A parallel reconstruction method, based on an iterative maximum likelihood (ML) algorithm, is developed to provide fast reconstruction for digital tomosynthesis mammography. Tomosynthesis mammography acquires 11 low-dose projections of a breast by moving an x-ray tube over a 50° angular range. In parallel reconstruction, each projection is divided into multiple segments along the chest-to-nipple direction. Using the 11 projections, segments located at the same distance from the chest wall are combined to compute a partial reconstruction of the total breast volume. The shape of the partial reconstruction forms a thin slab, angled toward the x-ray source at a projection angle 0°. The reconstruction of the total breast volume is obtained by merging the partial reconstructions. The overlap region between neighboring partial reconstructions and neighboring projection segments is utilized to compensate for the incomplete data at the boundary locations present in the partial reconstructions. A serial execution of the reconstruction is compared to a parallel implementation, using clinical data. The serial code was run on a PC with a single PentiumIV 2.2GHz CPU. The parallel implementation was developed using MPI and run on a 64-node Linux cluster using 800MHz Itanium CPUs. The serial reconstruction for a medium-sized breast (5cm thickness, 11cm chest-to-nipple distance) takes 115 minutes, while a parallel implementation takes only 3.5 minutes. The reconstruction time for a larger breast using a serial implementation takes 187 minutes, while a parallel implementation takes 6.5 minutes. No significant differences were observed between the reconstructions produced by the serial and parallel implementations.

PURPOSE: Rheumatoid arthritis (RA) of the hand is a significant healthcare problem. Techniques to accurately quantity the structural changes from RA are crucial for the development and prescription of therapies. Analysis of radiographic joint space width (JSW) is widely used and has demonstrated promise. However, radiography presents a 2D view of the joint. In this study we performed tomosynthesis reconstructions of proximal interphalangeal (PIP), and metacarpophalangeal (MCP) joints to measure the 3D joint structure. METHODS: We performed a reader study using simulated radiographs of 12 MCP and 12 PIP joints from skeletal specimens imaged with micro-CT. The tomosynthesis technique provided images of reconstructed planes with 0.75 mm spacing, which were presented to 2 readers with a computer tool. The readers were instructed to delineate the joint surfaces on tomosynthetic slices where they could visualize the margins. We performed a quantitative analysis of 5 slices surrounding the central portion of each joint. Reader-determined JSW was compared to a gold standard. As a figure of merit we calculated the average root-mean square deviation (RMSD). RESULTS: RMSD was 0.22 mm for both joints. For the individual joints, RMSD was 0.18 mm (MCP), and 0.26 mm (PIP). The reduced performance for the smaller PIP joints suggests that a slice spacing less than 0.75 mm may be more appropriate. CONCLUSIONS: We have demonstrated the capability of limited 3D rendering of joint surfaces using digital tomosynthesis. This technique promises to provide an improved method to visualize the structural changes of RA.

Multi-slice helical CT scanners allow for much faster scanning and better x-ray utilization than do their single-slice predecessors, but they engender considerably more complicated data sampling patterns due to the interlacing of the samples from different rows as the patient is translated. Characterizing and optimizing this sampling is challenging because the conebeam geometry of such scanners means that the projections measured by each detector row are at least slightly oblique, making it difficult to apply standard multidimensional sampling analyses. In this study, we seek to apply a more general framework for analyzing sampled imaging systems known as Fourier crosstalk analysis. Our purpose in this preliminary work is to compare the information content of the data acquired in three different scanner geometries and operating conditions with ostensibly equivalent volume coverage and average longitudinal sampling interval: a single-slice scanner operating at pitch 1, a four-slice scanner operating at pitch 3 and a 15-slice scanner operating at pitch 15. We find that moving from a single-slice to a multi-slice geometry introduces longitudinal crosstalk characteristic of the longitudinal sampling interval between periods of individual each detector row, and not of the overall interlaced sampling pattern. This is attributed to data inconsistencies caused by the obliqueness of the projections in a multi-slice/conebeam configuration. However, these preliminary results suggest that the significance of this additional crosstalk actually decreases as the number of detector rows increases.

In addition to a conventional Computed Tomography (CT) image, dual energy (dual kVp) imaging can be used to generate an image of the same anatomy that represents the equivalent density of a particular material, for example, calcium, iodine, water, etc. This image can be used to improve the differentiation of materials as well as improve the accuracy of absolute density measurements in a cross-sectional image. It is important to understand the certainty of the estimation of the density of the material. Both simulations and measurements are used to quantify these errors. Data are acquired using a flat-panel based volumetric CT system, by taking two scans and adjusting the maximum energy of the source spectrum (kVp). Physics based simulations are used to compare with the measurements. After validating the simulation algorithms, the accuracy of the dual kVp method is determined using the simulations in a perturbation study.

A complete 32 slice CT detector system has been constructed which uses back illuminated photodiodes (BIPs). Individual detector modules in the system incorporate the BIPs along with highly integrated A/D conversion electronics on the same substrate. A symmetrical mechanical structure allows the system to be compact and lightweight for use at high rotational speeds. The unique design also has the advantage of having no internal cables. The current BIP exhibits a higher level of crosstalk between photosensitive elements when compared to a conventional photodiode. Differences in the crosstalk level at detector module boundaries can cause artifacts unless the crosstalk can either be reduced or a software correction made. In order to show that the BIP is a viable technology for use in multislice CT, a performance evaluation of the complete BIP system along with its associated mechanical, electrical and software components is required. The 32 slice detector system has been mounted to a rotating CT scanner for image performance evaluations. Measurements of low contrast sensitivity, MTF, limiting resolution and other parameters have been done. A crosstalk correction algorithm has also been developed and evaluated under different conditions. Low contrast sensitivity, MTF and limiting resolution of the system match those of a current conventional CT scanner of similar geometry. The crosstalk correction effectively eliminates artifacts caused by non-uniform crosstalk at module boundaries. MTF and noise properties before and after crosstalk correction match theoretical values.

We report the implementation and first test results of a two-channel spectral Computed Tomography (CT) prototype. We use an energy-resolving CT detector with a sandwich-like two layer set-up. Compared to dual-energy approaches with tube voltage switching, it yields a low and a high energy channel in a one shot measurement. We explain the basic set-up of the system and its calibration. The effects of spectral weighting are examined and the weighting functions w(E) of the detector channels are calculated. We present spectral image data of a water phantom, a set of calibration materials and an organic sample. Finally, we show how the data can be used for quantitative CT measurements. The system is work in progress and currently not available in the United States.

In oncology applications, images obtained from CT scanners are often used to estimate the radiation dose delivered to a target organ. Since the patient needs to be positioned in a similar manner as in a therapy machine, a part of the patient is often outside the FOV defined by the CT scanner. The patient anatomy outside FOV may lie in the treatment beam path and must be considered when calculating dose delivery. In addition, truncated projections often produce image artifacts and make attenuation estimation difficult. In this paper, we propose a reconstruction algorithm that allows adequate estimation of the object outside the FOV. We make use of the fact that the total attenuation of each ideal projection in a parallel sampling geometry remains constant over all views. To overcome the small fluctuation resulting from the non-perfect calibrations and patient motion, we use projections of two neighboring non-truncated views as the basis of the attenuation estimation. We use the magnitudes and slopes of the samples at the location of truncation to estimate the water cylinders that can best fit to the projection data outside the FOV. To improve the robustness of the algorithm, continuity constraints are placed on the fitting parameters. Extensive phantom and patient experiments were conducted to test the robustness and accuracy of the proposed algorithm. Results show that CT number accuracy is fully recovered inside the scan FOV. In all cases, the shape of the truncated object outside the FOV is recovered to an accuracy of a few millimeters.

In this paper, the performance of focused lamellar anti-scatter grids, which are currently used in fluoroscopy, is studied in order to determine guidelines of grid usage for flat detector based cone beam CT. The investigation aims at obtaining the signal to noise ratio improvement factor by the use of anti-scatter grids. First, the results of detailed Monte Carlo simulations as well as measurements are presented. From these the general characteristics of the impinging field of scattered and primary photons are derived. Phantoms modeling the head, thorax and pelvis regions have been studied for various imaging geometries with varying phantom size, cone and fan angles and patient-detector distances. Second, simulation results are shown for ideally focused and vacuum spaced grids as best case approach as well as for grids with realistic spacing materials. The grid performance is evaluated by means of the primary and scatter transmission and the signal to noise ratio improvement factor as function of imaging geometry and grid parameters. For a typical flat detector cone beam CT setup, the grid selectivity and thus the performance of anti-scatter grids is much lower compared to setups where the grid is located directly behind the irradiated object. While for small object-to-grid distances a standard grid improves the SNR, the SNR for geometries as used in flat detector based cone beam CT is deteriorated by the use of an anti-scatter grid for many application scenarios. This holds even for the pelvic region. Standard fluoroscopy anti-scatter grids were found to decrease the SNR in many application scenarios of cone beam CT due to the large patient-detector distance and have, therefore, only a limited benefit in flat detector based cone beam CT.

Cascaded-systems analyses have been used successfully by many investigators to describe signal and noise transfer in quantum-based x-ray detectors in medical imaging. However, the Fourier-based linear-systems approach is only valid when assumptions of linearity and shift invariance are satisfied. Digital detectors, in which a bounded image signal is spatially integrated in discrete detector elements, are not shift invariant in their response. In addition, many detectors make use of fiber optics or structured phosphors such as CsI to pass light to a photodetector-both of which have a shift-variant response. These issues raise serious concerns regarding the validity of Fourier-based approaches for describing the signal and noise performance of these detectors. We have used a Monte Carlo approach to compare the image Wiener noise power spectrum (NPS) with that predicted using a Fourier-based approach when these assumptions fail. It is shown that excellent agreement is obtained between Monte Carlo results and those obtained using a Fourier-based wide-sense cyclostationary analysis, including the description of noise aliasing. A simple model of a digital detector coupled to a fiber optic bundle is described using a novel cascaded cyclostationary approach in which image quanta are integrated over fiber elements and then randomly re-distributed at the fiber output. While the image signal sometimes contains significant non-stationary beating artifacts, the Monte Carlo results generally show good agreement with Fourier models of the NPS when noise measurements are made over a sufficiently large region of interest.

While analysis of the image noise-power spectrum (NPS) and noise-equivalent quanta (NEQ) are important aspects of imaging system characterization, such metrics are in themselves insufficient descriptions of imaging performance in that they make no account of the imaging task. This paper seeks to quantitatively incorporate imaging task in imaging performance evaluation by combining NEQ with a variety of idealized spatial-frequency-dependent task functions to yield the model observer detectability index. The approach is applied to the case of fully 3D volumetric imaging by flat-panel detector cone-beam CT through analysis of 3D detectability for conditions of varying dose, reconstruction filter, and voxel size. Generalization of the NEQ through incorporation of background "anatomical" noise suggests significant degradation in model observer performance, and the effect is quantified for a variety of detection and discrimination tasks. For anatomical noise modeled according to 1/f b statistics in power-spectral density, the effect is shown to be most severe for low-frequency detection tasks, and somewhat less for mid-to-high spatial frequency tasks, such as discrimination and localization. By considering the fully 3D NEQ, which is known to be asymmetric for cone-beam CT, a compelling hypothesis is realized regarding the detection of structures in volumetric CT images - specifically, that the detectability is different in axial versus sagittal / coronal domains due to asymmetry in the NEQ between these domains, compared to the case in which 3D data are fully interrogated (e.g., by a machine algorithm). This has significant implications for 3-D imaging modalities, including flat-panel cone-beam CT, where the NPS exhibits asymmetric frequency characteristics [viz., high-pass (filtered-ramp) in the transverse domain and low-pass in the longitudinal]. The impact of asymmetric NPS characteristics on detectability was investigated by analysis of the 3D detectability for imaging tasks corresponding to detection and discrimination of fine and low-contrast structures.

Cascaded models have been used by a number of investigators to derive analytic expressions for the Wiener noise power spectrum (NPS) and detective quantum efficiency (DQE) based on design parameters to evaluate the performance of medical x-ray imaging systems. These analytic models are required to establish operating benchmarks and compare the performance of real detectors. Although application of the cascaded approach has had several successes, its contribution is often limited when applied to complex models. This is due to the fact that while final algebraic expressions can be relatively simple, the cascaded approach involves the manipulation of many hundreds of terms. To overcome this limitation a computational engine has been developed using Matlab's Simulink and symbolic math capabilities. Based on a recursive programming approach, this engine generates analytic expressions of NPS and DQE for cascaded models of arbitrary complexity. In order to validate the resulting expressions, a Monte Carlo (MC) simulation program has been developed that performs an analysis based on C-code generated by the computational engine for each model. The Monte Carlo code generates an incident quantum image as a Poisson distribution of quanta. This distribution is passed through appropriate serial and parallel cascades of modules representing elementary processes and is used to calculate the NPS for comparison with the analytic NPS. Results show excellent agreement between Monte Carlo and theoretical expressions. We are at the stage where complex cascaded modelling is becoming practical tool in the design of new detector systems.

An often neglected assumption related to detector performance metrics such as the modulation transfer function (MTF), noise power spectrum (NPS), and detective quantum efficiency (DQE) is that they only apply to a small region around the centre of an x-ray image. In the periphery of an image, image formation is from obliquely incident x rays. These off-axis x rays will introduce an additional degrading effect on the above detector performance metrics. In our study, we use Monte Carlo simulations to quantify the effects of off-axis radiation on the MTF, NPS, and DQE on common diagnostic x-ray detectors. In our simulations, we vary the incident angle of x rays between 0° and 12°, which is a typical range of divergence in diagnostic x-ray imaging. In the case of amorphous selenium, our results show that off-axis incident x rays degrade the MTF above 5 cycles/mm with increasing severity at higher incident angles and x-ray energy, and more importantly has very little effect on the NPS. Hence, the impact is more severe on the DQE due to the MTF squared dependency. For an incident x-ray angle of 12° (~13 cm from central axis or chest wall in mammography), the DQE falls to 50% of its initial value at 10 and 7 cycles/mm for x-ray energies of 20 and 40 keV, respectively. This loss of signal-to-noise ratio may be most significant near the skin line in mammography studies.

Recently developed flat-panel detectors have been proven to have a much better image quality than conventional electronic portal imaging devices (EPIDs) used in radiation therapy. They are, however, not yet ideal for portal imaging application primarily due to the low x-ray absorption for megavoltage(MV) x-rays, i.e., low quantum efficiency (QE), typically on the order of 2-4% as compared to the theoretical limit of 100%. A significant increase of QE is desirable for applications such as MV cone-beam computed tomography (MVCT) and MV fluoroscopy. Our goal is to develop a new generation of area detectors for radiotherapy treatment verification, with a QE an order of magnitude higher than that of current flat-panel systems and an equivalent spatial resolution. In this paper, we will first discuss the rationale and the challenges in designing a high QE detector for portal imaging application and give an overview of previous designs and their limitations. We will then introduce our novel design for a high QE detector, which has a thick, dense x-ray direct-conversion layer coupled to a 2D active matrix for image storage and readout. The conversion layer is made of high-density metal elements to convert x-rays to electrons and sub-pixel sized cavities filled with an ionization medium (e.g., gas or a-Se) to convert the electrons to free charges that are collected on electrodes connected to the active matrix. The QE, spatial resolution, and sensitivity of the proposed detector have been modeled, and results will be presented. It is shown that this new detector will be quantum noise limited and have both a high QE and a high resolution. Thus, further development based on this novel design is warranted.

The development of fluoroscopic imagers exhibiting performance that is primarily limited by the noise of the incident x-ray quanta, even at very low exposures, remains a highly desirable objective for active matrix flat-panel technology. Previous theoretical and empirical studies have indicated that promising strategies to acheiving this goal include the development of array designs incorporating improved optical collection fill factors, pixel-level amplifiers, or very high-gain photoconductors. Our group is pursuing all three strategies and this paper describes progress toward the systematic development of array designs involving the last approach. The research involved the iterative fabrication and evaluation of a series of prototype imagers incorporating a promising high-gain photoconductive material, mercuric iodide (HgI₂). Over many cycles of photoconductor deposition and array evaluation, improvements ina variety of properties have been observed and remaining fundamental challenges have become apparent. For example, process compatibility between the deposited HgI₂ and the arrays have been greatly improved, while preserving efficient, prompt signal extraction. As a result, x-ray sensitivities within a factor of two of the nominal limit associated with the single-crystal form of HgI₂ have been observed at relatively low electric fields (~0.1 to 0.6 V/μm), for some iterations. In addition, for a number of iterations, performance targets for dark current stability and range of linearity have been met or exceeded. However, spotting of the array, due to localized chemical reactions, is still a concern. Moreover, the dark current, uniformity of pixel response, and degree of charge trapping, though markedly improved for some iterations, require further optimization. Furthermore, achieving the desired performance for all properties simultaneously remains an important goal. In this paper, a broad overview of the progress of the research will be presented, remaining challenges in the development of this photoconductive material will be outlined, and prospects for further improvement will be discussed.

The dynamic range of many flat panel imaging systems are fundamentally limited by the dynamic range of the charge amplifier and readout signal processing. We developed two new flat panel readout methods that achieve extended dynamic range by changing the read out charge amplifier feedback capacitance dynamically and on a real-time basis. In one method, the feedback capacitor is selected automatically by a level sensing circuit, pixel-by-pixel, based on its exposure level. Alternatively, capacitor selection is driven externally, such that each pixel is read out two (or more) times, each time with increased feedback capacitance. Both methods allow the acquisition of X-ray image data with a dynamic range approaching the fundamental limits of flat panel pixels. Data with an equivalent bit depth of better than 16 bits are made available for further image processing. Successful implementation of these methods requires careful matching of selectable capacitor values and switching thresholds, with the imager noise and sensitivity characteristics, to insure X-ray quantum limited operation over the whole extended dynamic range. Successful implementation also depends on the use of new calibration methods and image reconstruction algorithms, to insure artifact free rebuilding of linear image data by the downstream image processing systems. The multiple gain ranging flat panel readout method extends the utility of flat panel imagers and paves the way to new flat panel applications, such as cone beam CT. We believe that this method will provide a valuable extension to the clinical application of flat panel imagers.

A new concept - an indirect flat-panel detector with avalanche gain - for low dose x-ray imaging has been proposed. The detector consists of an amorphous selenium (a-Se) photoconductor optically coupled to a structured cesium iodide (CsI) scintillator. Under an electric field E_Se, the a-Se is sensitive to light and converts the optical photons emitted from CsI into electronic signal. These signals can be stored and read out in the same fashion as in existing flat-panel detectors. When E_Se is increased to > 90 V/μm, avalanche multiplication occurs. The avalanche gain ranges between 1-800 depending on E_Se and the thickness of the a-Se layer d_Se. The avalanche a-Se photoconductor is referred to as HARP (High-gain Avalanche Rushing amorphous Photoconductor). A cascaded linear system model for the proposed detector was developed in order to determine the optimal CsI properties and avalanche gain for different x-ray imaging applications. Our results showed that x-ray quantum noise limited performance can be achieved at the lowest exposure level necessary for fluoroscopy (0.1 μR) and mammography (0.1 mR) with a moderate avalanche gain of 20 (d = 1-2 μm). A laboratory test system using an existing HARP tube optically coupled (through a lens) to a CsI layer was built and the advantage of avalanche gain in overcoming electronic noise was demonstrated experimentally. One of the advantages of the avalanche gain is that it will permit the use of high resolution (HR) CsI (which due to its low light output has not previously been used in flat-panel detectors) to improve DQE at high spatial frequencies.

In this paper, two new selenium large area detectors are introduced. The first detector is a 24x30cm detector suitable for full-field digital mammography applications. The second detector is a new 43x43cm detector for general radiographic applications. Both detectors are capable of static and dynamic imaging, and are quantum noise limited over the exposure ranges intended for their typical use. For static imaging applications, ghost and lag were compared on both detectors, and no measurable artefacts were reported. On dynamic imaging sequences, lag was shown to be significant on both detectors, and a method for reducing the artifact due to lag was presented.

The dependence of the x-ray sensitivity of a-Se based x-ray image detectors on repeated x-ray exposures and exposure history is studied by considering deep trapping of charge carriers, trapped charges due to previous exposures, bimolecular recombination, space charge effects and electric field dependent electron-hole pair creation energy. We numerically solve the continuity equations of both holes and electrons, trapping rate equations, and the Poison equation across the photoconductor for long pulse x-ray exposures. The electric field distribution across the photoconductor and the relative x-ray sensitivity as a function of cumulated x-ray exposure have been studied for both mammographic and chest radiographic applications. The electric field distribution across the photoconductor has been found to vary widely for high exposures. The relative x-ray sensitivity decreases with increasing cumulated x-ray exposure and tents to saturate. The sensitivity reduction at negative bias is more pronounced than at positive bias.

The DQEs of four digital X-ray detector systems have been measured in accordance with the new international standard IEC 62220-1: two CR detector systems of the same type, a CsI-based indirect flat panel detector and a selenium-based direct flat panel detector. A mobile measurement set-up complying with IEC 62220-1 has been realized. All equipment used was of a specific design, tested and calibrated. A standardized radiation quality (RQA5) was applied, and the air kerma at the detector entrance was varied between about 1 μGy and 20 μGy. The measurements of the two CR detector systems were performed at different sites using different X-ray generators/tubes and were in agreement within 0.02. The maximum DQE values were obtained for the lowest spatial frequency for which the DQE is required to be reported according to the IEC standard, i.e. at 0.5 mm^-1: The maximum DQE value measured was 0.21 for the CR systems, 0.42 for the indirect flat panel detector, and 0.31 for the direct Selenium-based detector. It has been demonstrated that the international standard IEC 62220-1 allows accurate and reliable measurements of the DQE to be conducted. It is now possible to objectively measure and compare DQE values of digital X-ray detector systems.

A flat X-ray detector with lead oxide (PbO) as direct conversion material has been developed. The material lead oxide, which has a very high X-ray absorption, was analysed in detail including Raman spectroscopy and electron microscopy. X-ray performance data such as dark current, charge yield and temporal behaviour were evaluated on small functional samples. A process to cover a-Si TFT-plates with PbO has been developed. We present imaging results from a large detector with an active area of 18 × 20 cm². The detector has 1080 × 960 pixels with a pixel pitch of 184 μm. The linearity of detector response was verified. The NPS was determined with a total dark noise as low as 1800 electrons/pixel. The MTF was measured with two different methods: first with the analysis of a square wave phantom and second with a narrow slit. The MTF at the Nyquist frequency of 2.72 lp/mm was 50 %. We calculated first DQE values of our prototype detector plates. Full size images of anatomic and technical phantoms are shown.

Photoconductive polycrystalline mercuric iodide deposited on flat panel thin film transistor (TFT) arrays is being developed for direct digital X-ray detectors that can perform both radiographic and fluoroscopic medical imaging. The mercuric iodide is either vacuum deposited by Physical Vapor Deposition (PVD) or coated onto the array by a wet Particle-In-Binder (PIB) process. The PVD deposition technology has been scaled up to the 20 cm x 25 cm size required in common medical imaging applications. A TFT array with a pixel pitch of 127 microns is used for these imagers. Arrays of 10 cm x 10 cm size have been used to evaluate performance of mercuric iodide imagers. Radiographic and fluoroscopic images of diagnostic quality at up to 15 pulses per second were demonstrated. As we previously reported, the resolution is limited to the TFT array Nyquist frequency of ~3.9 lp/mm (127 micron pixel pitch). Detective Quantum Efficiency (DQE) has been measured as a function of spatial frequency for these imagers. The DQE is lower than the theoretically calculated value due to some additional noise sources of the electronics and the array. We will retest the DQE after eliminating these noise sources. Reliability and stress testing was also began for polycrystalline mercuric iodide PVD and PIB detectors. These are simplified detectors based upon a stripe electrode or circular electrode structure. The detectors were stressed under various voltage bias, temperature and time conditions. The effects of the stress tests on the detector dark current and sensitivity were determined.

Columnar structured cesium iodide (CsI) scintillators doped with Thallium (Tl) have been used extensively for indirect x-ray imaging detectors. The purpose of this paper is to develop a methodology for systematic investigation of the inherent imaging performance of CsI as a function of thickness and design type. The results will facilitate the optimization of CsI design for different x-ray imaging applications, and allow validation of physical models developed for the light channeling process in columnar CsI layers. CsI samples of different types and thicknesses were obtained from the same manufacturer. They were optimized either for light output (HL) or image resolution (HR), and the thickness ranged between 150 and 600 microns. During experimental measurements, the CsI samples were placed in direct contact with a high resolution CMOS optical sensor with a pixel pitch of 48 microns. The modulation transfer function (MTF), noise power spectrum (NPS) and detective quantum efficiency (DQE) of the detector with different CsI configurations were measured experimentally. The aperture function of the CMOS sensor was determined separately, which allows estimation of the MTF of CsI alone. We also measured the pulse height distribution of the light output from both the HL and HR CsI at different x-ray energies, from which the x-ray quantum efficiency, Swank factor and x-ray conversion gain were determined. Our results showed that the MTF at 5 cycles/mm for the HR type is 50 % higher than for the HL, however at a cost of ~ 36 % reduction in light output. The Swank factor below K-edges of Cs and I is 0.91 and 0.93 for the HR and HL types, respectively, which means that their DQE(0) are essentially identical for the same thickness. The presampling MTF decreases as a function of thickness, and the rate of decrease drops as the thickness increases. This indicates that the light channeling process in CsI improves the MTF of thicker layers more significantly than it does for the thinner ones.

X-ray detection with luminescent screens requires optical signal transfer as an intermediate step between x-ray detection and conversion to an electronic signal. Luminescent screens may be granular (phosphor screens), structured (e.g. CsI) or transparent (scintillators). The optical signal is imaged with lenses, fibre optics, electron optics or by proximity focussing to an electronic detector. Poor focussing or poor optical contact may degrade the signal and noise transfer characteristics, i.e. modulation transfer function (MTF) and detective quantum efficiency (DQE). The case when x-rays are detected with granular luminescent screens, imaged onto flat panel electronic detectors is considered here. The detector assembly often requires layers of glue or protective thin films creating optical gaps, in which light is spread, hence spatial resolution is degraded. The noise spectrum is not necessarily changed the same way. Its exact shape depends on the dominant noise sources in a given detector configuration under the specific operating conditions: The noise of the primary x-ray quanta, noise aliasing and direct x-ray detection by the electronic detection layer are the main contributions in this investigation. Especially at high spatial frequencies small optical gaps in conjunction with white quantum noise from direct x-ray absorption of the electronic imager degrade DQE: A gap of 40 μm between luminescent screen and detector reduces the DQE by 33% at the Nyquist frequency. This was demonstrated with an a-Si imager of 143-μm pixel size and a Lanex Fine luminescent screen operated at 100 kV.

Pulse wave velocity (PWV) is widely used for estimating the stiffness of an artery. It is well known that a stiffened artery can be associated with various diseases and with aging and disease. Usually, PWV is measured using the “foot-to-foot” method. The “foot” of the pressure wave is not clear due to reflected waves and blood noise. Also, PWV is an average indicator of artery stiffness between the two measuring points, and therefore does not identify local stiffness variations. We propose producing a flexural wave in the arterial wall using low frequency localized ultrasound radiation force and measuring the wave velocity along the arterial wall. The wave velocity can be measured accurately over a few millimeters. A mathematical model for wave propagation along the artery is developed with which the Young’s modulus of the artery can be determined from measured wave velocities. Experiments were conducted on a pig carotid artery in gelatin. The wave velocity was measured by the phase change at a known distance for a given frequency. The measured wave velocity is about 3 m/s at 100 Hz and 6.5 m/s at 500 Hz. The real part of complex elastic modulus of the artery is estimated to be 300 kPa.

Urinary stone components can be characterized through their diffraction signatures using monoenergetic or polyenergetic x rays. Sharp characteristic diffraction peaks are observed under monoenergetic conditions, facilitating component separability in measurements of stones with mixed composition. This favors uniqueness of a materials analysis solution. However, these monoenergetic techniques are either impractical (as with synchrotron radiation) or impossible to perform in situ (as with conventional low energy diffractometry). Alternatively, our approach measures signals from x-ray diffraction, or coherent scatter (CS), for stones and their components using polyenergetic x rays from diagnostic equipment, allowing for in situ applications. Although the polyenergetic x-ray spectrum is the primary contributor to the angular broadening of diffraction peaks in our measured CS cross-sections, we show that it is possible to relate the polyenergetic and monoenergetic results through a “non-stationary” convolution operation. This requires the computation of a linear superposition integral of the monoenergetic cross-section with a function representative of the polyenergetic spectrum. Experimentally acquired diffractometry cross-sections of the seven major urinary stone components were subjected to this operation, revealing good agreement of diffraction features with CS. These results indicate that angular resolution is principally hindered by the energy spectrum in CS measurements. Nevertheless, distinct scatter patterns were observed using CS, suggesting that a strictly monoenergetic beam is not required for the depiction of pure stone components.

Here we report on the development of a new molecular imaging technique using inelastic scattering of fast neutrons. Earlier studies demonstrated a significant difference in trace element concentrations between benign and malignant tissue for several cancers including breast, lung, and colon. Unfortunately, the measurement techniques were not compatible with living organisms and this discovery did not translate into diagnostic techniques. Recently we have developed a tomographic approach to measuring the trace element concentrations using neutrons to stimulate characteristic gamma emission from atomic nuclei in the body. Spatial projections of the emitted energy spectra allow tomographic image reconstruction of the elemental concentrations. In preliminary experiments, spectra have been acquired using a 7.5MeV neutron beam incident on several multi-element phantoms. These experiments demonstrate our ability to determine the presence of Oxygen, Carbon, Copper, Iron, and Calcium. We describe the experimental technique and present acquired spectra.

The time resolved/time gated method can be used as a non-invasive technique for measuring changes in the concentration of oxy-and deoxy-hemoglobin in tissues by monitoring absorbed near infrared (NIR) signals in diffusion-reflectance and trans-illuminescence geometries. The objective is to develop novel technology that would enhance the efficiency of sensing in intact organism, in vivo, offering new capabilities for studying biological function by multispectral spectroscopy [1-4]. In this paper, we describe a novel scheme of an excitation source that enhances the spatial resolution and signal to noise ratio of detected NIR absorbed signals in diffusion reflectance spectroscopy technique used for measuring changes in the concentration of oxy-and deoxy-hemoglobin in tissues. We identify the properties and application of long visible and IR wavelength Vertical Cavity Surface Emitting Laser as an alternative 'Excitation Source'. We also investigate the Jitter performance of VCSELs in a high rate transmission systems like; the time gated reflectance absorbance spectroscopy system, as timing jitter of excitation source seriously degrades the temporal resolution of the collected absorbed luminescence at the detector end, which significantly effects the proper monitoring of the changes in the concentration of oxy- and deoxy-hemoglobin within the tissue.

We investigated ultrafast laser-based x-ray (ULX) source as an attractive alternative to a microfocal x-ray tube used in micro-CT systems. The laser pulse duration was in the 30 fs-200 fs range, the repetition rate in the 10 Hz - 1 kHz range. A number of solid targets including Ge, Mo, Rh, Ag, Sn, Ba, La, Nd with matching filters was used. We optimized conditions for x-rays generation and measured: x-ray spectra, conversion efficiency (from laser light to x-rays), x-ray fluence, effective x-ray focal spot size and spatial resolution, contrast resolution and radiation dose. Good quality projection images of small animals in single-and dual-energy mode were obtained. ULX generates narrow x-ray spectra that consist mainly of characteristic lines that can be easily tailored (by changing laser beam target) to the imaging task, (e.g. to maximize contrast while minimizing radiation dose). X-ray fluence can exceed fluence produced by conventional microfocal tube with 10 μm focal-spot hence allowing for faster scans with very high spatial resolution. Changing the laser target, and thus matching the characteristic emission lines with the investigated animal's thickness and composition, can be done quickly and easily. Using narrow emission lines for imaging, instead of broad bremsstrahlung, offers superior dose utilization and limits beam-hardening effects. Employing two narrow emission lines-above and below the absorption edge of a contrast agent-in quick succession allows dual-energy-subtraction micro-CT for imaging with a contrast medium. Dual-energy-subtraction is not practical with a microfocal tube. Compact, robust, ultrafast lasers are commercially available, and their characteristics are rapidly improving. We plan to construct a prototype in vivo ultrafast laser-based micro-CT system.

A reconstruction theory for intensity diffraction tomography (I-DT) has been proposed that permits for the reconstruction of a weakly-scattering object without explicity knowledge of phase information. In this work, we examine the noise properties of I-DT. An explicit expression for the variance of the estimated object function as a function of spatial frequency is derived and employed for understanding the noise properties of images reconstruction in I-DT. It is demonstrated analytically and numerically that the noise properties of I-DT are significantly different from those of conventional diffraction tomography (DT).

Recent results show that bone vasculature is a major contributor to local tissue porosity, and therefore can be directly linked to the mechanical properties of bone tissue. With the advent of third generation synchrotron radiation (SR) sources, micro-computed tomography (μCT) with resolutions in the order of 1 μm and better has become feasible. This technique has been employed frequently to analyze trabecular architecture and local bone tissue properties, i.e. the hard or mineralized bone tissue. Nevertheless, less is known about the soft tissues in bone, mainly due to inadequate imaging capabilities. Here, we discuss three different methods and applications to visualize soft tissues. The first approach is referred to as negative imaging. In this case the material around the soft tissue provides the absorption contrast necessary for X-ray based tomography. Bone vasculature from two different mouse strains was investigated and compared qualitatively. Differences were observed in terms of local vessel number and vessel orientation. The second technique represents corrosion casting, which is principally adapted for imaging of vascular systems. The technique of corrosion casting has already been applied successfully at the Swiss Light Source. Using the technology we were able to show that pathological features reminiscent of Alzheimer’s disease could be distinguished in the brain vasculature of APP transgenic mice. The third technique discussed here is phase contrast imaging exploiting the high degree of coherence of third generation synchrotron light sources, which provide the necessary physical conditions for phase contrast. The in-line approach followed here for phase contrast retrieval is a modification of the Gerchberg-Saxton-Fienup type. Several measurements and theoretical thoughts concerning phase contrast imaging are presented, including mathematical phase retrieval. Although up-to-now only phase images have been computed, the approach is now ready to retrieve the phase for a large number of angular positions of the specimen allowing application of holotomography, which is the three-dimensional reconstruction of phase images.

Conventional film-screen mammography is the most effective tool for the early detection of breast cancer currently available. However, conventional mammography has relatively low sensitivity to detect small breast cancers (under several millimeters) owing to an overlap in the appearances of benign and malignant lesions, and surrounding structure. The limitations accompanying conventional mammography is to be addressed by incorporating a cone beam CT imaging technique with a recently developed flat panel detector. Computer simulation and preliminary studies have been performed to prove the feasibility of developing a flat panel detector-based cone beam CT breast imaging (FPD-CBCTBI) technique. A preliminary system characterization study of flat panel detector-based cone beam CT for breast imaging was performed to confirm the findings in the computer simulation and previous phantom studies using the current prototype cone beam CT scanner. The results indicate that the CBCTBI technique effectively removes structure overlap and significantly improves the detectability of small breast tumors. More importantly, the results also demonstrate CBCTBI offers good image quality with the radiation dose level less than or equal to that of conventional mammography. The results from this study suggest that FPD-CBCTBI is a potentially powerful breast-imaging tool.

A pendant-geometry, cone-beam breast CT scanner has been constructed and is undergoing thorough testing in our facility. The system is capable of acquiring 30 frames/sec in 2×2 binning mode (1024×768 pixels) using a flat panel detector coupled to a thallium-doped cesium iodide scintillator. The DQE of the detector system for RQA5 and RQA9 x-ray beam qualities were computed, and the low frequency DQE values were 65% and 57% respectively at approximately 16 μR/frame. The results also shown that minor improvements in DQE are achieved for exposures greater than 16 μR/frame. It is expected that the scanner will be available for the imaging of human volunteers in the first half of 2004.

Simulation results from previous studies indicate that a quasi-monochromatic x-ray beam can be produced using a newly developed beam filtration technique. This technique utilizes heavy filtration with novel high Z filter materials having k-edges just above those of CsI, producing a near monochromatic beam with mean energy optimized for detection. The value of a near monochromatic x-ray source for a fully 3D tomography application is the expected improved ability to separate tissues with very small differences in attenuation coefficients for a range of uncompressed breast sizes while maintaining dose levels at or below existing dual view mammography. In this study, we experimentally investigate a set of filter materials (Al, Cu, Ag, Ce, W, and Pb), filter thicknesses (10th, 100th, and 200th VL), and tube potentials (40-80 kVp) using a newly constructed test apparatus. Initial experimental results corroborate simulations and indicate that this approach can improve image quality (SNR) at constant dose. Al, Cu, W, and Pb provide optimal exposure efficiency results at 60 kVp and above. Decreasing relative improvements are observed above 100th VL filter thickness at 78 cm SID. Results are obtained without significant tube heating (except at 40 kVp). In addition, simulations indicate significant reductions in beam hardening. This optimized beam will be incorporated into a novel cone-beam x-ray computed mammotomography sub-system together with an emission tomograph in a dual modality CT/SPECT application specific emission and transmission tomography system for fully 3D uncompressed breast imaging.

An inverse-geometry volumetric CT (IGCT) system for imaging in a single fast rotation without cone-beam artifacts is being developed. It employs a large scanned source array and a smaller detector array. For a single-source/single-detector implementation, the FOV is limited to a fraction of the source size. Here we explore options to increase the FOV without increasing the source size by using multiple detectors spaced apart laterally to increase the range of radial distances sampled. We also look at multiple source array systems for faster scans. To properly reconstruct the FOV, Radon space must be sufficiently covered and sampled in a uniform manner. Optimal placement of the detectors relative to the source was determined analytically given system constraints (5cm detector width, 25cm source width, 45cm source-to-isocenter distance). For a 1x3 system (three detectors per source) detector spacing (DS) was 18deg and source-to-detector distances (SDD) were 113, 100 and 113cm to provide optimum Radon sampling and a FOV of 44cm. For multiple-source systems, maximum angular spacing between sources cannot exceed 125deg since detectors corresponding to one source cannot be occluded by a second source. Therefore, for 2x3 and 3x3 systems using the above DS and SDD, optimum spacing between sources is 115deg and 61deg respectively, requiring minimum scan rotations of 115deg and 107deg. Also, a 3x3 system can be much faster for full 360deg dataset scans than a 2x3 system (120deg vs. 245deg). We found that a significantly increased FOV can be achieved while maintaining uniform radial sampling as well as a substantial reduction in scan time using several different geometries. Further multi-parameter optimization is underway.

In this paper, soft tissue contrast visibility in neural applications is investigated for volume imaging based on flat X-ray detector cone-beam CT. Experiments have been performed on a high precision bench-top system with rotating object table and fixed X-ray tube-detector arrangement. Several scans of a post mortem human head specimen have been performed under various conditions. Hereby two different flat X-ray detectors with 366 x 298mm² (Trixell Pixium 4700) and 176 x 176mm² (Trixell Pixium 4800) active area have been employed. During a single rotation up to 720 projections have been acquired. For reconstruction of the 3D images a Feldkamp algorithm has been employed. Reconstructed images of the head of human cadaver demonstrate that added soft tissue contrast down to 10 HU is detectable for X-ray dose comparable to CT. However, the limited size of the smaller detector led to truncation artifacts, which were partly compensated by extrapolation of the projections outside the field of view. To reduce cupping artifacts resulting from scattered radiation and to improve visibility of low contrast details, a novel homogenization procedure based on segmentation and polynomial fitting has been developed and applied on the reconstructed voxel data. Even for narrow HU-Windows, limitations due to scatter induced cupping artifacts are no longer noticeable after applying the homogenization procedure.

Sensor fill factor is one of the key pixel design requirements for high performance imaging arrays. In our conventional imaging pixel architecture with a TFT and a photodiode deposited in the same plane, the maximum area that the photodiode can occupy is limited by the size of the TFT and the surrounding metal lines. A full fill factor array design was previously proposed using a continuous sensor layer1. Despite the benefits of 100% fill factor, when applied to large-area applications, this array design suffers from high parasitic line capacitances and, thus, high line noise. We have designed and fabricated an alternative pixel structure in which the photodiode is deposited and patterned over the TFT, but does not overlap with the lines underneath. Separating the diode from the TFT plane allows extra space for an additional TFT which can be used for pixel reset and clipping excessive charge in the photodiode developed under high illumination. This reduces memory effect by 250%. The yield and the reliability are expected to improve as well since the TFTs and lines are buried underneath the diode. With the increased fill factor, we collect 56% more electrons per pixel, thereby improving the signal to noise ratio. The maximum signal to noise ratio is achieved when the increased signal and the undesirable parasitic capacitance on the data line are best optimized. Linearity, sensitivity, leakage, and MTF characteristics of a prototype X-ray imager based on this architecture are presented.

We study the properties of a new microangiographic system, consisting of a Region of Interest (ROI) microangiographic detector, x-ray source, and patient. The study was performed under conditions intended for clinical procedures such as neurological diagnostic angiograms as well as treatments of intracranial aneurysms, and vessel-stenoses. The study was performed in two steps; first a uniform head equivalent phantom was used as a “filter”. This allowed us to study the properties of the detector alone, under clinically relevant x-ray spectra. We report the detector MTF, NPS, NEQ, and DQE for beam energies ranging from 60-100kVp and for different detector entrance exposures. For the second step, the phantom was placed adjacent to the detector, allowing scatter to enter the detector and new measurements were obtained for the same beam energies and detector entrance exposures. Different radiation field sizes were studied, and the effects of different scatter amounts were investigated. The spatial distribution of scatter was studied using the edge-spread method and a generalized system MTF was obtained by combining the scatter MTF weighted by the scatter fraction with the detector MTF and focal spot unsharpness due to magnification. The NPS combined with the generalized MTF gave the generalized system NEQ and DQE. The generalized NEQ and the ideal object detectability were used to calculate the Dose Area Product to the patient for 75% object detection probability. This was used as a system optimization method.

Full-leg and full-spine imaging with standard computed radiography (CR) systems requires several cassettes/storage phosphor screens to be placed in a staggered arrangement and exposed simultaneously to achieve an increased imaging area. A method has been developed that can automatically and accurately stitch the acquired sub-images without relying on any external reference markers. It can detect and correct the order, orientation, and overlap arrangement of the subimages for stitching. The automatic determination of the order, orientation, and overlap arrangement of the sub-images consists of (1) constructing a hypothesis list that includes all cassette/screen arrangements, (2) refining hypotheses based on a set of rules derived from imaging physics, (3) correlating each consecutive sub-image pair in each hypothesis and establishing an overall figure-of-merit, (4) selecting the hypothesis of maximum figure-of-merit. The stitching process requires the CR reader to over scan each CR screen so that the screen edges are completely visible in the acquired sub-images. The rotational displacement and vertical displacement between two consecutive sub-images are calculated by matching the orientation and location of the screen edge in the front image and its corresponding shadow in the back image. The horizontal displacement is estimated by maximizing the correlation function between the two image sections in the overlap region. Accordingly, the two images are stitched together. This process is repeated for the newly stitched composite image and the next consecutive sub-image until a full-image composite is created. The method has been evaluated in both phantom experiments and clinical studies. The standard deviation of image misregistration is below one image pixel.

The GE Senographe 2000D, the first full field digital mammography system based on amorphous silicon (a-Si) flat panel arrays and a cesium iodide (CsI) scintillator, has been in clinical use for over five years. One of the major advantages of this technology platform over competing platforms is the inherent flexibility of the design. Specifically, it is possible to optimize the x-ray conversion layer (scintillator) independently of the light conversion layer (panel) and vice versa. This is illustrated by a new detector utilizing the same amorphous silicon (a-Si) flat panel design, but an optimized scintillator layer, which provides up to 15% higher DQE than the existing detector. By utilizing the existing flat panel with an optimized scintillator layer, it is possible to significantly boost performance without changes to the panel design. Future enhancements to both the panel and scintillator will raise the DQE at zero frequency to more than 80%. The a-Si/CsI platform is especially well suited to advanced applications utilizing very low doses.

We have developed a novel flat-panel detector with CsI:Tl scintillator. The detector consists of a single piece 43cm x 43cm amorphous silicon thin-film transistor (TFT) array with MIS (metal-insulator-semiconductor) photoelectric converter having a pixel pitch of 160μm coated with a needle-like crystal CsI:Tl scintillator. Signal chain was totally revised from current detector utilizing an innovative sensor technology. The novel detector and current detector were equipped to a digital radiography system allowing a quantitative and comparative study. Results show that the novel detector has a linear response covering the radiographic exposure range. It has a moderate modulation transfer function (MTF) sufficient to the radiography tasks and effective to suppress the aliasing. The detective quantum efficiency (DQE) was almost twice than the current detector. The result of contrast-detail phantom exposed with a 1/2x dose level is equivalent to that of current detector with a 1x dose level. These results show that performance of novel detector is superior to and expected to reduce the patient dose in half than current detector due to higher DQE and innovative sensor technology.

The requirements for medical X-ray detectors tend towards higher spatial resolution, especially for mammography. Therefore, we have investigated common absorber materials with respect to the possible intrinsic limitations of their spatial resolution. Primary interaction of an incident X-ray quantum is followed by a series of processes: Rayleigh scattering, Compton effect, or the generation of fluorescence photons and subsequent electrons. Lateral diffusion of carriers relative to their drift towards the electrodes also broadens the point-spread function. One consequence is that the spatial resolution of the detector, expressed in terms of the modulation transfer function (MTF), is reduced. Monte Carlo simulations have been carried out for spectra with tube voltages of 28-120 kV using the program ROSI (Roentgen Simulation) based on the well-established EGS4 algorithm. The lateral distribution of deposited energy has been calculated in typical materials such as Se, CdTe, HgI₂, and PbI₂ and used to determine the line spread function. The complex absorption process is found to determine the spatial resolution of the detector considerably. The spectrum at energies closely above the K-edge of the absorber material tends to result in a reduced MTF. At energies above 50 keV, electron energy loss increasingly reduces spatial resolution in the high frequency range. The influence of fluorescence is strongest in the 5-20 lp/mm range. If a very high spatial resolution is required, a well-adapted semiconductor should be applied.

The effect of polychromatic x-rays on image reconstruction in helical cone-beam computed tomography is investigated. A pre-reconstruction dual-energy technique is developed to reduce beam-hardening artifacts and enhance contrast in soft tissue. The effect of realistic signal noise on the dual-energy method is studied.

Dose is becoming increasingly important for computed tomography clinical practice. It is of general interest to understand the impact that system design can have on dose and image quality. This study addresses the effect of bowtie shape on the dose and contrast-to-noise across the field of view. Simulation of the CT acquisition is used to calculate the energy deposition throughout a numerical phantom for a set of relevant system operating parameters and bowtie shapes. Mean absorbed dose is calculated by summing over the phantom volume and is compared with other typical dose specifications. A more aggressive attenuation profile of the bowtie which offers higher attenuation in the periphery of the field of view can offer the benefit of lower dose but at the expense of reduced contrast-to-noise at the edge of the cross-sectional image.

The development of new digital mammography techniques such as dual-energy imaging, tomosynthesis and CT mammography often requires investigating optimal camera design parameters and imaging techniques. One tool that is useful for this purpose is Monte Carlo simulation. This paper presents a methodology for generating simulated images from a CsI-based, flat-panel imager by using the Geant 3 Monte Carlo code to model x-ray transport and absorption within the CsI scintillator, and the DETECT-II code to track optical photon spread within a columnar model of the CsI scintillator. The Monte Carlo modeling of x-ray transport and absorption within the CsI was validated by comparing to previously published values for the probability of a K-shell interaction, the fluorescent yield, the probability of a K-fluorescent emission, and the escape fraction describing the probability of a K x-ray escaping the scintillator. To validate the combined (Geant coupled with DETECT-II) Monte Carlo approach to form simulated images, comparison of modulation transfer functions (MTFs) and system sensitivity (electrons/mR/pixel) obtained from simulations were compared to empirical measurements obtained with different x-ray spectra and imagers with varying CsI thicknesses. By varying the absorption and reflective properties of the columnar CsI used in the DETECT-II code, good agreement between simulated MTFs and system sensitivity and empirically measured values were observed.

This paper examines the noise performance of an inverse-geometry volumetric CT (IGCT) scanner through simulations. The IGCT system uses a large area scanned source and a smaller array of detectors to rapidly acquire volumetric data with negligible cone-beam artifacts. The first investigation compares the photon efficiency of the IGCT geometry to a 2D parallel ray system. The second investigation models the photon output of the IGCT source and calculates the expected noise. For the photon efficiency investigation, the same total number of photons was modeled in an IGCT acquisition and a comparable multi-slice 2D parallel ray acquisition. For both cases noise projections were simulated and the central axial slice reconstructed. In the second study, to investigate the noise in an IGCT system, the expected x-ray photon flux was modeled and projections simulated through ellipsoid phantoms. All simulations were compared to theoretical predictions. The results of the photon efficiency simulations verify that the IGCT geometry is as efficient in photon utilization as a 2D parallel ray geometry. For a 10 cm diameter 4 cm thick ellipsoid water phantom and for reasonable system parameters, the calculated standard deviation was approximately 15 HU at the center of the ellipsoid. For the same size phantom with maximum attenuation equivalent to 30 cm of water, the calculated noise was approximately 131 HU. The theoretical noise predictions for these objects were 15 HU and 112 HU respectively. These results predict acceptable noise levels for a system with a 0.16 second scan time and 12 lp/cm isotropic resolution.

An ROC study was conducted to evaluate the usefulness of assessing breast lesion characteristics with stereomammography. Stereoscopic image pairs of 158 breast biopsy tissue specimens were acquired with a GE digital mammography system. Two stereo image pairs were taken at 1.8X magnification geometry and at approximately orthogonal orientations for each specimen. Display software was developed for a high resolution MegaScan CRT monitor driven by a DOME stereo display board. The specimens contained either a mass, microcalcifications, both, or normal tissue. About 40% of the specimens were found to contain malignancy by pathological analysis. Five MQSA radiologists participated in the observer performance experiment. The two views of each specimen were read independently and were separated by a large number of other specimen images to reduce any effects of memorization. Each observer read 316 specimen images in a systematically randomized order. The observer first read the monoscopic image and entered his/her assessment in terms of the confidence ratings on the presence of microcalcifications and/or masses, margin clearance, BI-RADS assessment, and the likelihood of malignancy. The corresponding stereoscopic images were then displayed on the same monitor and were viewed through stereoscopic LCD glasses. The observer was free to change the ratings in every category after stereoscopic reading. The ratings of the observers were analyzed by ROC methodology. For the 5 MQSA radiologists, the average Az value for estimation of the likelihood of malignancy of the lesions improved from 0.70 for monoscopic viewing to 0.72 (p<0.05) after stereoscopic viewing, and the average Az value for the presence of microcalcifications improved from 0.94 to 0.95 (p<0.05). The Az value for the presence of masses improved from 0.80 to 0.82 after stereoscopic viewing, but the difference fell short of statistical significance (p=0.08). The visual assessment of margin clearance was found to have very low correlation with pathological analysis. This study demonstrates the potential of using stereomammography to improve the detection and characterization of mammographic lesions.

Most current research in mammographic detector development has been focused on digital detectors. However, the vast majority of clinical practice still uses film/screen systems. This paper will report on the design and performance features of a new film/screen system for mammography that has the potential to become the new gold standard in image quality for this demanding application. This new system includes a new film and new intensifying screens. The new film has unique emulsions on each side of the base. It features novel fine-grain, metal ion-doped silver halide microcrystals. It is exposed with a single intensifying screen. Two different intensifying screens have been developed to create systems with different speeds. The new screens use an improved phosphor and new coating structures. The highfrequency MTF has been boosted. The system provides higher contrast, lower noise, and sharper images.

A method for the determination of optimal operating points of digital mammography systems is described. The digital mammography equipment uses a flat panel detector and a bi-metal molybdenum/rhodium x-ray tube. An operating point is defined by the selection of the x-ray tube target material, x-ray filtration, kVp and detector entrance dose. Breast thickness and composition are estimated from a low dose pre-exposure, then used to index tables containing sets of operating points. The operating points are determined using a model of the image chain, which computes contrast to noise ratio (CNR) and average glandular dose (AGD) for all possible exposure conditions and breast thickness and composition combinations. The selected operating points are those which provide the required CNR for the lowest AGD. An AGD reduction of 30% to 50% can be achieved for comparable Image Quality, relative to current operating points. Resulting from the optimization process, the rhodium target is used in more than 75% of cases. Measurements of CNR and AGD have been performed on various tissue equivalent materials with good agreement between calculated and measured values. The proposed method provides full Image Quality benefit of digital mammography while minimizing dose to patients in a controlled and predictive way.

The purpose of this study was to investigate if the glandular dose to the breast in mammography can significantly be reduced without compromising image quality, when using photon counting technology, in a multi-slit scanning photon counting detector, compared to a conventional film mammography system and commercial available digital mammography systems with TFT-array detectors. A CDMAM phantom study, with two different thicknesses of additional PMMA absorber, 4 cm and 7 cm respectively, has shown that multi-slit scanning photon counting detector technology can reduce the dose, without reducing the image quality. This comparison was made to two commercial available digital mammography systems Senographe 2000D (from GEMS) and Selenia (from Lorad). The results show that dose can be reduced with 63% to 77%, depending on object thickness, when using XCT for mammography. This dose reduction has also been verified clinically through a small pilot study with patients and specimen, where the comparison was made between XCT and film.

A simple x-ray detector that utilizes amorphous selenium (a-Se) directly deposited on a specially-designed CCD (charge coupled device) with a 25 micrometer del (detector element) pitch is described. This simple detector has been used to test the feasibility of creating digital mammography detectors. To enable the use of electron transport CCDs with a-Se, we have developed a-Se hole blocking layers to permit the transfer of electrons to the CCD while suppressing hole leakage current in the presence of the high negative bias (~1000 V) required to make the a-Se x-ray sensitive. We report measurements of the charge transfer efficiency (CTE), dark signal, x-ray sensitivity, x-ray signal linearity, and x-ray MTF (modulation transfer function) of the simple detector. As the thickness of the a-Se hole blocking layer was increased, the MTF decreased. For a thin (1 micrometer) blocking layer the MTF at a spatial frequency of 20 cycles/mm was 0.4.

Measurements and Monte Carlo simulations were used to investigate the scatter properties of a scanned multi-slit digital mammography system. Scatter to primary ratio (S/P) in the center of the image field was calculated for different thickness of breast equivalent material and different tube potentials. The simulated model also varied the angular acceptance, the number of slits and the distance between the slits of a dedicated scatter rejection device. In addition to the expected scatter from the breast equivalent material, scatter within the detector contributes to the S/P-ratio. The main part of the scatter is identified as coming from this process. Measured total S/P-ratios below 3% are reported for breast range 3-8 cm. The scatter-DQE is used as figure-of-merit for comparison to other imaging geometries and scatter rejection schemes.

A novel angiography system with cone-beam reconstruction using a large-area flat panel detector (FPD), with 40x30cm active area and 2048x1536 matrixes with a 194μm pixel pitch, has been developed. We present results on a basic performance, spatial resolution and contrast detectability obtained on this angiography system with cone-beam function using the FPD, and compare with a conventional angiography system with an image intensifier (I.I.) and charge-coupled device (CCD) camera. We’d achieved a fast acquisition, 15 seconds as for a subtraction mode by rotating a ceiling suspended C-arm at a speed of 40 degrees per second, and ensured a large reconstructed columnar volume, φ250mmx180mm, by using the large-area detector. As a result of the evaluation, the 3D image acquired from the FPD system has a high spatial resolution with no distortion and good contrast detectability.

The objective of this study was to identify the x-ray tube voltage that results in optimum performance for abdominal CT imaging for a range of imaging tasks and patient sizes. Theoretical calculations were performed of the contrast to noise ratio (CNR) for disk shaped lesions of muscle, fat, bone and iodine embedded in a uniform water background. Lesion contrast was the mean Hounsfield Unit value at the effective photon energy, and image noise was determined from the total radiation intensity incident on the CT x-ray detector. Patient size ranging from young infants (10 kg) to oversized adults (120 kg), with CNR values obtained for x-ray tube voltages ranging from 80 to 140 kV. Patients of varying sizes were modeled as an equivalent cylinder of water, and the mean section dose (D) was determined for each selected x-ray tube kV value at a constant mAs. For each patient size and lesion type, we identified an optimal kV as the x-ray tube voltage that yields a maximum value of the figure of merit (CNR2/D). Increasing the x-ray tube voltage from 80 to 140 kV reduced lesion contrast by 11% for muscle, 21% for fat, 35% for bone and 52% for iodine, and these reductions were approximately independent of patient size. Increasing the x-ray tube voltage from 80 to 140 kV increased a muscle lesion CNR relative to a uniform water background by a factor of 2.6, with similar trends observed for fat (2.3), bone (1.9) and iodine (1.4). The improvement in lesion CNR with increasing x-ray tube voltage was highest for the largest sized patients. Increasing the x-ray tube voltage from 80 to 140 kV increased the patient dose by a factor of between 5.0 and 6.2 depending on the patient size. For small sized patients (10 and 30 kg) and muscle lesions, best performance is obtained at 80 kV; however, for adults (70 kg) and oversized adults (120 kg), the best performance would be obtained at 140 kV. Imaging fat lesions was best performed at 80 kV for all patients except for oversized adults, where 140 kV offers the best imaging performance. For high Z lesions of bone and iodine, imaging performance generally degrades with increasing kV for all patient sizes, with the degree of degradation largest for the smallest patients. We conclude that 80 kV is optimal with respect to radiation dose in abdominal CT for all pediatric patients. For adults, 80 kV is the x-ray voltage of choice for high Z lesions, whereas 140 kV would generally be the voltage of choice of lesions that have an atomic number similar to that of water.

The emergence of helical tomotherapy has provided a unique opportunity to combine aspects of diagnostic computed tomography and radiation treatment. Daily megavoltage computed tomography (MVCT) scans of a patient in the treatment position provide an ideal input for adaptive radiation therapy, whereby the quantitative CT knowledge of a patient from a treatment fraction combined with the knowledge of the therapy dose distribution can be used to alter and correct for the dose delivery in subsequent fractions. In order for adaptive radiotherapy to be successful, the quantitative information from the CT scan must be as accurate as possible in geometric and dosimetric information. One potential impediment to the accuracy of the CT data values is x-ray scatter. In our study, we quantify the magnitude of x-ray scatter in the tomotherapy (fan-beam) MVCT system, based on Monte Carlo simulations of the scatter-to-primary ratio (SPR) as a function of incident x-ray energy, fan-beam slice thickness, patient size, and air gap distance. Furthermore, based on these SPR values, the impact on CT number accuracy is shown, and the implications for adaptive radiotherapy (i.e. dose reconstruction) are discussed. Under conditions common to tomotherapy MVCT scanning, SPR values range from 0.02 to 0.16 (depending on the size of the phantom), and are generally lower than those encountered in diagnostic cone-beam CT and megavoltage portal imaging. These SPR values are sufficient enough to introduce CT number errors as high as 5 HU in soft-tissue and 100 HU in bone. The implication of this inaccuracy for adaptive radiotherapy would be to cause potential dose calculation errors during dose reconstruction and treatment re-planning.

The real-time flat panel detector-based cone beam CT breast imaging (FPD-CBCTBI) has attracted increasing attention for its merits of early detection of small breast cancerous tumors, 3-D diagnosis, and treatment planning with glandular dose levels not exceeding those of conventional film-screen mammography. In this research, our motivation is to further reduce the x-ray exposure level for the cone beam CT scan while retaining acceptable image quality for medical diagnosis by applying efficient denoising techniques. In this paper, the wavelet-based multiscale anisotropic diffusion algorithm is applied for cone beam CT breast imaging denoising. Experimental results demonstrate that the denoising algorithm is very efficient for cone bean CT breast imaging for noise reduction and edge preservation. The denoising results indicate that in clinical applications of the cone beam CT breast imaging, the patient’s radiation dose can be reduced by up to 60% while obtaining acceptable image quality for diagnosis.

The dose efficiency index (DEI) is a dose independent measure that quantifies the low-contrast performance of CT scanners. The purpose of this paper is to compare the results of DEI analysis with those of ROC analysis. A custom-made phantom consisting of diluted contrast targets of various sizes and densities was scanned at 80 & 120 kV on a multislice CT scanner (Hispeed Advantage RP, GE). Eight radiographers reviewed the images and identified discernable targets. The likelihood of detection was measured as a function of the target size and scan conditions. The results of the DEI & ROC analyses showed that the low-contrast resolution was higher at 80kV than at 120kV. The p-value obtained by the paired-t test was 0.000 for both analyses, indicating that the difference was significant. Under the conditions used in this study, the DEI analysis was found to be an effective alternative to ROC analysis for characterizing the low-contrast performance of CT scanners. The evaluation time was about 1/6 compared with ROC analysis. It is much simpler to calculate than ROC and is useful in comparing scanners from different manufacturers or as part of ongoing quality assurance.

High-resolution computed tomography (CT) reconstructions currently require either full field of view (FOV) exposure, resulting in high dose, or region of interest (ROI) exposure, resulting in artifacts. To obtain high-resolution 3D reconstruction of an ROI with minimal artifiacts, we have developed a method involving a non-uniform ROI beam filter to reduce dose outside the ROI while acquiring the ROI at a higher dose. High-resolution, high-dose full-field projections ofa phontom were obtained. ROIs in the images were selected and the low-dose data outside the ROI were simulated by adding various levels of noise to the projection data corresponding to a dose of 1/16 and 1/256 of the original dose. For an ROI of 30% FOV, artifacts in the reconstructed ROI were minimal for both dose reduction levels. For an ROI of 10% FOV, artifacts remained minimal only for the 1/16^th dose case. The effect of the presence of a high contrast object outside the ROI was also studied. We found that the intensity of the artifacts increases with the contrast of the object, its size, and its distance from the axis of rotation. CT using an ROI filter provides a way to reconstruct an ROI with reduced integral dose and yet with minimal artifacts and improved spatial resolution.

A modified Feldkamp (FDK) 3-D cone beam algorithm was developed to conduct the half-scan scheme on a flat panel detector (FPD)-based prototype CBCT system which is available in our Lab. X-ray source scans the object in a circular trajectory in the range of 180 degrees plus full cone angle rather than 360 degrees for a full scan. The redundant data is weighted by a weighting function in 3-D case based on the ones proposed by Parker1, FDK algorithm is then adopted to get the reconstruction image. The breast phantom and breast specimen are used in the experiment and the objects reconstructed from half-scan CBCT are compared with those reconstructed from full scan. The applicability of the half-scan scheme is testified in terms of image noise level, contrast resolution. The half-scan scheme is expected to improve the temporal resolution and may also reduce the patients’ X-ray dose level. The result showed an encouraging potential usage of the Half-scan CBCT in the functional 4-D CT imaging.

A back illuminated photodiode (BIP) has been developed for multislice CT which enables a CT scanner to be made with a large number of slices. Light absorbed on one surface of the photodiode generates charge carriers which diffuse through the silicon bulk to charge collection anodes on the opposite side. It is this minority carrier diffusion through the bulk which governs the spatial and temporal properties of the photodiode. The diffusion of carriers in the silicon is described by a partial differential equation (PDE) which can be solved numerically for a given geometry. Modeling carrier diffusion allows one to determine the risetime and response profiles for a given photodiode design without the expense of fabricating an actual device. The PDE for carrier diffusion has been solved for different BIP geometries using the Matlab PDE toolbox. This software uses the finite element method in two dimensions. Risetime and response profiles have been simulated for the same geometries as BIP devices that have been fabricated. The simulation results are in good agreement with the measurements done on actual devices. New geometries are also simulated to show how the BIP can be optimized.

Dynamic X-ray Computed Tomography (CT) is an attractive imaging technology for the guidance in minimal invasive surgery. In this field, projection data simulation is an important tool to optimise scanner geometry and to validate reconstruction algorithms. A realistic simulation software, called “Sindbad” has been developed to compute 2D projections. It is based on an analytical model and allows simulating X-ray emission, attenuation through an examined object and photon detection. New functionality has been added to simulate a virtual scanner and to combine 2D projections from CAD phantoms with CT data volumes. Phantoms can be animated with independent motion and temporal evolution laws. CT data can be deformed over time by using a Free Form Deformation (FFD) technique. Encouraging results have been obtained for the simulation of a lung biopsy. To simulate breathing, CT lung data are animated by using a respiratory law. Biopsy needle was inserted along a straight line from an entry point to a target point at a regular speed. The guidance direction also varied with time according to the respiration law. Similar simulations are also used to validate dynamic reconstruction algorithm for radiotherapy planning.

The three-dimensional (3D) imaging performance of a cone-beam computed tomography (CBCT) system can be quantitatively characterized by its 3D point spread function (PSF). To avoid the pitfalls associated with 3D PSF measurement through the use of a micro point phantom, we adopt an edge-based technique, which iteratively blurs a step edge into a spread edge with a Gaussian blurring kernel. In experiment, a small Teflon solid ball (diameter ~6mm) is used to provide step edges along the scanlines across the ball center, in terms of x-ray linear attenuation coefficient. Correspondingly, the spread edge profiles are extracted from the digital volume ball reconstructed by a CBCT system. From a spread edge profile, a step edge can be established by a rectification procedure (essentially an edge identification). A 1D PSF along a scanline is modeled as a Gaussian distribution, which is determined by iterative edge-blurring technique. The 1D PSFs on a cross section of the ball constitute a 2D PSF, and the 2D PSFs at three orthogonal cross sections are used to represent the 3D PSF at the position of the ball. By repositioning the ball phantom and repeating the procedure, we measured the 3D PSFs at 126 positions over half of the support space. Experiment shows that a CBCT system is of spatial variance and anisotropy in terms of FWHMs (full width at half maximum) of the blurring kernel. Both numerical and graphical presentations of PSF results are provided. As a result, the FWHM of the 3D PSF in our CBCT system varies in the range between 0.66 mm and 1.39 mm.

The properties of 32x16 (and 16x16) element pin photodiode arrays with the element size of 1-mm square or smaller and with the gap between the adjacent elements as small as 90 μm are discussed. The arrays were built on a 30-μm thick single Si die. These arrays have superior optical and electrical characteristics by design, and work at zero Volts bias. The arrays are characterized with close to 100% internal quantum efficiency within the spectral range 500-800 nm. The cross talks are smaller than 0.5% within the spectral range 400 to 1000 nm. Very low leakage currents, high shunt resistance-above 5 GΩ at room temperature, low total capacitance, fast rise time, and excellent temperature stability of all parameters-provide for superior performance of such arrays in comparison with those currently used in many fields of medical imaging. The thin silicon die provides superior mechanical flexibility, which, combined with the flip chip packaging technology, allows for its application in extreme mechanical and thermal conditions. The arrays can be tiled facilitating the building of large-scale photodetector matrices. The paper discusses the main features of the structures that create this exceptional array performance.

We have developed a four-dimensional CT (4D CT) using continuous rotation of cone-beam x-ray. The maximum nominal beam width of the 4D CT is 128 mm at the center of rotation in the longitudinal direction. In order to obtain appropriate estimations of exposure dose, detailed single-slice dose profi les perpendicular to the rotation axis including scattered radiation were measured in PMMA cylindrical phantoms, which were cylindrical lucite phantoms of 160 mm and 320 mm diameter and 900 mm length. Dose profi les were measured with a pin photodiode detector at the center and a peripheral point of 10 mm depth. A pin silicon photodiode sensor with 3 × 3 × 3 mm sensitive region was used as an x-ray detector, which was scanned along longitudinal direction in the phantom for beam widths of 20, 42, 74, 106 and 138 mm. The dose profi les had long tails caused by scattered radiation more than 200 mm out of the beam width edge. The exposure dose covered 95 % was distributed along about 360 mm length at the center and about 310 mm at the periphery, which was independent of the beam width. Before the advent of multi-detector CT, CTDI100 was used to approximate integral dose for clinical scan conditions. However, for 4D CT employing a variable beam width, the standard CTDI was not a good estimation. This work was carried out to establish a method of the dose measurements including scattered radiation for cone-beam CT such as 4D CT. In order to perform the dose assessment including scattered radiation, dose measured length should be recommended to measure integral dose over beam widths plus at least 230 mm, which covered 95 % total exposure dose.

The purpose of this presentation is to show how a commercially available spiral CT has been modified for use as the electro-mechanical scanner hardware for a prototype flat panel detector-based cone beam computed tomography (FPD-CBCT) imaging system. FPD-CBCT has the benefits of isotropic high resolution, low contrast sensitivity and 3D visualization. In contrast to spiral CT, which acquires a series of narrow slices, FPD-CBCT acquires a full volume of data (limited by the cone angle and the FPD active area) in one <= 360° scan. Our goal was to use a GE HighSpeed Advantage (HSA) CT system as the basis for an FPD-CBVCT imaging prototype for performing phantom, animal and patient imaging studies. Specific electromechanical and radiographic subsystems controlled include: gantry rotation and tilt, patient table positioning, rotor control, mA control, the high frequency generator (kVp, exposure time, repetition rate) and image data acquisition. Also, a 2D full field FPD replaced the 1D detector, as well as the existing slit collimator was retrofitted to a full field collimator to allow x-ray exposure over the entire active area of the FPD. In addition, x-ray projection data was acquired at 30 fps. Power and communication signals to control modules on the rotating part of the gantry were transmitted through integrated slip rings on the gantry. A stationary host computer controlled mechanical motion of the gantry and sent trigger signals to on-board electronic interface modules to control data acquisition and radiographic functions. Acquired image data was grabbed to the system memory of an on-board industrial computer, saved to hard disk and downloaded through a network connection to the stationary computer for 3D reconstruction. Through the synchronized control of the pulsed x-ray exposures, data acquisition, and gantry rotation the system achieved a circle cone beam image acquisition protocol. With integrated control of the gantry tilt and of the position and translation speed of the patient table, spiral cone beam and circle-plus-arc cone beam image acquisition protocols will also be achieved. Performance is being evaluated with optical encoders, standard dosimetry equipment and phantom studies.

The purpose of this study was to perform a preliminary evaluation of a newly constructed s flat panel detector (FPD)-based system for cone beam CT imaging applications. A prototype flat-panel detector-based cone beam CT imaging system has been designed and constructed by modifying a GE HiSpeed Advantage CT scanner. The prototype consisted of a modified GE CT HiSpeed Advantage CT gantry, an x-ray tube, an 19.5 x 24.4 cm Varian Paxscan 2520 real time flat-panel detector mounted on the gantry, a CT table and an on-gantry PC to control image acquisition. Another PC workstation serving as an operating console controlled the CT gantry and CT table, and sent trigger pulses to dedicated electronic interface modules to control radiographic exposure and to initiate data acquisition. Captured image data sets were first stored in an on-gantry computer and then downloaded from the on-gantry PC to the operating console for high-speed 3D image reconstruction. During data acquisition, the x-ray tube and the FPD can be rotated on the gantry over Nx360 degrees due to integrated slip ring technology. With a single scan, this device was able to acquire up to 288 two-dimensional projections (960 x 768 x 16 bits) for direct 3D reconstruction within 9.6 seconds. This system was used for a series of preliminary phantom studies. Using the continuous scan mode of the scanner, 288 projections were acquired for all volume scans. Direct 3D reconstructions were obtained to evaluate the system for cone beam CT for lung imaging applications. The preliminary results indicate that the newly built flat panel detector-based cone beam CT scanner worked as expected.

Since the inception of Computerized Tomography (CT), tube thermal performance has been characterized by tube heat storage and dissipation. As the number of CT detector rows has increased from one to sixteen and higher, these parameters have become less relevant to clinical performance. In addition, peak power is quite dependent on focal spot size, so quoting one parameter without the other gives an incomplete picture. We propose a new set of specifications that more completely characterize a tube's thermal performance: the Scan Performance Index (SPI) and the Focal Spot Loadability (FSL). The Scan Performance Index (SPI) is a measure of clinical tube performance over a range of application parameters. The Focal Spot Loadability (FSL) is a measure of peak power as a function of focal spot size. This paper describes these figures of merit and provides some proposed parameter definitions. Calculations of SPI were made based on the mA vs. time performance curves under different assumptions of the thermal cycle repeat frequency and clinically relevant scan time range. Comparison of the results of SPI calculations with the traditional heat storage and dissipation characteristics vs. actual clinical capability leads to the conclusion that SPI is the better indicator of total throughput, especially as tube power and CT fan beam coverage increases, and total scan time decreases. FSL is shown to elucidate the inherent ability of a CT tube to handle high power loads for a given focal spot size. We conclude that these two new CT tube characteristics should be considered by clinicians in place of the traditional tube characteristics in order to benchmark the thermal performance capability of a given CT tube.

Digital x-ray imager known to flat-panel detector has been studied for the application of a various medical modalities. Currently, two types of detection methods have been realized in digital radiography. One is an indirect conversion method and the other is a direct conversion method. we have been developing a new x-ray detector that combines a columnar CsI:Na scintillation layer with a photosensitive a-Se with dielectric thin film. In this structure, an x-ray is converted to visible light in a CsI:Na scintillation layer and visible light is then converted to electric charges in a-Se layer. The electron-hole pairs can be also generated from x-ray interaction in the a-Se photoconductor, which can improve the detection efficiency of electric charge. We designed the thickness of CsI:Na scintillator by using computer simulation. MCNP is a general-purpose, continuous-energy, generalized-geometry, time-dependent, coupled neutron / photon / electron Monte Carlo transport code. The spectra of x-ray absorption was simulated by using MCNP 4C code. The morphology of the vacuum deposited CsI:Na scintillator and the parylene film were analyzed. Photoluminescence characterization of CsI:Na showed a light emission peak centered at 420nm as expected, which matched the absorption spectrum of amorphous selenium(a-Se). For an electric field of 10V/μm, the dark currents of our detector were below 370 pA/cm² and the SNR of CsI:Na coupled a-Se detector with a dielectric layer was 1.8 times greater than that without CsI:Na layer.

The contributors to image noise of two computed radiography (CR) detector systems-a state-of-the-art and a wellchosen laboratory CR image plate-were studied by two different methods. Method 1 analyzes the image noise content of a series of images obtained at a wide range of different X-ray exposure levels. It uses a model to fit the observed exposure dependence of the normalized noise power spectrum (NNPS): It distinguishes between an NNPS component that is independent of the exposure level and mainly due to correlated noise, and an NNPS component which is inversely proportional to the exposure level and consists mainly of quantum noise. Method 2 analyzes several images taken at the same exposure level and distinguishes between correlated noise, which remains unchanged in repeated exposures, and uncorrelated noise which is different in each image. The results of the two methods allowed the relevant noise contributions in CR images to be quantitatively determined. The novel laboratory image plate showed a significant reduction of correlated noise with an accompanying increase in the DQE. The results also served to estimate a possible improvement of DQE if an appropriate flat field correction is made for these CR systems.

This paper describes a new flat panel imager designed for use in cardiovascular and mobile C-arm imaging systems. The a-Si sensor array has a 1024 x 1024 matrix with a pixel pitch of 194 μm, resulting in an active area of 198.7 mm x 198.7 mm. The imager allows frame rates of up to 30 fps in full resolution fluoroscopy mode and up to 60 fps in a 2 x 2 binned low dose fluoroscopy mode. Typically, a 600 μm thick deposited columnar CsI(Tl) layer is used as the scintillator. Improvements in the pixel architecture, charge amplifier ASICs, and system level electronics resulted in a very low electronic noise floor, such that both the fluoroscopy and low dose fluoroscopy modes of the panel are x-ray quantum limited below 1 μR/frame. Low power consumption electronics combined with a mechanical design optimized for heat transfer and dissipation makes air-cooling sufficient for most environments. The small size of 24.1 x 24.1 x 6 cm and the weight of only 4.1 kg meet the requirements of C-Arm systems. Special consideration was given to the border around the active area, which has been reduced to 2 cm. Reported performance parameters include linearity, lag, contrast ratio, MTF, and DQE. For the full resolution mode, the MTF is greater than 0.53 and 0.21 at 1 and 2 lp/mm, respectively. DQE measured at 22 nGy/frame was greater than 0.68, 0.50, and 0.23 at 0, 1, and 2 lp/mm, respectively.

One approach to increase pixel signal-to-noise ratio (SNR) in low noise digital fluoroscopy is to employ in-situ pixel amplification via current-mediated active pixel sensors (C-APS). Experiments reveal a reduction in readout noise and indicate that an a-Si C-APS, coupled together with an established X-ray detection technology such as amorphous selenium (a-Se), can meet the stringent requirements (of < 1000 noise electrons) for digital X-ray fluoroscopy. A challenge with the C-APS circuit is the presence of a small-signal input linearity constraint. While using such a pixel amplifier for real-time fluoroscopy (where the exposure level is small) is feasible, the voltage change at the amplifier input is much higher in chest radiography or mammography due to the larger X-ray exposure levels. The larger input voltage causes the C-APS output to be non-linear thus reducing the pixel dynamic range. In addition, the resulting larger pixel output current causes the external column amplifier to saturate further reducing the pixel dynamic range. In this research, we investigate two alternate amplified pixel architectures that exhibit higher dynamic range. The test pixels are designed and simulated using an a-Si TFT model implemented in Verilog-A and results indicate a linear performance, high dynamic range, and a programmable circuit gain via choice of supply voltage and sampling time. These high dynamic range pixel architectures have the potential to enable a large area, active matrix flat panel imager (AMFPI) to switch instantly between low exposure, fluoroscopic imaging and higher exposure radiographic imaging modes. Lastly, the high dynamic range pixel circuits are suitable for integration with on-panel multiplexers for both gate and data lines, which can further reduce circuit complexity.

Vapor deposited lead iodide films show a wide range of physical attributes dependant upon fabrication conditions. High density is most readily achieved with films less than 100 μm. Thicker films, with lessening density, often show lower response (gain) as charge collection becomes less efficient. Lack of consistency in density throughout a deposition invariably leads to non-uniform electronic properties, which is challenging to both model and predict. To overcome this, tighter control of deposition parameters is required during the slow growth process (<10 μm/hour). Lead iodide films are characterized in forms of planar devices deposited onto conductive glass and active pixel arrays deposited onto a-Si TFT arrays1. Electronic properties (e.g. leakage current, gain) show little variation that can be traced to substrate choice. Films generally provide less than 100 pA/mm² leakage current as they show saturation in gain (at approximate fields of 1 V/μm). We recently modified our readout electronics to accept positive bias. Using positive bias on the top electrode provides better charge collection for the lower mobility electrons and (despite process variability) better quality films can provide sensitivities greater than 6 μC/R*cm², with only partial x-ray absorption, and show less than 20 pA/mm² dark current.

In this paper, we present a simulation methodology that allows computed tomography (CT) detector designer to assess the criteria and tolerance requirements of positioning and alignment of x-ray collimators to scintillation detector arrays. Based on x-ray and optical transport simulation, we have developed an analytical model to predict the response of the measurement chain (collimator-scintillator-photodiode) in CT detector. We present the resulting effects of misalignments of the collimator array to detector array in the imaging chain through the simulation of beam hardening (water equivalent material and/or Bone) and light output using 120 kVp bremsstrahlung x-ray spectrum from a tungsten target anode source.

We analyzed the relationship between the noise components and detectability of the various Imaging Plates to find the essential factors for improving Imaging Plate. CR system noise is classified into quantum noise and fixed noise. The dominant component of fixed noise is the structural noise of IP. The contribution of the structural noise was relatively high at higher exposure and at higher spatial frequency. For example the ratio of structural noise of early type of IP is 55% in case of 2 mR (5.16 × 10^-7 C/kg) exposure at 2 cycles/mm spatial frequency. On the other hand, the ratio of the latest type is about 20%. This improvement leads to 2.3 times NEQ combined with improvement of quantum noise. Threshold depth of the 0.5 mm diameter hole of CD diagram was improved from 1.4 mm to 0.8 mm according to the improvement of NEQ from 50,000 to 100,000 /mm² at the related spatial frequencies. Amount of improvement of threshold contrast was influenced especially by the NEQ at relatively high spatial frequencies. So improvement of the structural noise, which is dominant at higher frequency, is important as well as quantum noise.

We have developed a novel method to measure the presampling modulation transfer function (MTF) in digital radiography systems using a novel edge device and algorithm. It can simultaneously measure the presampling MTFs in horizontally and vertically by utilizing its four edges. Calculation algorithm is composed of six steps, which are detection of edge, determination of angle, differentiation, composition of line spread function (LSF), fast Fourier transform (FFT) and sinc correction, respectively. Verification of the accuracy was conducted comparing with the established slit method. The repeatability of the measurement and the dose dependence was also examined. The measured MTF of edge device was coincident to that of slit within 0.02 up to the Nyquist frequency (3.125 cycles/mm). The repeatability was within 0.002 up to the Nyquist frequency. It is also confirmed that the result is not affected by the alignment error against the x-ray axis. In conclusion, an accurate and feasible method to measure the presampling MTF was established using a novel edge test device and algorithm.

At CEA-LETI, a DEXA approach for systems using a digital 2D radiographic detector has been developed. It relies on an original X-rays scatter management method, based on a combined use of an analytical model and of scatter calibration data acquired through different thicknesses of Lucite slabs. Since Lucite X-rays interaction properties are equivalent to fat, the approach leads to a scatter flux map representative of a 100% fat region. However, patients’ soft tissues are composed of lean and fat. Therefore, the obtained scatter map has to be refined in order to take into account the various fat ratios that can present patients. This refinement consists in establishing a formula relating the fat ratio to the thicknesses of Low and High Energy Lucite slabs leading to same signal level. This proportion is then used to compute, on the basis of X-rays/matter interaction equations, correction factors to apply to Lucite equivalent X-rays scatter map. Influence of fat ratio correction has been evaluated, on a digital 2D bone densitometer, with phantoms composed of a PVC step (simulating bone) and different Lucite/water thicknesses as well as on patients. The results show that our X-rays scatter determination approach can take into account variations of body composition.

A silicon mixed-signal integrated circuit is needed to extract and process x-ray induced signals from a coated flat panel thin film transistor array (TFT) in order to generate a digital x-ray image. Indigo Systems Corporation has designed, fabricated, and tested such a readout integrated circuit (ROIC), the ISC9717. This off-the-shelf, high performance, low-noise, 128-channel device is fully programmable with a multistage pipelined architecture and a 9 to 14-bit programmable A/D converter per channel, making it suitable for numerous X-ray medical imaging applications. These include high-resolution radiography in single frame mode and fluoroscopy where high frame rates are required. The ISC9717 can be used with various flat panel arrays and solid-state detectors materials: Selenium (Se), Cesium Iodide (CsI), Silicon (Si), Amorphous Silicon, Gallium Arsenide (GaAs), and Cadmium Zinc Telluride (CdZnTe). The 80-micron pitch ROIC is designed to interface (wire bonding or flip-chip) along one or two sides of the x-ray panel, where ROICs are abutted vertically, each reading out charge from pixels multiplexed onto 128 horizontal read lines. The paper will present the design and test results of the ROIC, including the mechanical and electrical interface to a TFT array, system performance requirements, output multiplexing of the digital signals to an off-board processor, and characterization test results from fabricated arrays.

Experimental prototype of a photon counting scanning slit X-ray imaging system is being evaluated for potential application in digital mammography. This system is based on a recently developed and tested “edge-on” illuminated Microchannel Plate (MCP) detector. The MCP detectors are well known for providing a combination of capabilities such as direct conversion, physical charge amplification, pulse counting, high spatial and temporal resolution, and very low noise. However, their application for medical imaging was hampered by their low detection efficiency. This limitation was addressed using an “edge-on” illumination mode for MCP. The current experimental prototype was developed to investigate the imaging performance of this detector concept for digital mammography. The current prototype provides a 60 mm field of view, 200 kHz count rate with 20% non-paralysable dead time and >7 lp/mm limiting resolution. A 0.3 mm focal spot W target X-ray tube was used for image acquisition. The detector noise is 0.3 count/pixel for 50x50 micron pixels. The count rate of the current prototype is limited by the delay line readout electronics, which causes long scanning times (minutes) and high tube loading. This problem will be addressed using multichannel ASIC electronics for clinical implementation. However, the current readout architecture is adequate for evaluation of the performance parameters of the new detector concept. It is very simple and provides a maximum intrinsic resolution of 28 micron FWHM. The prototype was evaluated using resolution, contrast detail and breast Phantoms. The MTF and DQE of the system are being evaluated at different tube voltages. The design parameters of a scanning multiple slit mammography system are being evaluated. It is concluded that a photon counting, quantum limited and virtually scatter free digital mammography system can be developed based on the proposed detector.

Slot scanning imaging techniques allow for effective scatter rejection without attenuating primary x-rays. The use of these techniques should generate better image quality for the same mean glandular dose (MGD) or a similar image quality for a lower MGD as compared to imaging techniques using an anti-scatter grid. In this study, we compared a slot scanning digital mammography system (SenoScan, Fisher Imaging Systems, Denver, CO) to a full-field digital mammography (FFDM) system used in conjunction with a 5:1 anti-scatter grid (SenoGraphe 2000D, General Electric Medical Systems, Milwaukee, WI). Images of a contrast-detail phantom (University Hospital Nijmegen, The Netherlands) were reviewed to measure the contrast-detail curves for both systems. These curves were measured at 100%, 71%, 49% and 33% of the reference mean glandular dose (MGD), as determined by photo-timing, for the Fisher system and 100% for the GE system. Soft-copy reading was performed on review workstations provided by the manufacturers. The correct observation ratios (CORs) were also computed and used to compare the performance of the two systems. The results showed that, based on the contrast-detail curves, the performance of the Fisher images, acquired at 100% and 71% of the reference MGD, was comparable to the GE images at 100% of the reference MGD. The CORs for Fisher images were 0.463 and 0.444 at 100% and 71% of the reference MGD, respectively, compared to 0.453 for the GE images at 100% of the reference MGD.

The method of calculating DQE of a general digital imaging system is proposed by IEC and it is coming to the stage of final draft. However, about digital mammography, nothing is decided yet. This research examines the evaluation method for image quality of a digital mammography with clinical equipment through physical evaluation of the mammographic computed radiography (CR) systems under clinical conditions. We used two CR systems. One consisted of a single plate image reader (FCR PROFECT CS, Fuji), which includes dual-side reading and 50-micron pixels. Other consisted of a single plate image reader (FCR 5000H, Fuji), which includes single-side reading and 100-micron pixels. Digital characteristic curves, presampling MTFs and digital Wiener spectra were measured as indices of image quality. Presampling MTFs were measured from slit and edge images at 28kV. Digital Wiener spectra were measured at 28kV with breast equivalent filter. Presampling MTFs with both readings were almost the same. Digital Wiener spectra with dual side reading were superior to those with single side reading. NEQ of CR system with dual side reading was superior to that with single side reading because of the good efficiency of light condensing. New mammographic CR systems with dual side readings should be a further powerful tool for detecting low-contrast lesions in breast. Wiener spectra need to determine exposure conditions, in order to perform comparison between institutions, since it is strongly influenced of beam quality and a dose. We also compared overall characteristic curves, overall MTFs and overall Wiener spectra of a new CR system with them of a screen-film system. Although MTF was calculated by the slit method, it is necessary to examine another method in quest of MTF including the effect of image processing of CR system.

Small microcalcifications essential to the early detection of breast cancer may be obscured by overlapping tissue structures. Dual-energy digital mammography (DEDM), where separate low- and high-energy images are acquired and synthesized to cancel the tissue structures, may improve the ability to detect and visualize microcalcifications. The investigation of DEDM began with a signal-to noise ratio analysis to estimate and relate the noise level in the dual-energy calcification signals to the x-ray spectra, microcalcification size, tissue composition and breast thickness. We investigated various inverse-mapping functions, both linear and non-linear, to estimate the calcification thickness from low- and high-energy measurements. Transmission (calibration) measurements made at two different kVp values for variable aluminum thickness (to simulate calcifications) and variable glandular-tissue ratio for a fixed total tissue thickness were used to determine the coefficients of the inverse-mapping functions by a least-squares analysis. We implemented and evaluated the DEDM technique under narrow-beam geometry. Phantoms, used in the evaluation, were constructed by placing different aluminum strips over breast-tissue-equivalent materials of different compositions. The resulting phantom images consisted of four distinct regions, each with a different combination of aluminum thickness and tissue composition. DEDM with non-linear inverse-mapping functions could successfully cancel the contrast of the tissue-structure background to better visualize the overlapping aluminum strip. We are currently in the process of translating our DEDM techniques into full-field imaging. We have designed special phantoms with variable glandular ratios and variable calcification thicknesses for evaluation of the full-field dual-energy calcification images.

MTF is accepted as a measure for sharpness of a detector system, but analysis of one system by different researchers often results in differences. This can be due to differences in exposure setup or calculation algorithm. In this multicenter study, we investigate which options in the algorithm for the edge method result in differences in MTF. Three edge images were sent for analysis to nine participants, together with a questionnaire about different steps in their algorithm. One image was generated synthetically and is scatter-free and noise-free. The other images were created with an edge phantom between two slabs of 2 cm PMMA and were known to have a slight difference in MTF. The results were compared in both absolutely and relatively. All participants could calculate the MTF from the images. Although there are numerous differences between the different implementations, the results for the synthetical image are quite similar. This indicates that the algorithms perform similarly in noise-free and scatter-free conditions. With the real images, larger deviations are observed. The implementations can be divided in two groups according to their ability to reproduce a low frequency drop. The main difference between both groups was the use of data conditioning prior to the Fourier transform. In the group with low frequency drop, only slight absolute differences are observed. The other algorithms show larger differences. These differences underline the need for guidelines if the MTF curve gets a crucial role in the acceptance of a digital mammography system.

The purpose of this study is to describe a method that allows the calculation of a contrast-detail curve for a particular system configuration using simulated micro calcifications into clinical mammograms. We made use of simulated templates of micro calcifications and adjusted their x-ray transmission coefficients and resolution to the properties of the mammographic system under consideration (4). We expressed the thickness of the simulated micro calcifications in terms of Al equivalence. In a first step we validated that the thickness of very small Al particles with well known size and thickness can be calculated from their x-ray transmission characteristics at a particular X-ray beam energy. Then, micro calcifications with equivalent diameters in the plane of the detector ranging from 300 to 800 μm and thicknesses, expressed in Al equivalent, covering 77 to 800 μm were simulated into the raw data of real clinical images. The procedure was tested on 2 system configurations: the GE Senographe 2000 D and the Se based Agfa Embrace DM1000 system. We adapted the X-ray transmissions and spatial characteristics of the simulated micro calcifications such that the same physical micro calcification could be simulated into images with the specific exposure parameters (Senographe 2000D: 28 kVp-Rh/Rh, Embrace DM1000: 28 kVp-Mo/Rh), compressed breast thickness (42+/-5mm) and detector under consideration. After processing and printing, 3 observers scored the visibility of the micro calcifications. We derived contrast-detail curves. This psychophysical method allows to summarize the performance of a digital mammography detector including processing and visualization.

Although digital mammography is currently being used all over the world, most of the mammographic units are still based on screen-film systems. In these systems, the choice of the best combination between the screen and the film is important to assure image quality. This work presents a computer simulation method to help choosing the proper screen-film system to each application, showing sensitometric parameters for each combination, like speed, latitude and contrast determined by sensitometric curves film and screen-film combination, measured experimentally, using commercial calibrated sensitometer and densitometer, allowing further comparison between them according to the application required. Panthom images are presented showing the results that will be obtained in clinical use. The influence of each screen is also determined. Phantom images were obtained using a known screen-film combination. These images were digitized in a laser scanner. The exposures information is used to predict the final image by using the screen-film sensitometric curve, chosen by the user. The computer simulation was used for evaluating several mammographic films combined with different screens, currently used for mammography. Simulated results were in good agreement with values obtained experimentally. The results obtained with the proposed algorithm confirm the possibility of using this method for evaluating the performance of any screen-film combination considering the sensitometric curve. It can be an important tool for quick evaluation of a screen-film system as it predicts the image characteristics with no unnecessary X-ray exposition.

A method for measuring the characteristic curves generated by the mammography imaging systems has not yet been well established due to poor quality control over X-ray exposure in the range of kV values, which is lower than the conventional quality. In this paper, we proposed a bootstrap method using a “stepwedge” designed for characteristic curve measurement in mammography. A ten-step stepwedge containing calcium phosphate, with each step having a different density of material, was employed. In our experiment, the tube voltage and mA values were changed in the range of 25 to 32 kV at increments of 1 kV and in the range of 20 to 100 mAs at increments of 20 mAs, respectively. The results of the curve measurements indicated that our method might be useful to both screen-film mammography and computed radiography (CR), although additional experiments to evaluate the accuracy and precision of the acquired data are required.

In this work, we examine the application of intensity diffraction tomography (I-DT) for imaging three-dimensional (3D) phase objects. We develop and investigate two algorithms for reconstructing phase objects that utilize only half of the measurements that would be needed to reconstruct a complex-valued object function. Each reconstruction algorithm reconstructs the phase object by use of different sets of intensity measurements. We demonstrate that the numerical and noise propagation properties of the two reconstruction algorithms differ considerably.

Tomographic images of biological materials, for example, phantoms of Polyethylene, Polycarbonate, Nylon and Plexiglas with 10 mm diameter with internal bores of 2 to 5 mm diameter and length containing Intralipid-10% solution in various concentrations are obtained. The concentration is defined as the percentage of the volume of the Intralipd-10% against that of water, and the region of interest. In addition to these, the phantoms are filled with few other soft materials (external shell with internal soft material) and the images are obtained. The data acquisition and image reconstruction procedures of the laser CT system are similar to those of X-ray CT. The images are acceptably distinct. Based on experimental observation of the images, the present system can reconstruct images of small objects and soft materials in the trans-illumination mode. In addition, we have studied partially the dynamical properties of the embedded biological soft tissue of a snail.

μ-tomographic images are obtained for few soft materials with multi-structure, cylinder with holes of different diameter and biological soft tissue. 2D images are obtained in the transmission mode. 3D images are reconstructed with the use of the 2D slices for visualization of the internal structure. In addition, we used few simple geometrical approximations, for example, total geometrical efficiency, solid angle contribution and gradient. We obtained a series of images in the region 24-27 keV. 3D visualization of the materials is processed and analyzed the results. The present study is also focused on few geometrical considerations in order to design the collimators in front of the fluorescent source to improve the geometrical efficiency.

The heterodyne detection technique, on which the coherent detection imaging (CDI) method founds, can discriminate and select very weak, highly directional forward scattered, and coherence retaining photons that emerge from scattering media in spite of their complex and highly scattering nature. That property enables us to reconstruct tomographic images using the same reconstruction technique as that of X-Ray CT, i.e., the filtered backprojection method. Our group had so far developed a transillumination laser CT imaging method based on the CDI method in the visible and near-infrared regions and reconstruction from projections, and reported a variety of tomographic images both in vitro and in vivo of biological objects to demonstrate the effectiveness to biomedical use. Since the previous system was not optimized, it took several hours to obtain a single image. For a practical use, we developed a prototype CDI-based imaging system using parallel fiber array and optical switches to reduce the measurement time significantly. Here, we describe a prototype transillumination laser CT imaging system using fiber-optic based on optical heterodyne detection for early diagnosis of rheumatoid arthritis (RA), by demonstrating the tomographic imaging of acrylic phantom as well as the fundamental imaging properties. We expect that further refinements of the fiber-optic-based laser CT imaging system could lead to a novel and practical diagnostic tool for rheumatoid arthritis and other joint- and bone-related diseases in human finger.

There many instances where the use of a quasi-monochromatic x ray source (QMS) would be an advantage in radiology. One instance is dual-energy imaging, where two images may be processed in order to enhance visualization of a particular element, typically iodine, but also possibly gadolinium. Another instance is fluoroscopy, where the radiation and/or contrast doses may be minimized by using a radiation source with an energy peak that lies just above the contrast K edge. The most straightforward method of implementing a QMS is to incorporate appropriate heavy elements into an x ray tube anode and then to filter the resulting spectrum appropriately. Science Research Laboratory, Inc. has embarked on a program to develop anode materials that can withstand the power loadings encountered in a useful tube. Thus far, we have developed and tested one relevant anode material (erbia) and demonstrated that it could withstand a power of 0.25 MW/cm² at a tip speed of 3300cm/sec without harm. The destructive limit was determined to be at least a factor of two higher due to overwriting of the electron beam in certain regions. Under the condition of no overwriting, coating survivability should scale directly with tip speed. For futures studies we have designed an apparatus that increases the tip speed by a factor of 4.8. In the longer term, the tip speed could be increased by as much as a factor of 20.

Gibbs ringing is an inevitable artifact in MR Fourier Transform (FT) imaging caused by truncating k-space data via a rectangular window. A common practice is to multiply k-space data with a filter function prior to FT reconstruction for softening the amount of ringing and accepting the concomitant blur. Although by properly choosing the filter functions, the apodization approach is effective in removing the overshoot and ringing, it is at the price of both spatial correlation and spatial resolution of pixel intensities in the reconstructed image. Our study reported in this paper shows that the truncation and apodization introduce/change the spatial correlation and the spatial resolution of pixel intensities in FT imaging. We develop a 4-step procedure progressively prove the spatially asymptotic independence of pixel intensities and introduce a delta-train method to derive the spatial resolution in the image, for both the basic and the filtered FT imaging. The methods are intuitive without any assumption and approximation. Different filter functions have been investigated. We also discuss the relationship between the spatial resolution and pixel size and correct some misconceptions on these issues.

A computer-aided diagnosis scheme for the detection of interstitial disease in standard digital posteroanterior (PA) chest radiographs is presented. The detection technique is supervised-manually labelled data should be provided for training the algorithm-and fully automatic, and can be used as part of a computerized analysis scheme for X-ray lung images. Prior to the detection, a segmentation should be performed which delineates the lung field boundaries. Subsequently, a quadratic decision rule is employed for every pixel within the lung fields to associate with each pixel a probabilistic measure indicating interstitial disease. The locally obtained per-pixel probabilities are fused to a single global probability indicating to what extent there is interstitial disease present in the image. Finally, a threshold on this quantity classifies the image as containing interstitial disease or not. The probability combination scheme presented utilizes the quantiles of the local posterior probabilities to fuse the local probability into a global one. Using this nonparametric technique, reasonable results are obtained on the interstitial disease detection task. The area under the receiver operating characteristic equals 0.92 for the optimal setting.

Accurate computation of dose delivered to the patient depends on knowledge of the x-ray spectrum of therapy, x-ray beams. This work applies the expectation-maximization (EM) method to the problem of reconstructing the x-ray source spectrum from transmission data of Aluminum and Lead. The proposed EM method is compared to an algebraic approach for solving the corresponding linear system of equations.

This work presents a computer algorithm to evaluate film digitizers in terms of its spatial resolution by using image processing techniques. A testing pattern containing slits of different widths, not-equally spaced, was developed. When digitizing this pattern, the algorithm automatically analyses the digital image and evaluate spatial resolution capabilities of the digitizer. These analyses were made by calculating computationally digital image slit width and the distance between the slits. Sampling distances in both directions (parallel and perpendicular to scan direction) are determined by comparing calculated values with previous measurements made by using a calibrated microscope. Evaluation also includes the determination of the presampling modulation transfer function (MTF). The algorithm was used for the measurement of effective pixel size and presampling MTF of a Lumiscan 50 laser digitizer and an Umax Powerlook II optical scanner in both directions. Results showed that the laser digitizer presents significative difference between parallel and perpendicular MTF and the optical scanner presents better MTF considering high frequencies components. Results obtained with the proposed algorithm confirm the possibility of evaluating spatial resolution limitations of any film digitizer using an automatic and simple method.

The purpose of this study is to determine the effects of the working distance on the accuracy of confocal scanning laser tomography using the Heidelberg Retina Tomograph II. Twenty eyes of normal patients were imaged and the topographies of the retinal surfaces were recorded. Each eye was imaged first at the optimum working distance, establishing the baseline exam, and then re-imaged at four different working distances (one at a shorter distance than optimum, three more at longer distances than optimum, variation done in 2 mm increments). The recorded data at various working distances was compared to the baseline data. The deviation from the baseline was compared to the normal standard deviation for the instrument reported in the literature. Data is within the normal standard deviation when staying between -2 mm and +4 mm of optimum working distance. Some stereometric parameters vary greater than the normal standard deviation if working distance is more than +4 mm from optimum. To minimize error in recorded data, the operator of the Heidelberg Retinal Tomograph II should image the patient’s eye between -2 mm and +4 mm of optimum working distance. Staying in this range should provide results that vary within the normal standard deviation.

The Radiology Research Laboratory at the Henry Ford Hospital has been involved in modeling x-ray units in order to predict image quality. A critical part of that modeling process is the accurate choice of interaction coefficients. This paper serves as a review and comparison of existing interaction models. Our objective was to obtain accurate and easily calculated interaction coefficients, at diagnostically relevant energies. We obtained data from: McMaster, Lawrence Berkeley Lab data (LBL), XCOM and FFAST Data from NIST, and the EPDL-97 database via LLNL. Our studies involve low energy photons; therefore, comparisons were limited to Coherent (Rayleigh), Incoherent (Compton) and Photoelectric effects, which were summed to determine a total interaction cross section. Without measured data, it becomes difficult to definitively choose the most accurate method. However, known limitations in the McMaster data and smoothing of photo-edge transitions can be used as a guide to establish more valid approaches. Each method was compared to one another graphically and at individual points. We found that agreement between all methods was excellent when away from photo-edges. Near photo-edges and at low energies, most methods were less accurate. Only the Chanter (FFAST) data seems to have consistently and accurately predicted the placement of edges (through M-shell), while minimizing smoothing errors. The EPDL-97 data by LLNL was the best over method in predicting coherent and incoherent cross sections.

GATE (Geant4 Application for Tomographic Emission) was used to perform a Monte-Carlo simulation of a fully 3D clinical PET scanner. The Philips Allegro PET system was simulated in order (a). to allow a detailed study of the parameters affecting the system’s performance under various imaging conditions, and (b). to further validate the use of GATE for the simulation of clinical PET systems. A model of the detection system and its geometry was developed. The simulation of count rate related performance characteristics of the scanner was facilitated through the development of a dead time model to describe data flow and data loss at the level of detected single events and coincidences. The developed system design and associated dead time model were validated by comparing simulated with experimental measurements obtained with the Allegro PET system. These measurements included the use of point as well as distributed sources, allowing us to determine spatial resolution, scatter fraction, sensitivity, and count rate performance based on the NEMA NU2-1994 protocols. Using the NEMA phantom, simulated single and coincidence count rates were within 5% of the corresponding measured rates throughout a wide range of activity concentrations. Scatter fraction and random coincidences were also measured and combined with total recorded coincidence rates in order to validate simulated NEC rate curves. Differences between simulated and measured NEC curves were found to be within 7% and can be attributed to the approximations in the simulation including no photomultiplier tube response, gantry surroundings or the effects of pulse pile up in the modeling of the electronics. These results support an accurate modelling of the Philips Allegro PET system using GATE in combination with an appropriate dead time model.

Inter-crystal scatter could lead to loss of data or mispositioning of scintillation events, which is of a particular interest in imaging detectors used in Positron Emission Tomography. Because it is difficult to measure the scatter and to quantify its effects, a Monte Carlo simulation has been proposed to study inter-detector scattering and how it affects the spatial resolution of a BGO block. The MCNP4C2 code was used for simulations of the 8 x 4 crystal BGO block detector, the silicon layer covering its face and a line source. Accurate simulations were made for the grooves in the BGO block detector (light guides), which vary according to the location of the crystal within the block. In transaxial plane the cuts are symmetrical about the central cut and have a depths of 23.1, 24.4, 27.7 and 30mm respectively. In the axial plane there are three cuts, tow outer cuts are 30mm deep and the central cut is 23.1mm deep. The line spread function (LSF) was simulated for each column and row at a time by scanning the line source axially and transaxially across the collimated fall of the BGO block detector, which shielded all other columns and rows. Comparison was made between simulated inter-crystal scattering for neighbouring crystals

In this paper, we present the results and methods used to simulate the x-ray deposition and the light optical transport in x-ray detectors composed of scintillating and photodiodes arrays. Through this work, we have identified the main contributors of crosstalk and validated the model results through experimental measurements. We have assessed and quantified the effects of the reflectors, the pixel geometry and dimensions, the optical coupler between the scintillating material and the silicon photodiodes and optimized the design in order to meet the requirements of detector image quality in computer tomography. We describe some of the methods and techniques used to determine the optical properties of the material components in the simulation.

In this paper, the accuracy and precision of RSA analysis using a GE Innova^TM 4100 digital flat panel and a Siemens Multistar x-ray image intensifier (XRII) were evaluated and compared with that of a conventional film-screen system, in order to explore the possibility of real-time kinematic and dynamic RSA study. A phantom, having two rigid body segments with no movement, was constructed and imaged by the digital flat panel, XRII and conventional screen-film systems, respectively. The acquired images were measured and motions were derived. The mean and standard deviation of the repeated results were analyzed to determine the accuracy and precision, respectively. Comparing all three axes, the lowest rotational accuracy and precision were 0.008 ± 0.011°, 0.013 ± 0.015° and 0.006 ± 0.05° while the lowest translational accuracy and precision were 25 ± 28 mm, 17 ± 37 mm and 4 ± 6 mm for the film-screen, XRII and digital flat panel, respectively. The evaluation of the accuracy and precision of the RSA in this study confirms its place as a highly accurate method. The study shows that both digital flat panel and XRII systems have potential application to the kinematics and dynamics joint study.

Our goal in this paper is to evaluate the capability of real-time selenium-technology-based full-field digital mammography (FFDM) system in breast tomosynthesis. The objective of this study is to find out the present status of amorphous selenium technology in the sense of advanced applications in clinical use. We were using tuned aperture computed tomography (TACT+) 3-dimensional (3D) technology for reconstruction. Under evaluation were amorphous selenium signal-to-noise-ratio, flat panel image artefacts and acquisition time to perform full-field digital mammography 3D examination. To be able to validate the system we used a special breast phantom. We found out that 3D imaging technology provides diagnostic value and benefits over 2-dimensional (2D) imaging. 3D TACT advantages are to define if mammography finding is caused by a real abnormal lesion or by superposition of normal parenchymal structures, to be able to diagnose and analyze the findings properly, to detect changes in breast tissue which would otherwise be missed, to verify the possible multifocality of the breast cancers, to verify the correct target for biopsies and to reduce number of biopsies performed. Slice visualization and 3D volume model provide greater diagnostic information compared to 2D projection screening and diagnostic imaging.

Traditional Tomosynthesis requires the X-ray source to be parallel beams or cone beams constrained to the same plane above the object of interest. Commonly, the X-ray sources are placed uniformly around a circle or along a line in the same plane, which demands fixed and high-precision equipment. To eliminate the constraints on the X-ray sources and obtain fast and efficient reconstruction with adequate quality, a fast and unconstrained cone beam Tomosynthesis reconstruction method as well as the corresponding deblurring method is presented in this paper. In this method, two reference balls, whose connecting line is parallel to the detector, are placed above the detector. According to the information provided by these two balls, the X-ray source position and its relative motion are readily calculated for the reconstruction and the corresponding deblurring processes. A layer-by-layer, rather than a voxel-by-voxel, reconstruction is utilized in order to speed up the calculation. This fast and unconstrained cone beam Tomosynthesis has a large commercial value, allowing unconstrained projection acquisition and making portable and relatively cheap Tomosynthesis equipment possible.

A reconstruction method for transillumination laser CT using optical heterodyne detection is proposed. Laser CT does not obey the Radon transform due to surface effects, which occur as results of refractive index mismatch and roughness of surface, and bring about annulus artifact and impair quantitative accuracy of a reconstructed image. Optical attenuation with surface effects is described using a model function with respect to an incident angle. This reconstruction produces a least squares problem incorporating the surface effect, which is solved via the conjugate gradient method. The method is applied to data from a acrylic physical phantom containing Intralipid-10% solution in various concentrations for demonstrating the effectiveness of the proposed method in terms of morphological and quantitative information. The CDI method is proved to have an excellent quantitative accuracy.