3 Jun 2004 20:53 AR AR220-BE06-08.tex AR220-BE06-08.

sgm LaTeX2e(2002/01/18) P1: IKH

10.1146/annurev.bioeng.6.040803.140130

Annu. Rev. Biomed. Eng. 2004. 6:185–208

doi: 10.1146/annurev.bioeng.6.040803.140130
Copyright

c 2004 by Annual Reviews. All rights reserved

MICRO-COMPUTED TOMOGRAPHY—CURRENT
STATUS AND DEVELOPMENTS
Erik L. Ritman
Department of Physiology and Biomedical Engineering, Mayo Clinic College
of Medicine, Rochester, Minnesota 55905; email: elran@mayo.edu
Annu. Rev. Biomed. Eng. 2004.6:185-208. Downloaded from www.annualreviews.org
Access provided by University of Winnipeg on 01/23/19. For personal use only.

Key Words X-ray, synchrotron, contrast agents, three-dimensional, small animal

■ Abstract The recent rapid increase in interest in tomographic imaging of small
animals and of human (and large animal) organ biopsies is driven largely by drug
discovery, cancer detection/monitoring, phenotype identification and/or characteriza-
tion, and development of disease detection methods and monitoring efficacies of drugs
in disease treatment. In biomedical applications, micro-computed tomography (CT)
scanners can function as scaled-down (i.e., mini) clinical CT scanners that provide a
three-dimensional (3-D) image of most, if not the entire, torso of a mouse at image
resolution (50–100 µm) scaled proportional to that of a human CT image. Micro-CT
scanners, on the other hand, image specimens the size of intact rodent organs at spatial
resolutions from cellular (20 µm) down to subcellular dimensions (e.g., 1 µm) and fill
the resolution-hiatus between microscope imaging, which resolves individual cells in
thin sections of tissue, and mini-CT imaging of intact volumes.

CONTENTS
INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186
BASIC COMPONENTS OF MICRO-COMPUTED TOMOGRAPHY . . . . . . . . . . . . 190
X-Ray Source . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191
X-Ray to Light Conversion Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191
The Light to Electron Coupling/Conversion System . . . . . . . . . . . . . . . . . . . . . . . . 191
CAPABILITIES OF MINI- AND MICRO-CT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192
Resolution Requirements: Mini-CT Scanner . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192
Micro-CT Scanner . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192
Physics and Technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
Software-Based Extension of Micro-CT Capability . . . . . . . . . . . . . . . . . . . . . . . . . 195
Extension of Micro-CT Capabilities by Non-Scanner Methods . . . . . . . . . . . . . . . . 195
Radiopaque Indicators of Physiological Spaces and Processes . . . . . . . . . . . . . . . . 196
X-Ray Micro-CT as a Complementary Methodology . . . . . . . . . . . . . . . . . . . . . . . . 197
Radionuclide-Based Imaging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197
Histological Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198
Molecular Analysis Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199

1523-9829/04/0815-0185$14.00 185
 3 Jun 2004 20:53 AR AR220-BE06-08.tex AR220-BE06-08.sgm LaTeX2e(2002/01/18) P1: IKH

186 RITMAN

Data Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199

ONGOING DEVELOPMENTS OF X-RAY MICRO-CT . . . . . . . . . . . . . . . . . . . . . . 200
K-Edge Subtraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 200
X-Ray Phase Delay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201
X-Ray Scatter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201
X-Ray Fluorescence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201
Very High Resolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202
RADIATION DOSE LIMITATIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202
SUMMARY AND CONCLUSIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202
Annu. Rev. Biomed. Eng. 2004.6:185-208. Downloaded from www.annualreviews.org

INTRODUCTION
Access provided by University of Winnipeg on 01/23/19. For personal use only.

The rapid increase in interest in tomographic imaging of small animals and organ
biopsies is conveyed in recent publications (1), the appearance of dedicated spe-
cialty meetings, the NIH call for a number of Request for Applications (RFA)-type
grant proposals for developing small-animal imaging capabilities, and the growing
number of newly emerging companies that build such specialty imaging scanners.
Examples of the characteristics of some commercially available micro-computed
tomography (CT) scanners are summarized in Table 1 and a representative scanner
is illustrated in Figure 1. Note that these data are not exhaustive and their opera-
tional capabilities are also selected from the range of capabilities of each scanner.
This selection of scanners and their characteristics will likely be improved in the
next five years, but the physics and technological trade-offs that are involved in
the design of a small-animal CT scanner are expected to be common to future
systems.
The first clinically practical CT method was introduced in the early 1970s. The
state of the art in clinical helical CT scanners now involves volume imaging with
voxels that are of the order of 1 mm3 over an axial height of several centimeters that
is scanned within a half second. Although these scanners are also of some use for
small-animal imaging, special purpose, high-resolution CT scanners are necessary
for many studies of intact small animals and reasonably large (>1 cm3) tissue
specimens (2). A number of in-depth reviews of the technicalities and applications
of micro-CT have been published since the late 1980s (e.g., 3–8).
In the early 1980s, the first micro-CT scanners were developed (9–14) us-
ing bench-top X-ray CT sources. With the introduction of synchrotron radiation
sources (15) in a number of countries, material scientists used the high intensity
and brightness of these X-ray sources to do CT at micrometer resolutions. Sev-
eral landmark events occurred that laid the groundwork for the extension of the
method to biological specimens. Grodzins (16) presented the value of adjusting
the X-ray photon energy (E) to achieve an optimal, energy-dependent attenuation
[µ(E)] by the specimen [i.e., the X-ray pathlength or approximately the specimen
diameter (D)] such that µ(E) · D = 2, which maximizes CT image signal-to-noise
ratio. As illustrated in Figure 2, X-ray photons less than 25 keV interact mainly
via the photoelectric effect, and their attenuation varies approximately as E−3,
 Annu. Rev. Biomed. Eng. 2004.6:185-208. Downloaded from www.annualreviews.org
Access provided by University of Winnipeg on 01/23/19. For personal use only.
3 Jun 2004
20:53
AR

TABLE 1 Summary of selected, manufacturer-supplied, performance characteristics of several commercially available micro-CT
scanners. (Courtesy of data provided by G. Wang, University of Iowa.)
AR220-BE06-08.tex
AR220-BE06-08.sgm
LaTeX2e(2002/01/18)

MICRO-CT
P1: IKH

187
 3 Jun 2004 20:53 AR AR220-BE06-08.tex AR220-BE06-08.sgm LaTeX2e(2002/01/18) P1: IKH

188 RITMAN
Annu. Rev. Biomed. Eng. 2004.6:185-208. Downloaded from www.annualreviews.org
Access provided by University of Winnipeg on 01/23/19. For personal use only.

Figure 1 Photograph of a commercially available micro-CT scanner with opera-

tor and mouse subject. This particular scanner is designed for in vitro scanning in
preclinical and pharmaceutical research applications, especially for animal bone histo-
morphometry. It operates at 25–45 kV with anode current <0.6 mA, X-ray tube focal
spot size 50 µm × 150 µm, measurement diameter up to 50 mm, and length of scout
view up to 90 mm. Voxel size variable, from 30 µm to 300 µm with multislice capabil-
ity, variable 1 to 99 slices. The scan duration is <80 seconds. Courtesy of Orthometrix,
Inc., White Plains NY.

whereas at higher photon energies, at which Compton scatter is the main mecha-
nism of X-ray interaction with matter, the attenuation varies as E−0.5, indicating
the greater tissue contrast that could be achieved at low photon energies. A number
of investigators in physics and material sciences (18–24) published examples of
early implementations that built on Grodzins’ analysis. Another important land-
mark was the publication by Feldkamp (12), who was developing a CT scanner for
structural analysis of ceramics in turbine engines. In 1981, he built a scaled-down
CT scanner (L.A. Feldkamp, personal communication) using a microfocus X-ray
source, a sensitive video camera, and a fluorescent screen. By modifying Herman’s
fan-beam reconstruction algorithm (25) he, together with Davis, advanced from
fan beam to cone beam geometry. Davis subsequently formalized this approach
(12). In 1984, Feldkamp started a collaboration with investigators at Henry Ford
Hospital and at the University of Michigan resulting in several pioneering papers
on the use of bench-top micro-CT applied to bone microarchitecture (25a). This
 3 Jun 2004 20:53 AR AR220-BE06-08.tex AR220-BE06-08.sgm LaTeX2e(2002/01/18) P1: IKH

MICRO-CT 189
Annu. Rev. Biomed. Eng. 2004.6:185-208. Downloaded from www.annualreviews.org
Access provided by University of Winnipeg on 01/23/19. For personal use only.

Figure 2 Total mass attenuation coefficients, K(E), have several mechanism of X-

ray/matter interaction, such as the photoelectric effect, Compton scatter, and pair pro-
duction, each of which predominates over a defined energy band as shown here. These
different mechanisms have different implications for CT image contrast and artifacts
owing to beamhardening and scatter contamination. Reproduced with permission from
Reference 17.

reconstruction algorithm was central in enabling the use of large-angle cone beam
so that bench-top micro-CT scanners could now complete scans within reasonable
time periods. This is because a number of contiguous slices can be scanned simul-
taneously as compared to scanning one slice at a time, moving the specimen along
its axis of rotation by one slice thickness and repeating the scan and so forth until
the volume of interest is scanned (26, 27).
The cone-beam method, however, limits the axial extent of the specimen that
can be scanned in one scan. A third method, developed by Kalendar and others
(28) involves continuously translating the specimen along its axis of rotation. The
helical CT reconstruction algorithms are generalized versions of the cone beam
algorithm (29). This results in “helical” scanning (30), which has the advantage
 3 Jun 2004 20:53 AR AR220-BE06-08.tex AR220-BE06-08.sgm LaTeX2e(2002/01/18) P1: IKH

190 RITMAN

that very long specimens (such as an intact mouse or arterial segments) can be
scanned intact, independent of the imaging array’s axial length.
The current phase of micro-CT development, which started in the mid 1990s,
involves bench-top scanners that can scan intact small rodents in response to the
rapidly developing need to screen small animals for drug discovery, cancer de-
tection and monitoring, and genomics purposes (7, 8, 31–33). This need is well
met by use of special-purpose microimaging methods because it has the potential
for accurate measurement of organ anatomy and the spatiotemporal distribution
of various functions within the organs in a minimally invasive manner. The mini-
mally invasive aspect is required when longitudinal studies, which are statistically
Annu. Rev. Biomed. Eng. 2004.6:185-208. Downloaded from www.annualreviews.org

preferable to cross-sectional studies, are desired. Moreover, in genomics research,

Access provided by University of Winnipeg on 01/23/19. For personal use only.

this capability is particularly important because once the genetically expressed

modification of interest is identified, the (now) valuable animal is left intact for
more detailed study, use as a disease model, and/or reproduction.

BASIC COMPONENTS OF MICRO-COMPUTED

TOMOGRAPHY
All micro-CT scanners follow the same general plan: an X-ray source, a specimen
that is to be imaged, an X-ray-to-electronic signal-converting imaging array, and a
device that either rotates the specimen within the stationary scanner or rotates the
scanner around the stationary specimen. The method for providing high resolution
generally falls into one of three approaches, as illustrated in Figure 3.

Figure 3 All micro-CT scanners follow the same basic principle of X-ray source, specimen,
and X-ray imaging device. There are three main approaches to achieving the magnification
needed to exceed the inherent spatial resolution of the X-ray image conversion to signal
system. Each method has its different advantages and limitations. The left panel shows the
conventional use of the X-ray cone beam, which projects a magnified X-ray image onto a
large-area X-ray imaging system. The middle panel shows how a nonmagnified X-ray image
can be magnified by subsequent optical means either by a lens or by a tapered fiber-optic
coupling (not shown here). The right panel shows the use of a Bragg interferometer that can
be used to magnify the image by virtue of photon wavelength-specific diffraction rather than
specular reflection.
 3 Jun 2004 20:53 AR AR220-BE06-08.tex AR220-BE06-08.sgm LaTeX2e(2002/01/18) P1: IKH

MICRO-CT 191

X-Ray Source
Bench-top X-ray sources are generally microfocus (i.e., focal spot diameter
<100 µm). Line focus spectroscopy X-ray tubes can be used to produce a stack
of parallel fan beam exposures (34) or, when held at an acute angle, for use as a
“contact” illuminating system (35). The power output (P) of the microfocus tubes
is limited (36) and follows Pmax = 1.4 (x)0.88, where x equals the focal spot di-
ameter in µm. X-ray scatter has not been as serious a problem in micro-CT as it
is in clinical CT. However, to obtain maximum signal to noise, the scatter can be
greatly reduced by use of a diffraction X-ray optic placed between the specimen
and the detector when monochromatic radiation is used (37).
Annu. Rev. Biomed. Eng. 2004.6:185-208. Downloaded from www.annualreviews.org
Access provided by University of Winnipeg on 01/23/19. For personal use only.

X-Ray to Light Conversion Systems

A thin (<0.5 mm) fluorescent crystal plate [e.g., CsI(Tl)] that is cut from a large
crystal (or it is built up via vapor deposition) is very efficient at converting X-rays
to light. However, there are several trade-offs. A thicker crystal plate will convert
more X-ray photons to light, but the thicker crystal plate provides more opportu-
nity for the light to stray from the X-ray-to-light interaction site before it leaves the
crystal; hence, poorer spatial resolution. Some very efficient crystalline materials
have birefringence, which also degrades spatial resolution. Another concern of
any fluorescent material is that of “lag,” which is the duration during which fluo-
rescence continues to be emitted from in the crystal after the X-ray exposure has
ceased. This can result in superposition of different angles of view or in blurring
owing to cardiogenic or respiratory motion during the X-ray exposure.
A fluorescent screen made up of a compressed layer of small granules of fluo-
rescing material can be made very efficient as far as X-ray-to-light conversion is
concerned. Although spatial resolution is limited essentially by the granule size,
this medium is useful for conebeam-magnified images because the granule size
has diminished importance the greater the magnification.
Fiber optics doped with fluorescent material and/or tipped with fluorescent
material has great potential for increased efficiency owing to depth of fluorescent
material, and the fibers result in constrained light transmission, i.e., maintain spatial
resolution (38).

The Light to Electron Coupling/Conversion System

Image intensifiers are still used, but most micro-CT scanners use a fluorescent
image that is coupled to a charge-coupled device (CCD) imaging array [light
to electrons (39)] via an optical lens or a fiber optic (40). The lens is relatively
inefficient at optical coupling, but it has the advantage of variable magnification
(20). The fiber-optic coupling is much more efficient, but magnification is fixed.
The image intensifier does provide more rapid illumination of the imaging array
(but not necessarily more X-ray photon statistics), which allows for more rapid
bench-top CT scanning (41). The number of levels of analog-to-digital conversion
 3 Jun 2004 20:53 AR AR220-BE06-08.tex AR220-BE06-08.sgm LaTeX2e(2002/01/18) P1: IKH

192 RITMAN

needed depends on the noise in the image being digitized and the dynamic range
of signal within the image (42).

CAPABILITIES OF MINI- AND MICRO-CT

Of the three approaches to bench-top CT scanner design (Figure 3), the conebeam
magnification approach is preferred in miniscanners where speed is an important
factor, but the contact imaging approach has the advantage of high-resolution CT
down to approximately 1 µm owing to the limit of X-ray-to-light conversion res-
Annu. Rev. Biomed. Eng. 2004.6:185-208. Downloaded from www.annualreviews.org

olution (43). The Bragg diffraction approach magnifies the X-ray image (while
Access provided by University of Winnipeg on 01/23/19. For personal use only.

maintaining near-parallel X-ray beams), which then allows for submicron resolu-
tion (44, 45).

Resolution Requirements: Mini-CT Scanner

The volume of a mammal’s internal organs scales with its body weight (46).
Hence, if we wish to merely scale the current clinical images (an adult human is
approximately 75 kg with CT image voxel dimensions of the order of 1 mm3) to
a mouse, we need to have a scanner that has voxels of the order of (100 µm)3 for
a 25 g adult mouse or (30 µm)3 for a neonatal mouse. In addition to dimensional
scaling, which, broadly speaking, is proportional to body mass M0.3, the functional
aspects, such as heart and respiratory cycle durations, scale as M−0.25 so that scan
aperture durations should be reduced from <300 ms in humans to <75 ms in mice
if, for instance, the heart is to be imaged.
Because even small rodents have a large range in relative size as they mature
from neonate to adult, it would seem reasonable to match the scan resolution to
the smallest animal to be scanned, and then to retrospectively “bin” the CCD
imaging pixels in proportion to the size of the larger animal. Another method is
to adjust the angle of the detector array to the X-ray beam so that the projected,
transverse dimension of each detector element (which we might call a dexel so as
to distinguish it from the effective image pixel size) is reduced (47). The impact
of the different resolutions is illustrated in Figure 4.

Micro-CT Scanner
Although bench-top micro-CT scanners can achieve meaningful cubic voxel sizes
as small as 5 µm, the synchrotron radiation-based scanners provide the highest
resolution. This is because of the narrow bandwidth of the energy-selective X-ray
and intensity is sufficient to permit rapid scanning (49). Organs’ basic functional
units (BFUs, the smallest aggregation of diverse cells within an organ that functions
like the organ, such as an hepatic lobule or a nephron) are of the order of (100 µm)3
and are approximately the same size in humans and mice. Hence, with a voxel size
of less than (50 µm)3 it is possible to obtain information not achievable in clinical
scanners, such as the number, size, and packing of BFUs and membranes only
 3 Jun 2004 20:53 AR AR220-BE06-08.tex AR220-BE06-08.sgm LaTeX2e(2002/01/18) P1: IKH

MICRO-CT 193
Annu. Rev. Biomed. Eng. 2004.6:185-208. Downloaded from www.annualreviews.org
Access provided by University of Winnipeg on 01/23/19. For personal use only.

Figure 4 Computer-generated displays of mini- and micro-CT image data. The upper
two panels depict a resolution (40 µm) that might be expected from a mini-CT scanner,
which is a clinical CT scanner scaled down to match the mouse. The lower panels depict
higher-resolution (5 µm) micro-CT. These images clearly resolve the basic functional
units (approximately 0.001 mm3), the terminal bronchiole and its alveoli of the lung
(left) and the glomerulus and associated arterioles and tubule in the kidney (right).
Reproduced with permission from Reference 48.

a few cells thick (50). In addition to BFUs, there are structures, such as pores,
trabeculi, ducts, and blood vessels, that have a range of dimensions. Kinney et al.
(51) showed how CT voxel size plays a role in the ability to accurately measure
such features. With voxel sizes less than (10 µm)3, even cellular dimensions, such
as osteoclast erosions in the surface of bone (52) can be imaged, and with voxel
sizes less than (1 µm)3, subcellular dimensions can be imaged (45, 53, 54).

Physics and Technology

Although scaling down the voxel size involves machining the scanner compo-
nents down to a smaller size (which has technological implications for precision
tolerances and component packing), it also means that different aspects of the
 3 Jun 2004 20:53 AR AR220-BE06-08.tex AR220-BE06-08.sgm LaTeX2e(2002/01/18) P1: IKH

194 RITMAN

well-understood physics underlying CT scanning must be accommodated. Thus,

as pointed out previously (16, 55), if the X-ray photon-energy used is less than
25 keV, the X-ray interaction with matter is predominated by the photoelectric
effect, whereas in clinical scanners in which the photon energy generally exceeds
50 keV, it is predominantly Compton scatter (56). A desirable feature of the low
photon energy is that the X-ray attenuation is up to two decades higher than it is
in clinical CT images, thereby allowing better discrimination of soft tissue types.
Another consequence is, however, that if maximum signal-to-noise ratio is to be
achieved, the X-ray photon energy must be adjusted to match the object diame-
ter, and the radiation should be essentially monochromatic if excessive imaging
Annu. Rev. Biomed. Eng. 2004.6:185-208. Downloaded from www.annualreviews.org

artifacts, such as X-ray beam hardening (57), are to be minimized. This has the
Access provided by University of Winnipeg on 01/23/19. For personal use only.

implication that individual, scaled-down scanners must be used for individual mice
because the simultaneous scanning of multiple mice within, for instance, a clinical
scanner will result in suboptimal image signal-to-noise ratio and spatial resolution.
In addition, synchronization of the cardio-respiratory movements if multiple small
animals are being imaged at the same time is almost impossible.
In an attempt to generate quasi-monochromatic radiation in bench-top micro-CT
scanners, some X-ray source anode materials can be selected for their characteristic
Kα emission radiation (e.g., 8.0 keV for Cu, 17.5 keV for Mo, and 22.2 keV for Ag).
This radiation output can then be filtered through a thin metal foil selected to have
its K absorption edge at an energy just above the Kα of the respective anode material
(e.g., 8.3 keV for Ni, 18.0 keV for Zr, and 24.4 keV for Pa), which predominantly
attenuates the X-ray photons with energies above and below the Kα characteristic
radiation energy. Some promising approaches to bench-top monochromatic X-
ray sources suitable for micro-CT are in development, for example, the use of a
powerful laser to elicit characteristic Kα emission (e.g., 17.5 keV for Mo) from
a suitable target material, so that when combined with a suitable foil filter, little
bremsstrahlung contamination occurs and <400 eV bandwidth at full width at half
maximum (FWHM) of the spectrum of the Kα peak is possible (58).
For mini-CT scanners applied to living animals, the speed of scanning is criti-
cally important. This can be achieved by several mechanisms (49, 59, 60). One is
by use of higher X-ray photon energies (e.g., >50 keV) usually generated with a
tungsten anode with considerable filtering through an aluminum or a copper foil.
At these higher energies, the bandwidth is not as critical as at lower keV because
of the reduced impact of beamhardening, as indicated in Figure 2. This increased
speed can also be enhanced by bringing the X-ray source closer to the object, thus
increasing the cone beam angle together with all the difficulties that result (e.g.,
penumbral blurring owing to focal spot diameter and the reconstruction algorithm
needed). Scan speed can be further increased by using multi X-ray source/detector
array scanners so that the scan duration is proportionately decreased, such as
the micro-CT system built by Bioimaging Research, Inc. (BIR) (61). Another
approach to increasing the X-ray intensity at the specimen, and to some extent
reducing the photon energy bandwidth, is the use of polycapillary-focused X-ray
optics (62).
 3 Jun 2004 20:53 AR AR220-BE06-08.tex AR220-BE06-08.sgm LaTeX2e(2002/01/18) P1: IKH

MICRO-CT 195

For the micro-CT scanners, specimens are generally rotated around a vertical
axis within a fixed X-ray/detector system. However, anesthetized animals would
be physiologically challenged if held in this vertical orientation for any extended
period of time. Rotation of the animal horizontally around its cephalo-caudal
axis would result in gross distortions and movements of the animal during the
scan. Hence, in the mini-CT scanner, the torso of intact animals requires a fixed
horizontal support table for the animal, and the X-ray/detector system should rotate
around the animal, resulting in a true mini-clinical CT scanner, such as illustrated
in Figure 1.
Annu. Rev. Biomed. Eng. 2004.6:185-208. Downloaded from www.annualreviews.org

Software-Based Extension of Micro-CT Capability

Access provided by University of Winnipeg on 01/23/19. For personal use only.

Mini- and micro-CT scanners are limited by financial as well as physical con-
straints. For example, the area of the X-ray imaging detector array that restricts
the transverse diameter of the object that can be scanned, the size of the dexels
that limit the spatial resolution of the CT image that can be generated, the nonuni-
formity of dexel characteristic exposure/output curves that results in CT image
artifacts such as the so-called third generation ring artifact, and the limitations
in use of large-angle conebeam geometry to do the image magnification. These
limitations can, in part, be overcome by special scanning modifications and image
processing techniques. Thus, to avoid incomplete scan data sets resulting from too
small an imaging array, one approach is to place the axis of rotation at one edge of
the detector array (instead of the usual central location) so that the full transverse
profile can be recorded for a 180◦ range of angles of view from a 360◦ scan. If
this is not possible then profile extension techniques (63) or local reconstruction
algorithms can be used (64). Higher resolution can be achieved (than would be
expected for the dexel size used) by shifting the detector array by a fraction of the
dimension of a dexel in the transaxial direction, which can then be used to make
CT images with resolution higher than would be possible without this trick (65).
The third-generation CT ring artifacts can be minimized by shifting the detector
array during the scan so that the self-reinforcing ring artifact mechanism is inter-
rupted (66). High resolution obtained by use of conebeam magnification is limited
by the focal spot size. If it is too small, the heat generated by the electron beam
bombardment will melt the anode so that it needs to be operated at low flux (36),
resulting in very slow data acquisition, but if it is too large, then the increased flux
has to be traded-off against the penumbral blurring that occurs. The latter can be,
in part, overcome by point spread function (PSF) deconvolution, but this requires
that the PSF be measured at all dexels of the imaging array because it generally
differs depending on location within the array (67).

Extension of Micro-CT Capabilities by Non-Scanner Methods

When synchrotron radiation is used, which is very intense, rapid scanning is pos-
sible, so that in vivo scanning of limbs (68), lower torso, or head can be accom-
plished without significant motion blurring. With slower bench-top scanners, some
 3 Jun 2004 20:53 AR AR220-BE06-08.tex AR220-BE06-08.sgm LaTeX2e(2002/01/18) P1: IKH

196 RITMAN

cyclic motions owing to the heartbeat could be accommodated by acquiring the

necessary scan data incrementally at the same phase of the cardiac cycle of se-
quential heart beats, such as achieved for micro-MRI scanners (69). However,
noncyclic transient events, such as the progressive accumulation or washout of a
contrast agent in or from a physiological space, cannot be scanned fast enough, even
by the high-intensity synchrotron radiation or such a gated scanning method. The
transient process can, however, be literally snap-frozen within the tissue of interest,
and that frozen specimen can then be scanned at leisure while frozen (70, 71).

Radiopaque Indicators of Physiological Spaces and Processes

Annu. Rev. Biomed. Eng. 2004.6:185-208. Downloaded from www.annualreviews.org
Access provided by University of Winnipeg on 01/23/19. For personal use only.

Essentially, all current X-ray micro-CT scanners, and clinical CT scanners use the
attenuation of the X-ray by tissues as the signal for generating the x-ray images. In
micro-CT scanners, this means that for a given X-ray photon energy, the contrast
signal is most closely related to the atomic number (Z) and the concentration of that
element (72). At the X-ray exposures tolerated by living tissues, this means that the
signal-to-noise ratio in the CT image is adequate primarily for differentiating air,
fatty tissue (e.g., brain white matter), nonfatty tissue [e.g., muscle, brain gray matter
(73)], and bone [within bone, several layers at different stages of mineralization
can be distinguished (26)]. In some tissues, there are normally high concentrations
of heavy elements (e.g., iodine in the thyroid and iron in hemoglobin and in the
hemachromatotic liver) that just reach a level at which pathological increases, or
decreases, can be detected by change in CT image contrast (19, 74).
The most commonly used clinical contrast agents are based on the highly atten-
uating elements iodine or barium. Iodine is attractive because it is readily attached
to sugars, which are tolerated at relatively high concentrations in blood (75), and
barium is used because it forms insoluble salts and is therefore useful for opacify-
ing the gut or airway. Other elements, such as bromine (76) and xenon, which is
useful for aeration of the lung and increased solubility in fatty tissue such as brain
white matter (77), and for tissue perfusion. The relatively low sensitivity of X-ray
attenuation imaging to these agents requires that relatively high concentrations
are needed for quantitation of local concentrations of the contrast agent in small
regions of interest. Clinical CT scanners need at least 10 mg iodine per cubic cen-
timeter. This brings with it the problem of the physiological perturbation caused by
the relatively high concentration of the indicator as well as the perturbation caused
by the physical characteristics of the volume, viscosity, and specific gravity of the
injected contrast agent (75).
The discrimination of physiological spaces, other than the intravascular, can be
achieved by selectively opacifying those spaces with administered contrast agents.
As a fraction of the intravascular contrast agent passes through the vascular en-
dothelium (especially if it is impaired by reduced oxygen levels or because it is
newly formed, as is often the case in malignancies) and remains in the extravascular
space for up to several minutes, the extravascular space and endothelial perme-
ability can also be estimated from the images as has been shown in clinical CT
image analysis (78). Moreover, as these contrast agents are generally preferentially
 3 Jun 2004 20:53 AR AR220-BE06-08.tex AR220-BE06-08.sgm LaTeX2e(2002/01/18) P1: IKH

MICRO-CT 197

excreted through the kidney, the opacification of the nephrons (which occurs over
many minutes) could be used to quantitate several aspects of renal function as well,
as shown with clinical CT image analysis (79). An intraperitoneal injection of iod-
inated contrast agents has two effects: (a) the immediate ability to better delineate
the gut and liver within the abdominal cavity, and (b) because of the slow trans-
fer of the contrast medium to the blood stream, the cardiac chambers and large
vessels, as well as the filtration by the kidneys, are subsequently opacified (2).
Iodinated contrast agents encapsulated in micrometer-sized or submicrometer-
sized liposomes can be used as markers of the intravascular space or to be conveyed
into the lymphatic space (80, 81).
Annu. Rev. Biomed. Eng. 2004.6:185-208. Downloaded from www.annualreviews.org

For micro-CT, it is also possible to use heavy elements that are physiological
Access provided by University of Winnipeg on 01/23/19. For personal use only.

surrogates for the normal elements. For example, the use of strontium as a more
radiopaque surrogate for calcium in bone deposition (52). In this case, a short
duration of strontium loading, repeated at, for example, a month’s interval, will
provide two, separated, narrow bands of strontium laid down in the newly deposited
calcium mineral in bone. This could be used to quantitate local bone growth or
resorption and relate it to local 3-D microarchitecture. Another method, applied to
isolated specimens, is the use of osmium tetroxide staining because it preferentially
stains structures high in lipid content, such as cell walls, in high-resolution micro-
CT images (48).

X-Ray Micro-CT as a Complementary Methodology

No single imaging modality provides all the information about microstructure,
function, and molecular processes in one image. Indeed, scanners of each modal-
ity need special emphasis of a particular aspect of physics to optimize the imaging
of a particular aspect of organ microstructure and/or function. Consequently, no
individual CT image is likely to provide even the full range of the applications that
are possible for any one imaging modality. However, such limitations can often
be mitigated by combined use of micro-CT with another imaging modality, thus
greatly increasing the sensitivity and/or specificity of the resulting image data.
A necessary aspect of this synergistic use of multiple modalities is that accurate
spatial and temporal registration of the 3-D anatomy is achieved. This presents
a challenging technological problem that can, in part, be solved by software reg-
istration algorithms that warp, rotate, and/or translate an image within the 3-D
image, such as performed in the Gamma-Medica combined (but spatially sequen-
tial) micro-CT-SPECT scanner. Three examples of this combined approach follow.

Radionuclide-Based Imaging
Radionuclide-based imaging methods have great strengths the very high specificity
and sensitivity of the images (82, 83). This sensitivity allows the use of picomolar
concentrations of the radionuclide-labeled biologically active molecules and la-
beling of relatively sparsely distributed specific cells. Unfortunately, the methods
are limited by poor spatial resolution (>1 mm) and the poor photon statistics that
can be achieved within a reasonable time interval. However, X-ray CT can provide
 3 Jun 2004 20:53 AR AR220-BE06-08.tex AR220-BE06-08.sgm LaTeX2e(2002/01/18) P1: IKH

198 RITMAN

the high spatial resolution distribution of X-ray attenuation coefficients, which

can be used to improve the accuracy of the radionuclide-based imaging methods
by enabling correction for the gamma-ray attenuation through tissues. Similarly,
if there is a priori knowledge as to what physiological space the radionuclide is
likely to accumulate in, then radionuclide activity can be constrained to that space
(which is delineated by the X-ray CT image data so that the SPECT reconstruction
process can restrict the backprojected activity to the confines of that space) so that
the concentration of that radionuclide can be more accurately estimated.

Histological Techniques
Annu. Rev. Biomed. Eng. 2004.6:185-208. Downloaded from www.annualreviews.org

Histological techniques, with their broad spectrum of tissue element-specific stains

Access provided by University of Winnipeg on 01/23/19. For personal use only.

and immunohistochemical methods, can provide useful information about molec-

ular and cellular processes in cells and tissues. A weakness of the histological
method alone is that preparation and sectioning can damage and distort the sec-
tions as well as destroy the specimen for subsequent analysis by other method-
ologies. Consequently, it provides isotropic 3-D microstructural information only
with considerable difficulty. Figure 5 illustrates a 3-D micro-CT image of the

Figure 5 Left panel: a micro-CT image of a gerbil kidney impregnated with osmium tetrox-
ide. This particular section was found within the 3-D micro-CT image so as to match the (right
panel) histological (Methylene blue-stained) section that was performed after the micro-CT
scan was completed. The excellent registration enables us to integrate the histological in-
formation (e.g., cell type, immunological state) with the 3-D microarchitectural information
obtained from the micro-CT image, something that is virtually impossible to obtain from
any one imaging modality by itself. Courtesy of Dr. Michael D. Bentley, Minnesota State
University, Mankato, MN.
 3 Jun 2004 20:53 AR AR220-BE06-08.tex AR220-BE06-08.sgm LaTeX2e(2002/01/18) P1: IKH

MICRO-CT 199

microanatomy can be used to register the histological sections obtained after the
micro-CT scan is completed. This could greatly extend the information content
beyond that of either method used alone.

Molecular Analysis Methods

Molecular analysis methods, such as immunohistochemistry or mRNA quantita-
tion, often require fresh specimens and generally involve destruction of the speci-
men. Respectively, these methods provide either sparsely sampled information or
values averaged over an entire tissue sample. Some tissue analysis methods can
tolerate prior fixation, whereas others cannot. The 3-D microanatomy obtained
Annu. Rev. Biomed. Eng. 2004.6:185-208. Downloaded from www.annualreviews.org

via micro-CT can be used to determine the fraction of the tissue sample that has,
Access provided by University of Winnipeg on 01/23/19. For personal use only.

and/or the likely location where, the molecule of interest accumulates so that a
more meaningful comparison between animals can be made. The latter case can be
dealt with by doing a micro-CT scan of the freshly frozen specimen, while frozen,
so that the specimen can subsequently be processed for molecular analysis without
loss of the 3-D microanatomic information (84). Here, too, the average value of
the molecular concentration in the tissue sample can be corrected for the physi-
ological space’s volume. The 3-D micro-CT images can be used to direct where
the biopsy within the specimen should be obtained so as to enhance the likelihood
that the concentration of the molecule of interest is highest, or, conversely, how
the concentration distribution is related to microarchitectural features.

Data Processing
A major feature of the recent developments in genomics, drug discovery, and can-
cer research is the pressure to increase throughput, driven by the need to screen
many animals. Some automated procedures are needed for managing the acquisi-
tion analysis and archiving of thousands of images per year. The logistics of the
multitude of microimaging applications, however, may well be the Achilles’ heel
of the whole endeavor of rapid throughput mode of operation. With each image
containing 108–1010 voxels, just transmitting and archiving the data is a challenge,
let alone performing intelligent analysis within a reasonable time frame. Conse-
quently, one might ask if a smaller image data set would provide a statistically
adequate representation of the data for the entire organ or region within the body.
Often, this is not sufficient because a single, localized deficiency may initiate the
mechanical or functional breakdown of an organ (85, 86) and this means that the
entire organ must be inspected at the resolution necessary to detect this micro-
scopic nidus or weakest link (e.g., a hairline fracture). Similarly, the entire organ
or part of the body must be inspected if the early appearance of a malignant tumor
cell or some other cell is of interest.
It is also likely that the large number of animals to be investigated will require
computer-aided procedures to analyze the image data. Matching the recorded im-
ages to a standard atlas seems to be one such approach, a process currently un-
derway in the International Union of Physiological Sciences (IUPS) Physionome
 3 Jun 2004 20:53 AR AR220-BE06-08.tex AR220-BE06-08.sgm LaTeX2e(2002/01/18) P1: IKH

200 RITMAN

project (87). Additional quantifying tools also need to be designed to get accurate
and specific quantification for different organs or structures within the organs.
Thus, a tremendous effort has to be brought to bear on image analysis procedures
(such as segmentation), which have to be defined in close interaction between
computer science and biology investigators.

ONGOING DEVELOPMENTS OF X-RAY MICRO-CT

For mini- and micro-CT there will likely be continued incremental improvements
Annu. Rev. Biomed. Eng. 2004.6:185-208. Downloaded from www.annualreviews.org

of the technology of the scanners, and there are several potential physics aspects
Access provided by University of Winnipeg on 01/23/19. For personal use only.

that are being explored for the purposes of broadening the spectrum of contrast
mechanisms that can be utilized. The synchrotron radiation source, with its bril-
liance and energy-selective and narrow bandwidth photon generation capabilities,
provides a convenient and effective means to study nonattenuation-based X-ray
imaging, such as phase delay, fluorescence and K-edge subtraction, and X-ray
scatter-based micro-CT imaging. Much technological development is needed to
make such approaches suitable for routine bench-top CT imaging use. The follow-
ing physics principles are being explored for the purposes of extending the contrast
and sensitivity, primarily of micro-CT, at this stage. A particular advantage of the
fluorescent and scatter-based approaches is that the detected X-ray is the signal of
interest (so-called dark field acquisition) as compared to attenuation-based meth-
ods in which the signal is a reduction in the bright-field acquisition.
For the purposes of comparison with the methods described below, note that
attenuation-based CT using monchromatic synchrotron radiation has been shown
to discriminate 1 mg I/cm3 (74) and 30 mg Ca/cm3 (24). Strictly speaking, this
objective index of contrast resolution should be further normalized to the voxel
volume.

K-Edge Subtraction
K-edge subtraction has been explored by a number of investigators (72, 88, 89).
The method is based on the fact that every element has a well-defined, up to ten-
fold, energy-dependent transition of X-ray attenuation value. Examples of K-edge
transition energy use 7.1, 13.5, 15.2, 16.1, and 33.2 keV for Fe (hemoglobin),
Br (Cl surrogate), Rb (K surrogate), Sr (Ca surrogate), and I (in thyroid, but also
because it chemically binds readily to many organic compounds of interest), re-
spectively. This characteristic can be utilized by making a CT image at an X-ray
photon energy just below and again just above this transition photon energy. The
small difference between the X-ray attenuation values of other elements in various
tissues in these two images essentially eliminates the signal from these tissues if
the two CT images, obtained at the two slightly different energies, are subtracted.
The subtracted image therefore contains just the specific element that was selected
for imaging. A concentration of 15 µg iodine/cm3 has been detected with this
 3 Jun 2004 20:53 AR AR220-BE06-08.tex AR220-BE06-08.sgm LaTeX2e(2002/01/18) P1: IKH

MICRO-CT 201

method (74). This method has not been used as much as one might anticipate
because of the difficulty generating the tuned, narrow-spectral-bandwidth X-ray
photon energy with a bench-top X-ray source. Although synchrotron imaging
can overcome this difficulty, routine access to synchrotron is typically limited for
large-scale applications. This logistic bottleneck is being reduced somewhat by
several new synchrotron radiation sources that have recently been, or are almost
to be, added at several international locations. Another factor is that at the low
photon energies involved in K edge subtraction imaging of most biologically rel-
evant elements, only small-diameter specimens can be scanned at optimal image
quality.
Annu. Rev. Biomed. Eng. 2004.6:185-208. Downloaded from www.annualreviews.org
Access provided by University of Winnipeg on 01/23/19. For personal use only.

X-Ray Phase Delay

X-ray phase delay can also be used for X-ray imaging (90–93). X-rays travel at
slightly different velocities through different tissues, so that as an X-ray beam
passes through a particular tissue, a phase delay (or advance) relative to an X-ray
not passing through that tissue develops. Phase differences can be represented by
interference patterns in the X-ray intensities, which in turn can be mathematically
converted to actual phase delays. This method provides a tissue contrast higher than
that achieved with attenuation-based X-ray imaging. The concentration of iodine
that has been detected with this method is of the order of 250 µg/cm3, which is
four times more sensitive than attenuation-based, monochromatic synchronization
photon detector.

X-Ray Scatter
X-ray scatter at both small angle and large angle (to the direction of the X-ray beam)
contains much information about lower atomic number elements, and to some ex-
tent, the molecule bonds along the X-ray beam path (94–97). As organic molecules
have a characteristic angle-dependant X-ray intensity distribution, this becomes
a signature for that molecule. For highly periodic molecules, such as collagen,
distinct diffraction peaks can be identified (98–100). At >50 keV X-ray photon
energy, the ratio of Rayleigh-to-Compton scanner can be used to achieve increased
contrast for small changes in Z, such as 6 mg/cm−3 for iodine, 8 mg/cm−3 for iron,
and 10 mg/cm−3 for potassium (101). However, these data must be corrected for
attenuation, which can be done with the concurrently recorded attenuation image.
However, a difficulty with scatter imaging is the inefficient mechanisms currently
available for measuring, and localizing, the source of the scattered radiation.

X-Ray Fluorescence
X-ray fluorescence occurs when an electron that has been knocked out of its atomic
orbit by an X-ray photon results in a secondary X-ray photon being generated. This
photon’s energy is characteristic for the specific element, for example, 6.4, 11.9,
13.4, 14.2, and 28.6 keV for Fe, Br, Ru, Sr, and I, respectively. Hence, the identity,
 3 Jun 2004 20:53 AR AR220-BE06-08.tex AR220-BE06-08.sgm LaTeX2e(2002/01/18) P1: IKH

202 RITMAN

concentration, and location of that element can be accurately derived, for example,
a concentration of iodine down to 60 µg/cm3 has been detected with fluorescence
CT (73, 102, 103). A problem with this method is that the energy of fluorescent X-
ray photons is less than that of the illuminating X-rays, and hence they are generally
considerably more attenuated (104) and difficult to distinguish from photons that
have been scattered by other physics mechanisms.

Very High Resolution

X-ray optical methods, such as zone plates (53) or Bragg diffraction optics (19,
44, 45), can be used to magnify the X-ray projection image to the extent that sub-
Annu. Rev. Biomed. Eng. 2004.6:185-208. Downloaded from www.annualreviews.org

micrometer resolution can be achieved. At this magnification, additional physics

Access provided by University of Winnipeg on 01/23/19. For personal use only.

aspects as well as sensitivity to small temperature changes of the optics become

increasingly important issues.

RADIATION DOSE LIMITATIONS

A major limitation of X-ray attenuation imaging is its ionizing effect, which can
result in immediate radiation damage and long-term genetic damage (56). The
signal-to-noise ratio of X-ray attenuation imaging depends largely on the number
of X-ray photons detected per voxel of the 3-D image [i.e., the X-ray exposure
has to increase with increasing spatial resolution (105)]; therefore, its sensitivity
is ultimately limited by its potential for radiation damage. Grodzins showed that
for constant signal-to-noise ratio, the radiation dose necessary for small samples
was approximately proportional to the specimen diameter to the third power, so
that an X-ray exposure of 0.01 Gy (0.1 Rad) for a 10 cm diameter specimen
would require 102, 5 × 104, and 2.5 × 107 Gy for water-like specimens 1, 0.1, and
0.01 cm diameter, respectively. An exposure of 6 Gy is generally considered lethal
for a small rodent; therefore, exposures must be much less in animals required
to undergo multiple sequential scans. For mini-CT applications, a dose of 0.5–
1.0 Gy is generally required, which represents approximately 5% of LD 50/30
(50% of animals die by 30 days) in mature mice (2). But as longitudinal studies
require repeated exposures, and dose accumulation will present increasing risks
to the animal, efforts to achieve reduction in dose through instrumentation and
algorithms are essential for repeated animal studies.

SUMMARY AND CONCLUSIONS

Mini- and micro-CT imaging are rapidly developing fields stimulated by the recent
increased demand for small-animal imaging. These imaging methods are useful
because they nondestructively image a 3-D volume. However, they are limited by
relatively poor contrast of most biologically important molecules, as well as by
radiation damage. Current developments of micro-CT that are not based on X-ray
 3 Jun 2004 20:53 AR AR220-BE06-08.tex AR220-BE06-08.sgm LaTeX2e(2002/01/18) P1: IKH

MICRO-CT 203

attenuation have the potential to improve contrast. Synergistic use of micro-CT

with radionuclide or histological imaging methods can greatly expand the role for
micro-CT.

ACKNOWLEDGMENTS
The author acknowledges Dr. Brian P. Flannery and Mr. John H. Dunsmuir of
Exxon Mobil, who introduced him to the intricacies of micro-CT; the helpful
input from Professor Ulrich Bonse of Dortmund University for his insights into
synchrotron-based micro-CT; Dr. Ge Wang of the University of Iowa for providing
Annu. Rev. Biomed. Eng. 2004.6:185-208. Downloaded from www.annualreviews.org

the information about commercially available scanners (especially Table 1); and
Access provided by University of Winnipeg on 01/23/19. For personal use only.

Dr. Lee Feldkamp of the Ford Motor Company for his information that lead up to
the generation his conebeam algorithms. I also thank Ms. Julie M. Patterson and
Mr. Steven M. Jorgensen for the preparation of several of the illustrations, and Ms.
Delories C. Darling for typing this manuscript. The author’s micro-CT research is
supported by NIH grants EB000305 and HL65342, and research was carried out in
part at the National Synchrotron Light Source, Brookhaven National Laboratory,
which is supported by the US Department of Energy, Division of Materials Sci-
ences, and Division of Chemical Sciences, Contract No. DE-AC02-98CH10886.

The Annual Review of Biomedical Engineering is online at

http://bioeng.annualreviews.org

LITERATURE CITED
1. Service RF. 1999. Scanners get a fix in lab E Buratitni, A Baleran, pp. 155–170. Am-
animals. Science 286:2261–62 sterdam: IOS Press
2. Paulus MJ, Sari-Sarraff H, Gleason SS, 6. Bonse U, Busch F. 1996. X-ray com-
Bobreck M, Hicks JS, et al. 1999. A new puted microtomography (µCT) using syn-
X-ray computed tomography system for chrotron radiation (SR). Prog. Biophys.
laboratory mouse imaging. IEEE Trans. Mol. Biol. 65(1/2):133–69
Nucl. Sci. 46:558–64 7. Paulus MJ, Gleason SS, Kennel SJ, Hun-
3. Kinney JH, Johnson QC, Nichols MC, sicker PR, Johnson DK. 2000. High res-
Bonse U, Saroyan RA, et al. 1988. Three olution X-ray computed tomography: an
dimensional X-ray computed tomography emerging tool for small animal cancer re-
in material science. MRS Bull. XIII:13–17 search. Neoplasia 1–2:62–70
4. Graeff W, Engelke K. 1991. Micro radio- 8. Holdsworth DW, Thornton MM. 2002.
graphy and microtomography. In Hand- Micro-CT in small animal and speci-
book on Synchrotron Radiation, eds. S men imaging. Trends Biotech. 20(8):S34–
Ebashi, M Koch, E Rubinstein, pp. 361– 39
405. Amsterdam: Elsevier 9. Kujoori MA, Hillman BJ, Barrett H. 1980.
5. Graeff W. 1996. Biomedical application High resolution computed tomography of
of microtomography. In Proceedings of the normal rat nephrogram. Invest. Radiol.
International School of Physics “Enrico 15:148–54
Fermi” Course CXXVII, BiomedicalAap- 10. Sato T, Chieda O, Yamakashi Y, Tsub-
plications of Synchrotron Radiation, eds. ouchi M. 1981. X-ray tomography for
 3 Jun 2004 20:53 AR AR220-BE06-08.tex AR220-BE06-08.sgm LaTeX2e(2002/01/18) P1: IKH

204 RITMAN

microstructural objects. Appl. Opt. 20: tive imaging of materials microstructures

3880–83 using X-ray microscopy. Mat. Res. Soc.
11. Elliott JC, Dover SD. 1982. X-ray tomog- Symp. Proc. 217:81–95
raphy. J. Microsc. 126(2):211–13 23. Bonse U, Nußhardt R, Busch F, Pahl K,
12. Feldkamp LA, Davis LC, Kress JW. 1984. Kinney JH, et al. 1991. X-ray tomographic
Practical cone-beam algorithm. J. Opt. microscopy of fibre-reinforced materials.
Soc. Am. A 1:612–19 J. Mater. Sci. 26:4076–85
13. Burstein P, Bjorkholm PJ, Chase RC, 24. Salome M, Peyrin F, Cloetens P, Odet C,
Seguin FH. 1984. The largest and the Laval-Jeantet A-M, et al. 1999. A syn-
smallest X-ray computed tomography chrotron radiation microtomography sys-
systems. Nucl. Instrum. Methods Phys. tem for the analysis of trabecular bone
Annu. Rev. Biomed. Eng. 2004.6:185-208. Downloaded from www.annualreviews.org

Res. 221:207–12 samples. Med. Phys. 26(10):2194–204

Access provided by University of Winnipeg on 01/23/19. For personal use only.

14. Bowen DK, Elliott JS, Stock SR, 25. Herman GT. 1980. Image reconstruction
Dover SD. 1986. X-ray microtomography from projection. The Fundamentals of
with synchrotron radiation. Proc. SPIE Computerized Tomography, pp. 40–54.
691:94–98 New York: Academic Press
15. Margaritondo G. 1988. Introduction to 25a. Feldkamp LA, Goldstein SA, Parfitt AM,
Synchrotron Radiation. New York: Ox- Jesion G, Kleerekoper M. 1989. The direct
ford Univ. Press examination of three-dimensional bone
16. Grodzins L. 1983. Optimum energies architecture in vitro by computed tomog-
for X-ray transmission tomography of raphy. J. Bone Miner. Res. 4:3–11
small samples. Nucl. Instrum. Methods 26. Elliott JC, Dover SD. 1984. Three dimen-
206(3):541–45 sional distribution of mineral in bone at
17. Tonner PD, Stanley JH. 1992. Su- resolution of 15 µm determined by x-
pervoltage computed tomography for ray microtomography. Metab. Bone Re-
large aerospace structures. Mater. Evol. hab. Res. 5:219–21
50(12):1434–38 27. Rüegsegger P, Köller B. 1994. A micro CT
18. Cho ZH, Honmg KS, Nicioglu O. system for the nondestructive analysis of
1984. Three dimensional microtomog- bone samples. Bone Miner. 25:S4
raphy with synchrotron radiation using 28. Kalendar W, Klotz W, Vock E. 1990. Spi-
planar integral projection data. Nucl. In- ral volumetric CT with single-breathhold
strum. Methods Phys. Res. A 227(2):385– technique, continuous transport, and
92 continuous scanner rotation. Radiology
19. Bonse U, Johnson Q, Nichols M, 176:181–83
Nußhardt R, Krasnicki S, et al. 1986. 29. Wang G, Lin TH, Cheng PC, Shinozaki
High resolution tomography with chem- DM, Kim HG. 1991. Scanning cone-beam
ical specificity. Nucl. Instrum. Methods reconstruction algorithms for X-ray mi-
Phys. Res. A 246(1–3):644–48 crotomography. Proc. SPIE 1556:99–112
20. Flannery BP, Deckman HW, Roberge 30. Corrigan NM, Chavez AE, Wisner ER,
WG, D’Amico KL. 1987. Three dimen- Boone JM. 1999. A multiple detector ar-
sional X-ray microtomography. Science ray helical X-ray microtomography sys-
237:1439–44 tem for specimen imaging. Med. Phys.
21. Spanne P, Rivers M. 1987. Computerized 26(8):1708–13
microtomography using synchrotron ra- 31. Weber DA, Ivanovic M. 1999. Ultra-
diation from the NSLS. Nucl. Instrum. high-resolution imaging of small animals:
Methods Phys. Res. B 24/25:1063–67 implications for preclinical and research
22. Kinney JH, Nichols MC, Bonse U, Stock studies. J. Nucl. Med. 6:332–34
SR, Breunig TM, et al. 1991. Non destruc- 32. Wang G, Vannier MW. 2001. Micro-CT
 3 Jun 2004 20:53 AR AR220-BE06-08.tex AR220-BE06-08.sgm LaTeX2e(2002/01/18) P1: IKH

MICRO-CT 205

scanners for biomedical applications. An ray imaging X-ray detector for microto-
overview. Adv. Imaging 16:18–27 mography. Proc. Conf. Photoelectron. Im-
33. Paulus MJ, Gleason SS, Easterly ME, age Dev. Sect. 3:257–64
Foltz CJ. 2001. A review of high reso- 43. Stamapanoni M, Borchert GL, Wyss P,
lution X-ray computed tomography and Abela R, Patterson B, et al. 2002. High
other imaging modalities for small ani- resolution X-ray detector for synchrotron-
mals. Lab Anim. 30:36–45 based microtomography. Nucl. Instrum.
34. Kohlbrenner A, Hämmerle S, Laib A, Methods Phys. Res. A 491:291–301
Koller B, Rüegsegger P. 1999. Fast 3D 44. Nußhardt R. 1990. Elementspezifische
multiple fan-beam CT systems. Proc. und hoch ortsauflösende mikrotomogra-
SPIE 3772:44–53 phie mit synchrotronstrahlung. PhD the-
Annu. Rev. Biomed. Eng. 2004.6:185-208. Downloaded from www.annualreviews.org

35. Jorgensen SM, Demirkaya O, Ritman sis. Univ. Dortmund, Germany

Access provided by University of Winnipeg on 01/23/19. For personal use only.

EL. 1998. Three-dimensional imaging of 45. Stampanoni M, Borchert G, Abela R,

vascular and parenchyma inintact rodent Rüegsegger P. 2002. Bragg magnifier: a
organs with X-ray micro-CT. Am. J. Phys- detector for submicrometer X-ray com-
iol. Heart Circ. Physiol. 44:H1103–14 puter tomography. J. Appl. Phys. 92:
36. Flynn MJ, Hames, SM, Reimann DA, 7630–35
Wilderman SJ. 1994. Microfocus X-ray 46. Calder WA III. 1984. Size Function and
sources for 3D microtomography. Nucl. Life History. Cambridge: Harvard Univ.
Instrum. Methods Phys. Res. A 352(1– Press
3):312–15 47. DiBianca FA, Zou P, Jordan LM, Laugh-
37. Takeda T, Hai Y, Hayashi K, Nagata Y, Ya- ter JS, Zeman HD, Sebes J. 2000. A vari-
maji H, Hyodo K. 1994. High spatial res- able resolution X-ray detector for com-
olution CT with a synchrotron radiation puted tomography. II. Imaging theory and
system. J. Comp. Assist. Tomogr. 18:98– performance. Med. Phys. 27(8):1875–80
101 48. Ritman EL. 2002. Molecular imaging in
38. Bueno C, Rairden RL, Betz RA. 1996. Hy- small animals—roles for micro-CT. J.
brid scintilators for X-ray imaging. Proc. Cell. Biochem. 39:116–24
SPIE 2708:1–14 49. Wang Y, DeCarlo F, Foster I, Insley J,
39. Reimers P, Goebbels J, Weise HP, Wild- Kesselman C, et al. 1999. A quasi-real
ing K. 1984. Some aspects of industrial time X-ray microtomography system at
non-destructive evaluation by X-ray and the advanced photon source. Proc. SPIE
gamma ray computed tomography. Nucl. 3772:318–27
Instrum. Methods Phys. Res. A 221:201–6 50. van Spaendonck MP, Cryns K, van de
40. Castelli CM, Allinson NM, Moon KJ, Heyning PH, Scheuermann DW, van
Watson DL. 1994. High spatial resolution Camp G, Timmermans JP. 2000. High res-
scintillator screens coupled to CCD detec- olution imaging of the mouse inner ear by
tors for X-ray imaging applications. Nucl. microtomography: a new tool in inner ear
Instrum. Methods Phys. Res. A 348:649– research. Anat. Rec. 259:229–36
53 51. Kinney JH, Haupt DL, Nichols MC, Bre-
41. Johnson RH, Karau KL, Molther RC, unig TM, Marshall GW Jr, Marshall
Dawson CA. 1999. Quantification of pul- SJ. 1994. The X-ray tomographic micro-
monary arterial wall distensibility us- scope: three-dimensional perspectives of
ing parameters extracted form volumetric evolving microstructures. Nucl. Instrum.
micro-CT images. Proc. SPIE 3772:15– Methods Phys. Res. A 347(1–3):480–86
23 52. Ritman EL, Bolander ME, Fitzpatrick LA,
42. Dunsmuir JH, Ferguson SR, D’Amico Turner RT. 1998. Micro-CT imaging of
KL. 1992. Design and operation of an X- structure-to-function relationship of bone
 3 Jun 2004 20:53 AR AR220-BE06-08.tex AR220-BE06-08.sgm LaTeX2e(2002/01/18) P1: IKH

206 RITMAN

microstructure and associated vascular in- resolution computed tomography from ef-
volvement. Technol. Health Care 6:403– ficient sampling. J. Inverse Probl. 16:635–
12 50
53. Haddad WS, McNulty GI, Trebes JE, An- 66. Davis GR. 1997. Image quality in X-ray
derson EH, Levesque RA, Yang L. 1994. microtomography. Proc. SPIE Dev. X-Ray
Ultrahigh resolution X-ray tomography. Tomogr. 3149:213–21
Science 266:1213–15 67. Witt ST, Riedel CH, Gössl M, Chme-
54. Wang Y, Jacobson C, Maser J, Os- lik MS, Ritman EL. 2003. Point spread
anna A. 1999. Soft X-ray microcopy function deconvolution in 3D micro-CT
with cryoscanning transmission X-ray mi- angiography for multiscale vascular tree
croscopc: II. Tomography. J. Microsc. separation. Proc. SPIE Med. Imag. Phys.
Annu. Rev. Biomed. Eng. 2004.6:185-208. Downloaded from www.annualreviews.org

197(Pt. 1):80–93 Med. Imaging 5030:720–27

Access provided by University of Winnipeg on 01/23/19. For personal use only.

55. Spanne P. 1989. X-ray energy optimiza- 68. Kinney JH, Lane NE, Haupt DL. 1995.
tion in computed tomography. Phys. Med. In vivo three dimensional microscopy
Biol. 34(6):679–90 of trabecular bone. J. Bone Miner. Res.
56. Dowseth DJ, Kenny PA, Johnston RE. 10(2):264–70
1998. The Physics of Diagnostic Imaging. 69. Qui HH, Cofer GP, Hedlund LW, Johnson
London: Chapman Hall Med. GA. 1997. Automated feedback control of
57. Kak AC, Slaney M. 1988. Principles body temperature for small animal stud-
of Computerized Tomographic Imaging. ies with MR microimaging. IEEE Trans.
New York: IEEE Press Biomed. Eng. 44:1107–13
58. Chen LM, Forget P, Toth R, Kieffer JC, 70. Kantor B, Jorgensen SM, Lund PE, Chme-
Krol A, et al. 2003. Laser-based intense lik MS, Reyes DA, Ritman EL. 2002.
hard X-ray source for mammography. Cryostatic micro-computed tomography
Proc. SPIE 5030:923–28 imaging of arterial wall perfusion. Scan-
59. Kohlbrenner A, Koller B, Hammerle S, ning 24:186–90
Ruegsegger P. 2001. In vivo microtomog- 71. Gössl M, Beighley PE, Malyar NM, Rit-
raphy. Adv. Exp. Med. Biol. 496:213– man EL. 2004. Transendothelial solute
24 transport in the coronary vessel wall—
60. Sasov A. 2002. In vivo micro-CT for small role of vasa vasorum: a study of cryo-
animal imaging. Proc. IEEE. Int. Symp. gstatic micro-CT. Am. J. Physiol. In re-
Biomed. Imaging, Washington, DC, pp. view
377–80 72. Torikoshi M, Tsumoo T, Sasaki M, Endo
61. Bio-Imaging Research, Inc. (BIR). www. M, Noda Y, et al. 2003. Electron density
bio-imaging.com measurement with dual energy X-ray CT
62. Jorgensen SM, Reyes DA, MacDonald using synchrotron radiation. Phys. Med.
CA, Ritman EL. 1999. Micro-CT scanner Biol. 7:673–85
with a focusing polycapillary x-ray optic. 73. Takeda T, Momose A, Hirano K, Haraoka
Proc. SPIE 3772:158–66 S, Watannabe T, Itai Y. 2000. Human car-
63. Lewitt RM. 1979. Processing of incom- cinoma, early experience with phase con-
plete measurements data in computed to- trast X-ray CT with synchrotron radiation-
mography. Med. Phys. 6:412–17 comparative specimen study with optical
64. Ritman EL, Dunsmuir JH, Faridani A, microscopy. Radiology 214:298–301
Finch DV, Smith KT, Thomas PJ. 1997. 74. Dilmanian FA. 1992. Computed tomog-
Local reconstruction applied to X-ray mi- raphy with monochromatic X-rays. Am.
crotomography. IMA Vol. Math. Appl. In- J. Physiol. Imaging 3/4:175–93
verse Probl. Wave Propag. 90:443–52 75. Paroz Z, Moncada R, Sovak M. 1987.
65. Faridani A, Ritman EL. 2002. High- Contrast Media: Biological Effects and
 3 Jun 2004 20:53 AR AR220-BE06-08.tex AR220-BE06-08.sgm LaTeX2e(2002/01/18) P1: IKH

MICRO-CT 207

Clinical Applications,Vols. 1–3. Boca Ra- biopsies. J. Musculoskeletal. Neuronal.

ton, FL: CRC Press Interact. 3:83–88
76. Gonzalez CF, Osterholm JL, Triolo AJ, 85. Müller R, Gerber SC, Hayes WC. 1998.
Bell RD, Menghetti RA. 1986. Fluoro- Micro-compression: a novel technique
carbon emulsion as a potential contrast for the nondestructive assessment of lo-
medium in the subarachnoid space and cal bone failure. Technol. Health. Care
brain tissue. Experiments in cats. Acta Ra- 6:433–44
diol. Suppl. 369:554–57 86. Suki B, Lutchen KR, Ingenito EP. 2003.
77. Gur D, Yonas H, Jackson DL, Wolfson On progressive nature of emphysema:
SK Jr, Rockette H, et al. 1985. Simul- roles of proteases, inflammation, and me-
taneous measurements of cerebal blood chanical forces. Am. J. Respir. Crit. Care
Annu. Rev. Biomed. Eng. 2004.6:185-208. Downloaded from www.annualreviews.org

flow by xenon/CT method and the micro- Med. 168:516–21

Access provided by University of Winnipeg on 01/23/19. For personal use only.

sphere method. A comparison. Invest. Ra- 87. Cowley AW Jr. 1997. Genomics to physi-
diol. 20:672–77 ology and beyond. How do we get there?
78. Ritman EL. 1994. Myocardial capillary Physiologist 40:205–11
permeability to iohexol: evaluation with 88. Bonse U, Nußhardt R, Busch F, Pahl
fast X-ray computed tomography. Invest. R, Johnson QC, et al. 1989. Optimiz-
Radiol. 29:612–17 ing of CCD-based energy-modulated X-
79. Lerman LO, Rodriguez-Porcel M, ray micro-tomography. Rev. Sci. Instrum.
Romero JC. 1999. The development of 60(7, Pt. 2B):2478–81
X-ray imaging to study renal function. 89. Kinney JH, Johnson QC, Nichols MC,
Kidney Int. 55:400–16 Bonse U, Nußhardt R. 1986. Elemental
80. Ketai LH, Muggenberg BA, McIntire GL, and chemical state imaging using syn-
Bacon ER, Rosenberg R, et al. 1999. chrotron radiation. Appl. Opt. 25:4583–85
CT imaging of intrathoracic lymph nodes 90. Momose A. 1995. Demonstration of
in dogs with bronchoscopically adminis- phase-contrast X-ray computed tomogra-
trated iodinated nanoparticles. Acad. Ra- phy using an X-ray interferometer. Nucl.
diol. 6:49–54 Instrum. Methods Phys. Res. A 352:622–
81. Kao C-Y, Hoffman EA, Beck KC, Bil- 28
lamkonda RV, Annapragada AV. 2003. 91. Beckman F, Bonse U, Busch F, Gunnewig
Long residence time nano-scale liposomal O. 1997. X-ray microtomography (µCT)
iohexol for X-ray-based blood pool imag- using phase contrast for the investigation
ing. Acad. Radiol. 10:475–83 of organic matter. J. Comput. Assist. To-
82. Williams MB, Guimin Z, More MJ, mogr. 21:539–53
Goode AR, Majewski S, et al. 2000. In- 92. Beckman F, Heise K, Kölsch, Bonse
tegrated CT-SPECT system for small- U, Rajewsky MF, et al. 1999. Three-
animal imaging penetrating radiation sys- dimensional imaging of nerve tissue by X-
tems and application II. Proc. SPIE 4142: ray phase contrast microtomography. Bio-
265–74 phys. J. 76:98–102
83. Cherry SR, Shao Y, Silverman RW, 93. Takeda T, Momose A, Yu Q, Yuasa T, Dil-
Chatziioannou A, Meadors K. 1997. Mi- manian FA, et al. 2002. New types of X-
croPET: a high resolution PET scanner for ray computed tomography (CT) with syn-
imaging small animals. IEEE Trans. Nucl. chrotron radiation, fluorescent X-ray CT
Sci. 44:1161–66 and phase-contrast X-ray CT using inter-
84. Maran A, Khosla S, Riggs BL, Zhang M, ferometer. Cell. Mol. Biol. 46:1077–88
Ritman EL, Turner RT. 2003. Measure- 94. Harding G, Kosanetzky J. 1986. Chemi-
ment of gene expression following cryo- cal imaging with X-ray scatter. Proc. SPIE
genic µ-CT scanning of human iliac crest 671:81–88
 3 Jun 2004 20:53 AR AR220-BE06-08.tex AR220-BE06-08.sgm LaTeX2e(2002/01/18) P1: IKH

208 RITMAN

95. Schlomka JP, Harding A, van Stevendaal characterization of coronary atherosclero-

U, Grass M, Harding G. 2003. Coherent sis by synchrotron radiation X-ray micro-
scatter computed tomography—a novel tomography and highly localized X-ray
medical imaging technique. Proc. SPIE diffraction. Phys. Med. Biol. 47:4345–56
5030:256–65 101. Duvauchelle P, Peix G, Babot D. 1999.
96. Westmore MS, Fenster A, Cunningham Rayleigh to Compton ratio computed to-
IA. 1997. Tomographic imaging of the mography using synchrotron radiation.
angular-dependant coherent scatter cross Proc. SPIE 3772:268–78
section. Med. Phys. 24:3–10 102. Takeda T, Tu Q, Yashiro T, Yuasa T,
97. Wernick MN, Wirjadi O, Chapman D, Hasegawa Y, et al. 1999. Human thyroid
Oltulu O, Zhong Z, Yang Y. 2002. Prelim- specimen imaging by fluorescent X-ray
Annu. Rev. Biomed. Eng. 2004.6:185-208. Downloaded from www.annualreviews.org

inary investigation of a multiple-image ra- computed tomography with synchrotron

Access provided by University of Winnipeg on 01/23/19. For personal use only.

diography method. Proc. IEEE Int. Symp. radiation. Proc. SPIE 3772:258–67
Biomed. Imaging, July 7–10, Washington, 103. Takeda T, Maeda T, Yuasa T, Itai K,
DC, pp. 129–132 Sakamoto K, et al. 1996. Fluorescent
98. Harding G, Kosanetzky J, Neitzel U. scanning X-ray tomographic image with
1987. X-ray diffraction computed tomog- monochromatic synchrotron X-ray. Med.
raphy. Med. Phys. 14(4):515–25 Image. Tech. 14(2):183–94
99. Stock SR, Barss J, Dohl T, Veis A, Almer 104. Hogan JP, Gonsalves RA, Krieger AS.
JD. 2002. X-ray absorption microtomog- 1991. Fluorescent computed tomography:
raphy (micro-CT) and small beam diffrac- a model for correction of X-ray absorp-
tion mapping of sea urchin teeth. J. Instr. tion. IEEE Trans. Nucl. Sci. 38:1721–27
Biol. 139:1–12 105. Chesler DA, Riederer SJ, Pelc NJ. 1977.
100. Jin H, Hm K, Chan JY, Butler LG, Kurtz Noise due to photon counting statistics in
RL, Thiam S, et al. 2002. High resolu- computed X-ray tomography. J. Comput.
tion three-dimensional visualization and Assist. Tomogr. 1:64–77
P1: FRK
June 15, 2004 16:55 Annual Reviews AR220-FM

Annual Review of Biomedical Engineering

Volume 6, 2004

CONTENTS
FRONTISPIECE, Roderic Pettigrew xii
THE NATIONAL INSTITUTE OF BIOMEDICAL IMAGING AND
Annu. Rev. Biomed. Eng. 2004.6:185-208. Downloaded from www.annualreviews.org

BIOENGINEERING, Shu Chien 1

Access provided by University of Winnipeg on 01/23/19. For personal use only.

TISSUE ENGINEERING APPLICATIONS OF THERAPEUTIC CLONING,

Anthony Atala and Chester J. Koh 27
BIOMATERIALS: WHERE WE HAVE BEEN AND WHERE WE ARE GOING,
Buddy D. Ratner and Stephanie J. Bryant 41
TISSUE GROWTH AND REMODELING, Stephen C. Cowin 77
BREAST TISSUE ENGINEERING, Charles W. Patrick, Jr. 109
TISSUE ENGINEERING OF LIGAMENTS, G. Vunjak-Novakovic,
Gregory Altman, Rebecca Horan, and David L. Kaplan 131
ADVANCES IN HIGH-FIELD MAGNETIC RESONANCE IMAGING,
Xiaoping Hu and David G. Norris 157
MICRO-COMPUTED TOMOGRAPHY—CURRENT STATUS AND
DEVELOPMENTS, Erik L. Ritman 185
OPTICAL PROJECTION TOMOGRAPHY, James Sharpe 209
MECHANICAL BIOFFECTS OF ULTRASOUND, Diane Dalecki 229
OCULAR BIOMECHANICS AND BIOTRANSPORT, C. Ross Ethier,
Mark Johnson, and Jeff Ruberti 249
MECHANOTRANSDUCTION AT CELL-MATRIX AND CELL-CELL
CONTACTS, Christopher S. Chen, John Tan, and Joe Tien 275
FUNCTIONAL EFFICACY OF TENDON REPAIR PROCESSES,
David L. Butler, Natalia Juncosa, and Matthew R. Dressler 303
FLUID MECHANICS OF HEART VALVES, Ajit P. Yoganathan, Zhaoming He,
and S. Casey Jones 331
MOLECULAR MACHINES, C. Mavroidis, A. Dubey, and M.L. Yarmush 363
ENGINEERING SYNTHETIC VECTORS FOR IMPROVED DNA DELIVERY:
INSIGHTS FROM INTRACELLULAR PATHWAYS, Charles M. Roth and
Sumati Sundaram 397
FRACTAL ANALYSIS OF THE VASCULAR TREE IN THE HUMAN RETINA,
Barry R. Masters 427

v
P1: FRK
June 15, 2004 16:55 Annual Reviews AR220-FM

vi CONTENTS

ADVANCES IN QUANTITATIVE ELECTROENCEPHALOGRAM ANALYSIS

METHODS, Nitish V. Thakor and Shanbao Tong 453
ROBOTICS, MOTOR LEARNING, AND NEUROLOGIC RECOVERY,
David J. Reinkensmeyer, Jeremy L. Emken, and Steven C. Cramer 497

INDEXES
Subject Index 527
Cumulative Index of Contributing Authors, Volumes 1–6 549
Cumulative Index of Chapter Titles, Volumes 1–6 552
Annu. Rev. Biomed. Eng. 2004.6:185-208. Downloaded from www.annualreviews.org

ERRATA
Access provided by University of Winnipeg on 01/23/19. For personal use only.

An online log of corrections to Annual Review of Biomedical

Engineering chapters may be found at http://bioeng.annualreviews.org/