x-ray power. This shall lead to new approaches like the combined use of spiral CT and multi-row detectors and cone angle CT based on area detectors. We review respective efforts and their implications for potential industrial applications.

In x-ray computed tomography (CT) spiral/helical scanning is achieved by continuous and simultaneous source rotation, object translation and data acquisition. In medical imaging, fan-beam spiral CT has replaced conventional incremental CT. Also, cone-beam spiral CT is feasible and advantageous is important applications. In this paper, resolution characteristics, protocol optimization, and image restoration are discussed in fan-beam spiral CT. Then, generalized Feldkamp cone-beam reconstruction is described is spiral geometry. Based on Snyder's iterative deblurring theory, a unified iterative algorithm is reported to handle incomplete cone-beam data. Finally, several future possibilities in spiral CT are mentioned.

The term microtomography generally refers to x-ray computed tomography with a resolution of better than 100 microns. Because of the range of different specimens that can be studied, it is important that resolution, x-ray energy and exposure are matched to the composition and size of the specimen. Where control over specimen size is possible, it should be made as small as possible since there is a third order relationship between size and required exposure for a given resolution and contrast. In first generation microtomography systems the specimen is stepped though a single collimated x-ray beam and the attenuation recorded by a single detector. The ability of such systems to record both number and energy of transmitted photons facilitates quantitative evaluation of noise and errors in the detection system and their propagation through to the reconstructed image. Most microtomography scanners are similar in principle to third generation medical scanners, except that the specimen, rather than the source/detector system, rotates. Images from such scanners are subject to ring artifacts because of slight differences in response of the detector elements. Fourth generation medical scanners overcome this problem by using fixed ring of detectors with only the source rotating, but it is impractical to build a microtomography scanner using the same principle. We have designed a fourth generation microtomography scanner which uses a CCD camera operated in time delay integration mode. This overcome the problem of ring artifacts and allows specimens larger than the CCD imaging area to be scanned.

The use of soft x-ray nanotomography techniques for the evaluation and failure mode analysis of microchips was investigated. Realistic numerical simulations of the imaging process were performed and a specialized approach to image reconstruction from limited projection data was devised. Prior knowledge of the structure and its component materials was used to eliminate artifacts in the reconstructed images so that defects and deviations from the original design could be visualized. Simulated data sets wee generated with a total of 21 projections over three different angular ranges: -50 to +50, -80 to +80 and -90 to +90 degrees. In addition, a low level of illumination was assumed. It was shown that sub-micron defects within one cell of a microchip could be imaged in 3D using such an approach.

In forestry and tree sciences computerized tomography allows the quantitative determination of the locally varying absorption coefficients for penetrating radiation within a thin slice of the trunk. The tomogram shows not only hollows, rot, knots and other defects but also the distribution of water in the invisible interior of the stem. Portable systems have been developed and built for computerized tomography of standing trees in forests and parks. They use the radio nuclide Cesium-137 as source of radiation. The MCT-3 is based on the translation-rotation- method. A bearing ring carries the shielded source of 13 GBq of Cs-137 and three scintillation detectors. The MCT-F is based on the fan-beam method and has 30 detectors. It has an inner diameter of 100 cm and a stronger source of 185 GBq. Equipment was used in forestry sciences and in tree-care to obtain information about decay, checks, heartwood formation and moisture content, for the detection of interior decay by fungi and its spread in a horizontal and vertical direction, for determining sapwood area dependent on fertilization, for evaluating development and treatment of tree wounds and for studying the influence of resin tapping on the water supply of pines.

Coherently scattered x-rays are mainly confined to a forward peaked cone, which exhibits, due to their coherence, structural information of the atomic arrangement in the sample. Coherent scattering in amorphous materials, which are of random short range order, therefore results in board diffraction ring patter, whereas crystalline substance show more confined diffraction rings or even Brag spots. X-ray diffraction computed tomography (XRDCT) reconstructs the intensities diffracted from extended objects on a square image grid and thus retrieves the local structure. A short survey is presented about what information can be extracted from diffraction experiments. Hereby a new method is proposed to use the Rietveld refinement for quantitative XRDCT. Also the possible use of XRDCT to reconstruct the spatial distribution of preferred orientation axis is suggested. An imaging system for XRDCT, consisting of a medical image intensifier tube and CCD readout system, is presented, which includes a modified beam stop for recording the intensity of the transmitted beam. Depending on the application this imaging system cam work in first generation or second generation tomography mode. Furthermore a new approach for the reconstruction of the differential coherent cross-section is proposed. It includes an absorption correction based on weighted sinograms. The introduced reconstruction strategy is elucidated by experimental result from a simple phantom. The measured data also validate the simulation program, written to study more complex phantoms under different experimental conditions. Finally possible applications in medical and material science are discussed. A design for a mammography setup using x-ray diffraction is presented.

A high spatial-resolution transmission x-ray tomographic microscope (XTM) has been developed as a 3D visualization technique at the Advanced Proton Source at Argonne National Laboratory. Using brilliant synchrotron radiation, the XTM can collect a complete set of 2D data for a 3D reconstruction within a fairly short time. The 2D transmission data were taken with a cryogenically cooled CCD camera coupled with an optical microscope with 32x magnification and a single cadmium tungstate crystal screen with a thickness of approximately 60 micrometers . For experiments performed with 20 keV x-ray energy on gold specimens, the XTM produced 2D transmission data with a resolutions of approximately 1 micrometers . A series of 90 projections taken over a range of 180 degree projection angle was taken within a half hour. The reconstruction, with support of a filtered back-projection algorithm, demonstrates the ability of the XTM to provide fast 3D structural information about a specimen from 2D projection data with approximately 1 micrometers resolution and more than 15 percent local image contrast.

An x-ray tomography system is being developed for high resolution inspection of large objects. The goal is to achieve 25 micron resolution over object sizes that are tens of centimeters in extent. Typical objects will be metal in composition and therefore high energy, few MeV x-rays will be required. A proof-of-principle system with a limited field of view has been developed. Preliminary results are presented.