Cone beam computed tomography (CBCT) systems with rotational gantries that have standard flat panel detectors (FPD) are widely used for the 3D rendering of vascular structures using Feldkamp cone beam reconstruction algorithms. One of the inherent limitations of these systems is limited resolution (<;3 lp/mm). There are systems available with higher resolution but their small FOV limits them to small animal imaging only. In this work, we report on region-of-interest (ROI) CBCT with a high resolution CMOS detector (75 μm pixels, 600 μm HR-CsI) mounted with motorized detector changer on a commercial FPD-based C-arm angiography gantry (194 μm pixels, 600 μm HL-CsI). A cylindrical CT phantom and neuro stents were imaged with both detectors. For each detector a total of 209 images were acquired in a rotational protocol. The technique parameters chosen for the FPD by the imaging system were used for the CMOS detector. The anti-scatter grid was removed and the incident scatter was kept the same for both detectors with identical collimator settings. The FPD images were reconstructed for the 10 cm x10 cm FOV and the CMOS images were reconstructed for a 3.84 cm x 3.84 cm FOV. Although the reconstructed images from the CMOS detector demonstrated comparable contrast to the FPD images, the reconstructed 3D images of the neuro stent clearly showed that the CMOS detector improved delineation of smaller objects such as the stent struts (~70 μm) compared to the FPD. Further development and the potential for substantial clinical impact are suggested.
The linearity between CT numbers and iodine concentrations is proved analytically provided the correlations between
the atoms are negligible. This relationship is applied to correct the CT numbers in the ICA regions in monochromatic
images of dual energy CT with the slow kV-switching technique where one scan with a low/high tube voltage follows by
another scan with a high/low voltage. The iodine concentration may change significantly during the kV-switching. The
resultant CT numbers in ICA regions may not be meaningful in the monochromatic images from pre-reconstruction
decompositions because the data with the low/high voltages are not consistent. Using the linearity between CT numbers
and iodine concentrations, the CT numbers in ICA regions can be corrected by referring to the CT numbers in the
polychromatic images with the low/high voltages. A numerical simulation and a phantom study are performed to
examine the linearity between CT numbers and iodine concentrations. The CT number correction by use of the linearity
is tested in the numerical simulation study in slow kV-switching dual energy CT. The results show that the corrected CT
numbers by use of the linearity are accurate.
A method to remove stents and consequently to eliminate the blooming artifacts off the stents is proposed in dual energy
CT. The method could also reduce the blooming artifacts off calcified plaques. A phantom study is performed to test the
method. The phantom consists of a stainless steel stent and a dumbbell shaped plastic (Delrin) cylinders. With the dual
energy technique and the knowledge of the stent material, we separate the stent from the Delrin at each image voxels
accurately. The large and small diameters of the Delrin are measured from the images by the full width at half maximum
as 2.8 mm and 1.4 mm, respectively. They are very close to the true values of 2.4 mm and 1.2 mm. By respectively
discarding even and odd view data for the low and high voltages, we simulate the fast kV-switching acquisitions where
one view mis-registration exists between the low/high voltage scans. Comparing with the original images by the slow
kV-switching where a perfect registration is realized between the low/high voltage scans, the images from the fast kVswitching
show no significant differences except for the noise pattern.
Fast kV-switching is a dual energy acquisition technique in CT in which alternating views correspond to the low and
high tube voltages. Its high temporal resolution and its suitability to a variety of source trajectories make it an
attractive option for dual energy data acquisition. Its disadvantages include a one view mis-registration between the
data for high and low voltages, the potential for poor spectrum separation because the fast kV-switching waveform
may be more like a sine wave than the desired square wave, and the higher noise in the low voltage data because of the
technical difficulty of swinging the tube current to counter the loss of x-ray production efficiency and loss of
penetration at lower tube voltages. These issues are investigated with a recently developed pre-reconstruction
decomposition method by the authors. Results include that symmetric view matching eliminates streaks from the view
mis-registration, a sinusoidal waveform swinging between 80 and 135 kV gives sufficient spectrum separation, and
that contrast-to-noise for the simulated imaging task maximizes at monochromatic energy of 75 keV.
In this work we apply the circle-and-line acquisition for the 256-detector row medical CT scanner. Reconstruction is based on the exact algorithm of the FBP type suggested recently by one of the co-authors. We derived equations for the cylindrical detector, common for medical CT scanners. To minimize hardware development efforts we use ramp-based reconstruction of the circle data. The line data provides an additional term that corrects the cone beam artifacts that are caused by the incompleteness of the circular trajectory. We illustrate feasibility of our approach using simulated data and real scanned data of the anthropomorphic phantom and evaluate stability of reconstruction to motion and misalignments during the scan. The additional patient dose from the line scan is relatively low compared to the circle scan. The proposed algorithm allows cone beam artifact-free reconstruction with large cone angle.
Sinogram truncation is a common problem in tomographic reconstruction; it occurs when a scanned object or patient extends outside the scan field-of-view. The truncation artifact propagates from the edge of truncation towards the center, resulting in degraded image quality. Several methods have been proposed recently to reconstruct the image artifact-free within the scan FOV; however it is often necessary to recover image outside the scan FOV. We propose a novel truncation correction algorithm that accurately completes unmeasured data outside of the scan field-of-view, which allows us to extend the reconstruction field-of-view. Contrary to 1D extrapolation, we perform interpolation along the so-called sinogram curves. First, we propose an approach to parameterize the family of sinogram curves for efficient sinogram decomposition. Secondly, we propose two ways to estimate the truncated data outside the field-of-view. Both methods are combined for more accurate sinogram completion. Our evaluation shows the validity of our approach. Even objects completely outside the FOV can be accurately reconstructed using the proposed method. The proposed method can be used with any modality where sinogram truncation occurs, such as CT, C-arm, PET/CT, and SPECT.
Multi-slice helical CT-systems suffer from windmill artifacts: black/white patterns that spin off of features with high longitudinal gradients. The number of black/white pairs matches the number of slices (detector rows) in the multi-slive detector. The period of spin is the same as the helical pitch. We investigate the cause of the pattern by following the traces of selected voxels through the multi-slive detector array as a function of view position. This forms an "extracted sinogram" which represents the data used to reconstruct the specific voxel. Now we can determine the cause of the artifact by correlating the windmill streak in the image with the extracted data. The investigation shows that inadequate sampling along the longitudinal direction causes the artifact.
In many clinical applications, it is necessary to tilt the gantry of an X-ray CT system with respect to the patient. Tilting the gantry introduces no complications for single-slice fan-beam systems; however, most systems today are helical multislice systems with up to 16 slices (and this number is sure to increase in the future). The image reconstruction algorithms used in multislice helical CT systems must be modified to compensate for the tilt. If they are not, the quality of reconstructed images will be poor with the presence of significant artifacts produced by the tilt. Practical helical multislice algorithms currently incorporated in today’s systems include helical fan-beam, ASSR (Advanced single-slice rebinning), and Feldkamp algorithms. This paper presents the modifications necessary to compensate for gantry tilt for the helical cone-beam Feldkamp algorithm implemented by Toshiba (referred to as TCOT for true cone-beam tomography). Unlike some of the other algorithms, gantry tilt compensation is simple and straightforward to implement with no significant increase in computational complexity. It will be shown that the effect of the gantry tilt is to introduce a lateral shift in the isocenter of the reconstructed slice of interest, which is a function of the tilt, couch speed, and view angle. This lateral shift is easily calculated and incorporated into the backprojection algorithm. The tilt-compensated algorithm is called T-TCOT. Experimental tilted-gantry data has been obtained with 8- and 16 slice Toshiba Aquilion systems, and examples of uncompensated and tilt compensated images are presented.
A goal of multi-scale CT (MSCT) is to decrease the scan time needed to image a volume of the patient. Detectors with multiple rows of sensors enable higher helical pitches and thus shorter scan times. One of the difficulties in devising an efficient reconstruction method with good image quality and good dose utilization is that each image pixel is irradiated by the cone-beam for a different range of gantry orientations. We derive a new half-scan weighting scheme for a helical, cone-beam backprojection algorithm based on the virtual fan angle. The virtual fan angle, in turn, determines the gantry view range such that image quality is maintained by allowing only valid ray-sums while using the available dose. This restricts the virtual fan angle to be at least the true geometric fan angle but less than 180 degree(s). The result is a computational efficient and dose efficient reconstruction algorithm with a continuous range of field-of-view dependent helical pitches. A 43% higher helical pitch is possible for the smallest field-of-view compared with the largest field-of-view, using the parameters of a commercial MSCT-scanner. Dose efficiency is compared among the new method and standard half scan and full scan approaches.
Small pellets are often used as fiducial markers in a calibration phantom to estimate the geometrical parameters in 3D (three-dimensional) reconstruction. But calibration accuracy depends on the accuracy of locating the pellet centers. Here we describe a technique for fast and accurate detection of these centers. The phantom consists of tungsten carbide pellets arranged in a helical trajectory. The plastic holder mounting the pellets may cause unequal distribution of attenuation around edge pellets compared to the center ones. After log subtraction with flood frames the grayscale gradient in the background is derived within the mask for every point for a reliable background correction. The pellets are identified from the amplitude projections of each frame and a mask is used to refine its position. The grayscale gradient of the background is suitably estimated at each point by the equation of a plane. The center obtained after gradient filter correction is compared with manual measurement, and to measurement using a single background value for each mask. Gradient correction gives centers within 0.3 +/- 0.1 pixel of the manual measurements for the edge pellets, while a single value for background correction yields results within 0.6 +/- 0.3 pixel.
Digital subtraction angiographic image sequences of a calibration phantom are acquired at 30 frames per second from a C-arm gantry covering an angular arc of more than 180 degrees rotating at 40 degrees/s. For each frame, after XRII distortion correction, the relation between the source and its image plane orientation in 3D space is estimated from fiducial markers in the calibration phantom. This gives a mapping between the three-dimensional calibration object and its two- dimensional projection at each gantry angle. We derive eleven mapping coefficients as a function of gantry angle. We use the coefficients to backproject the contribution to any physical voxel. Thus the wobble correction is incorporated directly into cone-beam backprojection. In the absence of gantry wobble, this method is equivalent to the short-scan Feldkamp algorithm, any deviation of the coefficients from those perfect values can be taken as a measure of the gantry wobble. The mapping method requires no special knowledge of the system geometry and any wobble, twisting of the C-arm or XRII during rotation is automatically included. A phantom with tungsten- carbide beads is reconstructed. Accuracy is obtained by comparing reprojections of the center of the tungsten beads with their known values.
A distortion correction table compression method based on polynomial fitting has been developed for implementation in a commercial volume-CT system. To achieve the fastest processing rates, distortion correction tables must fit into the limited memory present in hardware. The number of elements in raw lookup tables is approximately 2 X Ni X Nj X N(theta ), where Ni X Ni is the image dimensions in pixels, and N(theta ) is the number of frames. Two- dimensional (2D) compression fits 4th-order polynomials to columns and rows of the raw table, reducing table size to 2 X 5 X 5 X Nf. Three-dimensional (3D) compression further compresses 2D tables in the angle dimension; reducing table size to 2 X 5 X 5 X 5. Tradeoffs between table size, accuracy, speed, and amount of distortion were investigated with data acquired from 7', 9', 10', 12', 14', and 16' IIs. The mean error was approximately 0.11, 0.20, and 0.20 pixels for raw table, 2D and 3D corrected data; with standard deviations of 0.08, 0.12, and 0.12 pixels.
Near-micrometer resolution, three-dimensional computed tomographic images were made of a test object using the hard x-ray microscope developed by the National Institute of Standards and Technology (NIST). The microscope uses a cooled CCD camera with direct conversion of the incident x rays by a 512 multiplied by 512 chip with 19 micrometer by 19 micrometer cells. Magnification by a factor of 20 is achieved using asymmetric Bragg diffraction from a pair of silicon crystals. The imaging system is designed for samples of the order of 0.50 mm diameter by 0.50 mm height. From beamline X23A3 at the National Synchrotron Light Source (NSLS), Brookhaven National Laboratory (BNL) 8.17 keV x rays were used. Two hundred, 512 multiplied by 512 two-dimensional projections were collected every 0.9 degrees about the test object using the NIST microscope. The projections were digitized and sent to a computer for volume tomographic reconstruction by a parallel-beam, convolution-backprojection algorithm into a 5123 image with (1 micrometer)3 voxels. The test object consisted of glass and nickel microspheres with distributions from about 4 t 40 micrometer (glass) or to 24 micrometer (nickel) diameters suspended in epoxy in order to demonstrate near one micrometer resolution in all three dimensions and probe contrast sensitivity. The effect and interplay of photon statistics and energy, and sample composition, density and size on tomographic performance are discussed as are resolution limitations and image artifacts from Fresnel diffraction.
High energy photon backscatter uses pair production to probe deep beneath surfaces with single side accessibility or to image thick, radiographically opaque objects. At the higher photon energies needed to penetrate thick and/or highly attenuating objects, Compton backscatter becomes strongly forward peaked with relatively little backscatter flux. Furthermore, the downward energy shift of the backscattered photon makes it more susceptible to attenuation on its outbound path. Above 1.022 MeV, pair production is possible; at about 10 MeV, pari production crosses over Compton scatter as the dominant x-ray interaction mechanism. The backscattered photons can be hard x rays from the bremsstrahlung of the electrons and positrons or 0.511 MeV photons from the annihilation of the positron. Monte Carlo computer simulations of such a backscatter system were done to characterize the output signals and to optimize a high energy detector design. This paper touches on the physics of high energy backscatter imaging and describes at some length the detector design for tomographic and radiographic imaging.
We interface a Toshiba LX-40A radiation therapy simulator with a custom designed data acquisition system and a 386-PC. The LX-40A includes a 14 inch (35.6 cm) x-ray image intensifier and a 1 inch saticon camera. Two-dimensional cone beam projections of an anthropomorphic chest/lung phantom are collected. The Feldkamp algorithm reconstructs the volume data set into a volume image of the phantom. We discuss several volume computed tomography issues including spatial distortions in the x-ray image intensifier, scatter, and veiling glare. We compare the volume computed tomography results with two-dimensional CT imaging of the same phantom scanned on the equipment.
The authors use a Toshiba LX-40A Radiation Therapy Simulator with a 14' (35.6 cm) image intensifier and a 1' saticon camera to collect data for computed tomography (CT) imaging. The custom designed data acquisition system is interfaced with a 386-PC and the LX-40A to allow the LX-40A to perform as a CT scanner under PC control. The motion versatility of the simulator allows fields of view (FOV) greater than 40 cm with a single 360 degree(s) rotation of the gantry. Distortion and other corrections are applied to give relatively artifact-free images. A visible resolution performance of 7 lp/cm is obtained throughout the FOV. One percent contrast targets are visible down to 3 mm for head-sized objects, though contrast sensitivity depends, of course, on many scan parameters. An objective of this research is to give CT functionality to a radiation therapy simulator; this eliminates the need for a conventional, diagnostic CT scanner for radiation therapy planning. Extensions to multislice and volume CT are possible.