This PDF file contains the front matter associated with SPIE Proceedings Volume 8313, including the Title Page, Copyright information, Table of Contents, Introduction, an Overview Paper of the 40 year history of this meeting, and the Conference Committee listing.

X-ray scatter is a common cause of image artifacts for cone-beam CT systems due to the expanded field of view and degrades the quantitative accuracy of measured Hounsfield Units (HU). Due to the strong dependency of scatter on the object being scanned, it is crucial to measure the scatter signal for each object. We propose to use a beam pass array (BPA) composed of parallel-holes within a tungsten plate to measure scatter for a dedicated breast CT system. A complete study of the performance of the BPA was conducted. The goal of this study was to explore the feasibility of measuring and compensating for the scatter signal for each individual object. Different clinical study schemes were investigated, including a full rotation scan with BPA and discrete projections acquired with BPA followed by interpolation for full rotation. Different sized cylindrical phantoms and a breast shaped polyethylene phantom were used to test for the robustness of the proposed method. Physically measured scatter signals were converted into scatter to primary ratios (SPRs) at discrete locations through the projection image. A complete noise-free 2D SPR was generated from these discrete measurements. SPR results were compared to Monte Carlo simulation results and scatter corrected CT images were quantitatively evaluated for "cupping" artifact. With the proposed method, a reduction of up to 47 HU of "cupping" was demonstrated. In conclusion, the proposed BPA method demonstrated effective and accurate objectspecific scatter correction with the main advantage of dose-sparing compared to beam stop array (BSA) approaches.

A system using a wide slot beam and simple anti-scatter grid has been designed to provide a localized map of tissue type that could be overlaid on the simultaneous conventional transmission image to provide an inexpensive, low dose adjunct to conventional screening mammography. The system was demonstrated to differentiate between scatter peak angles corresponding to adipose tissue and carcinoma. Adequate intensity in the coherent scatter image can be achieved at a dose commonly used for screening mammography. Depth information is obtainable from the stereoscopic viewing angles. Phantom imaging measurements and Monte Carlo simulations show good agreement.

Digital breast tomosynthesis (DBT) is a promising alternative approach to overcome the limitations of tissue superposition found in full-field 2D digital mammography. However, due to the absence of anti-scatter grids in DBT, accurate scatter estimation for each projection is necessary for modelling the image reconstruction stage. In this work we identify the limitations associated with scatter estimation using spatial invariant scatter kernels, in particular at the edge region where such methods result in scatter overestimation. Such approaches show an overestimation of scatter-to-primary ratio of over 50% at the edges when compared with results from direct Monte Carlo simulation. This problem was found to increase with projection angle. Simulation work presented here shows that this overestimation in scatter is largely due to air gap between the lower curved breast edge and the detector. We propose a new fast, accurate scatter field estimator for use in DBT which not only considers the breast thickness and primary incidence angle, but also accounts for scatter exiting the breast edge region and traversing an air gap prior to absorption in the detector. The new proposed scatter estimator represents an alternative approach to this problem which reduces discrepancies at the edge of a breast phantom. Moreover, the time required for generating scatter has dropped from approximately 12 hours using Monte Carlo simulations for 1010 photons to just a few minutes per projection. The insertion of scatter from the compression paddle to aforementioned methodologies is also discussed.

In digital breast tomosynthesis (DBT), a reconstruction of the breast is generated from projections acquired over a limited range of x-ray tube angles. There are two principal schemes for acquiring projections, continuous tube motion and step-and-shoot motion. Although continuous tube motion has the benefit of reducing patient motion by lowering scan time, it has the drawback of introducing blurring artifacts due to focal spot motion. The purpose of this work is to determine the optimal scan time which minimizes this trade-off. To this end, the filtered backprojection reconstruction of a sinusoidal input is calculated. At various frequencies, the optimal scan time is determined by the value which maximizes the modulation of the reconstruction. Although prior authors have studied the dependency of the modulation on focal spot motion, this work is unique in also modeling patient motion. It is shown that because continuous tube motion and patient motion have competing influences on whether scan time should be long or short, the modulation is maximized by an intermediate scan time. This optimal scan time decreases with object velocity and increases with exposure time. To optimize step-and-shoot motion, we calculate the scan time for which the modulation attains the maximum value achievable in a comparable system with continuous tube motion. This scan time provides a threshold below which the benefits of step-and-shoot motion are justified. In conclusion, this work optimizes scan time in DBT systems with patient motion and either continuous tube motion or step-and-shoot motion by maximizing the modulation of the reconstruction.

The stationary Digital Breast Tomosynthesis System (s-DBT) has the advantage over the conventional DBT systems as there is no motion blurring in the projection images associated with the x-ray source motion. We have developed a prototype s-DBT system by retrofitting a Hologic Selenia Dimensions rotating gantry tomosynthesis system with a distributed carbon nanotube (CNT) x-ray source array. The linear array consists of 31 x-ray generating focal spots distributed over a 30 degree angle. Each x-ray beam can be electronically activated allowing the flexibility and easy implementation of novel tomosynthesis scanning with different scanning parameters and configurations. Here we report the initial results of investigation on the imaging quality of the s-DBT system and its dependence on the acquisition parameters including the number of projections views, the total angular span of the projection views, the dose distribution between different projections, and the total dose. A mammography phantom is used to visually assess image quality. The modulation transfer function (MTF) of a line wire phantom is used to evaluate the system spatial resolution. For s-DBT the in-plan system resolution, as measured by the MTF, does not change for different configurations. This is in contrast to rotating gantry DBT systems, where the MTF degrades for increased angular span due to increased focal spot blurring associated with the x-ray source motion. The overall image quality factor, a composite measure of the signal difference to noise ratio (SdNR) for mass detection and the z-axis artifact spread function for microcalcification detection, is best for the configuration with a large angular span, an intermediate number of projection views, and an even dose distribution. These results suggest possible directions for further improvement of s-DBT systems for high quality breast cancer imaging.

Dual-energy contrast-enhanced digital breast tomosynthesis (DE CE-DBT) image quality is affected by a large parameter space including the tomosynthesis acquisition geometry, imaging technique factors, the choice of reconstruction algorithm, and the subject breast characteristics. The influence of most of these factors on reconstructed image quality is well understood for DBT. However, due to the contrast agent uptake kinetics in CE imaging, the subject breast characteristics change over time, presenting a challenge for optimization . In this work we experimentally evaluate the sensitivity of the reconstructed image quality to timing of the low-energy and high-energy images and changes in iodine concentration during image acquisition. For four contrast uptake patterns, a variety of acquisition protocols were tested with different timing and geometry. The influence of the choice of reconstruction algorithm (SART or FBP) was also assessed. Image quality was evaluated in terms of the lesion signal-difference-to-noise ratio (LSDNR) in the central slice of DE CE-DBT reconstructions. Results suggest that for maximum image quality, the low- and high-energy image acquisitions should be made within one x-ray tube sweep, as separate low- and high-energy tube sweeps can degrade LSDNR. In terms of LSDNR per square-root dose, the image quality is nearly equal between SART reconstructions with 9 and 15 angular views, but using fewer angular views can result in a significant improvement in the quantitative accuracy of the reconstructions due to the shorter imaging time interval.

Digital breast tomosynthesis (DBT) is a novel x-ray imaging technique that provides 3D structural information of the breast. In contrast to 2D mammography, DBT minimizes tissue overlap potentially improving cancer detection and reducing number of unnecessary recalls. The addition of a contrast agent to DBT and mammography for lesion enhancement has the benefit of providing functional information of a lesion, as lesion contrast uptake and washout patterns may help differentiate between benign and malignant tumors. This study used a task-based method to determine the optimal imaging approach by analyzing six imaging paradigms in terms of their ability to resolve iodine at a given dose: contrast enhanced mammography and tomosynthesis, temporal subtraction mammography and tomosynthesis, and dual energy subtraction mammography and tomosynthesis. Imaging performance was characterized using a detectability index d', derived from the system task transfer function (TTF), an imaging task, iodine contrast, and the noise power spectrum (NPS). The task modeled a 5 mm lesion containing iodine concentrations between 2.1 mg/cc and 8.6 mg/cc. TTF was obtained using an edge phantom, and the NPS was measured over several exposure levels, energies, and target-filter combinations. Using a structured CIRS phantom, d' was generated as a function of dose and iodine concentration. In general, higher dose gave higher d', but for the lowest iodine concentration and lowest dose, dual energy subtraction tomosynthesis and temporal subtraction tomosynthesis demonstrated the highest performance.

Digital breast tomosynthesis (DBT) is a three-dimensional (3D) x-ray imaging modality that has recently been employed to increase lesion conspicuity through the removal of overlying tissue. Recently, a great deal of work has been devoted to the development of contrast enhanced (CE) DBT. Radio-opaque contrast agents (e.g. iodine) are injected into patients with suspicious breast lesions, with the goal of differentiating malignant tumors from benign by imaging the contrast uptake signature associated with angiogenesis. Either temporal subtraction (TS) or dual energy (DE) subtraction may be performed to further remove structural noise from the images. The current work quantifies the change in power-law noise after either DE subtraction or TS using structured breast tissue equivalent phantoms. Additionally, iodine contrast filled phantoms were used to determine the effect of x-ray energy and image subtraction technique on the signaldifference- to-noise ratio (SDNR). Finally, we investigate the improvement in imaging performance of an amorphous selenium (a-Se) direct conversion flat panel detector with increased a-Se thickness.

Breast CT (BCT) using a photon counting detector (PCD) has a number of advantages that can potentially improve clinical performance. Previous computer simulation studies showed that the signal to noise ratio (SNR) for microcalcifications is higher with energy weighted photon counting BCT as compared to cesium iodide energy integrating detector (CsI-EID) based BCT. CsI-EID inherently weighs the incident x-ray photons in direct proportion to the energy (contradicting the information content) which is not an optimal approach. PCD do not inherently weigh the incident photons. By choosing optimal energy weights, higher SNR can be achieved for microcalcifications and hence better detectability. In this simulation study, forward projection data of a numerical breast phantom with microcalcifications inserted were acquired using CsI-EID and PCD. The PCD projections were optimally weighed, and reconstructed using filtered back-projection. We compared observer performance in identifying microcalcifications in the reconstructed images using ROC analysis. ROC based results show that the average area(s) under curve(s) (AUC) for AUC_PCD based methods are higher than the average AUC_CsI-EID method.

Mammography is currently the most widely accepted tool for detection and diagnosis of breast cancer. However, the sensitivity of mammography is reduced in women with dense breast tissue due to tissue overlap, which may obscure lesions. Digital breast tomosynthesis with contrast enhancement reduces tissue overlap and provides additional functional information about lesions (i.e. morphology and kinetics), which in turn may improve lesion characterization. The performance of such techniques is highly dependent on the structural composition of the breast, which varies significantly across patients. Therefore, optimization of breast imaging systems should be done with respect to this patient versatility. Furthermore, imaging techniques that employ contrast require the inclusion of a temporally varying breast composition with respect to the contrast agent kinetics to enable the optimization of the system. To these ends, we have developed a dynamic 4D anthropomorphic breast phantom, which can be used for optimizing a breast imaging system by incorporating material characteristics. The presented dynamic phantom is based on two recently developed anthropomorphic breast phantoms, which can be representative of a whole population through their randomized anatomical feature generation and various compression levels. The 4D dynamic phantom is incorporated with the kinetics of contrast agent uptake in different tissues and can realistically model benign and malignant lesions. To demonstrate the utility of the proposed dynamic phantom, contrast-enhanced digital mammography and breast tomosynthesis were simulated where a ray-tracing algorithm emulated the projections, a filtered back projection algorithm was used for reconstruction, and dual-energy and temporal subtractions were performed and compared.

This paper proposes a new method to improve contrast of a mammogram using multi-energy x-ray (MEX) images. The x-ray attenuation differences among breast tissues increase as incident photons have lower energy. Thus an image obtained by a narrow low energy spectrum has higher contrast than a full (wide) energy spectrum image. The proposed mammogram enhancement utilizes this fact using MEX images. Lowpass data of a low energy spectrum image and high frequency components of a wide energy spectrum image are combined to have high contrast and low noise. Nonsubsampled contourlet transform (NSCT) is employed to decompose image data into multi-scale and multidirectional information. The NSCT overcomes the shortage of directions of wavelet transform by expressing smoothness along contours sufficiently. The outcome of the transform is a lowpass subband and multiple bandpass directional subbands. First, the lowpass subband coefficients of a wide energy spectrum image are substituted by those of a low energy spectrum image. Before the coefficient modification, the low energy spectrum image is processed to have high contrast and sharp details. Next, for the bandpass directional subbands, the locally adaptive bivariate shrinkage of contourlet coefficients is applied to suppress noise. The bivariate shrinkage function exploits interscale dependency of coefficients. Local contrast of the resultant mammogram is considerably enhanced and shows clear fibroglandular tissue structures. Experimental results illustrate that the proposed method produces a high contrast and low noise level image, as compared to the conventional mammography based on a single energy spectrum image.

X-ray scatter leads to erroneous calculations of dual-energy digital mammography (DEDM). The purpose of this work is to design an algorithmic method for scatter correction in DEDM without extra exposures or lead sheet. The method was developed based on the knowledge that scatter radiation in mammograms varies slowly spatially and most pixels in mammograms are non-calcification pixels, and implemented on a commercial full-field digital mammography system with a phantom of breast tissue equivalent material. The pinhole-array interpolation scatter correction method was also implemented on the system. We compared the background dual-energy (DE) calcification signals in the DE calcification images. Results show that the background signal in the DE calcification image can be reduced. The rms of background DE calcification image signal of 1105μm with scatter-uncorrected data was reduced to 187μm and 253μm after scatter correction, using our algorithmic method and pinhole-array interpolation method, respectively. The range of background DE calcification signals using scatter-uncorrected data was reduced by ~80% with scatter-corrected data using algorithmic method. The proposed algorithmic scatter correction method is effective; it has similar or even better performance than pinhole-array interpolation method in scatter correction for DEDM.

Phase-contrast imaging is an emerging technology that may increase the signal-difference-to-noise ratio in medical imaging. One of the most promising phase-contrast techniques is Talbot interferometry, which, combined with energy-sensitive photon-counting detectors, enables spectral differential phase-contrast mammography. We have evaluated a realistic system based on this technique by cascaded-systems analysis and with a task-dependent ideal-observer detectability index as a figure-of-merit. Beam-propagation simulations were used for validation and illustration of the analytical framework. Differential phase contrast improved detectability compared to absorption contrast, in particular for fine tumor structures. This result was supported by images of human mastectomy samples that were acquired with a conventional detector. The optimal incident energy was higher in differential phase contrast than in absorption contrast when disregarding the setup design energy. Further, optimal weighting of the transmitted spectrum was found to have a weaker energy dependence than for absorption contrast. Taking the design energy into account yielded a superimposed maximum on both detectability as a function of incident energy, and on optimal weighting. Spectral material decomposition was not facilitated by phase contrast, but phase information may be used instead of spectral information.

Active matrix flat-panel imagers (AMFPIs) have become ubiquitous across a wide variety of medical x-ray imaging applications. While AMFPIs based on both direct and indirect detection of the incident radiation offer many advantages, their performance is limited by relatively modest system gain compared to electronic additive noise. The effects of this limitation upon imaging performance become particularly apparent at lower exposures and/or at higher spatial frequencies. One potential strategy for overcoming this limitation involves the use of a high gain photoconductor such as mercuric iodide (HgI₂) which has the potential to improve system gain by up to an order of magnitude compared to that provided by a-Se, the only photoconductor material presently used in direct detection AMFPIs. In this paper, preliminary results from an investigation of the performance of a prototype direct detection AMFPI employing a relatively thin layer of polycrystalline HgI₂ created through a screen-printing method and operated under mammographic irradiation conditions are presented. The results encourage further examination of this strategy to improve the performance of AMFPIs designed for mammography.

Effective detective quantum efficiency (eDQE) and effective noise equivalent quanta (eNEQ) were recently introduced to broaden the notion of DQE and NEQ by including system parameters such as focus blurring and system scatter rejection methods. This work investigates eDQE and eNEQ normalized for mean glandular dose (eNEQ_MGD) as a means to characterize and select optimal exposure parameters for a digital mammographic system. The eDQE was measured for three anode/filter combinations, with and without anti-scatter grid and for four thicknesses of poly(methylmethacrylate) (PMMA). The modulation transfer function used to calculate eDQE and eNEQ was measured from an edge positioned at 20,40,60,70 mm above the table top without scattering material in the beam. The grid-in eDQE results for all A/F settings were generally larger than those for grid-out. Contrarily, the eNEQ_MGD results were higher for grid-out than gridin, with a maximum difference of 61% among all A/F combinations and PMMA thicknesses. The W/Rh combination gave the highest eNEQMGD for all PMMA thicknesses compared to the other A/F combinations (for grid-in and grid-out), supporting the results of alternative methods (e.g. the signal difference to noise ratio method). The eNEQ_MGD was then multiplied with the contrast obtained from a 0.2mm Al square, resulting in a normalized quantity that was higher for the W/Rh combination than for the other A/F combinations. In particular, the results for the W/Rh combination were greater for the grid-in case. Furthermore, these results showed close agreement with a non-prewhitened match filter with eye response model observer (d') normalized for MGD.

We present a novel method for characterizing mammographic findings using spectral imaging without the use of contrast agent. Within a statistical framework, suspicious findings are analyzed to determine if they are likely to be benign cystic lesions or malignant tissue. To evaluate the method, we have designed a phantom where combinations of different tissue types are realized by decomposition into the material bases aluminum and polyethylene. The results indicate that the lesion size limit for reliable characterization is below 10 mm diameter, when quantum noise is the only considered source of uncertainty. Furthermore, preliminary results using clinical images are encouraging, but allow no conclusions with significance.

European Guidelines for quality control in digital mammography specify acceptable and achievable standards of image quality (IQ) in terms of threshold gold thickness using the CDMAM test object. However, there is little evidence relating such measurements to cancer detection. This work investigated the relationship between calcification detection and threshold gold thickness. An observer study was performed using a set of 162 amorphous selenium direct digital (DR) detector images (81 no cancer and 81 with 1-3 inserted calcification clusters). From these images four additional IQs were simulated: different digital detectors (computed radiography (CR) and DR) and dose levels. Seven observers marked and rated the locations of suspicious regions. DBM analysis of variances was performed on the JAFROC figure of merit (FoM) yielding 95% confidence intervals for IQ pairs. Automated threshold gold thickness (T_g) analysis was performed for the 0.25mm gold disc diameter on CDMAM images at the same IQs (16 images per IQ). T_g was plotted against FoM and a power law fitted to the data. There was a significant reduction in FoM for calcification detection for CR images compared with DR; FoM decreased from 0.83 to 0.63 (p≤0.0001). Detection was also sensitive to dose. There was a good correlation between FoM and T_g (R²=0.80, p<0.05), consequently threshold gold thickness was a good predictor of calcification detection at the same IQ. Since the majority of threshold gold thicknesses for the various IQs were above the acceptable standard despite large variations in calcification detection by radiologists, current EU guidelines may need revising.

Breast percent density (PD%), as measured mammographically, is one of the strongest known risk factors for breast cancer. While the majority of studies to date have focused on PD% assessment from digitized film mammograms, digital mammography (DM) is becoming increasingly common, and allows for direct PD% assessment at the time of imaging. This work investigates the accuracy of a generalized linear model-based (GLM) estimation of PD% from raw and postprocessed digital mammograms, utilizing image acquisition physics, patient characteristics and gray-level intensity features of the specific image. The model is trained in a leave-one-woman-out fashion on a series of 81 cases for which bilateral, mediolateral-oblique DM images were available in both raw and post-processed format. Baseline continuous and categorical density estimates were provided by a trained breast-imaging radiologist. Regression analysis is performed and Pearson's correlation, r, and Cohen's kappa, κ, are computed. The GLM PD% estimation model performed well on both processed (r=0.89, p<0.001) and raw (r=0.75, p<0.001) images. Model agreement with radiologist assigned density categories was also high for processed (κ=0.79, p<0.001) and raw (κ=0.76, p<0.001) images. Model-based prediction of breast PD% could allow for a reproducible estimation of breast density, providing a rapid risk assessment tool for clinical practice.

The aim of the present work was to develop a method for simulating breast lesions in digital mammographic images. Based on the visual appearance of real masses, three dimensional masses were created using a 3D random walk method where the choice of parameters (number of walks and number of steps) enables one to control the appearance of the simulated structure. This work is the first occasion that the random walk results have been combined with a model of digital mammographic imaging systems. This model takes into account appropriate physical image acquisition processes representing a particular digital X-ray mammography system. The X-ray spectrum, local glandularity above the insertion site and scatter were all taken account during the insertion procedure. A preliminary observer study was used to validate the realism of the masses. Seven expert readers each viewed 60 full field mammograms and rated the realism of the masses they contained. Half of the images contained real, histologically-confirmed masses, and half contained simulated lesions. The ROC analysis of the study (average AUC of 0.58±0.06) suggests that, on the average, there is evidence that the radiologists could distinguish, somewhat, between real and simulated masses.

The diagnostic quality of phase-contrast computed tomography (PCCT) is one the unexplored areas in medical imaging; at the same time, it seems to offer the opportunity as a fast and highly sensitive diagnostic tool. Conventional computed tomography (CT) has had an enormous impact on medicine, while it is limited in soft-tissue contrast. One example that portrays this challenge is the differentiation between benign and malignant renal cysts. In this work we report on a feasibility study to determine the usefulness of PCCT in differentiation of renal cysts. A renal phantom was imaged with a grating-based PCCT system consisting of a standard rotating anode x-ray tube (40 kV, 70 mA) and a Pilatus II photoncounting detector (pixel size: 172 μm). The phantom is composed of a renal equivalent soft-tissue and cystic lesions grouped in non-enhancing cyst and hemorrhage series and an iodine enhancing series. The acquired projection images (absorption and phase-contrast) are reconstructed with a standard filtered backprojection algorithm. For evaluation both reconstructions are compared in respect to contrast-to-noise ratio (CNR), signal-to-noise ratio (SNR), and subjective image quality. We found that with PCCT a significantly improved differentiation between hemorrhage renal cysts from contrast enhancing malignant cysts is possible. If comparing PCCT and CT with respect to CNR and SNR, PCCT shows significant improvements. In conclusion, PCCT has the potential to improve the diagnostics and characterization of renal cysts without using any contrast agents. These results in combination with a non-synchrotron setup indicate a future paradigm shift in diagnostic computed tomography.

X-ray tomography of small animals is an important tool for medical research. For high-resolution x-ray imaging of few-cm-thick samples such as, e.g., mice, high-brightness x-ray sources with energies in the few-10-keV range are required. In this paper we perform the first small-animal imaging and tomography experiments using liquid-metal-jet-anode x-ray sources. This type of source shows promise to increase the brightness of microfocus x-ray systems, but present sources are typically optimized for an energy of 9 keV. Here we describe the details of a high-brightness 24-keV electron-impact laboratory microfocus x-ray source based on continuous operation of a heated liquid-In/Ga-jet anode. The source normally operates with 40 W of electron-beam power focused onto the metal jet, producing a 7×7 μm² FWHM x-ray spot. The peak spectral brightness is 4 × 10⁹ photons / ( s × mm² × mrad² × 0.1%BW) at the 24.2 keV In K_α line. We use the new In/Ga source and an existing Ga/In/Sn source for high-resolution imaging and tomography of mice.

Carbon nanotube (CNT) micro-focus x-ray tubes have been demonstrated as a novel technology for in-vivo small animal imaging. It enables simultaneous respiratory and cardiac gated prospective CT imaging of free breathing animals with high temporal resolution. Operating the micro-focus CNT x-ray source at high power is required to achieve high temporal resolution. The thermal loading of the anode focal spot is a limiting factor in determining the maximum power of an x-ray tube. In this paper, we developed a reliable simulation model to quantitatively analyze the anode heat load of the CNT x-ray source operating in both DC and pulse modes. The anode temperature distribution is simulated using finite element analysis. The model is validated by comparing simulation results for the micro-focus x- ray tube with reported experimental results. We investigated the relationship between the maximum power and the effective focal spot size for CNT micro-CT system running in both DC and pulse modes. Our results show that when operating in pulse mode, the maximum power of the CNT x-ray source can be significantly higher than when operating in DC mode. In DC mode, we found that the maximum power scales non-linearly with the effective focal spot size as P(in W) = (0.25/ sin θ+1.6)f^0.73 _s (in μm), where 1/sin θ is the projection factor for a given anode angle θ. However, in pulse mode the maximum power linearly increases with the effective focal spot size asP(in W) = (0.20/ sin θ+0.35)f_s(in μm), and is significantly higher than that in the DC mode. This implies that it is feasible to improve the micro-CT temporal resolution further without sacrificing the image quality. The simulation method developed here also enables us to analyze the thermal loading of the other CNT x-ray sources for other applications, such as the stationary digital breast tomosynthesis scanner and the CNT microbeam radiation therapy system.

Computerized phantoms are finding an increasingly important role in medical imaging research. With the ability to simulate various imaging conditions, they offer a practical means with which to quantitatively evaluate and improve imaging devices and techniques. This is especially true in CT due to the high radiation levels involved with it. Despite their utility, due to the time required to develop them, only a handful of computational models currently exist of varying detail. Most phantoms available are limited to 3D and not capable of modeling patient motion. We have previously developed a technique to rapidly create highly detailed 4D extended cardiac-torso (XCAT) phantoms based on patient CT data [1]. In this study, we utilize this technique to generate 58 new adult XCAT phantoms to be added to our growing library of virtual patients available for imaging research. These computerized patients provide a valuable tool for investigating imaging devices and the effects of anatomy and motion in imaging. They also provide the essential tools to investigate patient-specific dose estimation and optimization for adults undergoing CT procedures.

Phantom equivalents of different human anatomical parts are routinely used for imaging system evaluation or dose calculations. The various recommendations on the generic phantom structure given by organizations such as the AAPM, are not always accurate when evaluating a very specific task. When we compared the AAPM head phantom containing 3 mm of aluminum to actual neuro-endovascular image guided interventions (neuro-EIGI) occurring in the Circle of Willis, we found that the system automatic exposure rate control (AERC) significantly underestimated the x-ray parameter selection. To build a more accurate phantom for neuro-EIGI, we reevaluated the amount of aluminum which must be included in the phantom. Human skulls were imaged at different angles, using various angiographic exposures, at kV's relevant to neuro-angiography. An aluminum step wedge was also imaged under identical conditions, and a correlation between the gray values of the imaged skulls and those of the aluminum step thicknesses was established. The average equivalent aluminum thickness for the skull samples for frontal projections in the Circle of Willis region was found to be about 13 mm. The results showed no significant changes in the average equivalent aluminum thickness with kV or mAs variation. When a uniform phantom using 13 mm aluminum and 15 cm acrylic was compared with an anthropomorphic head phantom the x-ray parameters selected by the AERC system were practically identical. These new findings indicate that for this specific task, the amount of aluminum included in the head equivalent must be increased substantially from 3 mm to a value of 13 mm.

X-ray fluorescence molecular imaging (XFMI) can be a potential alternative to existing molecular imaging modalities providing high sensitivity and good spatial resolution. However, high sensitivity at a few tens of μg/mL can be reached only with monochromatic synchrotron x-rays; in typical laboratory setting using conventional x-ray sources XFMI has been reported to reach only about 10 mg/mL sensitivity. In this paper, we demonstrated the feasibility of simultaneously detecting x-ray fluorescence signals from a mouse-sized object containing iodine and indium solutions at 50 μg/mL concentration using a carbon nanotube (CNT) x-ray source. XFMI has the high potential to provide molecular imaging capability in small-animal models with high sensitivity, high spatial and temporal resolution, high multiplexing capacity, and at low radiation dose.

X-ray Luminescence CT (XLCT) is a hybrid imaging modality combining x-ray and optical imaging in which x-ray luminescent nanophosphors (NPs) are used as emissive imaging probes. NPs are easily excited using common CT energy x-ray beams, and the NP luminescence is efficiently collected using sensitive light-based detection systems. XLCT can be recognized as a close analog to fluorescence diffuse optical tomography (FDOT). However, XLCT has remarkable advantages over FDOT due to the substantial excitation penetration depths provided by x-rays relative to laser light sources, long-term photo-stability of NPs, and the ability to tune NP emission within the NIR spectral window. Since XCLT uses an x-ray pencil beam excitation, the emitted light can be measured and back-projected along the x-ray path during reconstruction, where the size of the x-ray pencil beam determines the resolution for XLCT. In addition, no background signal competes with NP luminescence (i.e., no auto fluorescence) in XLCT. Currently, no small animal XLCT system has been proposed or tested. This paper investigates an XLCT system built and integrated with a dual source micro-CT system. A novel sampling paradigms that results in more efficient scanning is proposed and tested via simulations. Our preliminary experimental results in phantoms indicate that a basic CT-like reconstruction is able to recover a map of the NP locations and differences in NP concentrations. With the proposed dual source system and faster scanning approaches, XLCT has the potential to revolutionize molecular imaging in preclinical studies.

Background: Micro-computed tomography offers numerous advantages for small animal imaging, including the ability to monitor the same animals throughout a longitudinal study. However, concerns are often raised regarding the effects of x-ray dose accumulated over the course of the experiment. In this study, we scan C57BL/6 mice multiple times per week for six weeks, to determine the effect of the cumulative dose on pulmonary tissue at the end of the study. Methods/Results: C57BL/6 male mice were split into two groups (irradiated group=10, control group=10). The irradiated group was scanned (80kVp/50mA) each week for 6 weeks; the weekly scan session had three scans. This resulted in a weekly dose of 0.84 Gy, and a total study dose of 5.04 Gy. The control group was scanned on the final week. Scans from weeks 1 and 6 were reconstructed and analyzed: overall, there was no significant difference in lung volume or lung density between the control group and the irradiated group. Similarly, there were no significant differences between the week 1 and week 6 scans in the irradiated group. Histological samples taken from excised lung tissue also showed no evidence of inflammation or fibrosis in the irradiated group. Conclusion: This study demonstrates that a 5 Gy x-ray dose accumulated over six weeks during a longitudinal micro-CT study has no significant effects on the pulmonary tissue of C57BL/6 mice. As a result, the many advantages of micro- CT imaging, including rapid acquisition of high-resolution, isotropic images in free-breathing mice, can be taken advantage of in longitudinal studies without concern for negative dose-related effects.

We are investigating a liquid xenon (LXe) gamma ray detector-based PET system for small animals. The proposed system consists of 12 modules arranged into a ring. Each detector module is a trapezoidal LXe time projection chamber (TPC) viewed by two arrays of large area avalanche photodiodes (LAAPD). We developed a Geant4-based Monte Carlo code to model the LXe PET system and study its imaging performance. Events depositing energy in multiple locations were reconstructed with a Compton reconstruction algorithm. The simulated data were stored in a list-mode file and reconstructed with the maximum likelihood expectation maximization iterative algorithm (MLEM). Simulation results indicate an absolute sensitivity at the center of the field of view (FOV) of 12.6% and a 3D position resolution ≤ 0.8 mm (FWHM) throughout the FOV. A simulated image of a micro-Derenzo phantom shows that rods of 0.6 mm diameter are visible.

Dual energy imaging enables material decomposition and requires attenuation measurements with at least two different energies. Today's clinically available implementations use two separate exposures, at a low and high tube voltage (dual kV). Photon counting x-ray detectors (PCXDs) are an alternative technology that takes advantage of an x-ray source's broad spectrum by counting the number of transmitted photons at each energy from a single exposure. The richness of the information contained in these measurements can depend heavily on the detector's energy response, itself dependent on count rate. We compare the material decomposition precision of dual kV with energy integrating detectors to that of realistic PCXDs. We model the three primary effects that degrade PCXD performance: count rate limitations, energy resolution, and spectrum tailing. For example, the high flux rates required for clinical imaging pose a serious challenge for PCXDs. The Cram´er-Rao Lower Bound is used to predict the best possible material decomposition variance as a function of these detector imperfections. For dual kV and photon counting, we determined the optimal kV, mAs, and filtration for a broad range of imaging tasks, subject to dose and tube power constraints. We found that a well-optimized dual kV protocol performs on par with the estimated performance of today's PCXDs. For dual kV protocols, it is helpful to increase the energy separation between the spectra by increasing the kV separation and adding filtration. For PCXDs, the detector's count rate capabilities must be increased and the spectrum tailing reduced for photon counting to become a competitive technology at the high intensities of clinical imaging.

μWe introduce a novel hybrid prototype scanner built to explore benefits of the quantum-counting technique in the context of clinical CT. The scanner is equipped with two measurement systems. One is a CdTe-based counting detector with 22cm field-of-view. Its revised ASIC architecture allows configuration of the counter thresholds of the 225m small sub-pixels in chess patterns, enabling data acquisition in four energy bins or studying high-flux scenarios with pile-up trigger. The other one is a conventional GOS-based energy-integrating detector from a clinical CT scanner. The integration of both detection technologies in one CT scanner provides two major advantages. It allows direct comparison of image quality and contrast reproduction as well as instantaneous quantification of the relative dose usage and material separation performance achievable with counting techniques. In addition, data from the conventional detector can be used as complementary information during reconstruction of the images from the counting device. In this paper we present CT images acquired with the hybrid prototype scanner, illustrate its underlying conceptual methods, and provide first experimental results quantifying clinical benefits of quantum-counting CT.

We have developed an ultra-fast photon-counting energy-resolved application specific integrated circuit (ASIC) for spectral computed tomography (CT). A comprehensive characterization has been carried out to investigate the performance of the ASIC in terms of energy resolution under different photon flux rates and the count rate linearity in photon-counting mode. An energy resolution of 4.7 % has been achieved for 59.5 keV low flux photons. The count rate performance of the ASIC was measured with 120 kVp polychromatic x-rays. The results indicate that the count rate linearity can be kept for a flux rate up to 150 Mphotons s^-1 mm^-2 with retained energy information, and this value is increased to be 250 Mphotons s^-1 mm^-2 in photon-counting mode.

Photon counting statistics with pulse pileup effects (PPE) have been investigated for detection schemes with two assumptions: a fixed deadtime (i.e., non-paralyable and paralyzable detection schemes) and a delta pulse shape. Analytical expressions have been developed which shed light on interesting findings: (1) the variance becomes smaller than mean with PPE; and (2) the variance-to-mean ratio (VMR) of narrow energy windows remains very close to 1. In this study, we experimentally investigate VMR with PPE with a variable deadtime (i.e., pulse height analysis) and bipolar and unipolar pulse shapes with a long tail in addition to the above cases. We will use Monte Carlo simulation for a systematic study and a physical photon counting detector to confirm the results.

The prospect of single-photon counting (SPC) detectors for x-ray image acquisition has identified a number of potential benefits over the usual approach in which the detector signal is proportional to total energy deposited during an image-acquisition interval. While a generalized approach to describing transfer of signal and noise through medical imaging systems has been developed over the past several years, in which image-forming processes are represented in terms in cascades of amplified point processes, theis approach describes conventional detectors. We extend cascaded systems analysis (CSA) to the description of signal and noise in SPC x-ray detectors. Expressions describing transfer of the mean value and noise power spectrum (NPS) through a signal thresholding stage are derived and used to describe the effects of conversion to secondary quanta, collection of secondary quanta, additive noise, and thresholding on the zero-frequency DQE of a hypothetical flat-panel SPC detector. We found that when the mean number of secondary quanta collected is 2 to 3 times greater than the threshold value, the zero-frequency DQE for an SPC detector was 5-10% greater than that of a conventionaldetector. We also found that when additive noise is not thresholded out, DQE(0) can be degraded 30-50 %. However, when thresholding techniques are implemented, in some situations this reduction can be largely recovered. It is concluded that in some situations SPC x-ray imaging has the potential to provide equal or better DQE values than conventional energy-integrating detectors.

The x-ray spectrum recorded by a photon-counting x-ray detector (PCXD) is distorted due to the following physical effects which are independent of the count rate: finite energy-resolution, Compton scattering, charge-sharing, and Kescape. If left uncompensated, the spectral response (SR) of a PCXD due to the above effects will result in image artifacts and inaccurate material decomposition. We propose a new SR compensation (SRC) algorithm using the sinogram restoration approach. The two main contributions of our proposed algorithm are: (1) our algorithm uses an efficient conjugate gradient method in which the first and second derivatives of the cost functions are directly calculated analytically, whereas a slower optimization method that requires numerous function evaluations was used in other work; (2) our algorithm guarantees convergence by combining the non-linear conjugate gradient method with line searches that satisfy Wolfe conditions, whereas the algorithm in other work is not backed by theorems from optimization theory to guarantee convergence. In this study, we validate the performance of the proposed algorithm using computer simulations. The bias was reduced to zero from 11%, and image artifacts were removed from the reconstructed images. Quantitative K-edge imaging in possible only when SR compensation is done.

Energy-resolving photon-counting detectors have the potential for improved material decomposition compared to dual-kVp approaches. However, material decomposition accuracy is limited by the nonideal spectral response of the detectors. This work proposes an empirical method for correcting the nonideal spectral response, including spectrum-tailing effects. Unlike previous correction methods which relied on synchrotron measurements, the proposed method can be performed on the scanner. The proposed method estimates a spectral-response matrix by performing x-ray projection measurements through a range of known thicknesses of two or more calibration materials. Once estimated, the spectral-response matrix is incorporated into conventional material decomposition algorithms. A simulation study investigated preliminary feasibility of the proposed method. The spectral-response matrix was estimated using simulated projection measurements through PMMA, aluminum, and gadolinium. An energy-resolved acquisition of a thorax phantom with gadolinium in the blood pool was simulated assuming a five-bin detector with realistic spectral response. Energy-bin data was decomposed into Compton, photoelectric, and gadolinium basis projections with and without the proposed correction method. Basis images were reconstructed by filtered backprojection. Results demonstrated that the nonideal spectral response reduced the ability to distinguish gadolinium from materials such as bone, while images reconstructed with the proposed correction method successfully depicted the contrast agent. The proposed correction method reduced errors from 9% to 0.6% in the Compton image, 90% to 0.6% in the photoelectric image and from 40% to 6% in the gadolinium image when using a three-material calibration. Overall, results support feasibility of the proposed spectral-response correction method.

Limitations to the spatial resolution of current digital x-ray systems are bounded by the physical characteristics of the xray detector. However, the need to image smaller structures provides motivation to develop high-resolution x-ray detector systems for use with computed radiographic, and tomographic x-ray systems. We report the implementation of a tilted detector technique (TDT) to attain near isotropic resolution enhancement by combining two orthogonal image views, acquired with existing detector hardware tilted at a fixed angle. Images were acquired using a ceiling-mounted x-ray unit (Proteus XR/a, GE Medical Systems, 50kVp, 250mAs). Images were digitized using a Fujifilm Capsula X CR system, from a 35×43cm detector cassette placed on an angulated stand, featuring a 3520×4280 image matrix with an in-plane pixel spacing of 0.1mm. Three images were acquired: two for use with our TDT; and one for comparison, with no detector tilt. Performance was determined by using two line-pair phantoms (Models 07-521 and 07-533, Nuclear Associates) placed orthogonally to each other in the field of view. Custom software corrected for perspective distortion, co-registered and combined the tilted-detector images into a single higher-resolution image. Following unwarping and co-registration, the limiting spatial resolution of an image obtained via the weighted combination of the two orthogonal views (8 lp/mm) is found to be superior to that of a single view acquired with no detector tilt (5 lp/mm). This novel technique shows significant improvement in the spatial resolution of x-ray image acquisitions, using existing x-ray components and detector hardware.

The detective quantum efficiency (DQE) is widely accepted as a primary measure of x-ray detector performance in the scientific community. A standard method for measuring the DQE, based on IEC 62220-1, requires the system to have a linear response meaning that the detector output signals are proportional to the incident x-ray exposure. However, many systems have a non-linear response due to characteristics of the detector, or post processing of the detector signals, that cannot be disabled and may involve unknown algorithms considered proprietary by the manufacturer. For these reasons, the DQE has not been considered as a practical candidate for routine quality assurance testing in a clinical setting. In this article we described a method that can be used to measure the DQE of both linear and non-linear systems that employ only linear image processing algorithms. The method was validated on a Cesium Iodide based flat panel system that simultaneously stores a raw (linear) and processed (non-linear) image for each exposure. It was found that the resulting DQE was equivalent to a conventional standards-compliant DQE with measurement precision, and the gray-scale inversion and linear edge enhancement did not affect the DQE result. While not IEC 62220-1 compliant, it may be adequate for QA programs.

Intracranial aneurysm (IA) embolization using Gugliemi Detachable Coils (GDC) under x-ray fluoroscopic guidance is one of the most important neuro-vascular interventions. Coil deposition accuracy is key and could benefit substantially from higher resolution imagers such as the micro-angiographic fluoroscope (MAF). The effect of MAF guidance improvement over the use of standard Flat Panels (FP) is challenging to assess for such a complex procedure. We propose and investigate a new metric, inter-frame cross-correlation sensitivity (CCS), to compare detector performance for such procedures. Pixel (P) and histogram (H) CCS's were calculated as one minus the cross-correlation coefficients between pixel values and histograms for the region of interest at successive procedure steps. IA treatment using GDC's was simulated using an anthropomorphic head phantom which includes an aneurysm. GDC's were deposited in steps of 3 cm and the procedure was imaged with a FP and the MAF. To measure sensitivity to detect progress of the procedure by change in images of successive steps, an ROI was selected over the aneurysm location and pixel-value and histogram changes were calculated after each step. For the FP, after 4 steps, the H and P CCSs between successive steps were practically zero, indicating that there were no significant changes in the observed images. For the MAF, H and P CCSs were greater than zero even after 10 steps (30 cm GDC), indicating observable changes. Further, the proposed quantification method was applied for evaluation of seven patients imaged using the MAF, yielding similar results (H and P CCSs greater than zero after the last GDC deposition). The proposed metric indicates that the MAF can offer better guidance during such procedures.

Due to the need for high-resolution angiographic and interventional vascular imaging, a Micro-Angiographic Fluoroscope (MAF) detector with a Control, Acquisition, Processing, and Image Display System (CAPIDS) was installed on a detector changer, which was attached to the C-arm of a clinical angiographic unit at a local hospital. The MAF detector provides high-resolution, high-sensitivity, and real-time imaging capabilities and consists of a 300 μm thick CsI phosphor, a dual stage micro-channel plate light image intensifier (LII) coupled to a fiber optic taper (FOT), and a scientific grade frame-transfer CCD camera, providing an image matrix of 1024×1024 35 μm effective square pixels with 12 bit depth. The changer allows the MAF region-of-interest (ROI) detector to be inserted in front of the Image Intensifier (II) when higher resolution is needed during angiographic or interventional vascular imaging procedures, e.g. endovascular stent deployment. The CAPIDS was developed and implemented using Laboratory Virtual Instrumentation Engineering Workbench (LabVIEW) software and provides a user-friendly interface that enables control of several clinical radiographic imaging modes of the MAF including: fluoroscopy, roadmapping, radiography, and digital-subtraction-angiography (DSA). The total system has been used for image guidance during endovascular image-guided interventions (EIGI) for diagnosing and treating artery stenoses and aneurysms using self-expanding endovascular stents and coils in fifteen patient cases, which have demonstrated benefits of using the ROI detector. The visualization of the fine detail of the endovascular devices and the vessels generally gave the clinicians confidence on performing neurovascular interventions and in some instances contributed to improved interventions.

A novel cone-beam CT (CBCT) system has been developed with promising capabilities for musculoskeletal imaging (e.g., weight-bearing extremities and combined radiographic / volumetric imaging). The prototype system demonstrates diagnostic-quality imaging performance, while the compact geometry and short scan orbit raise new considerations for scatter management and dose characterization that challenge conventional methods. The compact geometry leads to elevated, heterogeneous x-ray scatter distributions - even for small anatomical sites (e.g., knee or wrist), and the short scan orbit results in a non-uniform dose distribution. These complex dose and scatter distributions were investigated via experimental measurements and GPU-accelerated Monte Carlo (MC) simulation. The combination provided a powerful basis for characterizing dose distributions in patient-specific anatomy, investigating the benefits of an antiscatter grid, and examining distinct contributions of coherent and incoherent scatter in artifact correction. Measurements with a 16 cm CTDI phantom show that the dose from the short-scan orbit (0.09 mGy/mAs at isocenter) varies from 0.16 to 0.05 mGy/mAs at various locations on the periphery (all obtained at 80 kVp). MC estimation agreed with dose measurements within 10-15%. Dose distribution in patient-specific anatomy was computed with MC, confirming such heterogeneity and highlighting the elevated energy deposition in bone (factor of ~5-10) compared to soft-tissue. Scatter-to-primary ratio (SPR) up to ~1.5-2 was evident in some regions of the knee. A 10:1 antiscatter grid was found earlier to result in significant improvement in soft-tissue imaging performance without increase in dose. The results of MC simulations elucidated the mechanism behind scatter reduction in the presence of a grid. A ~3-fold reduction in average SPR was found in the MC simulations; however, a linear grid was found to impart additional heterogeneity in the scatter distribution, mainly due to the increase in the contribution of coherent scatter with increased spatial variation. Scatter correction using MC-generated scatter distributions demonstrated significant improvement in cupping and streaks. Physical experimentation combined with GPU-accelerated MC simulation provided a sophisticated, yet practical approach in identifying low-dose acquisition techniques, optimizing scatter correction methods, and evaluating patientspecific dose.

Clinical applications of CBCT imaging are still limited by excessive imaging dose from repeated scans and poor image quality mainly due to scatter contamination. Compressed sensing (CS) reconstruction algorithms have shown promises in recovering faithful signals from low-dose projection data, but do not serve well the needs of accurate CBCT imaging if effective scatter correction is not in place. Scatter can be accurately measured and removed using measurement-based methods. However, in conventional FDK reconstruction, these approaches are considered unpractical since they require multiple scans or moving the beam blocker during the data acquisition to compensate for the inevitable primary loss. In this work, we combine the measurement-based scatter correction and CS-based iterative reconstruction algorithm, such that scatter-free images can be obtained from low-dose data. We lower the CBCT dose by reducing the projection number and inserting lead strips between the x-ray source and the object. The insertion of lead strips also enables scatter measurement on the measured samples inside the strip shadows. CS-based iterative reconstruction is finally carried out to obtain scatter-free and low-dose CBCT images. Simulation studies are designed to optimize the lead strip geometry for a certain dose reduction ratio. After optimization, our approach reduces the CT number error from over 220HU to below 5HU on the Shepp-Logan phantom, with a dose reduction of ~80%. With the same dose reduction and the optimized method parameters, the CT number error is reduced from 242HU to 20HU in the selected region of interest on Catphan©600 phantom.

The multi-source inverse-geometry CT(MS-IGCT) system is composed of multiple sources and a small 2D detector array. An experimental MS-IGCT system was built and we report initial results with 2×4 x-ray sources, a 75 mm inplane field-of-view (FOV) and 160 mm z-coverage in a single gantry rotation. To evaluate the system performance, experimental data were acquired from several phantoms and a post-mortem rat. Before image reconstruction, geometric calibration, data normalization, beam hardening correction and detector spectral calibration were performed. For reconstruction, the projection data were rebinned into two full cone beam data sets, and the FDK algorithm was used. The reconstructed volumes from the upper and lower source rows shared an overlap volume which was combined in image space. The reconstructed images of the uniform cylinder phantom showed good uniformity of the reconstructed values without any artifacts. The rat data were also reconstructed reliably. The initial experimental results from this rotating-gantry MS-IGCT system demonstrated its ability to image a complex anatomical object without any significant image artifacts and to ultimately achieve large volumetric coverage in a single gantry rotation.

Between 2006 and 2008, the business volume of the top 20 orthopedic manufacturers increased by 30% to about $35 Billion. Similar growth rates could be observed in the market of neurological devices, which went up in 2009 by 10.9% to a volume of $2.2 Billion in the US and by 7.0% to 500 Million in Europe.* These remarkable increases are closely connected to the fact that nowadays, many medical procedures, such as implantations in osteosynthesis or the placement of stents in neuroradiology can be performed using minimally-invasive approaches. Such approaches require elaborate interoperative imaging technology. C-arm based tomographic X-ray region-of-interest (ROI) tomography can deliver suitable imaging guidance in these circumstances: it can offer 3D information in desired patient regions at reasonably low X-ray dose. Tomographic ROI reconstruction, however, is in general challenging since projection images might be severely truncated. Recently, a novel, truncation-robust algorithm (ATRACT) has been suggested for 3D C-arm ROI imaging. In this paper, we report for the first time on the performance of ATRACT for reconstruction from real, angiographic C-arm data. Our results indicate that the resulting ROI image quality is suitable for intraoperative imaging. We observe only little differences to the images from a non-truncated acquisition, which would necessarily require significantly more X-ray dose.

While the Medtronic O-arm was developed for image-guidance applications during orthopedic procedures, it has potential to assist in cardiac surgical and electrophysiological applications; the purpose of this study was to evaluate the feasibility of using a mobile conebeam imaging system (O-arm) for gated cardiac imaging. In an in vivo study (two pigs), projection data from four independently acquired breath-held scans were combined to obtain cardiac gated 3D images. Projection images were acquired during the infusion of contrast agent and while tracking the ECG. Both standard and high-definition modes of the O-arm were evaluated. Projection data were retrospectively combined to generate images corresponding to systole and diastole; different acceptance windows were investigated. The contrast to noise ratio (CNR) between blood and myocardium was compared for the different gating strategies. Gated cardiac images were successfully reconstructed with as few as two scans combined (CNR = 2.5) and a window of 200 ms. Improved image quality was achieved when selecting views based on the minimum time from the selected phase point in the cardiac cycle, rather than a fixed window; in this case the effective temporal window increased to 475 ms for two scans. The O-arm has the potential to be used as a mobile cardiac imaging system, capable of three-dimensional imaging.

C-arm based tomographic 3D imaging is applied in an increasing number of minimal invasive procedures. Due to the limited acquisition speed for a complete projection data set required for tomographic reconstruction, breathing motion is a potential source of artifacts. This is the case for patients who cannot comply breathing commands (e.g. due to anesthesia). Intra-scan motion estimation and compensation is required. Here, a scheme for projection based local breathing motion estimation is combined with an anatomy adapted interpolation strategy and subsequent motion compensated filtered back projection. The breathing motion vector is measured as a displacement vector on the projections of a tomographic short scan acquisition using the diaphragm as a landmark. Scaling of the displacement to the acquisition iso-center and anatomy adapted volumetric motion vector field interpolation delivers a 3D motion vector per voxel. Motion compensated filtered back projection incorporates this motion vector field in the image reconstruction process. This approach is applied in animal experiments on a flat panel C-arm system delivering improved image quality (lower artifact levels, improved tumor delineation) in 3D liver tumor imaging.

In recent years, iterative image reconstruction algorithms have received much interest in x-ray CT imaging. Images reconstructed by the conventional filtered backprojection algorithm are often used as seed images to start the iterations. This paper presents a new consistency property of the measured CT projection data and the forward projected data from a motion contaminated image reconstructed using an analytical image reconstruction algorithm. It is theoretically proven and numerically validated that, when the measured projection data is not redundant, the measured projection data is consistent with the forward projected data of a reconstructed image using analytical image reconstruction algorithms, no matter whether motion artifacts are present in the image or not. However, when there is redundancy in the measured projection data, the consistency depends on the choice of weighting function used for the redundant data. The forward projected data is always consistent with those measured projections with the weight 1.0, no matter what weighting schemes are used. For the measured projection data with a weight smaller than 1.0, the forward projected data is consistent with the weighted sum of redundantly measured projection data. With this new property, some potential issues of directly using FBP images as seed images for iterative algorithms are discussed.

The dose efficiency of dual energy imaging can be improved if the spectra are filtered to increase spectral separation. Our preliminary simulations showed that a fixed Gd filter can be efficacious in rapid kVp-switching systems. We conducted physical experiments on a table top x-ray system to verify the performance improvement before moving forward to a real CT scanner. We chose a commercial Gd₂O₂S screen as our filter and used an acrylic-copper step wedge phantom. Data were collected at 70 and 125 kVp, with and without the filter. The tube current was adjusted to make the exposure rate with and without the filter to be roughly the same. The data were decomposed into basis material images and the variance of the decomposition was measured for each acrylic-copper thickness pair. Simulations were done with the same experimental settings for comparison and validation. The experiments verified that a Gd filter can reduce the variance at fixed dose. The variance reduction is monotonically stronger as the object becomes more attenuating. This study demonstrates the potential of fixed Gd filtration to improve the dose efficiency and material decomposition precision for rapid kVp-switching dual energy systems.

While the imaging protocol determines radiation dose and image quality, it is difficult to find the lowest dose protocol (kVp, mAs, filtration) that provides appropriate diagnostic quality images. Recently, we developed a method for retrospectively synthesizing CT scans of arbitrary protocols using a previously acquired dual energy scan that relies on projection space data. Here, we propose a new variant of synthetic CT that only requires reconstructed dual energy images to simulate realistic images from arbitrary low dose protocols. Axial scans of a phantom were acquired on a GE CT750 HD system at 80 kVp and separately at 140 kVp, enabling material decomposition. Additional scans at 100 and 120 kVp and at different exposures were made to compare with synthesized results. Raw data for any spectrum can be estimated by forward transmission through the material decomposition images, but these have degraded spatial resolution. To avoid blurring, the synthesized image is represented as a linear combination (i.e., mixed or blended image) of the 80/140 kVp images. Noise with the correct statistics is then added so that the total noise matches the expected noise of the simulated protocol (estimated from forward transmission). For the studied object, the resulting synthesized images are indistinguishable from the actual images. Synthetic CT enables users to visualize the impact of protocol changes on the contrast and noise of CT scans, which can be used to develop lower dose protocols by demonstrating dose/noise/protocol trade-offs. Our new image domain implementation significantly increases the accessibility of synthetic CT to potential users.

Radiation dose from CT scans is an increasing health concern in the practice of radiology. Higher dose scans can produce clearer images with high diagnostic quality, but may increase the potential risk of radiation-induced cancer or other side effects. Lowering radiation dose alone generally produces a noisier image and may degrade diagnostic performance. Recently, CT dose reduction based on non-local means (NLM) filtering for noise reduction has yielded promising results. However, traditional NLM denoising operates under the assumption that image noise is spatially uniform noise, while in CT images the noise level varies significantly within and across slices. Therefore, applying NLM filtering to CT data using a global filtering strength cannot achieve optimal denoising performance. In this work, we have developed a technique for efficiently estimating the local noise level for CT images, and have modified the NLM algorithm to adapt to local variations in noise level. The local noise level estimation technique matches the true noise distribution determined from multiple repetitive scans of a phantom object very well. The modified NLM algorithm provides more effective denoising of CT data throughout a volume, and may allow significant lowering of radiation dose. Both the noise map calculation and the adaptive NLM filtering can be performed in times that allow integration with the clinical workflow.

The recent CT systems yield high spatial resolution in all directions of volumetric images in clinical routine. The quantitative characterization of the performance of CT systems is important for comparing the effects of different scan and reconstruction parameters, for comparing between different CT systems, and for evaluating the accuracy of size and density measurements of fine details in CT images. This paper presents a method to determine the modulation transfer function (MTF) in the scan plane obtained by CT system from profiles of human anatomical structures such as blood vessel measured by clinical measurement conditions without magnified reconstruction. MTF estimations are performed for cylindrical tube phantoms with three different diameters (1 mm, 2 mm, and 3 mm) injected by solution of contrast material and human blood vessels measured by the clinical measurement conditions. We demonstrate the potential usefulness of the method for estimating the MTF from blood vessel profiles measured in CT systems.

The optimization of dual-energy computed tomography (DE-CT) is challenged by the lack of a theoretical foundation for image quality. This work reports a cascaded systems analysis model that was used to derive signal and noise propagation in DE-CBCT in prevalent Fourier metrics such as the noise-power spectrum (NPS) and noise-equivalent quanta (NEQ). The model was validated in comparison to measurements of the 3D NPS and NEQ in DE-CBCT images acquired using an experimental imaging bench. Task-based detectability index was derived using DE-NPS and NEQ as an objective function in optimizing DE imaging parameters such as the dose allocation factor (D_A) and kVp pair. The resulting dose allocation optimization is in agreement with the practice of assigning more dose to the high-energy image (D_A < 0.5), and the model provides a quantitative basis for examining the optimal dose allocation as a function of total dose, kVp pair, the presence of electronics noise, and the imaging task. An example optimization is shown for a breast tumor detection task. Using DE decomposition to cancel fibroglandular tissue (rendering a DE-CBCT image of breast tumor against an adipose tissue background) and assuming a total dose of 15mGy, the optimal kVp pair is identified at [45, 105]kVp with DA=0.46. The model is sufficiently general for applications beyond this example, demonstrating utility in the optimization in a broad range of imaging parameters. The model provides a new, valuable framework for understanding the theoretical limits of DE-CBCT imaging performance and maximizing image quality while minimizing radiation dose.

The increasing availability of iterative reconstruction (IR) algorithms on clinical scanners is creating a demand for effectively and efficiently evaluating imaging performance and potential dose reduction. In this study, the location- and task-specific evaluation was performed using detectability index (d') by combining a task function, the task transfer function (TTF), and the noise power spectrum (NPS). Task function modeled a wide variety detection tasks in terms of shape and contrast. The TTF and NPS were measured from a physical phantom as a function of contrast and dose levels. Measured d' values were compared between three IRs (IRIS, SAFIRE3 and SAFIRE5) and conventional filtered back-projection (FBP) at various dose levels, showing an equivalent performance of IR at lower dose levels. AUC further calculated from d' showed that compared to FBP, SAFIRE5 may reduce dose by up to 50-60%; SAFIRE3 and IRIS by up to 20-30%. This study provides an initial framework for the localized and task-specific evaluation of IRs in CT and a guideline for the identification of optimal operating dose point with iterative reconstructions.

X-ray phase sensitive imaging methods have seen tremendous growth and increased interest in recent years. Each method has its advantages and disadvantages, but all have shown the ability to improve the detection of various objects because of the additional phase measurements. Of the various methods, grating-based differential phase contrast computed tomography (DPC-CT) imaging has shown greater quantitative and diagnostic capabilities than traditional absorption CT. Although it has been shown that DPC-CT provides superior contrast of certain materials, one question has not been fully addressed to date is whether DPC-CT can provide improved accuracy in detecting low contrast masses using the same radiation dose as that given in absorption CT. The detectability is not only related to contrast to noise ratio, but also to the noise texture. The purpose of this study is to investigate how the peculiar noise texture found in cone-beam DPC-CT affects low contrast objects' detectability through human observer ROC analysis. Studies for both axial and sagittal planes were carried out, as both could potentially be used in clinical practice for a 3D image. The results demonstrate that noise texture found in conebeam DPC-CT strongly affects human visual perception, and that object detectabilities in axial and sagittal images of DPC-CT are different.

The purpose of this study is to investigate how well model observer can correlate with human observer in the lesion detection and localization task when the location of lesion is uncertain in CT imaging. A 35 × 26 cm oblong-shaped water phantom was scanned with and without two cylindrical rods (3 mm and 5 mm diameters) to simulate lesions with - 15HU contrast. Scans were repeated 100 times with the rods and 100 times without for each of 4 dose levels. Signal and background images were generated by selecting ROIs with 128x128 pixels, with the location of signal in each ROI randomly distributed. Human observer studies were conducted as three medical physicists identified the presence or absence of lesion, indicated the lesion location in each image and scored confidence level with a 6-point scale. ROC curves were fitted and area under curve (AUC) was calculated. The same data set was also analyzed using a Channelized Hottelling model observer with Gabor channels. Internal noise was added to the test variables for model observer study. AUC of ROC and LROC curves were calculated using non-parametric approach. The performance of human observer and model observer was compared. The Peason's product-moment correlation coefficients were 0.994 and 0.998 for 3mm and 5mm diameter lesions in ROC analysis and 0.987 and 0.999 in LROC analysis, indicating that model observer performance was highly correlated with the human observer performance for different size of lesions and different dose levels when signal location is uncertain. These results provide the potential of using model observer that correlates with human observer to assess CT image quality, optimize scanning protocol and reduce radiation dose.

The current version of the real-time skin-dose-tracking system (DTS) we have developed assumes the exposure is contained within the collimated beam and is uniform except for inverse-square variation. This study investigates the significance of factors that contribute to beam non-uniformity such as the heel effect and backscatter from the patient to areas of the skin inside and outside the collimated beam. Dose-calibrated Gafchromic film (XR-RV3, ISP) was placed in the beam in the plane of the patient table at a position 15 cm tube-side of isocenter on a Toshiba Infinix C-Arm system. Separate exposures were made with the film in contact with a block of 20-cm solid water providing backscatter and with the film suspended in air without backscatter, both with and without the table in the beam. The film was scanned to obtain dose profiles and comparison of the profiles for the various conditions allowed a determination of field non-uniformity and backscatter contribution. With the solid-water phantom and with the collimator opened completely for the 20-cm mode, the dose profile decreased by about 40% on the anode side of the field. Backscatter falloff at the beam edge was about 10% from the center and extra-beam backscatter decreased slowly with distance from the field, being about 3% of the beam maximum at 6 cm from the edge. Determination of the magnitude of these factors will allow them to be included in the skin-dose-distribution calculation and should provide a more accurate determination of peak-skin dose for the DTS.

The purpose of this study is to investigate the absorbed dose and energy by iodinated contrast medium in diagnostic radiology. A simulation geometry in which an inner sphere (d = 0.2cm, 1cm, 5cm) filled with iodinated contrast medium (or water) is located at the center of a 20cm diameter water sphere was used in simulations performed with MCNP5 codes. Monoenergetic x-rays with energies ranging from 40 to 80keV from a cone beam source were utilized and contrast medium concentration ranged from 100 to 1mg/ml. Absorbed dose ratio (R_D) to inner sphere and total absorbed energies ratio (R_E) to the whole phantom with and without iodinated contrast medium were investigated. The maximum R_D was ~13 for the 0.2cm diameter sphere with 100mg/ml contrast medium. The maximum R_E was ~1.05 for the 5cm diameter contrast sphere at 80keV with 100mg/ml contrast medium. Under the same incident photon energy, increasing the inner sphere size from 0.2cm to 5cm caused a ~63% increase in the R_D on average. Decreasing the contrast medium concentration from 100 to 10 mg/ml caused a decrease of R_D of ~ 76%. A conclusion was reached that although local absorbed dose increase caused by iodinated contrast agent could be high; the increase in total absorbed energy is negligible.

Reliable estimates for patient doses for CT examinations are desirable for the patients themselves as well as for new epidemiological studies. It has been shown that dose conversion coefficients normalized to CTDI_vol provide rather scanner independent quantities. In this work, it is demonstrated that this normalization provides also tube voltage independent values by simulating axial CT scans of a seven-year old infant and an adult. The differences in the effective dose conversion coefficients per CTDIvol between 80 and 120 kV is for most body heights below 5%. Only at the height of the testes and the thyroid the difference can be as large as 15%. This results in differences of the effective dose conversion coefficient per CTDIvol between 80 and 120 kV of less than 6-7% for typical CT examinations.

Recently published AAPM Task Group 204 developed conversion coefficients that use scanner reported CTDIvol to estimate dose to the center of patient undergoing fixed tube current body exam. However, most performed CT exams use TCM to reduce dose to patients. Therefore, the purpose of this study was to investigate the correlation between organ dose and a variety of patient size metrics in adult chest CT scans that use tube current modulation (TCM). Monte Carlo simulations were performed for 32 voxelized models with contoured lungs and glandular breasts tissue, consisting of females and males. These simulations made use of patient's actual TCM data to estimate organ dose. Using image data, different size metrics were calculated, these measurements were all performed on one slice, at the level of patient's nipple. Estimated doses were normalized by scanner-reported CTDIvol and plotted versus different metrics. CTDI_vol values were plotted versus different metrics to look at scanner's output versus size. The metrics performed similarly in terms of correlating with organ dose. Looking at each gender separately, for male models normalized lung dose showed a better linear correlation (r²=0.91) with effective diameter, while female models showed higher correlation (r²=0.59) with the anterior-posterior measurement. There was essentially no correlation observed between size and CTDI_vol-normalized breast dose. However, a linear relationship was observed between absolute breast dose and size. Dose to lungs and breasts were consistently higher in females with similar size as males which could be due to shape and composition differences between genders in the thoracic region.

The purpose of this work was twofold: (a) to estimate patient- and cohort-specific radiation dose and cancer risk index for abdominopelvic computer tomography (CT) scans; (b) to evaluate the effects of patient anatomical characteristics (size, age, and gender) and CT scanner model on dose and risk conversion coefficients. The study included 100 patient models (42 pediatric models, 58 adult models) and multi-detector array CT scanners from two commercial manufacturers (LightSpeed VCT, GE Healthcare; SOMATOM Definition Flash, Siemens Healthcare). A previously-validated Monte Carlo program was used to simulate organ dose for each patient model and each scanner, from which DLP-normalized-effective dose (k factor) and DLP-normalized-risk index values (q factor) were derived. The k factor showed exponential decrease with increasing patient size. For a given gender, q factor showed exponential decrease with both increasing patient size and patient age. The discrepancies in k and q factors across scanners were on average 8% and 15%, respectively. This study demonstrates the feasibility of estimating patient-specific organ dose and cohort-specific effective dose and risk index in abdominopelvic CT requiring only the knowledge of patient size, gender, and age.

Because tomographic reconstructions are ill-conditioned, algorithms that incorporate additional knowledge about the imaging volume generally have improved image quality. This is particularly true when measurements are noisy or have missing data. This paper presents a general framework for inclusion of the attenuation contributions of specific component objects known to be in the field-of-view as part of the reconstruction. Components such as surgical devices and tools may be modeled explicitly as being part of the attenuating volume but are inexactly known with respect to their locations poses, and possible deformations. The proposed reconstruction framework, referred to as Known-Component Reconstruction (KCR), is based on this novel parameterization of the object, a likelihood-based objective function, and alternating optimizations between registration and image parameters to jointly estimate the both the underlying attenuation and unknown registrations. A deformable KCR (dKCR) approach is introduced that adopts a control pointbased warping operator to accommodate shape mismatches between the component model and the physical component, thereby allowing for a more general class of inexactly known components. The KCR and dKCR approaches are applied to low-dose cone-beam CT data with spine fixation hardware present in the imaging volume. Such data is particularly challenging due to photon starvation effects in projection data behind the metallic components. The proposed algorithms are compared with traditional filtered-backprojection and penalized-likelihood reconstructions and found to provide substantially improved image quality. Whereas traditional approaches exhibit significant artifacts that complicate detection of breaches or fractures near metal, the KCR framework tends to provide good visualization of anatomy right up to the boundary of surgical devices.

Photon counting detectors are an emerging technology for spectral computed tomography. They have the potential to improve tissue contrast and specificity, reduce dose, and enable novel applications for K-edge and functional imaging. In this presentation various non-linear distortions were investigated that affect the image quality in photon-counting spectral CT: deadtime losses and spectral response, inherent to the new technology, and beam hardening artifacts that stem from the use of a polychromatic x-ray source. These effects were corrected or compensated for by performing calibration measurements. Techniques from material decomposition were applied to reconstruct images at a desired energy. Two methods were compared to synthesize a single monoenergetic image from photon counting data with multiple energy bins. The parameters were optimized to maximize a given image quality index. The procedures were evaluated on phantom data acquired on an experimental CT scanner with photon-counting detectors with two energy thresholds.

A new class of fully 4-dimensional image reconstruction algorithm for cardiac CT was developed. The proposed method is an iterative algorithm that alternates two methods, motion estimation (ME) method and motion compensated reconstruction (MCR) method. The ME method estimates the cardiac MVF using elastic image registration between the reference phase and other phases. The motion of heart was modeled by the linear combination of cubic B-spline basis function. The sum of squared difference and spatial and temporal regularization terms were chosen as the cost function, which is minimized by a nested conjugate gradient method. The MCR method (Schafer's method) reconstructs cardiac images using the MVF estimated by the ME method. The reconstructed images will be fed to ME in the next iteration. The ME and MCR were performed alternately till convergence was achieved. Accuracy of the proposed method was evaluated using 3 patient data acquired by a 64-slice CT scanner. The heart rates of the patients ranged between 52 and 88 beats-per-minute. The motion artifacts were significantly decreased by the proposed method, and the degree of improvement increased as the iteration progressed.

The noise properties of differential phase contrast CT (DPC-CT) demonstrate some peculiar features. It has been both theoretically and experimentally demonstrated that the noise variance of DPC-CT scales with spatial resolution following an inverse first order relationship. This is in stark contrast to absorption CT, where the noise variance scales with spatial resolution following an inverse third power. In addition to the scaling relationship, the noise power spectrum (NPS) of DPC-CT is dominated by low spatial frequencies and demonstrates a singular behavior when approaching zero frequency. This focuses the peak noise power within low spatial frequencies while high-frequency noise is suppressed. This is again in contrast to the absorption CT case where the NPS smoothly transitions to zero at zero frequency. The singular behavior of the DPC-CT NPS visually affects image noise texture and may hinder observer perception. In this paper, a method is proposed to improve the noise properties in DPC-CT and potentially improve observer performance. Specifically, the low frequency component of the filtering kernel used in reconstruction has been regularized to modify the noise power at low spatial frequencies. This results in a high-pass filtering of the image. The high-pass filtered image is combined with the original image to generate the final image. As a result of these two operations, the noise power is shifted to the high spatial frequency direction, improving visual perception, while image reconstruction accuracy is maintained. Experimental phantom results are presented to validate the proposed method.

Dedicated breast CT has great potential for improving the detection and diagnosis of breast cancer. In this study, statistical iterative reconstruction with a penalized likelihood objective function and a Huber prior are investigated for use with breast CT. This prior has two free parameters, the penalty weight and the edgepreservation threshold, that need to be evaluated to determine those values that give optimal performance. Computer simulations with breast-like phantoms were used to study these parameters using various figuresof- merit that relate to performance in detecting microcalcifications. Results suggested that a narrow range of Huber prior parameters give optimal performance. Furthermore, iterative reconstruction provided improved performance measures as compared to conventional filtered back-projection.

Conventional ordered-subsets (OS) methods for regularized image reconstruction involve computing the gradient of the regularizer for every subset update. When dealing with large problems with many subsets, such as in 3D X-ray CT, computing the gradient for each subset update can be very computationally expensive. To mitigate this issue, some investigators use unregularized iterations followed by a denoising operation after updating all subsets.¹ Although such methods save computation, their convergence properties are uncertain, and since they may not be minimizing any particular cost function it becomes more difficult to design regularization parameters. Furthermore, it is known that inserting filtering steps into unregularized algorithms can lead to undesirable spatial resolution properties.² Our goal here is to reduce the computational cost without inducing such problems. We propose a new OS-type algorithm that is derived using optimization transfer principles. The proposed method allows the gradient of the regularizer to be updated less frequently, and thus reduces the computational expense when many subsets are used. Our derivation leads to a correction term that accounts for the fact that the regularizer gradient is updated less frequent than every sub-iteration. Simulations and a phantom experiment show that the proposed method reconstructed images with compatible image quality within reduced computation time.

Accurately targeting of small lesions for success is crucial in breast biopsy. In this paper, we proposed a new 3D tomobased biopsy, which is characterized in being more accurate, easier to perform, lower in dose, and free of metal artifact. In the scout phase, a conventional tomosynthesis scan is performed, and the reconstructed 3D image is then used for radiologists to accurately localize target volume and determine optimized needle path. In the prefire phase, two prefire stereotactic images are obtained at +24° and -24° angular levels for retrieving needle and shifted lesion locations. By combining the reconstructed 3D tomosynthesis image, needle location and lesion location, synthetic prefire and postfire images are generated for radiologists' reference before firing the real needle. The proposed scheme not only improves the biopsy accuracy but also reduces dose by 3.7-5.6 times compared to conventional mammo-based stereotactic biopsy. A simulation using anthropomorphic phantom was conducted to verify our method. Both needle and lesion were precisely recovered just based on two tomo angled images. For the needle registration, the sum of translation discrepancy is less than 3 pixels, and the sum of rotation discrepancy is less than 3 degrees. For the lesion registration, the sum of coordinate discrepancy is less than 4 pixels. The predicted 3D prefire and postfire images exhibited more intuitive spatial relationship of the shifted lesion and biopsy needle tip than mammo-based stereotactic biopsy.

The development of differential phase contrast imaging using conventional x-ray tubes has spurred great interest in the medical imaging community. It has been shown to provide higher contrast than absorption imaging in some cases, and in this work we translate these advantages to tomosynthesis imaging. A general framework for reconstruction of images from differential phase contrast projection data has been proposed and implemented using data from a grating-based x-ray phase contrast tomosynthesis system. Reconstructed tomosynthesis images from differential phase contrast data are shown, using both a direct backprojection (BP) technique and a filtered backprojection (FBP) reconstruction method. From the results it is seen that phase contrast tomosynthesis can separate superimposed phase objects while providing complementary information to absorption tomosynthesis.

Filtered backprojection (FBP) has been commonly used as an efficient and robust reconstruction technique in tomographic X-ray imaging during the last decades. For standard geometries like circle or helix it is known how to efficiently filter the data. However, for geometries with only few projection views or with a limited angular range, the application of FBP algorithms generally provides poor results. In digital breast tomosynthesis (DBT) these limitations give rise to image artifacts due to the limited angular range and the coarse angular sampling. In this work, a generalized FBP algorithm is presented, which uses the filtered projection data of all acquired views for backprojection along one direction. The proposed method yields a computationally efficient generalized FBP algorithm for DBT, which provides similar image quality as iterative reconstruction techniques while preserving the ability for region of interest reconstructions. To demonstrate the excellent performance of this method, examples are given with a simulated breast phantom and the hardware BR3D phantom.

Improved visibility of microcalcifications (MCs) and masses in tomographic breast imaging is a major concern in the medical imaging community, with intense research activity considering both hardware and processing approaches to the problem. Much of the research involves digital breast tomosynthesis (DBT). In this paper, we present results of human-observer studies that investigated the effects of postreconstruction filter strength on MC detection in DBT images generated at various dose levels. The use of human observers poses severe limitations on objective-assessment studies involving multiple parameters and this paper also discusses our continued development of a visual-search mathematical model observer as a substitute for humans. In this work, DBT images were created using a rigorous computer simulation applied to realistic breast phantoms. Acquisitions with 0.7, 1.0 and 1.5 mGy doses were modeled and the Feldkamp FBP algorithm was used for reconstructions. A set of 3D Butterworth filters with cutoffs representing moderate (0.2 cycles/pixel, with pixel size = 100 microns) to no (0.5 cycles/pixel) postfiltering were tested. LROC studies were conducted with four observers. As expected, MC detectability fell off with reduced dose. At the same time, the best MC detection for a given dose was obtained with unfiltered images, suggesting that the increased noise levels associated with lower dose cannot be overcome with postfiltering. The model observer showed promising results in terms of agreement with the human observers. The causes for some points of disagreement merit examination.

Digital breast tomosynthesis (DBT) holds strong promise for improving the sensitivity of detecting subtle mass lesions. Detection of microcalcifications is more difficult because of high noise and subtle signals in the large DBT volume. It is important to enhance the contrast-to-noise ratio (CNR) of microcalcifications in DBT reconstruction. A major challenge of implementing microcalcification enhancement or noise regularization in DBT reconstruction is to preserve the image quality of masses, especially those with ill-defined margins and subtle spiculations. We are developing a new multiscale regularization (MSR) method for the simultaneous algebraic reconstruction technique (SART) to improve the CNR of microcalcifications without compromising the quality of masses. Each DBT slice is stratified into different frequency bands via wavelet decomposition and the regularization method applies different degrees of regularization to different frequency bands to preserve features of interest and suppress noise. Regularization is constrained by a characteristic map to avoid smoothing subtle microcalcifications. The characteristic map is generated via image feature analysis to identify potential microcalcification locations in the DBT volume. The MSR method was compared to the non-convex total pvariation (TpV) method and SART with no regularization (NR) in terms of the CNR and the full width at half maximum of the line profiles intersecting calcifications and mass spiculations in DBT of human subjects. The results demonstrated that SART regularized by the MSR method was superior to the TpV method for subtle microcalcifications in terms of CNR enhancement. The MSR method preserved the quality of subtle spiculations better than the TpV method in comparison to NR.

X-ray CT data collected in fan-beam geometry is often rebinned to parallel-beam geometry before image reconstruction. In this work, we propose an exact, efficient method to analytically calculate the covariance between image pixels for an FBP reconstruction when such a rebinning step is performed. As a special case, the method also enables calculation of image variance. Our approach builds on a previous investigation of noise covariance for direct fan-beam FBP reconstruction, and encompasses commonly used variants of parallel-beam rebinning algorithms that include data interleaving steps to process fan-beam data acquired with either a quarter-detector offset or a flying focal spot. The accuracy of the variance and covariance calculation results is validated through comparison with Monte Carlo estimates.

Many imaging scenarios involve a sequence of tomographic data acquisitions to monitor change over time - e.g., longitudinal studies of disease progression (tumor surveillance) and intraoperative imaging of tissue changes during intervention. Radiation dose imparted for these repeat acquisitions present a concern. Because such image sequences share a great deal of information between acquisitions, using prior image information from baseline scans in the reconstruction of subsequent scans can relax data fidelity requirements of follow-up acquisitions. For example, sparse data acquisitions, including angular undersampling and limited-angle tomography, limit exposure by reducing the number of acquired projections. Various approaches such as prior-image constrained compressed sensing (PICCS) have successfully incorporated prior images in the reconstruction of such sparse data. Another technique to limit radiation dose is to reduce the x-ray fluence per projection. However, many methods for reconstruction of sparse data do not include a noise model accounting for stochastic fluctuations in such low-dose measurements and cannot balance the differing information content of various measurements. In this paper, we present a prior-image, penalized-likelihood estimator (PI-PLE) that utilizes prior image information, compressed-sensing penalties, and a Poisson noise model for measurements. The approach is applied to a lung nodule surveillance scenario with sparse data acquired at low exposures to illustrate performance under cases of extremely limited data fidelity. The results show that PI-PLE is able to greatly reduce streak artifacts that otherwise arise from photon starvation, and maintain high-resolution anatomical features, whereas traditional approaches are subject to streak artifacts or lower-resolution images.

Analytic-based algorithms such as the FDK algorithm is used currently for image reconstruction from data acquired with prototypes of dedicated breast CT scanners. In general, analytic-based algorithms require data collected at a large number (~500) of views. In current breast-CT scans, imaging dose delivered to the patient is about the same as that used in a typical two-view mammography exam. This highly limited total imaging dose, when distributed over a large number of views in breast CT, can result in low-SNR data. There exists a renewed interest in developing optimization-based (i.e., iterative) algorithms for image reconstruction from low-SNR data and/or from sparse-view data collected at a reduced number of views. Results of recent studies on optimization-based algorithms from CT data suggest that the algorithms may reconstruct images of quality higher than than analytic-based algorithms from low-SNR data and/or from sparse-view data. In this work, we investigated image reconstruction from low-SNR patient-breast-CT data collected at a large number (~500), as well as at reduced numbers, of views. The result of the study appears to indicate that optimization-based reconstructions can yield breast-CT images from low-SNR data comparable to, or better than, the corresponding FDK reconstructions.

Although statistical image reconstruction methods for X-ray CT can provide improved image quality at reduced patient doses, computation times for 3D axial and helical CT are a challenge. Rapidly converging algorithms are needed for practical use. Augmented Lagrangian methods based on variable splitting recently have been found to be effective for image denoising and deblurring applications.⁵ These methods are particularly effective for non-smooth regularizers such as total variation or those involving the 1 norm. However, when standard "split Bregman" methods⁶ are applied directly to 3D X-ray CT problems, numerous auxiliary variables are needed, leading to undesirably high memory requirements.⁷ For minimizing regularized, weighted least-squares (WLS) cost functions, we propose a new splitting approach for CT, based on the alternating direction method of multipliers (ADMM)^1,5 that has multiple benefits over previous methods: (i) reduced memory requirements, (ii) effective preconditioning using modified ramp/cone filters, (iii) accommodating very general regularizers including edge-preserving roughness penalties, total variation methods, and sparsifying transforms like wavelets. Numerical results show that the proposed algorithm converges rapidly, and that the cone filter is particularly effective for accelerating convergence.

Determination of the rotation axis position for tomographic projection images is critical to perform an accurate reconstruction. Rotational centers in micro-tomography may shift by several microns after the initial calibration due to various factors such as temperature variation, sample system stability and sample loading procedures. Automatic detection of rotational centers after data acquisition is therefore crucial for accurate and efficient reconstructions, and it is commonly implemented at various synchrotron facilities. We propose to implement a reliable cross correlation method on the projections of 0 and 180 degree to automatically re-align the rotation axis at data collection time. For this purpose, several issues, such as the flat-field correction for the imaging system and the irregular data near projection boundaries, are handled to increase the stability achieving subpixel alignments. The method is shown from experimental results to be accurate, efficient and stable. The results from automatic detections are mostly within one pixel difference from manual/operator detection results. Following the data collection we developed an automatic sub-pixel rotational centering method. Intermediate results from this final process are generated for user inspection. The proposed method is able to detect rotational center shifts within 7 seconds for high-resolution projections of size 2048×2048. It is shown to be stable for static samples in complicated cases. GPU is utilized to fasten the cross correlation computation in the space domain, which achieves about 10 times speedup. The proposed method fits seamlessly into the current framework of beamline 2-BM at the Advanced Photon Source, Argonne National Laboratory. It may save 5 minutes for partial reconstructions and 5-10 minutes for manual detections without sacrificing accuracy.

Ring artifacts are common in tomography reconstructions. These artifacts may arise, in microtomography using synchrotron radiation, from dead pixels in CCD detectors, damaged scintillator screens and instabilities of the synchrotron beam. Ring artifact removal is therefore an important task. We apply an image inpainting algorithm on the sinogram data to reduce ring artifacts in reconstructions. The applied image inpainting algorithm is advantageous in two aspects. First, the proposed method is local, and it affects only the user-specified corrupted region. Second, it utilizes the directional information of level lines, as well as the intensity information, in the sinogram data. Experimental results show that the proposed method is able to compensate both the missing directional information and the distorted intensity information, which reduces ring artifacts in reconstructions. An improvement in reconstruction image qualities is achieved for micro-tomography using the proposed method. A comparison shows that the proposed method may outperform existing methods in terms of reconstruction accuracy.

Image noise is an important aspect when optimizing CT system parameters. For example in multi-energy CT the statistical error in material images is of high interest. It has been shown that the statistical error of individual line integrals can be quantified by the Cramér-Rao lower bound, but it is not straightforward to estimate the noise in the corresponding material images in a time-efficient way, since for an exact calculation a large number of projections has to be taken into account. This paper introduces a way to approximate the image noise by using noise estimates of only a few line integrals. The method is exemplified using simulated dual energy CT data, and the results are shown to closely correspond with Monte Carlo outcomes. The method provides a fast alternative to Monte Carlo simulations in order to calculate the image noise depending on various system parameters, which is very helpful, e.g., for system design.

The multi-source inverse-geometry CT(MS-IGCT) system is composed of multiple sources and a small 2D detector array. Each source is activated sequentially and covers a small portion of the field-of-view (FOV) and a full FOV reconstruction is acquired by combining projection data from all sources. During the data acquisition, the intensity of each x-ray source could change, e.g. because of instability in the power supply, leading to artifacts in the reconstructed image. To reduce the image artifacts, we developed a data normalization algorithm for the MS-IGCT system. The projection data of each source shares an overlap region in 2D Radon space with another source. Thus, substantially same projection data can be generated from different sources at different gantry positions. Since at least one source can illuminate a reference channel and therefore its data can be easily normalized. This normalized projection data can be used to normalize the raw data of another source with which it shares an overlap region. By performing this normalization process sequentially, the intensity variations of all sources can be corrected. The proposed method was tested with Shepp-Logan phantom using 10% random source intensity fluctuations. While the reconstructed image showed image artifacts that result from uncorrected fluctuations, after applying the proposed normalization algorithm, image artifacts were removed.

Helical computed tomography (HCT) has demonstrated the effectiveness in virtual colonoscopy (VC) or CTcolonography (CTC). One major concern with this clinical application is associated with the risk of high radiation exposure, especially for its use for screening purpose at a large population. In this work, we presented an improved Karhunen-Loeve (KL) domain penalized weighted least-squares (PWLS) strategy which considers the data correlations among the projection rays mainly due to partially overlap while system rotates. Two 1-dimensional (1D) projections, which called coupled projections (CPs), are composed according to the geometry. Each element of the 1D projection is carefully selected for a specific point within 2π angle along the system rotates and thus a highly correlation can be observed between any specific projection and the CPs. These highly correlated projections can be treated by an adaptive KL-PWLS strategy for accurate noise reduction. This method has been implemented and tested on computer simulated sinograms which mimic low-dose CT scans. The reconstructed images by the presented strategy demonstrated the potential of ultra low-dose CT application.

Image reconstruction in helical CT assumes that the source position at which each view is acquired is accurately reported to the image reconstruction algorithm. However, in practical situations this is not always the case. Hardware, instrumentation and bandwidth limitations along with manufacturing tolerances limit the accuracy at which source positions may be recorded. This study simulates scans under the conditions of positional reporting errors in combination with errors in the time moments at which positions are reported. The resulting image quality is evaluated and the deviations from images obtained under ideal conditions are characterized using pixel to pixel differences and mutual information between image pairs.

In this paper, a 4D iterative reconstruction scheme is developed to improve noise characteristics and/or reduce radiation exposure in multi-slice cardiac CT. In our implementation, image volume datasets are reconstructed at adjacent temporal positions with respect to an optimized cardiac phase (best phase). Nonlinear regularization priors operate on a 4D cube surrounding each voxel in 4D space, reducing image noise while maintaining temporal and spatial image sharpness. The temporal resolution of image data is maintained despite the usage of temporal data that can substantially exceed the reconstruction range of the 'best phase' reconstruction. Consequently, the noise statistics is significantly improved because non-correlated image data at different temporal positions are utilized. To reduce the high computational load, the iterative regularization in 4D can be transferred into image space. Raw data based Iterative Reconstruction reducing artifacts due to the non-exactness of the backprojector is decoupled from regularization and restricted to only those projection data belonging to the 'best phase' reconstruction. Finally, the image formation is achieved by a normalized combination of the low frequency part of the raw data based Iterative Reconstruction at cardiac best phase and the high frequency part of the image based regularization image at best phase. We demonstrate the potential of noise reduction on basis of clinical cardiac CT data. As an example for cardiac dual source CT (DSCT) data a noise reduction up to 70% was achieved. Even in case of a very high and irregular heart beat with an average heart rate of 115 bpm the high temporal resolution of DSCT could be maintained.

The appeal of compressed sensing (CS) in the context of medical imaging is undeniable. In MRI, it could enable shorter acquisition times while in CT, it has the potential to reduce the ionizing radiation dose imparted to patients. However, images reconstructed using a CS-based approach often show an unusual texture and a potential loss in spatial resolution. The prior image constrained compressed sensing (PICCS) algorithm has been shown to enable accurate image reconstruction at lower levels of sampling. This study systematically evaluates an implementation of PICCS applied to myocardial perfusion imaging with respect to two parameters of its objective function. The prior image parameter α was shown here to yield an optimal image quality in the range 0.4 to 0.5. A quantitative evaluation in terms of temporal resolution, spatial resolution, noise level, noise texture, and reconstruction accuracy was performed.

Low-dose x-ray computed tomography (CT) is clinically desired. However, the quality of low-dose CT image is severely degraded due to excessive photon quantum noise and electronic noise. It is known that accurate noise modeling is a fundamental issue for low-dose CT imaging, such as statistical iterative image reconstruction and statistics-based singoram restoration. In this paper, we first studied the statistical moment properties of the noise model in CT projection domain wherein the noise of detected signal is considered as quantum photon noise plus background electronic noise. More importantly, we derived a new formula to estimate the mean-variance relationship in Radon domain by using Taylor explanation. To test the presented variance estimation formula, an anthropomorphic torso phantom was scanned repeated by a commercial scanner at five different mAs levels from 100 down to 17. The experimental results demonstrate that the electronic noise is significant when low-dose scan is performed or the number of detected photons is limited. As an important conclusion of the presented study, electronic noise effect should be considered in low-dose CT image reconstruction.

Sub-second multi-detector computed tomography systems (MDCTs) offer great potentials for improving cardiac imaging. However, since the temporal resolution of such CT systems is not sufficient, blurring and artifacts produced by fast cardiac motion are still problematic. The purposes of this study were to investigate the accurate method for measurement of temporal resolution (TR) of the cardiac CT and required TR for obtaining better CT coronary angiography (CTCA). We employed a dual source CT system (Somatom Definition, Siemens), which has various temporal resolution modes (83, 125, and 165 msec) for electro-cardiogram (ECG)-gated scanning. The temporal sensitivity profiles (TSPs) were measured by a new method using temporal impulse generated by metal ball (impulse method). The CTCA images of 200 patients with heart rates (HRs) ranging from 36 to 117 beat per minute (bpm) were visually evaluated using a 4-point scale. The 165-msec TR mode, which is mostly available on recent MDCTs, showed a sufficient image quality only at low HR (≤ 60 bpm) for all 3 arteries. The image quality of 125-msec TR mode was acceptable at low to intermediate HRs (< 80 bpm) for LADs and LCXs, and insufficient for the RCAs in cases with HR more than 71 bpm. The 83-msec TR mode demonstrated excellent image quality except for cases with very quick motion of the RCAs at a high HR (>80 bpm).

The grating-based x-ray differential phase contrast (DPC) CT is emerging as a new technology with the potential for extensive preclinical and clinical applications. In general, the performance of an imaging system is jointly determined by its signal property (modulation transfer function-MTF(k)) and noise property (noise power spectrum-NPS(k)), which is characterized by its spectrum of noise equivalent quanta. As reported by us previously, owing to an adoption of the Hilbert filtering for image reconstruction in the fashion of filtered backprojection (FBP), the noise property of DPC-CT characterized by its NPS(k) differs drastically from that of the conventional attenuation-based CT (1/|k| trait vs. |k| trait). In this work, via system analysis, modeling and simulated phantom study, we initially investigate the signal property of DPC-CT characterized by its MTF(k) and compare it with that of the conventional CT. In addition, we investigate the DPC-CT's spectrum of noise equivalent quanta NEQ(k) - the most important figure of merit (FOM) in the assessment of an imaging system's performance - by taking the MTF(k) and NPS(k) jointly into account. Through such a thorough investigation into both the signal and noise properties, the imaging performance of DPC-CT and its potential over the conventional attenuation-based CT can be fully understood and appreciated.

Ring artifacts appear in computed tomography images if there are too large inhomogeneities between different detector elements. The question of how large inhomogeneities are acceptable is gaining in importance due to the development of energy discriminating photon counting CT, where detector homogeneity is an important design parameter. We propose using the systematic-to-statistical error quotient q, defined as the variance of the expected log-normalized count number between detector elements (dels) divided by the variance of log-normalized count numbers measured with the same del, as a metric of ring artifact visibility. With a simple observer study using simulated images, it is shown that rings are visible in the reconstructed image if q exceeds a threshold which lies close to 1.2 • 10^-3 for 1500 detector elements and 2000 projection angles. It is also shown by visual inspection of simulated images that the threshold value is, to a good approximation, inversely proportional to the number of angle measurements and independent of the number of detector elements. The results suggest that a simple oberver study, together with these scaling relationships, is sufficient for establishing sinogram homogeneity requirements for a particular reconstruction method.

Non-linear iterative reconstruction (IR) algorithms have shown promising improvements in image quality at reduced dose levels. However, IR images sometimes may be perceived as having different image noise texture than traditional filtered back projection (FBP) reconstruction. Standard linear-systems-based image quality evaluation metrics are limited in characterizing such textural differences and non-linear image-quality vs. dose trade-off behavior, hence limited in predicting potential impact of such texture differences in diagnostic task. In an attempt to objectively characterize and measure dose dependent image noise texture and statistical properties of IR and FBP images, we have investigated higher order moments and Haralicks Gray Level Co-occurrence Matrices (GLCM) based texture features on phantom images reconstructed by an iterative and a traditional FBP method. In this study, the first 4 central order moments, and multiple texture features from Haralick GLCM in 4 directions at 6 different ROI sizes and four dose levels were computed. For resolution, noise and texture trade-off analysis, spatial frequency domain NPS and contrastdependent MTF were also computed. Preliminary results of the study indicate that higher order moments, along with spatial domain measures of energy, contrast, correlation, homogeneity, and entropy consistently capture the textural differences between FBP and IR as dose changes. These metrics may be useful in describing the perceptual differences in randomness, coarseness, contrast, and smoothness of images reconstructed by non-linear algorithms.

We model the scatter radiation from the dose compensator in computed tomography (CT) using Monte Carlo (MC) simulation. Dose compensator reduces the patient radiation dose but also adds some scatter radiation which can reach detectors and degrade image quality. This scattered radiation impacts high contrast edges such as bone-soft tissue and tissue-air interfaces, similar to the well known off-focal radiation from the X-ray tube. To the best of our knowledge, this is the first study that uses MC simulation to model and investigate the compensator scatter radiation. Since detectors are far from the scatter source and have small solid angle, traditional MC is not computationally efficient. We have implemented variance reduction technique called forced detection (FD) to improve the computational efficiency, and achieved an improvement factor of 42,800. The simulated scatter to primary ratio (SPR) ranges from a few percent to 11% across the detectors. Simulations of water phantoms indicate that such scatter can lead to significant effects along edges as large as 20-30 HU, necessitating the need for correction. We demonstrate a correction in FBP reconstruction, somewhat similar to off-focal correction, and show that it significantly reduces the level of the artifact along the edge of the phantom.

In order to decrease the patient's radiation exposure from Computed Tomography, the new CT geometry CTDOR has been invented. It consists of two data sets: A conventional arc or flat panel detector and a mask ring with shieldings on the outside and detectors on the inside separated by windows. Combined with the reconstruction algorithm OPED, it has the theoretical potential to decrease the dose about 50% while providing the same image quality as conventional systems. First steps to evaluate this theory were done with a mask ring demonstrator combined with a conventional C-arm device. Although the quality of the demonstrator is limited, this set-up was supposed to demonstrate how the combination of the two data sets works in principle. Preliminary results from earlier studies, however, provided images of rather poor quality. This work presents better images obtained with an optimized data treatment. We showed that most artifacts are eliminated and that we get sharper images with higher contrast compared to the images reconstructed from the single data sets and compared to the earlier study. Regarding the limitations of the set-up, the resulting images were remarkably good. CTDOR is therefore a promising method, which is worth to perform further studies.

Recently iterative reconstruction algorithms with total variation (TV) regularization have shown its tremendous power in image reconstruction from few-view projection data, but it is much more demanding in computation. In this paper, we propose an accelerated augmented Lagrangian method (ALM) for few-view CT reconstruction with total variation regularization. Experimental phantom results demonstrate that the proposed method not only reconstruct high quality image from few-view projection data but also converge fast to the optimal solution.

The quantification of lung nodule volume based on CT images provides valuable information for disease diagnosis and staging. However, the precision of the quantification is protocol, system, and technique dependent and needs to be evaluated for each specific case. To efficiently investigate the quantitative precision and find an optimal operating point, it is important to develop a predictive model based on basic system parameters. In this study, a Fourier-based metric, the estimability index (e') was proposed as such a predictor, and validated across a variety of imaging conditions. To first obtain the ground truth of quantitative precision, an anthropomorphic chest phantom with synthetic spherical nodules were imaged on a 64 slice CT scanner across a range of protocols (five exposure levels and two reconstruction algorithms). The volumes of nodules were quantified from the images using clinical software, with the precision of the quantification calculated for each protocol. To predict the precision, e' was calculated for each protocol based on several Fourier-based figures of merit, which modeled the characteristic of the quantitation task and the imaging condition (resolution, noise, etc.) of a particular protocol. Results showed a strong correlation (R²=0.92) between the measured and predicted precision across all protocols, indicating e' as an effective predictor of the quantitative precision. This study provides a useful framework for quantification-oriented optimization of CT protocols.

Flat panel cone-beam computed tomography (CBCT) is developing to the state-of-the-art technique in several medical disciplines such as dental and otorhinolaryngological imaging. Dental and otorhinolaryngological CBCT systems offer a variety of different field-of-view sizes from 6.0 to 17.0 cm. Standard phantoms are only designed for the use in multi-slices CT (MSCT) and there is no phantom which provides detail structures for all common characteristic values and Hounsfield calibration. In this study we present a new phantom specially designed for use with MSCT and CBCT systems providing detail structures for MTF, 3D MTF, NPS, SNR, geometric distortion and HU calibration. With this phantom you'll only need one acquisition for image quality investigation and assurance. Materials and methods: The phantom design is shown in figure 1. To investigate the practicability, the phantom was scanned using dedicated MSCT-scanners, 3D C-arms und digital volume tomographs. The acquired axial image stacks were analyzed using a dedicated computer program, which is provided as an ImageJ plugin. The MTF was compared to other methodologies such as a thin wire, a sphere or noise response [10, 13, 14]. The HU values were also computed using other common methods. Results: These results are similar to the results of others studies [10, 13, 14]. The method has proven to be stable and delivers comparable results to other methodologies such as using a thin wire. The NPS was calculated for all materials. Furthermore, CT numbers for all materials were computed and compared to the desired values. The measurement of geometric deformation has proven to be accurate. Conclusion: A unique feature of this phantom is to compute the geometric deformation of the 3D-volume image. This offers the chance to improve accuracy, e.g. in dental implant planning. Another convenient feature is that the phantom needs to be scanned only once with otorhinolaryngological volume tomographs to be fully displayed. It is shown, that this new phantom provides amazing potential for image quality assurance in MSCT as well as in CBCT. It contains detail structures for all characteristic image quality values. In combination with a software tool it is the silver bullet for image quality assurance.

Onboard CBCT for radiation linear accelerators suffers from limited longitudinal coverage and various image quality problems, especially at wider cone angles. Such problems prevent CBCT being applied in full potentials for many clinical cancer sites, including head-neck, and for many quantitative applications, including tumor response evaluation and daily radiation dose computation. We propose to use CBCT with flexible X-ray source trajectories to overcome these limitations. The core idea is to combine gantry rotation with simultaneous couch motion. Longitudinal coverage can therefore be extended without limitation. Image quality can be enhanced by applying advanced exact CBCT reconstruction algorithm. However, unlike diagnostic CT where helical CBCT is widely used, LINAC onboard CBCT because gantry can only rotate within 360 degrees and couch table cannot move during gantry rotation. To solve the hardware problem, we program the new Varian TrueBeam LINAC machine in developer mode to realize simultaneous gantry and couch motion so to simulate any flexible scan trajectories. We also implemented CBCT simulation algorithms with digital phantoms to support any flexible source trajectories. We implemented and improved Katsevich exact reconstruction algorithm for image reconstruction from projection data obtained in phantom simulations. We have studied a few different source trajectory models including double circle, helical and saddle. The initial digital phantom results were encouraging. The longitudinal coverage was extended. Image quality has been improved using Katsevich reconstruction algorithm. Physics phantom studies on TrueBeam LINAC machine is our next step.

The phase stepping algorithm is commonly used for phase retrieval in grating-based differential phase-contrast (DPC) imaging, which requires multiple intensity images to compute one DPC image. It is not efficient for data acquisition, especially in the case of dynamic imaging using either DPC imaging or DPC-based come beam CT (DPC-CBCT) imaging. A Fourier transform-based approach has been developed for fringe pattern analysis in optics, and it was recently implemented into a synchrotron-based DPC tomography system. In this research, this approach is further developed for a bench-top DPC-CBCT imaging system with a hospital-grade x-ray tube. The key idea is to separate carrier fringes and object information in Fourier domain of the interferogram and to reconstruct the differentiated phase information using the object information. Only one interferogram is required for phase retrieval at a cost of spatial resolution. The fringes of moiré patterns are used as the carrier fringes, and a phantom is scanned to evaluate the approach. Various interferograms with different carrier fringe frequencies are investigated and the reconstruction image quality is evaluated in terms of contrast, noise and sharpness. The results indicated that the DPC images can be effectively retrieved using the Fourier transform-based approach and the reconstructed phase coefficient showed better contrast compared to that of attenuation-based contrast. The spatial resolution is acceptable in the phantom studies although it is not as good as the results of phase-stepping approach. The Fourier transform-based phase retrieval approach is able to greatly simplify data acquisition, to improve the temporal resolution and to make it possible for dynamic DPC-CBCT imaging. It is promising for perfusion imaging where spatial resolution is not a concern.

Differential phase contrast (DPC) imaging, which utilizes phase shift information of X-ray, has the potential of dramatically increasing the contrast in biological sample imaging compared to attenuation-based method that relies on X-ray absorption information, since the X-ray phase is much more sensitive than the attenuation during transmission. In a DPC imaging system, the phase stepping method is widely used to obtain DPC images: at each angle the phase grating is shifted incrementally to produce a set of images and then the so obtained images are used to retrieve DPC image. However, DPC imaging requires a high mechanical precision to perform phase stepping, which is generally one order higher than the period of phase grating. Given that phase grating period is generally 2-4 um, the requirement of mechanical accuracy and stability are very demanding (<0.5um) and difficult to meet in a system with rotating gantry. In this paper, we present a method that is able to greatly relax the requirement of mechanical accuracy and stability by stepping the source grating rather than the analyzer grating. This method is able to increase the system's mechanical tolerance without compromising image quality and make it feasible to install the system on a rotating gantry to perform differential phase-contrast cone beam CT (DPC-CBCT). It is also able to increase the grating shifting precision and as a result improve the reconstructed image quality. Mechanical tolerance investigation and image quality investigation at different phase stepping schemes and different dose levels will be carried out on both the original modality and the new modality, the results will be evaluated and compared. We will deliberately create random mechanical errors in phase stepping and evaluate the resulting DPC images and DPC-CBCT reconstructions. The contrast, noise level and sharpness will be evaluated to assess the influence of mechanical errors. By stepping the source grating, the system is expected to tolerate an error of 6-7 times bigger than that with analyzer grating stepping.

In this paper we present a novel geometric calibration procedure for cone-beam computed tomography (CBCT) devices with arbitrary geometry using a calibration phantom containing steel beads. In contrast to typical calibration procedures the position of the beads does not have to be known precisely as it is also recovered during calibration. In addition, the arrangement of the beads inside the phantom is very flexible and does not have to follow hard constraints. The bead centers are extracted with subpixel precision from the projection images while taking the absorption properties of the calibration phantom into account. Based on the recovered center positions and phantom geometry, the projection geometry is computed for every projection image. This geometry can be arbitrary and does not have to lie on a specific path, e.g. a circle. This allows to calibrate devices with reproducible mechanical errors in the gantry movement. We present an evaluation of the point extraction and the calibration procedure on ground-truth data and show reconstruction results on a device calibrated using the proposed calibration method.

The 'central artefact' is commonly seen in cone beam computed tomography (CBCT) images in case of an offset scanning geometry. This study evaluates the hypothesis that such artefact is caused by the rotationally asymmetric distribution of scatter. Predictions were examined by use of a hybrid simulation model in conjunction with a simple experimental design. The model accounted for the entire imaging chain of CBCT and was parameterized according to the Scanora 3D CBCT system (Soredex-PaloDEx, Finland). Computational cylinder phantoms of different cross-sections and internal structures were simulated under various symmetric and asymmetric settings. The integrated commercial software tool was employed for image reconstruction. Results confirmed that the scatter was the sole source for the central artefact. It was also found that the artefact was particularly apparent with the algebraic iterative reconstruction. The degree of asymmetry, by either the object's structure or eccentric placement, was much more influential than the effect of the truncated imaging field of view. The findings have also been further proved with a computational anatomical phantom of an adult human head. The computer simulation approach, as adopted in this study, is efficient and has broad applications to CBCT.

The contrast of iodine to soft tissue (water) decreases with higher tube voltage in reconstructed 3D X-ray images. Improved acquisition protocols with a tube voltage of about 80 kV for imaging iodine have been proposed earlier for diagnostic CT imaging. We investigate the contrast-to-noise ratio (CNR) and the CNR-to-dose ratio (CDR) for different concentrations of iodinated contrast agent inserts in water background. The tube voltage of the protocol is lowered from 123 kV to 83 kV in 10 kV steps. Measurements with 16 different settings of tube voltage, current and Copper (Cu) filter settings are investigated. The weighted computed tomography dose index (CTDIW) for the new protocol settings is measured. Four protocols with tube voltages between 83 kV and 103 kV and similar X-ray dose are compared to the original protocol. An iodine-water contrast phantom, containing a water filled cylinder with 5 tubes of different mixtures of iodine contrast agent inside a 32 cm PMMA ring, is imaged with each protocol. Increased contrast of the iodine filled tubes to the water background is clearly visible in the reconstructed volumes for lower tube voltage and less copper filtering. The CNR increased by about 100% for the 83 kV protocols compared to 123 kV and the dose utility in terms of the CDR even by 350%. The best results are obtained with the (83 kV, 561 mA, 0.4 mm Cu) - protocol. This protocol may improve iodine contrast agent visibility in various 3D imaging applications. For large patients a higher tube voltage, e.g. the (103 kV, 325 mA, 0.4 mm Cu) - protocol, may be used to avoid tube power limitations at 83 kV. This protocol still has improved iodine contrast compared to the 123 kV protocol and a larger tube power reserve.

Dual energy computed tomography (DECT) has been widely used in clinical practice and has been particularly effective for tissue diagnosis. In DECT the difference of two attenuation coefficients acquired by two kinds of X-ray energy enables tissue segmentation. One problem in conventional DECT is that the segmentation deteriorates in some cases, such as bone removal. This is due to two reasons. Firstly, the segmentation map is optimized without considering the Xray condition (tube voltage and current). If we consider the tube voltage, it is possible to create an optimized map, but unfortunately we cannot consider the tube current. Secondly, the X-ray condition is not optimized. The condition can be set empirically, but this means that the optimized condition is not used correctly. To solve these problems, we have developed methods for optimizing the map (Method-1) and the condition (Method-2). In Method-1, the map is optimized to minimize segmentation errors. The distribution of the attenuation coefficient is modeled by considering the tube current. In Method-2, the optimized condition is decided to minimize segmentation errors depending on tube voltagecurrent combinations while keeping the total exposure constant. We evaluated the effectiveness of Method-1 by performing a phantom experiment under the fixed condition and of Method-2 by performing a phantom experiment under different combinations calculated from the total exposure constant. When Method-1 was followed with Method-2, the segmentation error was reduced from 37.8 to 13.5 %. These results demonstrate that our developed methods can achieve highly accurate segmentation while keeping the total exposure constant.

The dual-energy computed tomography (CT) techniques can be adopted to separate the materials having similar Houndsfield Unit (HU) value such as tissues. In the technique, CT image values can be described as effective atomic number and electron density using the dual-energy equation. In this work, we measured effective atomic number and electron density using dual-energy CT images and assessed the image quality in vascular application. For the effective atomic number assessment, the measurements of a Polymethyl methacrylate (PMMA) and water demonstrated small discrepancies of 3.28 % and 5.56 %, respectively. For electron density measurement, the experimental errors of PMMA and water were 7.83 % and 4.00 %, respectively. The trend obtained when comparing the HU values and absolute values such as effective atomic number and electron density demonstrates that the CNR of the HU values is higher than that of the absolute values such as effective atomic number and electron density. With contrast media having low concentration, it is remarkable that the effective atomic number image occasionally has higher CNR values than the HU images. In this study, small discrepancies between the experimental values and known values were obtained. The CNR values provided meaningful results for the absolute measurements in a dual-energy CT technique.

Dual-energy computed tomography is a novel imaging tool that has the potential to reduce beam hardening artifacts and enhance material separation over conventional imaging techniques. Dual-energy acquisitions can be performed by using a fast kVp technology to switch between acquiring adjacent projections at two distinct x-ray spectra (80 and 140 kVp). These datasets can be used to further compute material density and monochromatic images for better material separation and beam hardening reduction by virtue of the projection domain process. The purpose of this study was to evaluate the feasibility of using dual-energy in cardiac imaging for myocardial perfusion detection and coronary artery lumen visualization. Data was acquired on a heart phantom, which consisted of the chambers and aorta filled with Iodine density solution (500 HU @ 120 kVp), a defect region between the aorta and chamber (40 HU @ 120 kVp), two Iodinefilled vessels (400 HU @ 120 kVp) of different diameters with high attenuation (hydroxyapatite) plaques (HAP), and with a 30-cm water equivalent body ring around the phantom. Prospective ECG-gated single-energy and prospective ECG-gated dual-energy imaging was performed. Results showed that the generated monochromatic images had minimal beam hardening artifacts which improved the accuracy and detection of the myocardial defect region. Material density images were useful in differentiating and quantifying the actual size of the plaque and coronary artery lumen. Overall, this study shows that dual-energy cardiac imaging will be a valuable tool for cardiac applications.

Clinical CT applications such as oncology follow-up using iodine maps require accurate contrast agent (CA) quantification within the patient. Unfortunately, due to beam hardening, the quantification of CA materials like iodine in dual energy systems can vary for different patient sizes and surrounding composition. In this paper we present a novel method that handles this problem which takes into account properly the CA energy dependent attenuation profile. Our method is applicable for different dual energy scanners, e.g. fast kVp switching or dual layer detector array and is fully compatible with image domain material analysis. In this paper we explain the concept of so called landmarks used by our method, and give the mathematical formulation of how to calculate them. We demonstrate by scans of various phantom shapes and by simulations, the robustness and the accuracy of the iodine concentration quantification obtained by our method.

Scan time of spectral-CTs is much longer than conventional CTs due to limited number of x-ray photons detectable by photon-counting detectors. However, the spectral pixel information in spectral-CT has much richer information on physiological and pathological status of the tissues than the CT-number in conventional CT, which makes the spectral- CT one of the promising future imaging modalities. One simple way to reduce the scan time in spectral-CT imaging is to reduce the number of views in the acquisition of projection data. But, this may result in poorer SNR and strong streak artifacts which can severely compromise the image quality. In this work, spectral-CT projection data were obtained from a lab-built spectral-CT consisting of a single CdTe photon counting detector, a micro-focus x-ray tube and scan mechanics. For the image reconstruction, we used two iterative image reconstruction methods, the simultaneous iterative reconstruction technique (SIRT) and the total variation minimization based on conjugate gradient method (CG-TV), along with the filtered back-projection (FBP) to compare the image quality. From the imaging of the iodine containing phantoms, we have observed that SIRT and CG-TV are superior to the FBP method in terms of SNR and streak artifacts.

The image quality entitlement is evaluated for multi-energy bin photon counting (PC) spectral CT relative to that of energy integration an dual kVp (dkVp) imaging. Physics simulations of X-ray projection channel data and CT images are used to map contrast-to-noise metrics for simple numerical phantoms objects with soft tissue, calcium and iodine materials. The benefits are quantified under ideal detector conditions. Spectral optimization yields on the order of 2X benefit for iodine visualization measured by CNR^2/dose in two different imaging modes: optimal energy weighting, and optimal mono-energy imaging. In another case studied, strict dose equivalence is maintained by use of a composite spectrum for PC simulation that combines simultaneously the two kVp excitations used sequentially for dkVp. In this case, mono-energetic imaging of iodine contrast agent is shown to achieve 40% higher dose efficiency for photon counting compared to dual kVp although non-ideal characteristics of the photon counting response can eliminate much of this benefit.

Micro-CT has become a powerful tool for small animal research. Many micro-CT applications require exogenous contrast agents, which are most commonly based on iodine. Despite advancements in contrast agents, single-energy micro-CT is sometimes limited in the separation of two different materials that share similar grayscale intensity values as in the case of bone and iodine. Dual energy micro-CT offers a solution to this separation problem, while eliminating the need for pre-injection scanning. Various dual energy micro-CT sampling strategies are possible, including 1) single source sequential scanning, 2) simultaneous dual source acquisition, or 3) single source with kVp switching. But, no commercial micro-CT system exists in which all these sampling strategies have been implemented. This study reports on the implementation and comparison of these scanning techniques on the same small animal imaging system. Furthermore, we propose a new sampling strategy that combines dual source and kVp switching. Post-sampling and reconstruction, a simple two-material dual energy decomposition was applied to differentiate iodine from bone. The results indicate the time differences and the potential problems associated with each sampling strategy. Dual source scanning allows for the fastest acquisition, but is prone to errors in decomposition associated with scattering and imperfect geometric alignment of the two imaging chains. KVp switching prevents these types of artifacts, but requires more time for sampling. The novel combination between the dual source and kVp switching has the potential to reduce sampling time and provide better decomposition performance.

Spectral reconstruction algorithms for x-ray CT require an accurate spectral model of the system, including the spectrum of the photons emitted by the source and the spectral sensitivity of the detector. Although these components of the spectral model have been characterized in previous studies, there might be additional components that are unaccounted for, such as the inherent filtration of the x-ray source and the detector. In this study, we present a technique for measuring the inaccuracies in the spectral model and accounting for them. This technique entails the acquisition of photon measurements with several materials placed between the source and the detector, and the solution of a linear system of equations. We test the accuracy of this technique in simulations, and demonstrate its potential to improve the results of spectral reconstruction.

In photon counting multi-bin CT, both projection based weighting and image based weighting (designed to maximize the signal-difference-to-noise ratio SDNR) in the final image, are pixel based and do not account for any spatial frequency dependency of signal and noise among the bins. The same weighting scheme will be used when imaging objects with a large fraction of high spatial frequencies and those with predominantly low spatial frequency content. Any effect on the detectability of a certain target due to correlation between detector elements and bins that might arise in pulse height discriminating systems will not be captured using such an approach. We show how to take the spatial frequency dependency of signal and noise transfer for each bin, and the spatial frequency composition of a target, into account when determining optimal task based weights for photon-counting multi-bin CT imaging using the 2D slice detectability index by applying cascaded system analysis to a spectrally resolved photon counting CT detector system with multiple bins.

Interior tomography problem can be solved using the so-called differentiated backprojection-projection onto convex sets (DBP-POCS) method, which requires a priori information within a small area interior to the region of interest (ROI) to be imaged. In theory, the small area wherein the a priori information is required can be in any shape, but most of the existing implementations carry out the Hilbert filtering either horizontally or vertically, leading to a vertical or horizontal strip that may be across a large area in the object. In this work, we specifically re-derive the reconstruction formula in the DBP-POCS fashion with radial Hilbert filtering (namely radial DBP-POCS method henceforth). We implement the radial DBP-POCS method, and thus the small area with the a priori information can be roughly circular (e.g., a sinus or ventricles among other anatomic cavities in human or animal body). We also conduct an experimental evaluation to verify the performance of this practical implementation. The performance of the radial DBP-POCS method with the a priori information in a small circular area is evaluated with projection data of the standard Shepp-Logan phantom simulated by computer. The preliminary performance study shows that, if the a priori information in a small circular area is available, the radial DBP-POCS method can solve the interior tomography problem in a much more practical way at high accuracy. In comparison to the implementations of DBP-POCS method demanding the a priori information in horizontal or vertical strip, the radial DBP-POCS method requires the a priori information within a small circular area only. Such a relaxed requirement on the availability of a priori information can be readily met in practice, because a variety of small circular areas (e.g., air-filled sinuses or fluid-filled ventricles among other anatomic cavities) exist in human or animal body. Therefore, the radial DBP-POCS method with a priori information in a small circular area is feasible in clinical and preclinical practice.

The finite Hilbert transform plays an essential role in the recently developed derivative back projection (DBP) image reconstruction methods. The accuracy of the numerical methods in implementing the finite Hilbert transform has a significant impact on the reconstruction algorithms' performance. A number of numerical implementations using linear and sinc interpolation kernels have been developed to carry out the Hilbert transform. In this study, we propose a new numerical method using a non-linear interpolation kernel to implement the finite Hilber transform. We evaluate the performance of the proposed non-linear implementation of the finite Hilbert transform in DBP image reconstruction and compare it with that of the implementations using the linear and sinc interpolation kernels, in which the contrast-to-noise ratio, noise power spectrum and modulation transfer function are investigated. The preliminary results show that the finite Hilbert transform implemented with the nonlinear interpolation kernel exhibits slightly better noise performance than the ones implemented with the linear or sinc interpolation.

Image reconstruction from sparse-view data in 2D fan-beam CT is investigated by constrained, total-variation minimization. This optimization problem exploits possible sparsity in the gradient magnitude image (GMI). The investigation is performed in simulation under ideal, noiseless data conditions in order to reveal a possible link between GMI sparsity and the necessary number of projection views for reconstructing an accurate image. Results are shown for two, quite different phantoms of similar GMI sparsity.

Myocardial perfusion scans are an important tool in the assessment of myocardial viability following an infarction. Cardiac perfusion analysis using CT datasets is limited by the presence of so-called partial scan artifacts. These artifacts are due to variations in beam hardening and scatter between different short-scan angular ranges. In this research, another angular range dependent effect is investigated: non-uniform noise spatial distribution. Images reconstructed using filtered backprojection (FBP) are subject to this effect. Statistical image reconstruction (SIR) is proposed as a potential solution. A numerical phantom with added Poisson noise was simulated and two swines were scanned in vivo to study the effect of FBP and SIR on the spatial uniformity of the noise distribution. It was demonstrated that images reconstructed using FBP often show variations in noise on the order of 50% between different time frames. This variation is mitigated to about 10% using SIR. The noise level is also reduced by a factor of 2 in SIR images. Finally, it is demonstrated that the measurement of quantitative perfusion metrics are generally more accurate when SIR is used instead of FBP.

This report develops a new strategy for the acceleration of a maximum likelihood (ML) iterative reconstruction algorithm for CT, by selecting a starting image which is closer to the solution of the ML algorithm than the commonly used filtered backprojection image. The starting image is obtained by passing both the acquired projection data and the reconstructed volume though a novel de-noising algorithm which uses the same image penalty function as the ML reconstruction. Clinical examples suggest that a savings of 5-10 iterations of the separable paraboloidal surrogates algorithm per volume is possible when using this new acceleration strategy.

Metal implants in the field of measurement lead to strong artifacts in CT images and reduce the image quality and the diagnostic value severely. We introduce frequency split metal artifact reduction (FSMAR), a conceptually new MAR method which is designed to reduce metal artifacts and preserve details and edges of structures even close to metal implants. There are many MAR methods which simply replace unreliable parts of the projection data by inpainting. FSMAR is a combination of an inpainting-based MAR method with a frequency split approach. Normalized metal artifact reduction (NMAR) is chosen as the inpainting-based MAR method in this work. The high frequencies of the original image, where all rawdata were used for the reconstruction, are combined with an NMAR-corrected image. NMAR uses a normalization step to reduce metal artifacts without introducing severe new artifacts. Algorithms using a frequency split were already used in CT for example to reduce cone-beam artifacts. FSMAR is tested for patient datasets with different metal implants. The study includes patients with hip prostheses, a neuro coil, and a spine fixation. All datasets were scanned with modern clinical dual source CT scanners. In contrast to other MAR methods, FSMAR yields images without the usual blurring close to metal implants.

This paper investigates a common metal artifacts problem in X-ray computed tomography (CT). The artifacts in reconstructed image may render image non-diagnostic because of inaccuracy beam hardening correction from high attenuation objects, satisfactory image could not be reconstructed from projections with missing or distorted data. In traditionally analytical metal artifact reduction (MAR) method, firstly subtract the metallic object part of projection data from the original obtained projection, secondly complete the subtracted part in original projection by using various interpolating method, thirdly reconstruction from the interpolated projection by filtered back-projection (FBP) algorithm. The interpolation error occurred during the second step can make unrealistic assumptions about the missing data, leading to DC shift artifact in the reconstructed images. We proposed a differentiated back-projection (DBP) type MAR method by instead of FBP algorithm with DBP algorithm in third step. In FBP algorithm the interpolated projection will be filtered on each projection view angle before back-projection, as a result the interpolation error is propagated to whole projection. However, the property of DBP algorithm provide a chance to do filter after the back-projection in a Hilbert filter direction, as a result the interpolation error affection would be reduce and there is expectation on improving quality of reconstructed images. In other word, if we choose the DBP algorithm instead of the FBP algorithm, less contaminated projection data with interpolation error would be used in reconstruction. A simulation study was performed to evaluate the proposed method using a given phantom.

This paper focuses on fan-beam image reconstruction from free-form X-ray source trajectory in computed tomography (CT). As in major standard image reconstruction methods, the weighting function of the redundant projection data should be carefully considered. The data redundancy reduction principle aims to compute the projection data only one time during image reconstruction. Usually, an X-ray line that passes through the object point intersects with the radiation source trajectory more than once time. For the symmetric closed circular trajectory, because each X-ray line passing through an object point would intersect with the source trajectory twice, the weighting function of 0.5 is used to handle this problem. However, normally there is no known symmetry property for the free-form X-ray source trajectory. In order to estimate the weighting function we have to calculate the number N of intersection between each X-ray line and every object points, then the weighting function is set to be 1/N such that the summation of weighting operator corresponding to a single X-ray line is the unity. However, the calculation of N is difficult and computationally expensive. In this paper we proposed a new scheme for a robust control of the redundant projection data for both closed and open trajectories. Instead of calculation of N, we assign each intersection point a plus or minus sign according to the proposed weighting function. As a result of summation of successive signs, they cancel out each other and finally equal to the unity. Numerical study was performed to evaluate the proposed weighting scheme using standard Shepp-Logan phantom with butterfly trajectory.

The fundamental principle of fan-beam CT is to reconstruct the object from its known projection by using two basic types of analytical reconstruction formulas filtered back-projection (FBP) and differentiated back-projection (DBP). Both approaches are used for various applications, and each type has its own advantage properties. Discussing their theory analysis and application introduction appeared in several papers. However, there is no previous discussion about the internal mathematical relation between them. In this paper we present a derivation of the fan-beam reconstruction formula from the starting point of parallel-beam projection by using the equality between the Hilbert transform of object`s 2D Radon transform and the Hilbert transform of fan-beam projection. Because the Ramp filter can be divided into a single Hilbert filtering step and a derivatives step, we rewrite the derived reconstruction formula by exchanging the calculation order of the Hilbert filter step and the derivatives step. As a result, not only the mathematical relation between FBP and DBP algorithm is clarified, but also present six formulas that include all type of analytical fan-beam reconstruction algorithms.

Over the last decade, significant progress has been made in terms of treatment of diseases using minimallyinvasive procedures. This progress was facilitated through multiple refinements of the imaging capabilities of C-arm systems in the interventional room, and more sophisticated procedures may become feasible by further refining the performance of these systems. Our primary focus is to eliminate two strong limitations of the current circular cone-beam imaging approach: cone-beam artifacts and limited extent of the volume covered in the direction of the patient bed. To solve this problem, we seek a source trajectory that (i) is complete in terms of Tuy's condition, (ii) can be periodically-repeated without discontinuities to allow long-object imaging, (iii) is practical, and (iv) offers full R-line coverage (an R-line is a line that connects any two source positions). A trajectory that satisfies all of our constraint is the Arc-Extended-Line-Arc(AELA) trajectory. Unfortunately, this trajectory does not allow smooth, continuous scanning at reasonable dose. In this work, we propose a new data acquisition geometry: the Ellipse-Line-Ellipse (ELE) trajectory. This geometry satisfies all of our constraints along with the attractive feature that smooth, continuous scanning at reasonable dose is enabled.

Adaptive statistical iterative reconstruction (ASiR) is a new reconstruction algorithm used in the field of medical X-ray imaging. This new reconstruction method combines the idealized system representation, as we know it from the standard Filtered Back Projection (FBP) algorithm, and the strength of iterative reconstruction by including a noise model in the reconstruction scheme. It studies how noise propagates through the reconstruction steps, feeds this model back into the loop and iteratively reduces noise in the reconstructed image without affecting spatial resolution. In this paper the effect of ASiR on the contrast to noise ratio is studied using the low contrast module of the Catphan phantom. The experiments were done on a GE LightSpeed VCT system at different voltages and currents. The results show reduced noise and increased contrast for the ASiR reconstructions compared to the standard FBP method. For the same contrast to noise ratio the images from ASiR can be obtained using 60% less current, leading to a reduction in dose of the same amount.

A statistical noise reduction approach is proposed for x-ray computed tomography imaging. The isotropic diffusion partial differential equation is derived from image space to projection space in the equi-angular fan-beam geometry and then used to obtain a projection space multi-scale decomposition by iteratively approximating a Gaussian convolution kernel function with an expected variance. Subsequently, the penalized weighted least square methods with three different objective functions are developed and implemented to reduce quantum noise in the projection data. Experimental results of computer simulated projection data have demonstrated the performance of the proposed approach.

This paper introduced a novel method to reconstruct low-dose computed tomography (LDCT) images based on weighted total variation minimization. Previous work has shown that a CT image can be reconstructed from under-sampling data by minimizing the total variation (TV) of the image with some constraints. However, due to the piecewise constant assumption, the conventional TV minimization algorithm often suffers over-smooth on the edges of the resulted image. Considering the anisotropic property of the edge voxels in an image, we proposed a weighted total variation (WTV) minimization algorithm to achieve high-quality LDCT image reconstruction from under-sampling sinogram data. The proposed algorithm was evaluated by computer simulation, where the Shepp-Logan phantom was used. In noise free cases, 20 projection views were enough to accurately reconstruct the phantom image by the WTV approach while the filtered backprojection (FBP) reconstruction failed. In noise cases, different levels of Poisson noise were added to the noise-free sinogram. The local signal to noise ratio (SNR) and global SNR were computed to evaluate the robustness of the WTV algorithm to the noise. The simulation results showed that the WTV algorithm can have much higher global SNR compared to the conventional TV. The relationship between the number of views of WTV algorithm and the mean square error (MSE) of the reconstructed images was also discussed in this paper in order to find the minimum projection numbers required to obtain an adequate reconstruction image for low-dose CT applications.

In computerized tomography, it is important to reduce the image noise without increasing the acquisition dose. Extensive research has been done into total variation minimization for image denoising and sparse-view reconstruction. However, TV minimization methods show superior denoising performance for simple images (with little texture), but result in texture information loss when applied to more complex images. Since in medical imaging, we are often confronted with textured images, it might not be beneficial to use TV. Our objective is to find a regularization term outperforming TV for sparse-view reconstruction and image denoising in general. A recent efficient solver was developed for convex problems, based on a split-Bregman approach, able to incorporate regularization terms different from TV. In this work, a proof-of-concept study demonstrates the usage of the discrete shearlet transform as a sparsifying transform within this solver for CT reconstructions. In particular, the regularization term is the 1-norm of the shearlet coefficients. We compared our newly developed shearlet approach to traditional TV on both sparse-view and on low-count simulated and measured preclinical data. Shearlet-based regularization does not outperform TV-based regularization for all datasets. Reconstructed images exhibit small aliasing artifacts in sparse-view reconstruction problems, but show no staircasing effect. This results in a slightly higher resolution than with TV-based regularization.

Four dimensional computed tomography (4D-CT) is very important for treatment planning in thorax or abdomen area, e.g. for guiding radiation therapy planning. The respiratory motion makes the reconstruction problem illposed. Recently, compressed sensing theory was introduced. It uses sparsity as a prior to solve the problem and improves image quality considerably. However, the images at each phase are reconstructed individually. The correlations between neighboring phases are not considered in the reconstruction process. In this paper, we propose the spatial-temporal total variation regularization (STTVR) method which not only employs the sparsity in the spatial domain but also in the temporal domain. The algorithm is validated with XCAT thorax phantom. The Euclidean norm of the reconstructed image and ground truth is calculated for evaluation. The results indicate that our method improves the reconstruction quality by more than 50% compared to standard ART.

Previously the Orthogonal Polynomial Expansion on the Disk (OPED) algorithm was presented. Further, in prototype experiments in combination with the CT D`or geometry feasibility was demonstrated. In this study we implemented OPED with a clinical Scanner, and evaluated the potential using phantom studies. All studies were acquired on a Siemens Somatom 64 (Erlangen, Germany) scanner, where raw projection data were reconstructed with the conventional FBP reconstruction and the OPED algorithm. OPED allows one to use fan beam geometry directly without any additional procedures such as interpolation or rebinning if using the CT D`or geometry. In particular, OPED describes an approximation of the image function as a sum of polynomials using Chebychev polynomials. For performance evaluation, the Catphan phantom 600 was imaged. OPED Images where reconstructed using C++ and MATLAB® .We measured uniformity, MTF and CNR for different dose levels and compared these to standard FBP images reconstructions with different filter kernels. The integration and interpretation of the MDCT projection data for the OPED algorithm was accomplished. Reconstruction time is about 6 s on Quad-Core 3 GHz Intel Xeon processor. Typical artifacts are reduced when applying OPED. Using OPED the MTF maintains constant over the whole FOV. Uniformity and CNR are equal compared to FBP. Advantages of OPED were demonstrated by applying the algorithm to projections images from a clinical MDCT scanner. In the future, we see OPED applications for low-dose or limited angle geometries to reduce the radiation dose while improving diagnostic quality of the reconstructed slices.

Advantages of iterative reconstruction (IR) algorithms over standard filtered backprojection (FBP) algorithms include improved resolution and better noise performance, and many IR algorithms have been described in the literature. More recently model-based IR algorithms (MBIR) have been developed, which incorporate accurate system models into IR, resulting in better image quality than IR algorithms without a system model. This work investigates the resolution improvement achieved when a system optics model (SOM) has been included in a standard OS-SART algorithm. Three OS-SART algorithms have been compared: (1) "Pencil beam" (IR-P) with no system optics; (2) system optics included in forward projection (IR-SOM-FP), and (3) system optics included in both forward and backprojection (IRSOM-FPBP). Simulated reconstructions of a 0.2 mm bead show that IR-SOM-FPBP produced a FWHM resolution of 0.41 mm, considerably better than FBPJ (0.87 mm), IR-P (0.63 mm), and IR-SOM-FP (0.59 mm).

Iterative reconstruction methods have emerged as a promising avenue to reduce dose in CT imaging. Another, perhaps less well-known, advance has been the development of inverse geometry CT (IGCT) imaging systems, which can significantly reduce the radiation dose delivered to a patient during a CT scan compared to conventional CT systems. Here we show that IGCT data can be reconstructed using iterative methods, thereby combining two novel methods for CT dose reduction. A prototype IGCT scanner was developed using a scanning beam digital X-ray system - an inverse geometry fluoroscopy system with a 9,000 focal spot x-ray source and small photon counting detector. 90 fluoroscopic projections or "superviews" spanning an angle of 360 degrees were acquired of an anthropomorphic phantom mimicking a 1 year-old boy. The superviews were reconstructed with a custom iterative reconstruction algorithm, based on the maximum-likelihood algorithm for transmission tomography (ML-TR). The normalization term was calculated based on flat-field data acquired without a phantom. 15 subsets were used, and a total of 10 complete iterations were performed. Initial reconstructed images showed faithful reconstruction of anatomical details. Good edge resolution and good contrast-to-noise properties were observed. Overall, ML-TR reconstruction of IGCT data collected by a bench-top prototype was shown to be viable, which may be an important milestone in the further development of inverse geometry CT.

Stationary source inverse-geometry CT (SS-IGCT) has been proposed as a new system architecture that has several key advantages over traditional cone beam CT (CBCT). One advantage is the potential for acquiring a large volume of interest with minimal cone-beam artifacts and with very high temporal resolution. We anticipate that SS-IGCT will use large, stationary source arrays, with gaps in between separate source array modules. These gaps make reconstruction challenging because most analytic reconstruction algorithms assume a continuous source trajectory. SS-IGCT is capable of producing the same dataset as a traditional scanner taking multiple overlapping axial scans, but with segments of the views missing from each axial scan because of gaps. We propose the following, two-stage volumetric reconstruction algorithm. In the first stage, the missing rays are estimated in a spatially varying fashion using available data and geometric considerations, and reconstruction proceeds with standard algorithms. The missing data are then re-estimated by a forward projection step. These new estimates are quite good and the reconstruction can be performed again using any algorithm that supports multiple parallel axial scans. Although inspired by iterative reconstruction, our algorithm only needs one "iteration" of forward- and back-projection in practice and is efficient. Simulations of a thorax phantom were performed showing the efficacy of this technique and the ability of SS-IGCT to suppress cone-beam artifacts compared to conventional CBCT. The noise and resolution characteristics are comparable to that of CBCT.

Model-based iterative reconstruction (MBIR) methods based on maximum a posteriori (MAP) estimation have been recently introduced to multi-slice CT scanners. The model-based approach has shown promising image quality improvement with reduced radiation dose compared to conventional FBP methods, but the associated high computation cost limits its widespread use in clinical environments. Among the various choices of numerical algorithms to optimize the MAP cost function, simultaneous update methods such as the conjugate gradient (CG) method have a relatively high level of parallelism to take full advantage of a new generation of many-core computing hardware. With proper preconditioning techniques, fast convergence speeds of CG algorithms have been demonstrated in 3D emission and 2D transmission reconstruction. However, 3D transmission reconstruction using preconditioned conjugate gradient (PCG) has not been reported. Additional challenges in applying PCG in 3D CT reconstruction include the large size of clinical CT data, shift-variant and incomplete sampling, and complex regularization schemes to meet the diagnostic standard of image quality. In this paper, we present a ramp-filter based PCG algorithm for 3D CT MBIR. Convergence speeds of algorithms with and without using the preconditioner are compared.

Despite the advance in iterative reconstruction methods for reducing metal artifacts, Feldkamp (FDK) based algorithms continue to be the most widely used CT reconstruction in medicine. While computationally efficient, FDK performs poorly in the presence of metallic objects. Projection-completion algorithms have been used to suppress metallic artifacts and, hence, improve image quality in FDK reconstruction. Here, we present a threedimensional adaptive filtering method that performs projection completion. It takes into account the metal content fraction in the voxel and applies a correction in the projections. This algorithm uses a fast, simple, forward-projection method to obtain accurate metal probability regions in the projection space. We compare our results with those obtained using projection completion by linear interpolation on a dental cone-beam CT.

The scattering noise artifact resulted in low-dose projection in repetitive cone-beam CT (CBCT) scans decreases the image quality and lessens the accuracy of the diagnosis. To improve the image quality of low-dose CT imaging, the statistical filtering is more effective in noise reduction. However, image filtering and enhancement during the entire reconstruction process exactly may be challenging due to high performance computing. The general reconstruction algorithm for CBCT data is the filtered back-projection, which for a volume of 512×512×512 takes up to a few minutes on a standard system. To speed up reconstruction, massively parallel architecture of current graphical processing unit (GPU) is a platform suitable for acceleration of mathematical calculation. In this paper, we focus on accelerating wavelet denoising and Feldkamp-Davis-Kress (FDK) back-projection using parallel processing on GPU, utilize compute unified device architecture (CUDA) platform and implement CBCT reconstruction based on CUDA technique. Finally, we evaluate our implementation on clinical tooth data sets. Resulting implementation of wavelet denoising is able to process a 1024×1024 image within 2 ms, except data loading process, and our GPU-based CBCT implementation reconstructs a 512×512×512 volume from 400 projection data in less than 1 minute.

In this paper, we express a concept that the complete reconstruction process should include the computation part on GPUs and the data storage part. We propose a multi-thread scheduling (MTS) method to implement the FDK algorithm, to coordinate?? the computing and storage time. In this method we use multi0-threads to control the GPUs and a separate thread to accomplish data storage, so as to cover the calculation and data storage in time process. In addition, we use the four-channel texture to maintain symmetrical projection data in CUDA framework, which can reduce the calculation time significantly.. Numerical experiment shows that the time cost of the whole process with our method is almost the same as the data storage time cost.

Since the introduction of ASiR, its potential in noise reduction has been reported in various clinical applications. However, the influence of different scan and reconstruction parameters on the trade off between ASiR's blurring effect and noise reduction in low contrast imaging has not been fully studied. Simple measurements on low contrast images, such as CNR or phantom scores could not explore the nuance nature of this problem. We tackled this topic using a method which compares the performance of ASiR in low contrast helical imaging based on an assumed filter layer on top of the FBP reconstruction. Transfer functions of this filter layer were obtained from the noise power spectra (NPS) of corresponding FBP and ASiR images that share the same scan and reconstruction parameters. 2D transfer functions were calculated as sqrt[NPSASiR(u, v)/NPSFBP(u, v)]. Synthesized ACR phantom images were generated by filtering the FBP images with the transfer functions of specific (FBP, ASiR) pairs, and were compared with the ASiR images. It is shown that the transfer functions could predict the deterministic blurring effect of ASiR on low contrast objects, as well as the degree of noise reductions. Using this method, the influence of dose, scan field of view (SFOV), display field of view (DFOV), ASiR level, and Recon Mode on the behavior of ASiR in low contrast imaging was studied. It was found that ASiR level, dose level, and DFOV play more important roles in determining the behavior of ASiR than the other two parameters.

We propose a new nonlocal regularization method for PET image reconstruction with the aid of high-resolution anatomical images. Unlike conventional reconstruction methods using prior anatomical information, our method using nonlocal regularization does not require additional processes to extract anatomical boundaries or segmented regions. The nonlocal regularization method applied to anatomy-based PET image reconstruction is expected to effectively reduce the error that often occurs due to signal mismatch between the PET image and the anatomical image. We also show that our method can be useful for enhancing the image resolution. To reconstruct the high-resolution image that represents the original underlying source distribution effectively sampled at a higher spatial sampling rate, we model the underlying PET image on a higher-resolution grid and perform our nonlocal regularization method with the aid of the side information obtained from high-resolution anatomical images. Our experimental results demonstrate that, compared to the conventional method based on local smoothing, our nonlocal regularization method enhances the resolution as well as the reconstruction accuracy even with the imperfect prior anatomical information or in the presence of signal mismatch between the PET image and the anatomical image.

Single photon emission computed tomography (SPECT) is an important nuclear medicine imaging technique and has been using in clinical diagnoses. The SPECT image can reflect not only organizational structure but also functional activities of human body, therefore diseases can be found much earlier. In SPECT, the reconstruction is based on the measurement of gamma photons emitted by the radiotracer. The number of gamma photons detected is proportional to the dose of radiopharmaceutical, but the dose is limited because of patient safety. There is an upper limit in the number of gamma photons that can be detected per unit time, so it takes a long time to acquire SPECT projection data. Sometimes we just can obtain highly under-sampled projection data because of the limit of the scanning time or imaging hardware. How to reconstruct an image using highly under-sampled projection data is an interesting problem. One method is to minimize the total variation (TV) of the reconstructed image during the iterative reconstruction. In this work, we developed an OSEM-TV SPECT reconstruction algorithm, which could reconstruct the image from highly under-sampled projection data with non-uniform attenuation. Simulation results demonstrate that the OSEM-TV algorithm performs well in SPECT reconstruction with non-uniform attenuation.

Single photon emission computed tomography (SPECT) provides functional information of the interested organs. Such information is very helpful in the early detection of diseases. However, the resolution of SPECT is low. The cone-beam SPECT reconstruction can improve the photon density and spatial resolution of the reconstructed image. In practice, because of the data sufficiency and scanning efficiency, helical cone-beam geometry is preferred. The iterative helical cone-beam reconstruction is time-consuming; the analytical methods are considerably faster and more efficient in clinic. Due to the attenuation of gamma photons, the analytical SPECT reconstruction problem is more complicated. Attenuation should be compensated for to obtain quantitative results. In this paper, based on the Novikov's reconstruction formula and our Ray-driven Technology, we present an analytical SPECT reconstruction algorithm for helical cone-beam geometry using Ray-driven Technology. The simulation results demonstrate the accuracy and robustness of our method.

In this paper we explore the capabilities of a graphical processing unit (GPU) for the fast calculation of a tomographic projection operator in content-adaptive mesh models (CAMM). We explore the use of two distinct methods, ray-tracing and mesh element projection, both implemented on classical computers (CPU) and GPU. Both methods have already been proposed in the literature for 2D and 3D emission tomography (EM) image reconstruction using a CAMM, however, there was no clear comparison between both methods in terms of computational efficiency, which is an aim of this paper.

Diffuse optical tomography (DOT) is one of the ways to probe highly scattering media such as tissue using low-energy near infra-red light (NIR) to reconstruct a map of the optical property distribution. The interaction of the photons in biological tissue is a non-linear process and the phton transport through the tissue is modelled using diffusion theory. The inversion problem is often solved through iterative methods based on nonlinear optimization for the minimization of a data-model misfit function. The solution of the non-linear problem can be improved by modeling and optimizing the cost functional. The cost functional is f(x) = x^TAx - b^T x + c and after minimization, the cost functional reduces to Ax = b. The spatial distribution of optical parameter can be obtained by solving the above equation iteratively for x. As the problem is non-linear, ill-posed and illconditioned, there will be an error or correction term for x at each iteration. A linearization strategy is proposed for the solution of the nonlinear ill-posed inverse problem by linear combination of system matrix and error in solution. By propagating the error (e) information (obtained from previous iteration) to the minimization function f(x), we can rewrite the minimization function as f(x; e) = (x + e)^TA(x + e) - b^T (x + e) + c. The revised cost functional is f(x; e) = f(x) + e^T Ae. The self guided spatial weighted prior (e^T Ae) error (e, error in estimating x) information along the principal nodes facilitates a well resolved dominant solution over the region of interest. The local minimization reduces the spreading of inclusion and removes the side lobes, thereby improving the contrast, localization and resolution of reconstructed image which has not been possible with conventional linear and regularization algorithm.

Among many molecular imaging modalities, Bioluminescence tomography (BLT) is an important optical molecular imaging modality. Due to its unique advantages in specificity, sensitivity, cost-effectiveness and low background noise, BLT is widely studied for live small animal imaging. Since only the photon distribution over the surface is measurable and the photo propagation with biological tissue is highly diffusive, BLT is often an ill-posed problem and may bear multiple solutions and aberrant reconstruction in the presence of measurement noise and optical parameter mismatches. For many BLT practical applications, such as early detection of tumors, the volumes of the light sources are very small compared with the whole body. Therefore, the L1-norm sparsity regularization has been used to take advantage of the sparsity prior knowledge and alleviate the ill-posedness of the problem. Iterative shrinkage (IST) algorithm is an important research achievement in a field of compressed sensing and widely applied in sparse signal reconstruction. However, the convergence rate of IST algorithm depends heavily on the linear operator. When the problem is ill-posed, it becomes very slow. In this paper, we present a sparsity regularization reconstruction method for BLT based on the two-step iterated shrinkage approach. By employing Two-step strategy of iterative reweighted shrinkage (IRS) to improve IST, the proposed method shows faster convergence rate and better adaptability for BLT. The simulation experiments with mouse atlas were conducted to evaluate the performance of proposed method. By contrast, the proposed method can obtain the stable and comparable reconstruction solution with less number of iterations.

Photoacoustic imaging is a noninvasive biomedical imaging modality for visualizing the internal structure and function of soft tissues. In this paper we present a combined iterative reconstruction algorithm to recover the distribution of optical absorption coefficients. Such a combination of the simultaneous algebraic reconstruction technique (SART) algorithm and the filtered back projection (FBP) algorithm bears the promise to improve the image quality of fast but not accurate. The different between the present proposal and traditional ART algorithm is that we computer a full error image with FBP, which is used to projection differences between original and reconstruction image, and to update the solution in each iteration step. The combined iterative reconstruction algorithm works well in rectification of the measurement and converges quickly onto an accurate estimate of the distribution of absolute absorption coefficients. Numerical simulations demonstrate that the proposed algorithm effectively reduces the artifacts and blurs, and yields better quality of reconstruction image than that with FBP algorithm in terms of both accuracy and robustness to noise.