In recent years, as a direct result of extensive research and development, a growing number of x-ray imagers based on the powerful concept of active matrix addressing have been introduced to clinical environments worldwide. These imagers consist of thin, large area, glass substrates onto which 2D, pixelated arrays are deposited. Each pixel comprises an amorphous silicon thin-film switch coupled to some form of pixel storage capacitor. The radiation is detected indirectly, by means of an overlying scintillators, or directly, by means of a photoconductor layer. While existing active matrix imagers offer many advantages, the present technology suffers form a variety of distinct limitations and disadvantages that pertain to imaging performance under conditions of low exposure per image frame, as well as to issues of mechanical robustness and cost. In this paper, the possibility that these restrictions could be removed through the adoption of novel, emerging technologies is discussed. The focus of the discussion is primarily on technologies under independent development for large-area electronics and other applications and for which a plausible chance of convergence with x-ray imaging technology exists.

Our goal is to understand the temporal response of amorphous selenium (a-Se) x-ray imaging detectors. The temporal response caused by charge trapping and release gives rise to the imaging properties of lag and ghosting. These imaging properties depend on both the design and operational parameters of the detectors as well as the material properties of a-Se. Our approach was to measure the x-ray photocurrent of electroded a-Se detectors as a function of x-ray exposure and correlate these measurements with an experimental investigation of charge trapping by use of the 'time-of-flight' method. These experimental results are compared with models based on previously published values of material properties such as electron and hole trapping and recombination cross-sections. The resulting model can be used to predict lag and ghosting in radiographic and fluoroscopic flat panel detectors. We also investigated the effect of charge trapping between the pixel electrodes of an a-Se flat-panel detector on lag and ghosting. Our method for quantitative evaluation of lag and ghosting in a-Se will facilitate the optimization of imaging performance. This is achieved by determining the combination of materials properties, system design and operational parameters which minimize artifacts arising from either lag or ghosting in a-Se flat-panel detectors.

We describe a high-resolution digital x-ray detector suitable for producing high quality mammographic images. The detector consists of an array of 3584 by 4096 pixels on 70 micrometer centers covering an area of 25 cm by 29 cm. The conversion layer of the detector is 250 micrometer thick amorphous selenium. Each pixel of the array consists of a storage capacitor for collecting x-ray signals and an amorphous silicon switching transistor. The signal is read out by custom high-speed, low-noise electronics. The integration of this detector with a mammographic x-ray system and acquisition console is described, as well as algorithms for calibration of the full system. We review characterization of the imaging performance of our system based on quantitative analyses of MTF and DQE data, and compare experimental results with theoretical calculations. We compare the performance of our direct conversion system with that of screen/film analog systems and indirect conversion digital detectors, such as LORAD's CsI/CCD detector, operated under similar conditions. MTF degradation mechanisms and system noise sources and their effect on DQE are discussed. We review qualitative aspects of image quality from our detector and present preliminary observer performance characteristics on clinical studies run with our system.

This paper describes our investigation of the X-ray detective characteristics of a thick polycrystalline CdZnTe film deposited on a large-area substrate. We deposited a polycrystalline CdZnTe film on a 9 inch by 9 inch substrate, and investigated its quality. It was verified to be quite uniform within the substrate. We also cut the film and connected it to a 3 inch by 3 inch TFT panel for evaluating the X-ray imaging performance. The TFT array format was 512 by 512 pixels with a pixel pitch of 150 micrometers . The thickness of the CdZnTe film was about 300 micrometers after lapping and polishing, and the film density per unit area was higher than 170 mg/cm². The average sensitivity was 1.5E9 e-/mR/mm²; the beam condition was 80 kV with 26-mm Al filtration. The MTF measured at 1 lp/mm was 0.82. The time response and uniformity of X-ray sensitivity were not still adequate, and further improvements are in progress. In conclusion, we have demonstrated the applicability of the polycrystalline CdZnTe film to a large-area detector, although further investigations and improvements are needed.

The effects of charge carrier trapping (i.e. incomplete charge collection) on the detective quantum efficiency (DQE) of a photoconductive detector are studied by using a cascaded linear system model. The model includes signal and noise propagations in the following stages: (1) x-ray attenuation, (2) conversion gain, (3) charge collection, (4) the addition of electronic noise. We examine the DQE(0) of a-Se for fluoroscopy application as a function of photoconductor thickness with varying amounts of electronic noise under (a) constant field, and (b) constant voltage operating conditions. We show that there is an optimum photoconductor thickness, which maximizes the DQE(0) under a constant voltage operation. The optimum thickness depends on the added electronic noise, x-ray exposure, bias voltage and polarity. The actual broad x-ray spectrum emitted from a typical x-ray tube is used in the calculation. The DQE for the negative bias is significantly lower than that of the positive bias, and the diversity in DQE, as expected, increases with the photoconductor thickness because of the asymmetric transport properties of holes and electrons in a-Se. The present results show that the DQE generally does not continue to improve with greater photoconductor thickness in the presence of added electronic noise because of charge transport and trapping effects.

Probability distribution functions (pdfs) of x-ray computed tomography (CT) signals form the basis for statistical reconstruction algorithms and for noise-simulation experiments. The conventional model for pdf's assumes a quanta-counting process obeying a discrete Poisson distribution. In reality, CT scanners employ energy-integrating sensors detecting a polyenergetic X-ray beam, with data quantized for digital reconstructions. A model was developed for the CT signal consisting of an energy spectrum of x-ray quanta (individually obeying Poisson statistics), contributing an incremental signal, proportional to their energy, to an analog-to-digital converter. Using a moment generating function approach, the pdf is shown to be a Compound Poisson process that is functionally dependent on the x-ray energy spectrum, flux level, and quantization step size. Pdfs were computed numerically by a Fourier transform of the characteristic function and compared to experimental pdfs collected from phantom scans. For exposures encountered in normal clinical usage, the pdf is similar to the conventional Gaussian approximation, with rescaling for quantization and polyenergetic spectra. For low intensities, the deviation from the conventional photon-counting (Poisson) model is significant and may have implications for statistical reconstruction algorithms.

Theoretical models of the detective quantum efficiency (DQE) provide insight into fundamental performance limitations and standards to which particular systems can be compared. Over the past several years, cascaded models have been developed to describe the DQE of several flat panel detectors. This article summarizes the governing principles of cascaded models, and conditions that must be satisfied to prevent misuse. It is shown how to incorporate: a) poly-energetic x rays; b) Swank noise; c) the Lubberts effect; d) reabsorption of K x rays from photo-electric interactions; e)secondary quantum noise; and, f)noise aliasing. Cascaded models involve cascading theoretical expressions of the noise-power spectrum (NPS) through multiple stages. Most expressions involve two or three terms, requiring the manipulation of algebraic expressions consisting of hundreds of terms. This practical limitation is alleviated using MATLAB's Simulink programming environment and symbolic math manipulations. It is shown that even for an 'indirect' detector, noise aliasing reduces the DQE by up to 50 percent at the cut-off frequency. Secondary quantum noise is generally a small effect, but reabsorption can reduce the DQE by 20-25 percent over a wide range of spatial frequencies.

Accurate attenuation correction in dual isotope single photon emission computed tomography (SPECT) requires spill-down and scatter correction. Convolution methods that allow such correction were investigated. Three energy windows were used: 1) a Compton scatter window placed below the Tc-99m photopeak window, 2) the Tc-99m photopeak, and 3) both In-111 photopeaks. Spill-down of In-111 into the Tc-99m window was measured using a small source of In-111 at depths of 5, 7, 9, and 11cm. From these data, the convolution kernel relating the In-111 image and the Tc-99m window spill-down image was computed for each depth. Using a small Tc-99m source at the same depths, the convolution kernels describing its spill-down into the Compton window were computed. A factor to scale the Tc-99m spill-down estimate by for obtaining the self-scatter estimate was determined from the total counts in the Tc-99m photopeak vs. the Compton scatter window when primary photons were blocked with a small lead disk. Depths of 5 and 9cm were optimal for spill-down and Tc-99m self-scatter correction, respectively. It was demonstrated that the spill-down and scatter correction increased effective attenuation coefficient towards its 'no-scatter' value and dramatically decreased the intensity of an In-111 source in Tc-99m projection images.

Physical and mathematical procedures of x-ray CT and MRI have been extensively studied during the past three decades. The statistical methods of image reconstruction, processing and analysis of x-ray CT and MRI have been widely used in the various fields. However, the lack of a complete description and a rigorous proof of their statistical properties become a common concern. This study attempts to fill this gap and discusses a framework for studying the statistical aspects of x-ray CT and MRI.

The influence of phosphor screens on the digital system image quality has been studied in a number of papers. However, there has been no detailed description of the effect of depth of x-ray interaction on the blur characteristics of the phosphor and on optical collection efficiency for both powder and structured screens. We present an analysis based on optical Monte Carlo simulations of the depth-dependent phosphor blur of two classes of single-layer phosphor screens: homogeneous and columnar. The spectral sensitivity of the optical sensor is modeled according to a typical a-Si:H photodiode absorption profile. We used Gd₂O₂S:Tb and CsI:Tl emission spectra respectively for the powder and columnar phosphor models. We present line-spread (LSF) and modulation transfer (MTF) functions associated with the spread of signal in the phosphor, and optical collection efficiencies. We find good agreement between the Monte Carlo estimates of the MTF and the analytical solutions available in the literature. Our optical collection efficiency results show depth dependence only for the screens with highly scattering and absorptive phosphor with reflective backing, and for the case of scattering phosphor with absorptive backing.

Terahertz imaging is an emerging modality, with potential for medical applications, using broadband sub-picosecond electromagnetic pulses in the range of frequencies between 100 GHz and 100 terahertz (THz). Images can be formed using parameters derived from the time domain, or at the range of frequencies in the Fourier domain. The choice of frequency at which to image may be an important factor for clinical applications. Image quality as a function of frequency was assessed for a terahertz pulsed imaging system by means of; (i) image noise measurements on a specially designed step wedge, and (ii) modulation transfer functions (MTF) derived from a range of spatial frequency square wave patterns. It was found that frequencies with larger signal magnitude gave lower image noise, measured using relative standard deviation (standard deviation divided by mean) for uniform regions of interest of the step wedge image. MTF results were as expected, with higher THz frequency signals demonstrating a consistently higher MTF and higher spatial frequency limiting resolution than the lower THz frequencies. There is a trade-off between image noise and spatial resolution with image frequency. Higher frequencies exhibit better spatial resolution than lower frequencies, however the decrease in signal power results in a degradation of the image.

Linear digital tomosynthesis consists in acquiring on a digital detector a few projections, at different view angles and for a linear X-ray source path. A simple shift and add process reconstructs planes parallel to the detector plane but with a low vertical resolution. To improve it, we propose to use Algebraic Reconstruction Technique (ART) and Half Quadratic Regularization (HQR) methods, which are based on an iterative process. In such a linear digital tomosynthesis context, we propose to reconstruct independent 2D tilted planes passing through the linear source trajectory. Thus, computation time is reduced and it becomes possible to regularize in an anisotropic way to adapt the regularization process to the reconstructed volume sampling. We validate our approach with experimental data acquired on a digital detector. ART significantly improves the vertical resolution in comparison with usual process. However, ART is sensitive to noisy projections and may produce poor quality reconstructions. The HQR piecewise smoothness constraint stabilizes the reconstruction process. With a total angular range of 40 degrees, we can reach a vertical resolution lower than 1 cm, while it is superior to 3 cm with the usual process. Furthermore, HQR method significantly reduces truncation artifacts due to high projection angles.

Stereoscopic viewing of radiographic images is advantageous in reducing ambiguity due to background anatomic noise. This advantage is usually considered to be balanced by doubling the dose required for acquisition of a stereo image pair. In the case of a quantum-noise limited detector, detection theory suggests a possible decrease of the dose by half. We tested this assumption by a series of contrast-detail observer experiments, using phantom images acquired over a range of exposures. The number of visible details, the effective reduction of the dose, and the effective decrease in the threshold SNR were compared for the acquired images viewed mono- and stereoscopically. Experimental results support our hypothesis for the images acquired at lower exposures. The quantum noise is more evident in the lower exposure images, corresponding to the quantum noise limited detection case. With increasing exposure, the observed dose benefit of stereoradiography decreased, but was always positive. Potential reasons for the reduced benefit observed with higher exposures are discussed.

A realistic numerical environment for simulating three-dimensional (3D) angiographic imaging of the coronary arteries has been developed. Through numerical simulation we propose to optimize acquisition and gating strategies, aiding in the design of 3D coronary imaging systems. We have previously developed a dynamic model of the coronary arteries, based on a high-resolution 3D image of an excised human heart, which was perfused with iodinated contrast agent. To mimic the motion of the arteries during the cardiac cycle, the motion of the vessel branch points was determined from cine bi-plane coronary angiograms of a patient with vessel anatomy similar to the excised heart. The static image was then non-linearly deformed to produce a sequence of volumetric images, with isotropic 0.4-mm resolution, representing the motion of the coronary arteries throughout the cardiac cycle. To simulate different acquisition strategies we have developed an algorithm to forward project through the volume data sets. The geometry of the CT system used to acquire the original 3D image of the static heart is mimicked in the re-projection algorithm. Thus, prospective radiographic projections corresponding to any projection-angle can be produced for any time -point throughout the cardiac cycle. Combining re-projections from selected time-points and view angles enables the evaluation of various acquisition and gating strategies.

By analyzing the noise properties of calibrated low-dose Computed Tomography (CT) projection data, it is clearly seen that the data can be regarded as approximately Gaussian distributed with a nonlinear signal-dependent variance. Based on this observation, the penalized weighted least-square (PWLS) smoothing framework is a choice for an optimal solution. It utilizes the prior variance-mean relationship to construct the weight matrix and the two-dimensional (2D) spatial information as the penalty or regularization operator. Furthermore, a K-L transform is applied along the z (slice) axis to further consider the correlation among different sinograms, resulting in a PWLS smoothing in the K-L domain. As a tool for feature extraction and de-correlation, the K-L transform maximizes the data variance represented by each component and simplifies the task of 3D filtering into 2D spatial process slice by slice. Therefore, by selecting an appropriate number of neighboring slices, the K-L domain PWLS smoothing fully utilizes the prior statistical knowledge and 3D spatial information for an accurate restoration of the noisy low-dose CT projections in an analytical manner. Experimental results demonstrate that the proposed method with appropriate control parameters improves the noise reduction without the loss of resolution.

Using helical, multi-detector computed tomography (CT) imaging technology operating at sub-second scanning speeds, clinicians are investigating the capabilities of CT for cardiac imaging. In this paper, we describe the application of novel modeling tools to assess CT system capability. These tools allow us to quantify the capabilities of both hardware and software algorithms for cardiac imaging. The model consists of a human thorax, a dynamic model of a human heart, and a complete physics-based, CT system model. The use of the model to predict image quality is demonstrated by varying both the reconstruction algorithm (half-scan, sector-based) and CT system parameters (axial detector resolution). The mathematical tools described provide a means to rapidly evaluate new reconstruction algorithms and CT system designs for cardiac imaging.

We have previously reported the effects of stereo angle and exposure on depth discrimination and the use of virtual cursors for absolute depth measurements in digital stereomammography. The current study further investigates the effects of magnification and zooming on depth perception. Stereoscopic image pairs of a phantom were acquired with a full-field digital mammography system. The modular phantom contained 25 crossing fibril pairs with depth separations between each pair ranging from 2 to 10 mm. Three phantom (fibril) configurations were imaged using techniques of 30 kVp, Rh/Rh, +/- 3 degrees stereo angle, , contact and 1.8X magnification geometry, and 4 to 63 mAs exposure range. The images were displayed on a Barco monitor driven by a Metheus stereo graphics board and viewed with LCD stereo glasses. Five observers participated in the study. Each observer visually judged if the vertical fibril was in front of or behind the horizontal fibril in each pair. It was found that the accuracy of depth discrimination increased with increasing fibril depth separation and x-ray exposure. Zooming the contact stereo images by 2X did not improve the accuracy. Under conditions of high noise (low mAs) and small depth separation between the fibrils, depth discriminations were significantly better in stereo images acquired with geometric magnification than in images acquired with a contact technique and displayed with or without zooming. This study indicates that stereoscopic imaging, especially with magnification, may be useful for visualizing the spatial distribution of microcalcifications in a cluster and differentiating overlapping tissues from masses on mammograms.

Diffraction Enhanced Imaging (DEI) can significantly improve the expressiveness of mammography radiographs. Whereas the contrast in conventional radiographs is based on small X-ray absorption differences of tissues, the contrast mechanism of DEI is, in addition, partially related to the differences in X-ray refraction properties. DEI has been successfully applied to in-vitro mammography studies where little absorption tissue differentiation is present. In this paper we will present work on high-energy DEI mammography, which has been carried out by utilizing a tunable monochromatic X-ray beam. Since the refraction characteristics of soft tissues are much less energy dependent than absorption, the use of high energy X-rays is favoured. They can be employed in mammographic imaging without reducing the image contrast, while getting the benefit of reduced dose since the X-ray absorption falls off considerably. In-vitro images of an American College of Radiology (ACR) mammographic phantom using monochromatic X-rays through 50 keV have been obtained with a digital detector. High-energy mammography has been successfully performed at a significantly lower dose than that usually applied in clinical mammography without important contrast loss.

The introduction of the Full Field Digital Mammography (FFDM) opens the way to a large range of future advanced applications. Among them, Contrast Enhanced Digital Mammography (CEDM) could be a fast and less expensive alternative to Magnetic Resonance Imaging (MRI) for breast lesion characterization. In this work, we have investigated, first on phantom then on patients, the capability of a modified FFDM system to show the contrast enhancement of lesions after intra-venous injection of iodine. The uptake has been estimated from the difference between pre- and post-contrast images. Phantom results showed that 1) detectability thresholds of the contrast media were compatible with clinical conditions; 2) breast radiological thickness has a low impact on uptake detectability; 3) spatial and temporal analysis showed delayed margin contrast uptake of the simulated lesion and slow increase of contrast in the background. Preliminary results on patients have confirmed the phantom results and have shown a contrast uptake in all malignant lesions despite the observed patient motion artifacts, and some moderate signal variability. This study demonstrated the feasibility of the Contrast Enhanced Digital Mammography technique. Further investigations and clinical validations will have to be completed before it can be widely used in a daily routine practice.

Although screen-film mammography is the current standard for detecting abnormalities in the breast, the sensitivity can range from 40-95 percent. One of the reasons this occurs is because surrounding dense tissues obscure the visualization of cancers. To address this issue, we are investigating contrast digital mammography (CDM), which involves imaging the uptake of a non-ionic iodinated contrast agent with full-field digital mammography to detect and characterize masses in the breast. The technical aspects, including spectral optimization and image processing, for the implementation of CDM have been investigated and a clinical trial is being carried out to study the technique. This article describes an experimental study using an iodine detail phantom in combination with various breast equivalent plastics to evaluate technique parameters used when imaging patients. It is desirable to find the technique parameters that will maximize iodine contrast while delivering a reasonably low dose to the patient. It was found that the contrast can be increased up to 53 percent for thinner breasts by increasing the dose to a level that still remains lower than the dose that would be received by the breast from conventional screening images.

A novel x-ray source, providing dichromatic beams for the application of dual-energy radiography, has been assembled and studied. The system works via Bragg diffraction, by monochromatizing the beam produced by a conventional W-anode x-ray tube with a mosaic crystal monochromator. The source generates a laminar beam (10 cm-high and 0.8 cm-wide), composed by two spatially superimposed quasi-monochromatic beams. The characteristics of the radiation field in terms of energy resolution and fluence have been reported, for three pairs of energies. A study of the spectra attenuated by several phantoms of breast equivalent tissue of different thicknesses shows that the optimal energy of the dichromatic beam for dual energy mammography application ranges between 18/36 keV and 18.6/37.2 keV and may be set as a function of the thickness or density of investigated tissue.

The most natural way of digital X-ray imaging is photon counting as the photon flux in itself is digital. In photon counting, the information in the X-ray flux is used more efficiently as the information carrying low-energy photons are given the same weight as higher energy photons carrying less image information. This is in contrast to all existing X-ray instruments, which are energy-integrating systems where the highest energy photons are given the highest weight. A novel technique for high resolution digital X-ray imaging, using gaseous avalanche detectors for photon counting with high signal-to-noise ratios for single X-ray photons, has been developed. The performance of this detector has been studied and compared to analogue film-screen system by imaging phantoms. Our results show that this technology can improve image quality while decreasing the glandular dose to the patient.

Flat-panel imager (FPI) based cone-beam computed tomography (CBCT) is a strong candidate technology for intraoperative imaging in image-guided procedures such as brachytherapy. The soft-tissue imaging performance and potential navigational utility have been investigated using a computer-controlled benchtop system. These early results have driven the development of an isocentric C-arm for intraoperative FPI-CBCT, capable of collecting 94 projections over 180 degrees in 110 seconds. The C-arm system employs a large-area FPI with 400 micron pixel pitch and Gd₂O₂S:Tb scintillator. Image acquisition, processing and reconstruction are orchestrated under a single Windows-based application. Reconstruction is performed by a modified Feldkamp algorithm implemented on a high-speed reconstruction engine. Non-idealities in the source and detector trajectories during orbital motion has been quantified and tested for stability. Cone-beam CT imaging performance was tested through both quantitative and qualitative methods. The system MTF was measured using a wire phantom and demonstrated frequency pass out to 0.6 mm^-1. Voxel noise was measured at 2.7 percent in a uniform 12 cm diameter water bath. Anatomical phantoms were employed for qualitative evaluation of the imaging performance. Images of an anaesthetized rabbit demonstrated the capacity of the system to discern soft-tissue structures within a living subject while offering sub-millimeter spatial resolution. The dose delivered in each of the imaging procedures was estimated from in-air exposure measurements to be approximately 0.1 cGy. Imaging studies of an anthropomorphic prostate phantom were performed with and without radioactive seeds. Soft-tissue imaging performance and seed detection appear to satisfy the imaging and navigation requirements for image-guided brachytherapy. These investigations advance the development and evaluation of such technology for image-guided surgical procedures, including brachytherapy, vertebroplasty and neurosurgery. The demonstrated soft-tissue visibility, excellent spatial resolution, low imaging dose, and convenient from factor make C-arm based cone-beam CT a powerful new technology for image-guidance applications.

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 volume CT imaging technique with a recently developed flat panel detector. A computer simulation study has been performed to prove the feasibility of developing a flat panel detector-based cone beam volume CT breast imaging (FPD-CBVCTBI) technique. In this study, a phantom and specimen experiment is performed to confirm the findings in the computer simulation using the current prototype cone beam volume CT scanner. The results indicate that the CBVCTBI technique effectively removes structure overlap and significantly improves the detectability of small breast tumors. More importantly, the results also demonstrate the patient dose level required for FPD-based CBVCTBI to detect a small tumor (under 5 mm) and a small calcification is less than or equal to that of conventional mammography. The results from this study suggest that FPD-CBVCTBI is a potentially powerful breast-imaging tool.

A feasibility study of breast CT with synchrotron radiation is currently being carried on at Elettra, the Trieste synchrotron radiation facility. Breast CT cannot be implemented easily with conventional radiographic tubes, due to the high dose that would be delivered to the breast by a polychromatic X-ray spectrum. The possibility of tuning the beam energy, available at a synchrotron radiation beamline, allows a significant reduction in the delivered dose, and at the same time the use of monochromatic beams avoids beam hardening artifacts on the reconstructed image. Images of in vitro breast tissue samples have been acquired by means of a high efficiency linear array detector coupled to a VLSI single photon counting readout electronics. The pixel width, determining the pixel size of the reconstructed image, is 200 micrometers , while the pixel height, determining the CT slice thickness, is 300 micrometers . Tomograms have been reconstructed by means of standard filtered backprojection algorithms. Images of normal and pathologic breast tissue samples show a good visibility of glandular structure. The delivered dose was in all cases comparable to the one delivered in clinical planar mammography. Due to the promising results we obtained, in vivo studies are under evaluation.

This paper presents a wavelet analysis-based multi-resolution cone-beam volume CT breast imaging technique that is adaptive for high-resolution and ultra-high resolution reconstructions. Wavelet analysis-based de-noising techniques are employed to improve image quality and further reduce the required absorbed dose. The following steps can summarize this technique. First, in the high-resolution mode, the high spatial resolution projections are rebinned into lower resolution projections through a wavelet decomposition/synthesis procedure while in the ultra-high-resolution mode the original spatial resolution of the projection data is kept. Second, a wavelet analysis-based de-noising technique is applied upon the projection data with quantum fluctuations to suppress the noise level in the reconstructed images. Third, a de-noising method through an adaptation of a wavelet shrinkage approach for noise reduction is utilized in the reconstructed data to improve the image quality in terms of the signal-to-noise ratio and dose efficiency. The computer simulations show that the wavelet analysis-based multi-resolution rebinning approach provides the flexibility to adjust spatial the reconstruction resolution and noise level for various imaging tasks. Also, the wavelet analysis-based de-noising technique efficiently suppresses the quantum mottle induced noise, and contributes to a better low contrast object reconstruction in terms of the signal-to-noise ratio (SNR) improvement. In addition, the reconstruction of a high contrast object, for example, a tiny calcification grain, is obtained with less density spread. The noise level in the reconstructed image is reduced, which means the necessary dose level can be further reduced while the image quality is not compromised.

The signal-to-noise ratio (SNR) characteristics of volumetric flat-panel cone-beam CT images are investigated theoretically and experimentally. Analytical models of the noise-power spectrum (NPS) and voxel noise developed in the context of conventional, slice-based CT are extended to the fully 3D case. While early models describe well the noise characteristics of 2D tomographic reconstructions for the case of a 1D detector with constant NPS, the fully 3D approach extends classical descriptions to quantify the 3D NPS and voxel noise in a manner that includes: a 2D detector with generalized blue and NPS characteristics, secondary quantum noise in the conversion of incident x-rays, electronics noise, and 3D aliasing. Classical relationships between image SNR and voxel size and 'slice thickness' are found inaccurate in describing the noise characteristics of 3D flat-panel one-beam CT. The relationship is shown instead to depend on the characteristics of the 3D bandwidth integral associated with the 2D detector modulation transfer function and choice of reconstruction filter. Image SNR is investigated experimentally for a low-frequency soft-tissue visualization task using a prototype system for FPI-CBCT and a low-contrast phantom consisting of tissue-equivalent inserts in water. The contrast-to-noise ratio in images of the low-contrast phantom consisting of tissue-equivalent inserts in water. The contrast-to-noise ratio in images of the low-contrast phantom is measured as a function of scatter-to-primary ratio, imaging dose, and spatial resolution in value reconstructions and compared to theoretical expectations. The adaptability of flat-panel cone-beam CT is quantitatively revealed in a task-directed approach that seeks to 'tune' image SNR through knowledgeable selection of dose and spatial resolution in a manner that is consistent with the imaging task and clinical constraints.

A standardized metrics program was developed and implemented by Advanced Diagnostics, Inc. (ADI) to quantify Diffractive Energy Imaging (DEI) system performance. DEI is a medical imaging technology which uses primarily diffracted wave information obtained form passing ultrasound energy through the anatomy. The resulting real-time, large field-of-view, high-resolution images are currently undergoing clinical evaluation for detection and management of breast disease. This unique imaging modality required novel modifications to conventional measurement and evaluation tools. Measurement tools were developed and metrics procedures were standardized to ensure measurement accuracy and repeatability during performance testing. The system modulation transfer function, effective size of the field- of-view, spatial resolution across the effective field-of- view, contrast resolution, minimum detection, and field uniformity were quantified. Improved system metrics monitor improvements to DEI image quality and are supported through images of target standards and subjective validation using images of human anatomy.

Diffraction Enhanced Imaging (DEI) is a powerful X-ray imaging technique that allows the visualization of structures having different refraction and/or absorption properties with respect to the background. In DEI, the sample is irradiated with a monochromatic and highly collimated X-ray beam, and the outgoing beam is analyzed by means of a perfect crystal. A comparison was drawn among DEI images of a standard (ACR) and a custom phantom using different harmonic diffraction orders. Images were obtained at two different synchrotron beamlines, the SYRMEP beamline at Elettra and the X15A beamline at the NSLS (Brookhaven, NY), utilizing a double-crystal Si monochromator and a single-crystal Si analyzer, operated in the symmetric, non-dispersive Bragg configuration. The harmonic order was separated by placing a refractive prism between the two crystals of the monochromator. The use of the and the reflections resulted in a 5-fold improvement in the analyzer angular sensitivity, consequently enhancing the extinction and refraction contrasts with respect to the reflection. The detail visibility was improved by 1-2 orders of magnitude. By means of the refractive prism technique, even higher harmonics might be used, thus promising even better image quality.

An x-ray multilayer mirror, specially designed to produce resonant absorption at a definite angle of incidence, may be used as an angular dispersive element for refractive x-ray radiography. In this method the signal-to-noise ratio can be significantly enhanced due to suppression of the shot noise produced by the direct beam. Refraction contrast of a copper wire 75 microns in diameter and a human hair was observed using Ni/C multilayer mirror with resonant absorption at CuKa radiation. The multilayer structure consisting of 30 bilayers was designed for CuKa radiation so as to have absorbing resonance of the width of about several arc seconds at a grazing angle of 0.8 degrees. A monochromatic probe x-ray beam with a divergence of approximately 5 arc seconds was obtained from a conventional x-ray tube and a double crystal monochromator set in a strongly dispersive configuration. We have developed theoretical basis for this method, and have experimentally proven that it is possible to create critical components for its practical implementation: a multilayer mirror with resonant absorption, an x-ray imaging photon-counting detector with spatial resolution of about several micrometers, and a probe beam with the divergence of several arc seconds. This result proves the feasibility of x-ray refraction radiography using resonantly absorbing multilayer mirrors manufactured by conventional magnetron sputtering technology.

Conventional x-ray imaging relies almost entirely on differences in the absorption of x-rays between tissues to produce contrast. While these differences are substantial between bone and soft tissue, they are very small between different soft tissue types resulting in poor visualization of soft tissues. Diffraction enhanced imaging (DEI) is currently in development by several groups as a new imaging modality that exploits information contained within the x- ray scattering distribution at low angles. We have used the SYRMEP beam line at the Elettra Synchrotron facility in Trieste, Italy to image a variety of tissue specimens, together with several phantoms. Mono-energetic photons in the range 17 keV to 25 keV were used with an analyzer crystal which diffracted the x-rays onto a detector. We have obtained some spectacular images which display remarkable contrast and resolution. The images can be processed to separate the pure absorption and pure refraction effects in a quantitative manner. These images demonstrate that DEI provides tissue morphology information not accessible with conventional radiographic imaging. The contrast caused primarily by refraction as the x-ray passes from one tissue type to another in the specimen is evident. Since x-ray refraction is much less energy dependent than absorption there is considerable potential for extremely low dose imaging. We believe that the potential of this technique is considerable and we present dat to illustrate the quality of the images.

Fundamental study on parallel beam radiography using a polycapillary plate is described. The x-ray generator used in this experiment is NST-1005 made by Sofron Inc. with maximum tube voltage and current of 100kV and 5.0 mA, respectively. In this experiment, the tube voltage was regulated from 20 to 30 kV, and the tube current had a constant value of 4.0 mA. The exposure time is regulated in order to control optimum film density. The polycapillary plate is J5022-21 made by Hamamatsu Photonics, and the outside and effective diameters are 87 and 77 mm, respectively. The thickness and the whole diameter of the polycapillary are 1.0 mm and 25 micrometers , respectively. The x- rays from the tube are formed to parallel beam by the polycapillary, and the radiogram is taken using an industrial x-ray film of Fuji IX 100 without using a screen. In the measurement of image resolution, we employed three brass spacers of 2, 30, and 60 mm in height. By the test chart, the resolution decreased according to increases in the spacer height without using the polycapillary. In contrast, the resolution seldom varied when the polycapillary was employed. In the polycapillary radiography of four tungsten wires, higher-contrast images of 50 micrometers wire were observed, and the line width seldom varied according to increases in the spacer height.

In angiography practice an iodate contrast medium is injected in patient vessels with catheters. The absorption of x-rays raises immediately after the iodine K-edge energy. In digital subtraction angiography, two images are used, acquired before and after the injection of the contrast medium, respectively. The vessels morphology result from the difference of images so obtained. This technique involves a non-negligible risk of morbidity or mortality, due to high concentration of injected contrast agent. We are investigating a new source which produces two thin parallel quasi-monochromatic beams - having peak energies centered before and after the iodine K-edge energy, respectively - by using a conventional x-ray tube and a highly oriented pyrolytic graphite mosaic crystal. The polychromatic x-rays incident on the crystal are monochromatized by Bragg diffraction and split in two thin parallel beams, by means of a collimating system. These two beams impinge on the phantom simulating patient vessels and are detected with solid-state array detectors. The image results as difference between the remaining intensities of two beams. We report a preliminary study of the new technique performed both with theoretical stimulations and experimental measurements. Results of computer simulation give information about characteristics as size and quality of the quasi- monochromatic beams, that should be considered in detail to design a system dedicated to the clinical practice. Experimental measurements have been performed on a small- field detector in order to shows the enhancement of image contrast obtained with the application of the new technique.

In a real fluoroscopic system experimental evaluations of the DQE may pretty soon run into difficulties. Easy as it might be to satisfy the need for linearity by means of correction look-up tables, the evaluation of the NPS is more tricky, because of various time integration mechanisms. In order to deal with such effects in a quantitatively correct manner the concept of a spatial-temporal1 DQE has been suggested. We have performed computer-aided DQE-evaluations 2,4,5 on a surgical C-arm, using MTF and NPS. Furthermore, we have attempted to estimate the time behavior of the spatial-temporal system transfer function. Using X-ray pulses in the ms regime, we have generated nearly 'lag-free' flat-field images. Our experiments showed two interesting results. The comparison of flat-field images in the continuous 'Fluoro' mode and the 'lag-free' mode revealed the theoretically expected highly overestimated DQE in the first case. The corresponding scaling factor could be derived quantitatively from the motion experiments with an X-ray contrast pulse (Cu-rod). More worth while noticing is the fact that we observed structural anomalies in the two-dimensional NPS that could not compensated for by a simple scaling factor but vanished only in the 'lag-free' mode. This can be explained theoretically by taking into account a mixing behavior between the spatial and temporal NPS components, i.e. the failure of the spatial-temporal separability of the system transfer function.

The purpose of this research is to establish the technical and clinical image quality of a 30 X 40 cm² dynamic flat detector (FD) compared to state-of-the-art IITV technology. A Trixell detector for vascular and RF applications is designed for a mixed use of fluoroscopy as well as exposure series and a range of radiographic applications. An RF system has been built which comprises both the FD as well as an IITV detector. This system enables a direct comparison of technical image quality measurements and patient images under exactly the same X-ray conditions. Image quality measurements comprise Detective Quantum Efficiency including transfer characteristics, Modulation Transfer Function, Noise Power Spectrum, lag, Low Frequency Drop and residual signals. Observation tests, using Threshold Contrast Detail Detectability (TCDD) techniques, are performed in order to confirm the results of the technical measurements. Results show a DQE (f) of the flat detector that is higher compared to IITV and above all constant over a wide dose range, the IITV DQE (f) drops at higher dose range due to fixed structure. Furthermore the Low Frequency Drop is substantially smaller in the FD-based system. The TCDD subjective tests show improved system performance in favor of the FD system.

New neuro-interventional devices such as stents require high spatial-resolution image guidance to enable accurate localization both along the vessel axis as well as in a preferred rotational orientation around the axis. A new high-resolution angiographic detector has been designed with capability for micro-angiography at rates exceeding the 5 fps of our current detector and, additionally, with noise low enough and gain high enough for fluoroscopy. Although the performance requirements are demanding and the detector must fit within practical clinical space constraints, image guidance is only needed within a approximately 5 cm region of interest at the site of the intervention. To achieve the design goals, the new detector is being assembled from available components which include a CsI(Tl) phosphor module coupled to a fiber-optic taper assembly with a two stage light image intensifier and a mirror between the output of the fiber taper and the input to a conventional high performance optical CCD camera. Resulting acquisition modes include 50-micron effective pixels at up to 30 fps with the capability to adjust sensitivity for both fluoroscopy and angiography. Estimates of signal at the various stages of detection are made with quantum accounting diagrams (QAD).

Fluoroscopic images are degraded by scattering of x-rays from within the patient and by veiling glare in the image intensifier. Both of these degradations are well described by a response function applied to primary intensity. We can automatically estimate the parameters of the response function with the aid of a reference object placed in the imaging field. Subtraction of the true reference signal yields artifacts unless proper scatter-glare correction is performed. We adjust the scatter-glare parameters in order to minimize these artifacts. We demonstrate this technique using an anthropomorphic phantom plus additional scattering material. Root mean square error in densitometric measurements of an x-ray phantom is reduced by 54 percent compared with no correction and by 36 percent compared with subtraction of uniform scatter measured under a beam stop.

A new DR system using a large-area flat panel detector (FPD) with a 40 by 30 cm active area and a 194 micrometers pixel pitch, has been developed to compare with a conventional image intensifier and charge-coupled device camera type DR system. After measuring basic characteristics of the new DR system such as signal-to-noise ratio, modulation transfer function, and detective quantum efficiency, we applied the FPD to a Gastro-Intestinal study with contrast media, and discussed its potential for clinical use with a medical doctor. In radiography mode, the new DR system with a large-are FPD has superior image quality compared with the conventional I.I.- CCD camera type DR system because of high SNR and DQE. In fluoroscopy mode, the SNR of the new DR system at the exposure range of over 2(mu) R/frame is similar with the conventional I.I.-CCD camera type DR system. As a result, we considered that new DR system with a large-area FPD could be applied to a clinical study replacing an I.I.-CCD camera type. In the evaluation using various clinical images taken with the new DR system by a medical doctor, the new DR system with a large-are FPD performed sufficiently for a GI study.

Recent advances in mouse genomics, including the production of transgenic mouse models, have created an interest in developing non-invasive imaging techniques for small-animal imaging applications. X-ray computed tomography (CT) can provide images with high-resolution isotropic voxels and low noise in relatively short acquisition times. In addition, CT provides volume data set, which allows the viewer to clearly visualize the spatial orientation of tissues within the mouse. We propose a model for an ideal, quantum-noise limited CT scanner for small-animal orientation of tissues within the mouse. We propose a model for an ideal, quantum- noise limited CT scanner for small-animal imaging with the objective of examining the fundamental limits of precision as a function of resolution and dose to the animal. The variance was calculated for several doses and voxel sizes to determine the precision in the linear attenuation coefficient values for the idealized small-animal volume CT scanner. For whole-body exposure of 1.5 Gy, our study predicts precision of +/- 5.8 percent in linear attenuation coefficient, with (0.1 mm)³ isotopic voxels. This work shows the effect of photon noise on the precision that can be expected for micro-computed tomography of small animals in vivo for a given isotopic voxel size and x-ray dose to the animal. The predictions of this work ca be used to design novel imaging systems for use in small-animal research.

We develop a new projection weighting function for interpolation and reconstruction of multi-slice helical computed tomography data with the hope of reducing longitudinal aliasing in reconstructed volumes. The weighting function is based on the application of the Papoulis generalized sampling theorem to the interlaced longitudinal samples acquired by the multi-slice scanner. We call the approach 180MAA, for multi-slice anti-aliasing. For pitch 3, the 180MAA approach yields high-quality images of the 3D Shepp-Logan phantom as well as a longitudinal MTF superior to that of the 180MLI approach, which is based on the use of linear interpolation. However, it is not as successful at mitigating aliasing as had been doped due to the presence of a significant and unexpected aliasing component that can be attributed to the small cone angle in multi-slice helical CT. The presence of this effect is interesting and significant in its own right, however.

Detection of lesions in planar mammograms is a difficult task, predominantly due to the masking effect of superimposed parenchymal breast patterns. Tomographic imaging of the breast can provide image slices through the breast, possibly reducing this masking effect. In recent years, there has been interest in developing CT mammography using flat-panel digital detectors in a truncated cone-beam geometry. In this study, we have developed a framework for determining optimal design and acquisition parameters for such a CT mammographic system. The ideal observer SNR is used as a figure-of-merit, under the assumptions that the imaging system is linear and shift-invariant, and that the noise is stationary. The ideal observer calculation uses mathematical models of signal and noise propagation through the flat-panel detector, and realistic models of the lesion detection task in breast imaging. It is used to investigate optimal kVp settings of a tungsten anode spectra for CT imaging of the uncompressed breast, given the constraint of an average glandular dose approximately equivalent to that of a two-view planar mammography study. It is observed that modeling a realistic mammographic background structure into the detection task can affect the optimal kVp settings suggested by the ideal observer SNR. Since the exposure/view in flat-panel CT mammography is considerably lower than for planar mammography, it is observed that electronic additive noise can also affect the optimal kVp setting. In general, the optimal kVp settings for the tungsten anode spectra studied here were in the range of 30-50 kVp.

The macroscopic static field inhomogeneity is not only the source of MR signal loss in gradient echo based imaging techniques, but also the source of geometrical image distortion as well as a limitation of spectral resolution. This piece of information is useful for both active shimming coil design and clinical imaging application. In order to further understand the spatial variation of the macroscopic background static field in human brain during MRI examination, this static field inhomogeneity was measured from the adult human volunteers with a volumetric imaging scheme, which was based on a 3D gradient echo technique with two consecutive gradient echoes. All the human volunteers were scanned in supine position using a birdcage headcoil on a 1.5 T clinical whole body scanner. We have constructed a high resolution 3D static field map over the brain volume. All experimental results have shown consistently that there are mainly two spots in the brain tissue volume exhibiting relatively severe static magnetic field inhomogeneity . They are normally located in the brain areas in the inferior frontal lobe immediately anterior to the nasal cavity and in the inferior temporal lobe above the ear canals, where air spaces exist in the vicinity. At those locations, the observed offset frequency in the proton resonance reached about 50 Hz over 5 mm distance along the z direction at 1.5 Tesla, corresponding to 1.5 ppm/cm locally.

The use of x-ray microtomography to evaluate 3D volumes of bone has become wide spread with the introduction of numerous commercial systems intended for the examination of small animals and specimens. However, a consistent method for describing the spatial resolution has not been employed in characterizing the performance of these units. In this paper, we report the blur factors which contribute to a system's performance, methods to experimentally assess the performance of microtomography system components, and an experimental method to assess tomographic resolving power. By scanning a 1.5 mm diameter ruby sphere, the overall system resolving power of a classic microtomography system was measured to have a 98 micron FWHM of a determined plane spread function, while the FWHM for a contemporary system was determined to be 23 microns.

The x-ray response of polycrystalline HgI2 for direct detection x-ray imagers, is studied using test arrays with 512 X 512 pixels of size 100 micron. We quantify the contributions to the x-ray sensitivity from electron and hole charge collection, x-ray absorption, effective fill factor and image lag, for x-ray energies from 25-100 kVp. The data analysis compares the measured sensitivity to the theoretical limit and identifies the contributions from various loss mechanisms. The sensitivity is explained by the ionization energy of approximately 5 eV, coupled with small corrections arising from incomplete x-ray absorption, incomplete charge collection, and image lag. Hence, imagers with HgI2 approach the theoretical maximum response for semiconductor detectors, with external array sensitivity demonstrated to within 50 percent of the limit.

The purpose of this study was to compare a large-area direct read-out flat-panel detector system with a speed class 200 screen-film system, a storage-phosphor system and a mammography screen-film system with regard to the detection of simulated rheumatoid erosions and to assess its diagnostic performance with decreasing exposure dose. The performance of a flat-panel system in such small lesions was considered especially interesting, as the spatial resolution of this system, limited by its pixel size, is considerably lower than that of conventional screen-film systems. An animal model with 160 joint specimens from 20 monkey paws was used. 640 regions were defined in these 160 meta- carpophalangeal and proximal interphalangeal joint specimens. Simulated rheumatoid erosions were created in 320 of these 640 regions. Specimens were enclosed in containers filed with water to obtain absorption and scatter radiation conditions similar to a human hand. Imaging was performed using a flat-panel system, a sped class 200 screen-film system, a mammography screen-film system and a storage- phosphor system under exactly matched conditions. Different exposure doses equivalent to speed classes of S equals 100, 200, 400, 800, 1600 and 3200 were used. Presence or absence of a lesion was assessed by three radiologists using a five level confidence scale. Receiver operating characteristic analysis was performed for a total of 21,120 observations and diagnostic performance estimated by the area under the ROC curve. The significance of differences between A_z values was tested with analysis of variance. ROC-analysis showed A_z values of 0.809, 0.768, 0.737, 0.710 and 0.685 for the flat-panel system, 0.770 for the screen-film system, 0.781, 0.739, 0.724 and 0.680 for the storage-phosphor system, and 0.798 for the mammography screen-film system. Analysis of variance showed significant differences for certain combinations of imaging modalities and exposure doses. The diagnostic performance of the flat-panel detector system is superior to that of a screen-film system and a storage-phosphor system for the detection of erosive lesions at clinical exposure settings. Using the flat-panel system the exposure does can be reduced by 50 percent to obtain a diagnostic performance comparable to a speed class 200 screen-film system.

Flat X-ray detectors require a systematic calibration and correction of image artifacts. Based on an analysis of the physics of the image generation chain, this work presents a unified framework for the correction of these artifacts. Algorithms for the correction steps are presented, including a new method for the calibration and correction of the intertwined offset, gain, and non-linearity as well as an improved method for the interpolation of defects, where the interpolation direction is chosen based on a novel method. Experiments using a hand phantom without and with a wire, imaged on a flat detector, demonstrate that line artifacts in Digital Subtraction Angiography (DSA) applications due to differences in non-linearity between adjacent amplifiers are significantly reduced by applying the non-linearity, offset, and gain correction in the correct order, as proposed in this work. For the defect interpolation investigations, we used medical images of angiographic image subtraction sequences, containing small vessels. Artificial clusters of pixel defects were added to these images and subsequently corrected. The experimental verification clearly demonstrates the robustness and superior performance of the new interpolation scheme, especially for clusters of defects.

Due to spatial gain differences of the photo diodes and inhomogeneities in the converter (CsI) a gain calibration is usually applied for flat dynamic X-ray detectors. This calibration is calculated from X-ray images. Using the reset light, integrated in the detector, a calibration of the photo diode gain is possible. Since neither the reset light intensity nor the X-ray field distribution in combination with the converter efficiency are spatially homogeneous the ratio of these two effects has to be measured and stored once in an X-ray reset-light map. In a reset light calibration the photo diode gain will be estimated and the final calibration is then calculated from this gain image and the stored X-ray reset-light map. The reset light gain image contains the same information as the X-ray image except the influence of the scintillator which should be very stable over time. Changes in the photo diode gain can easily and automatically be corrected using the reset light calibration. Defect pixels can be determined from the reset light gain images. This method would allow a continuous calibration during the lifetime of the detector without the need for any user interaction.

Quantitative analysis of a prototype photostimulable phosphor system for digital mammography was performed. The pre-sampled MTF, noise power spectrum (NPS), noise equivalent quanta (NEQ), and detective quantum efficiency (DQE) were measured at 26 and 32 kVp to assess the imaging performance of a commercial computed radiography system dedicated for mammographic imaging. The pre-sampled MTF demonstrated 5 percent modulation at 8 lp/mm with a small dependence on kVp, and noise power estimates indicated x-ray quantum-limited spectral characteristics from 2 mR up to approximately 30 mR incident exposure. Maintenance of x-ray information content up to approximately 500,000 quanta/mm2 based upon NEQ measurements was demonstrated. DQE (0 mm-1) was 30-50 percent, DQE (2.5 mm-1) was 15-25 percent, and DQE (4 mm-1) was 5-15 percent, depending on kVp, incident exposure, and readout direction. A significant increase in DQE compared to previous CR mammography implementations was found. In addition to the quantitative measurements, qualitative experience suggests that CR mammography is essentially equivalent to state-of-the-art mammography screen-film detector systems.

The FDA has approved the SenoScan slot-scanning Full-field Digital Mammography system. A high power Tungsten-target x- ray tube enables breast imaging with 0.22 s effective exposure time. A 21-cm X 29-cm area is scanned in less than 6 seconds, at a typical clinical technique of 30 kVp, 170 mA. The detector comprises a Thalium-doped Cesium Iodide scintillator coupled to a combination of four CCDs abutted along their narrow dimension to from a 10-mm by 210-mm slot. With the CCDs operated in time-delay-and-integration mode along the narrow dimension, the system functions in a continuous scanning mode. The MTF in the standard and high- resolution modes extend to 10-cycles/mm and beyond 14 cycles/mm respectively. The Detective Quantum Efficiency curve starts at 50 percent at DC and extends to 10 cycles/mm in Standard model. Accordingly the SenoScan system enables screening and diagnostic breast imaging with a limiting resolution approaching that of film-based systems. The overall system design and intrinsic scatter rejection efficiency directly translate in high DQE characteristics that enable screening at a significantly reduced patient dose.

We have implemented a scatter-correction algorithm (SCA) for digital mammography based on an iterative restoration filter. The scatter contribution to the image is modeled by an additive component that is proportional to the filtered unattenuated x-ray photon signal and dependent on the characteristics of the imaged object. The SCA's result is closer to the scatter-free signal than when a scatter grid is used. Presently, the SCA shows improved contrast-to-noise performance relative to the scatter grid for a breast thickness up to 3.6 cm, with potential for better performance up to 6 cm. We investigated the efficacy of our scatter-correction method on a series of x-ray images of anthropomorphic breast phantoms with maximum thicknesses ranging from 3.0 cm to 6.0 cm. A comparison of the scatter-corrected images with the scatter-free signal acquired using a slit collimator shows average deviations of 3 percent or less, even in the edge region of the phantoms. These results indicate that the SCA is superior to a scatter grid for 2D quantitative mammography applications, and may enable 3D quantitative applications in X-ray tomosynthesis.

For x-ray detectors, Compton interactions deposit photon energies along the paths of recoil electrons, which are not isotropic about the primary interaction sites. Light from each interacting x-ray is only generated near the path of a recoil electron. In this study, Compton scatter is modeled as an input-labeled cascade of the amplification and scattering processes to describe the transfer relationship of signal and noise in frequency space. The output of the model is the spatial distribution of secondary quanta generated by Compton recoil electrons. We determine the spatial dependence and statistical correlation of secondaries from the initial energy of the recoil electron and its range, resulting in the 'Compton' modulation transfer function (MTF) and noise power spectrum (NPS), respectively. Then the 'Compton' MTF and NPS are used to calculate the 'Compton' detective quantum efficiency (DQE). The probability density function of scattering angle of Compton recoil electron is developed using the Klein-Nishina coefficients. Results are applied to the description of a portal imaging system at 6 approximately MV where non-Compton interactions can be ignored. The MTF results are compared with a Monte Carlo calculation. This is the first model of how Compton interactions in the metal-plate/phosphor combination degrade image quality in terms of signal and noise. It is shown that Compton MTF depends on energy of x-ray photon in a complex way, and Compton scatter imposes a fundamental limitation on both the MTF and DQE of x-ray imaging system.

The lead disc method is a commonly used technique to measure scatter and glare in digital x-ray imaging1,2,3,4, however it is well known that the measurements are dependent upon the size of the disc. A common procedure is to take a series of measurements for a range of disc sizes and to extrapolate them to zero disc size, however the exact technique is not uniform of how this is done. Some use disc radius/diameter as the independent variable with extrapolation, some use non-linear extrapolation, while others use disc area and either linear or non-linear extrapolation. This paper presents a simple geometric model of x-ray scatter based on the ray tracing to calculate an estimate of scatter in the presence of lead discs. Under the experimental conditions used, the model predicted a non-linear relationship with both disc radius and area. A Toshiba mobile surgical x-ray apparatus, model SXT - 6 - 11, was used to measure scatter and glare with different field sizes for a range of disc diameters from 0.17 cm to 12 cm, for 16 cm of lucite. The model was found to be within 2 percent of measured estimates in the case of the 12 cm diameter field of view (FOV) for disc sizes greater than1cm, and greater than 3 cm for the case of the 5 cm FOV.

Real time flat panel detectors based on amorphous selenium (a-Se) have demonstrated to be the most advanced technology for direct conversion X-ray imaging in various medical applications. In continuation of real time detector development, ANRAD Corporation introduce in this paper a large size 14 inches X 14 inches active area detector built with an amorphous selenium (a-Se) converter coated on a TFT array. This new detector is a scaled up version of the 9 inches X 9 inches presented last year based on a TFT array with 150 um x 150 um pixel and a 1000 mm thick a-Se PIN structure operated at 10V/um. DQE(f=0) measurements were performed in low dose range and demonstrated to be in agreement with a linear model including 2500e of electronic noise. It is also shown that the spatial resolution (MTF) could be controlled by selenium coating process and can almost reach the theoretical limit defined by the pixel pitch. Finally, the first 14 inches X 14 inches chest image is presented.

The natural luminescence decay time of storage phosphors limits the scan speed of today's flying-spot scanners, since this 'afterglow' emission degrades the MTF along the fast-scan direction. Higher scan speeds can also decrease the total amount of signal captured at each point, that is, the read-out depth. One may reduce these problems by scanning entire lines at once, rather than points. However, this can create other problems related to stimulation and collection efficiency, system gain, sharpness, and noise. We have developed a new scanning engine containing a linear, laser-diode-based stimulation source and a light collection system with specially designed optics and multiple, linear, asymmetric CCDs. The stimulation/collection systems are housed in a compact, movable read head that scans quickly over a stationary image plate (IP). We evaluated this so-called 'ScanHead' with conventional powder (IPs) and with new CsBr:Eu²⁺ needle-based IPs.We also compared it to a conventional flying-spot scanner. Image quality results differed depending on plate type. DQE values for conventional powder IPs in the prototype scanner were comparable to today's state-of-the-art commercial systems (DQE(0)~0.25-0-27). DQE values for needle IPs were significantly better than those for powder IPs (DQE(0) ~0.58-0-60) and comparable to those quoted for flat-panel DR systems based on CsI:Tl.

We have completed the design and testing of a thermoacoustic computed tomography scanner for whole-breast imaging. We report on the technical changes in this design form our previous TCT scanner, and how these design changes have improved image quality. Improvements to the design include: greater angular coverage of TCT measurements, increased sensitivity of the ultrasound detector array, and improved delivery of radio wave energy. These improvements resulted in higher fidelity 3D reconstructions, reduced scan time, and fewer image artifacts. These improvements were documented by imaging simple, 3D phantoms, formulated from salinated agar spheres. We confirmed improvements in breast image quality by comparing images of patient volunteers taken with our previous TCT scanner and this new TCT scanner.

The objective of this work was to acquire co-registered digital tomosynthesis mammograms and 3-D breast ultrasound images of breast phantoms. A prototype mammography compression paddle was built for this application and installed on an x-ray tomosynthesis prototype system (GE). Following x-ray exposure, an automated two-dimensional ultrasound probe mover assembly is precisely positioned above the compression plate, and an attached high-frequency ultrasound transducer is scanned over the acoustically coupled phantom or localized region of interest within the phantom through computerized control. The co-ordinate system of one of the two data sets is then transformed into that of the other, and matching regions of interest on either image set can be simultaneously viewed on the x-ray and ultrasound images thus enhancing qualitative visualization, localization and characterization of regions of interest. The potentials of structured noise reduction, cyst versus solid mass differentiation and full 3-D visualization of multi-modality registered data sets in a single automated combined examination are realized for the first time. Elements of system design and required image correction algorithms will be described and phantom studies with this prototype, automated system on an anthropomorphic breast phantom will be presented.

High-intensity quasi-monochromatic x-ray irradiation from the linear plasma target is described. The plasma x-ray generator employs a high-voltage power supply, a low- impedance coaxial transmission line, a high-voltage condenser with a capacity of about 200 nF, a turbo-molecular pump, a thyristor pulse generator as a trigger device, and a flash x-ray tube. The high-voltage main condenser is charged up to 55 kV by the power supply, and the electric charges in the condenser are discharged to the tube after triggering the cathode electrode. The flash x-rays are then produced. The x-ray tube is of a demountable triode that is connected to the turbo molecular pump with a pressure of approximately 1 mPa. As the electron flows from the cathode electrode are roughly converged to the molybdenum target by the electric field in the tube, the plasma x-ray source, which consists of metal ions and electrons, forms by the target evaporating. Both the tube voltage and current displayed damped oscillations, and their peak values increased according to increases in the charging voltage. In the present work, the peak tube voltage was almost equal to the initial charging voltage of the main condenser, and the peak current was about 20 kA with a charging voltage of 55 kV. When the charging voltage was increased, the linear plasma x-ray source formed, and the characteristic x-ray intensities of K-series lines increased. The quasi- monochromatic radiography was performed by as new film-less computed radiography system.

Portal images allow the physician to visualize and quantify the position of anatomical structures within the radiation field during the treatment of cancer with radiation therapy. In this project, we exploit the presence of the low intensity bremsstrahlung photons present in the electron beam, and the high sensitivity of the new technology of flat panel based on amorphous-silicon arrays to generate images of the electron beam treatment field. This opens the possibility of routine on-line electron beam treatment verification. A large-scale array of 1024 X 1024 pixels (41 x 41 cm2) was used to acquire images from electron beams with energies from 6 to 21 MeV. For each energy, a gain correction image was acquired to compensate for the bremsstrahlung angular dependence. Several integration time factors were tested to obtain verification images within 30 monitor units, a low number for treatments with electron beams. Images of the head sections of a Rando phantom with 50 MU or less were acquired. Anatomical structures present in the phantom are clearly seen. Parameters influencing the quality of images acquired with electron beams, such as the detector integration time and the beam energy will be discussed. Examples of clinical images acquired with electron beams will also be presented.

The purpose of this study is to develop a full-field digital mammography system utilizing capillary optics. Specific aims are to identify optic properties that affect image quality and to optimize those properties in the design of a multi-element capillary array. It has been shown that polycapillary optics significantly improve mammographic image quality through increased resolution and reduced x-ray scatter. For practical clinical application much larger multi-element optics will be required. This study quantified the contributing factors to the multi-element optic MTF and investigated methods to determine optimal parameters for a practical design. Individual and a prototype multi-element array of linearly tapered optics with a common focal point were investigated. A conventional (MO/MO) mammography tube and computed radiography system were used. The system and optic MTF were measured using the angled slit method with a slit camera (10 micron slit). MTF measurements were performed with both stationary and scanned optics. Contributions to MTF included: distortion within individual optics, misalignment between optics, capillary channel size, and vibration. Measurement techniques used to identify and quantify the contributions to optic MTF included a phantom chosen specifically for polycapillary optics. This phantom provided a method for assessing the coherence among capillaries within an optic as well as the relative alignment of the optics within the array. In addition, modifications to the scanning procedure allowed for the isolation and quantification of several contributors to the system MTF. Specifically, measurements were made using a stationary optic, a scanning optic, and an optic placed at multiple locations within the imaged field of view. These techniques yielded the optic MTF, the degradation of MTF due to loss of coherence within the optic, and the degradation of MTF due to vibration of the scanning mechanism. Distortion within individual optics was, typically, quite small. However, MTF degradation resulting from twist was significant in some optics. MTF degradation due to misalignment was relatively large in the prototype triad. Modeling found that misalignment up to 50 microns reduced MTF by less than 10 percent up to 3 cycles/mm. Channel diameters of 52 microns and 85 microns reduced MTF by 9 percent to 20 percent at 5 cycles/mm and provided an optimal tradeoff between transmission and MTF. Vibration was identified as a significant degradation to MTF but can easily reduced with simple modifications. In spite of some reduced optic MTF values, system MTF has always been significantly improved - in some cases almost by the magnification ratio. These results allow for accurate modeling of optic performance and optimization of design parameters. This study demonstrates that a multi-element array can be produced with nearly optimal properties. A large area array suitable for clinical trial is feasible and is the next step in this program.

Scanning X-ray imaging systems provide significant reduction in the detected scatter radiation, cover large areas, and potentially offer high spatial resolution. Applications of one dimensional (1D) gaseous detectors and 'edge-on' illuminated silicon strip detectors for scanning slit imaging systems are currently under intensive investigation. In this work we investigate an 'edge-on' illuminated Porous Plate (PP) detector concept for applications in diagnostic X-ray imaging. As opposed to the existing X-ray imaging detectors, 'edge-on' PP detectors can provide a combination of high stopping power, high physical charge amplification, superior spatial resolution and flexible pixel shape. One common type of PP is Microchannel Plate (MCP). It has previously been investigated as a detector in surface-on illumination mode for medical X-ray imaging. However, its detection efficiency was determined to be too low for medical imaging applications. Using 'edge-on' illumination mode for MCPs with optimized structural parameters, it is possible to reach high detection efficiency. The characteristics of 'edge-on' MCP detectors are compared with the currently available X-ray imaging detectors. Possible use of other PP materials such as Porous Dielectrics, Microspheric Plates, a-Si based MCPs, Micro columnar and Micro granular Dielectrics are discussed. An 'edge-on' illuminated MCP detector for scanning X-ray imaging system is being developed in our laboratory. The details of this system and the read out electronics will be described.

This paper describes the noise properties of a Se-based flat-panel X-ray detector (FPD) with 128-channel CMOS readout integrated circuits (ICs). The detector used in this study is a 23 cm X 23 cm direct detection FPD utilizing a thick a-Se film. The TFT array format is 1536 X 1536 pixels with a pixel pitch of 150 micrometers . The key component of this FPD is a newly developed CMOS readout IC, which integrates 128 readout channels. Each channel consists of a charge-sensitive amplifier (CSA), a low-pass filter (LPF), a correlated double sampling circuit (CDS), and an analog sample-hold circuit. To evaluate the noise properties, theoretical noise models, including a precise expression for the CDS, were constructed and compared with the measured results, which were obtained by analyzing the noise power spectra (NPS) of the dark images taken without X-ray exposure. In addition, the detective quantum efficiency (DQE) was measured at several fluoroscopic dose ranges to demonstrate the overall signal-to-noise performance. As a result, a total readout electronic noise of 2,100 e- rms was obtained, in which thermal noise and 1/f noise of the amplifier was dominant. The DQE(0) measured was about 0.5 at 1 (mu) R.

Single crystals of mercuric iodide have been studied for many years for nuclear detectors. We have investigated the use of x-ray photoconductive polycrystalline mercuric iodide coatings on amorphous silicon flat panel thin film transistor (TFT) arrays as x-ray detectors for radiographic and fluoroscopic applications in medical imaging. The mercuric iodide coatings were vacuum deposited by Physical Vapor Deposition (PVD). This coating technology is capable of being scaled up to sizes required in common medical imaging applications. Coatings were deposited on 4 inches X 4 inches TFT arrays for imaging performance evaluation and also on conductive-coated glass substrates for measurements of x-ray sensitivity, dark current and image lag. The TFT arrays used included pixel pitch dimensions of both 100 and 139 microns. Coating thickness between 150 microns and 250 microns were tested in the 25 kVp-100 kVp x-ray energy range utilizing exposures typical for both fluoroscopic, and radiographic imaging. X-ray sensitivities measured for the mercuric iodide samples and coated TFT detectors were superior to any published results for competitive materials (up to 7100 ke/mR/pixel for 100 micron pixels). It is believed that this higher sensitivity, can result in fluoroscopic imaging signal levels high enough to overshadow electronic noise. Image lag characteristics appear adequate for fluoroscopic rates. Resolution tests on resolution target phantoms showed that resolution is limited to the Nyquist frequency for the 139 micron pixel detectors. The ability to operate at low voltages gives adequate dark currents for most applications and allows low voltage electronics designs. Mercuric Iodide coated TFT arrays were found to be outstanding candidates for direct digital radiographic detectors for both static and dynamic (fluoroscopic) applications. Their high x-ray sensitivity, high resolution, low dark current, low voltage operation, and good lag characteristics provide a unique combination of desirable imaging performance parameters.

Wide-band-gap semiconductor detectors are recently in spotlight for various applications because of their good performances, such as the high energy resolution, the compactness in array geometry, and the room temperature operation. The performance of these detectors, for example, CdZnTe, is mainly limited by the charge transport properties. Especially, the dispersive nature of trapping and detrapping process affects on the detector performance resulting in random fluctuations in the current flowing. Based on the spectroscopic measurement, in this study, a simple analytical model is developed to investigate the charge transport characteristics for planar semiconductor detectors, especially for CdZnTe of m-i-m (metal-intrinsic-metal) diode structure. The model can take the input variables of material properties, as well as the operation parameters, such as the applied bias voltage, the pulse shaping time, the incident direction and the energy of gamma-rays. The measured gamma spectra from CdZnTe for Co57 showed excellent agreement with the simulation results from our model, and the parameters governing detector performance were analyzed. We expect that this model will be very useful to understand the charge transport mechanism in the wide-band-gap semiconductor detectors, and to optimally design the detector geometry for various applications.

The solid-state detector(SSD) for X-CT consists of photodiode coupled to CdWO_4$(CWO. It is important to maximize the light collection in respect of a patient's dose, radiation effect and X-ray efficiency. The factors affecting the light collection efficiency are analyzed and optimized by using experimental data and appropriate simulation code. Quantum nomogram is used to investigate the signal propagation characteristics of optimally designed solid-state detector and to ensure at which stage quantum sink occurs. This paper shows that the part of SSD, the CWO of treatment with ground top/ground side yields higher quanta than that of ground top/polish side, which is different from the result of previous studies. We also shows that optimum thickness of SiN passivation and p-layer is 0.12mm and 0.1mm, respectively. From the quantum nomogram calculated for optimal design, it is predicted that the most serious signal degradation occurs at the photodiode.

In this work we exploit the advantages of using a bi-chromatic X-rays source coupled with a single photon counting pixel detector to perform a feasibility study for dual energy mammography. This technique allows enhancing the contrast between different breast tissues by composing two images acquired at two different energies. The high and low energy images have been acquired by a single X-ray shot. The bi-chromatic beam has been produced per diffraction of polychromatic photons by a monochromator crystal. The imaging system is based on a single photon counting silicon pixel detector. The data read-out is performed by a VLSI Integrated Circuit bump-bonded to the sensor. The energy threshold of each electronics channel can be individually trimmed. We set the threshold of one pixel below 16 keV while the threshold of the neighboring pixel between 16 and 32 keV. With a single exposure the information from both energies is recorded. After separation between low and high threshold pixels, we obtained two independent images. We acquired radiographs of phantoms made of three different materials. Appling a dual energy algorithm, we obtained synthesized images where any of the three materials is removed from the radiograph, enhancing the contrast between the two remaining.

In this study, we investigated which photon energy results in the lowest patient does when the image contrast to nosie ratio (CNR) is kept constant. This optimum photon energy was obtained for a range of patient sizes, as well as for the detection of lesions with atomic number ranging from 6.5 to 53. Mono-energetic photons from 20 keV to 140 keV were investigated with an x-ray detector that had 100 percent quantum detection efficiency. Patients were modeled as slabs of water with a thickness that ranged from 5 to 30 cm. Image contrast was computed from the x-ray attenuation of small lesions consisting of low Z materials as well as higher Z materials. Relative values of the CNR as a function of the x-ray photon energy were obtained by assuming that the image noise was proportional to the square root of the incident number of x-ray photons under scatter free conditions. The energy imparted to the patient as a function of photon energy was obtained using published data of the absorbed percentage of the incident energy fluence. For each patient thickness and lesion composition, the CNR was kept constant by appropriate adjustment of the x-ray beam intensity, and the corresponding x-ray photon energy that resulted in the minimum patient dose was determined. The optimum photon energy for detection ga low Z lesion increased monotonically from approximately 62 keV for small patients to approximately 78 keV for large patients. For high Z lesions, optimum photon energies were approximately 34 keV for small patients, and increased to approximately 40 keV for large patients. The optimum photon energy was found to vary by about a factor of two over the range of patient thickness investigated. The optimum photon energy also varied by a factor of two for the range of detection tasks investigated.

X-ray detector systems can be characterized by their measured or estimated detective quantum efficiency (DQE). Assessment of DQE includes a measurement of the modulation transfer function (MTF) and the normalized noise power spectrum (NNPS). The incoming X-ray quantum flux has to be estimated. In this paper, the influence of the different possibilities regarding the measurement methods and phantoms, the X-ray quantum flux estimation models and the exposure geometry on the DQE of a full field digital mammography detector is assessed. Physical models were used to fit MTF measurements from bar-pattern and edge phantoms. The NNPS was calculated by 2D-FFT on a large number of flat-field subimages. The flux was calculated using anode spectra models (Boone, 1997) and attenuation data (NIST). We compared the influence of scattered radiation MTF calculations of both phantoms were similar. The edge method is preferred for practical reasons. NNPS data were similar to 1D synthetic-slit measurements. DQE data compared well with literature. Different exposure geometry conditions (with scattered radiation) showed similar results but a siginificantly lower DQE than in absence of scattered radiation. DQE assessment is feasible using normal exposure conditions, an edge phantom and calculated estimations of the flux.

The LODOX (Low Dose X-ray) Scanner, created by De Beers, is currently being clinically tested at the Trauma Unit of Groote Schuur Hospital and the University of Cape Town in South Africa. High quality images with exceedingly low radiation suggest that the technology may also be used to identify breast cancer lesions and microcalcifications. The measured LODOX modulation transfer function averages 6 percent at 10 cycles per millimeter, while the detected quantum efficiency is approximately 25 percent at 1 cycle per millimeter. The mean glandular doses calculated for a breast thickness of 4 cm at various intensities -- ranging from 0.022 rad at 70mAs to 0.043 rad at 125mAs -- were approximately 10 times less than the value designated by the American College of Radiology (0.3 rad per breast image). At 40kV, LODOX exhibits an average half value layer of 1.59 mm of Al (compared to 0.3 to 0.4 mm recommended for mammography), illustrating the unfavorable higher penetration of LODOX X-rays. The extremely low radiation dose delivered by the LODOX suggests that the technology would be feasible for detecting and diagnosing cancers in the sensitive tissue of the breast, once adjustments to X-ray range and beam hardness had been accomplished.

Previously we used a simple 2-D model to evaluate the imaging performance of a digital radiographic system while varying input parameters such as transducer blur and signal size. We extend this work using a realistic phosphor simulation to explore the effect of the incident x-ray spectrum and the depth dependence of the point spread function and optical collection efficiency. Initially we investigate one Swank screen type representative of modern powder phosphor design. Images resulting from these simulations are used to get an estimate of the impact of these factors on lesion detectability. Results show that the simple 2-D model gives optimistic estimates of detectability.

The purpose of this study is to investigate the influence of the scattered x rays on the signal sharpness on the radiographs produced by using a computed radiography (CR) system by measuring the spatial frequency spectra of the signal image. By using a 0.1 mm slit on the polymethyl methacrylate (PMMA) for thicknesses of 0.5 cm to 20.5 cm, the slit images were acquired as a signal by use of imaging plates at tube voltages of 50 kV to 120 kV. The relative exposure profiles for the slit images were Fourier transformed to obtain the spatial frequency spectra. For comparison of the frequency spectra with and without the scattered x rays, we defined the scattered x-ray influence factor (SIF) representing the magnitude of the influence of the scattered x rays on the spatial frequency spectra of the signal image. To investigate the contribution of the primary and scatter components to the degradation of the signal sharpness, we proposed a method for separating the spatial frequency spectrum of the signal image into the primary and scatter components. By obtaining the SIF, we found that, for very lower frequencies (less than about 0.3 mm^-1), the shape of the spatial frequency spectra of the signal image depends on the scattered x rays, but, for higher frequencies, hardly depends. As a result of the separation of the frequency spectra of the signal image, we found that the contribution of the scatter component for very lower frequencies (less than about 0.2 mm^-1) to the total spectrum of the signal image was not negligible and became greater as the scattering material thickness and the tube voltage increased. On the contrary, for higher frequencies, the primary component was dominant compared with the scatter component for all thicknesses and tube voltages.

Terahertz (THz) radiation has a frequency of the order of 10¹²12 Hz. This lies between the infrared and microwave regions of the electromagnetic spectrum; a section labeled the 'THz gap'. Infrared and microwave radiation is used in the medical field; research is underway for an application for THz radiation. At present no formal safety analysis of a THz pulsed imaging (TPI) system has been performed. This will be necessary for future in vivo studies. The radiation is delivered in a train of femtosecond pulses. International guidelines on exposure to non-ionizing radiation, and supporting literature, were reviewed to determine the Maximum Permissible Exposure (MPE) for radiation of this range of wavelengths, both for a single pulse and continuous wave exposure. Two methods of deriving the MPE were identified. Current guidelines for infrared and microwave regions of the electromagnetic spectrum incorporate the THz region. Using conservative parameter estimation an MPE per pulse, over the area of the beam, of 94 W was calculated. At present THz pulsed imaging systems produce pulses of power approximately 1 mW; this lies within the limit calculated using the published guidelines. There are, however, areas requiring further investigation before the technique becomes routine in clinical practice.

Direct conversion imaging panels have pixels that need to be protected from excessive voltage across the FET switch resulting from excess X-ray exposure. Protection can be achieved by placing a dielectric layer between the charge generator and the high voltage bias electrode. The charges that accumulate at the interface create a counter electric field so that the electron-hole pairs recombine in the selenium, thus limiting the charge accumulated at the corresponding pixels. A variety of means have been used to neutralize the charges accumulated at the interface, a so called erase step. We have demonstrated a novel charge transport layer (CTL) that protects the FETs from transient overexposure and eliminates the need for a separate erase step. By selection of the time constant of the CTL, the panel can be used for static and dynamic imaging. The operating principle of this new structure and the performance of prototypes compared to conventional will be reported.

This paper describes the results of the comparison of 5 film digitizers designed for medical application to find an appropriate film digitizer for a project involving display of mammograms on a high performance CRT. Two of the digitizers were of the laser-scanning type while 3 were based on CCD technology. The comparison included estimates of linearity, uniformity, signal-to-noise ratio and spatial resolution in terms of the Modulation Transfer Function (MTF). The digitizers investigated show vastly different results. No system excelled in all performance parameters. A laser scanning digitizer was chosen for the planned display project because (1) it had the best linearity (down to OD 4); (2) it had the highest SNR at OD 4; (3) it had a comparatively good MTF.

The formation and development of plaques in the arterial wall is a direct consequence of atherosclerosis. The composition of a plaque is of particular interest as it is thought to be an important indicator of vulnerability, or risk of rupture and thrombosis. Current diagnostic methods do not yet have the ability to fully characterize plaque composition. Coherent-scatter imaging, a technique being developed in our laboratory, produces images based on the low-angle scattering properties of tissue. As these properties depend on molecular structure, material-specific maps of the different components in a tissue can be created. Material-specific images were produced for an atherosclerotic carotid artery. The image distributions of fatty and calcified deposits agreed with visual examination of the specimen. Preliminary results indicate that fat and calcifications, two typical plaque constituents, can be identified and distinguished from the undiseased vessel wall using coherently scattered x rays.

Phase contrast x-ray imaging has been studied intensively using x-rays from synchrotron radiation and micro-focus x- ray rubes. However, these studies have revealed the difficulty of this technique's application to practical medical imaging. We have created a phase contrast imaging technique using practical x-ray tubes with small focal spot sizes. In a previous study, we identified the radiographic conditions for phase contrast magnification mammography with screen-film systems, where the edge effect due to phase contrast outweighs the geometrical unsharpness introduced by the 0.1mm-focal spot of a molybdenum-anode x-ray tube. In the present study, phase contrast images of a hand and a chest phantoms were obtained using computed radiography and a geometrical magnification of two using a tungsten-anode x- ray tube with a 0.1mm-focal spot. The digital images were printed so that objects were reproduced at their original size. Compared to conventional digital contact images obtained at the same x-ray dose, the life-sized digital phase contrast x-ray images displayed improved sharpness and resolution, and reduced grain. We describe the empirical results of digital phase contrast imaging, and discuss these performance improvements through simulation including x-ray scatter.

We have examined the characteristic of a flat-panel detector used for our cone-beam CT scanner. We calculated detection efficiency, the modulation transfer function and the noise characteristic of flat-panel detector by the Monte Carlo method, and compared the results with experiment data. From these data, we estimate the relation between the patient dose and signal-to-noise ratio of projection data of CT image. Furthermore, to reduce the patient dose, we have examined the effect of removing the noise from projection degraded by quantum mottle, by the wavelet analysis. Our preliminary results show that de-noising of projection data with wavelet analysis has an effect to reduce the patient dose to less than 1/10, without decreasing the quality of CT image.

Region-of-interest (ROI) fluoroscopy has previously been investigated as a method to reduce x-ray exposure to the patient and the operator. This ROI fluoroscopy technique allows the operator to arbitrarily determine the shape, size, and location of the ROI. A device was used to generate patient specific x-ray beam filters. The device is comprised of 18 step-motors that control a 16 X 16 matrix of pistons to form the filter from a deformable attenuating material. Patient exposure reductions were measured to be 84 percent for a 65 kVp beam. Operator exposure reduction was measured to be 69 percent. Due to the reduced x-ray scatter, image contrast was improved by 23 percent inside the ROI. The reduced gray level in the periphery was corrected using an experimentally determined compensation ratio. A running average interpolation technique was used to eliminate the artifacts from the ROI edge. As expected, the final corrected images show increased noise in the periphery. However, the anatomical structures in the periphery could still be visualized. This arbitrary shaped region of interest fluoroscopic technique was shown to be effective in terms of its ability to reduce patient and operator exposure without significant reduction in image quality. The ability to define an arbitrary shaped ROI should make the technique more clinically feasible.

We propose a new technique for the correction of eddy- current effects in the k-space domain using the non- diffusion weighted image as a reference. To avoid affecting the contrast properties constant masks were derived form the original images by thresholding and averaging. A 2D polynom was calculated in the Fourier domain that describes a distribution of phase gradients which contains the information how to correct the misregistered pixels. The technique does not need any field map or additional calibration scans within the measurement sequence. The distortions in the fiber tracts, the bias and the variance in the anisotropy values were minimized while preserving an acceptable signal to noise ratio. The algorithm is computationally not very intensive and easy to implement.

It is commonly thought that improvements in image S/N translate directly to improvements in fMRI sensitivity. This study demonstrates that improvements in image S/N by means of increased field strength, and the use of surface coils, does not translate to similar gains in temporal S/N, due to physiological noise. An analysis of the S/N dependence on signal strength for both the imaginary and real noise components is presented. The real noise component, which is lower than the imaginary component, is a significant contributor to temporal variations in single-shot fMRI procedures. The imaginary component becomes an important contributor to time series fluctuations in multi-shot or 3D techniques. In the experiments presented here, the relationship between image and temporal S/N is examined by modulating the signal strength by means of echo time stepping, field strength modulation, and RF coil comparison. The spatial and temporal noise contributions of the resting brain are characterized by comparing phantom, subject, and thermal noise measurements. The sensitivity of both spiral and EPI single-shot acquisition methods to physiologic and systematic noise is characterized. The results suggest that the fMRI sensitivity plateaus as image S/N is increased unless physiological noise is filtered out.

The purpose of this study is to develop an x-ray detector system for multislice CTs with 0.5 slice thickness and 0.5- second revolution, and to evaluate clinical advantages. The detector adopts a newly developed ceramic scintillator having a high light-output efficiency as well as a high light-transmittance. By employing high precision machining technology to fabricate the detector, it is not necessary to mask the gap from incident x-ray in the longitudinal direction between neighboring cells. The scintillator properties and the geometry of the detector provide a high x-ray-to-light conversion efficiency, and realize a good low-contrast detectability for CTs. The detector structure allows 0.5 mm slice thickness. The afterglow intensity of the scintaillator is less than 0.1 percent at 3 ms, facilitating 1800 views per second for 0.5-second revolution. This detector improves the performance of CT scanners.

Efficient and high performance magnetic field gradient and active shim coils have always been desirable in today's molecular imaging applications using magnetic resonance imaging (MRI). Various MR techniques, such as high spatial resolution single shot echo-planar imaging (EPI), MR diffusion imaging and MR microscopy as well as cardiac imaging, do require efficient and high performance coils in order to be feasible for imaging small animals. Exact solutions for current density on a spherical surface were explored for generating various orders of field harmonics within a spherical volume. Such compact expressions can be very useful for further coil optimization. After carrying through tedious mathematical expansions and algebraic reductions, exact surface current solutions for spherical field coils have been obtained for various zonal orders and validated. In the paper, the current density on a spherical surface is expanded in associated Legendre polynomial series of order 1, the resulting vector potential of magnetic field was evaluated, and the analytical expressions for both magnetic field and stored magnetic energy were obtained. Then, the continuous current density expressions for various orders of harmonics were identified analytically. Applying the stream function technique, the discrete current wire model can be generated from that of the continuous. The analytical predictions were in an excellent agreement with numerical results. These exact analytical solutions offer a useful mathematical relationship for future magnetic field coil design involving spherical geometry.

It has been known that the actual slice profile of 2D MRI is often far form that of a perfect rectangular shape especially when the flip angle is large. This can be the source of error and image artifact in many MRI experiments. To carefully study this imperfection in RF excitation for various numerically optimized RF pulses, we have implemented an efficient numerical algorithm for simulating the evolution of the magnetization during a MR experiment. The scheme solves the Bloch equation via a numerical procedure that involves only the successive matrix multiplications, which are initialized by the values of RF amplitude and the frequency offset due to the companioning magnetic gradient field. It permits the consideration of the magnetization relaxation processes. The actual slice profile was numerically simulated by solving the Bloch equation for a given RF pulse shape and slice selection field gradient. The simulation results shows that the slice profile is often far from perfect for many commonly used frequency selective RF pulses. As a result of this, the apparent image intensity is an integrated signal over the slice envelope for the transverse magnetization for each pixel, which is an averaged contribution over the entire slice thickness, and can be significantly different than that predicted by the ideal case. As expected, this discrepancy worsens when the flip angle exceeds the flip angle for which the RF pulse was optimized. Furthermore, in the routine diagnostic imaging, because of the non-uniform excitation angle along the slice profile, it generates non-uniform steady state profile and can even alter the resulting image contrast.

Developing and optimizing an x-ray scatter control and reduction technique is one big challenge for cone beam volume computed tomography (CBVCT) because CBVCT will be much less immune to scatter than fan-beam CT. X-ray Scatter reduces image contrast, increases image noise and introduces reconstruction error into CBVCT. To reduce scatter interference, a practical algorithm that is based upon the beam stop array technique and image sequence processing has been developed on a flat panel detector-based CBVCT prototype scanner. This paper presents beam stop array-based scatter correction algorithm and the evaluation results through phantom studies. The results indicate that the beam stop array-based scatter correction algorithm is practical and effective to reduce and correct x-ray scatter for a CBVCT imaging task.

One of factors that more affect the radiographic image quality is the scatter radiation produced by interaction between the x-ray and the radiographed object. Recently the Computer Aided Diagnosis (CAD) Systems are coming to aid the detection of breast small details. Nevertheless, we not sure how much the scatter radiation decrease the efficiency of this systems. This work presents a model in order to quantify the scatter radiation and find it relation between CAD's results used for the microcalcification detection. We simulated scatter photons that reaches the film and we added it to the mammography image. The new images were processed and the alterations of the CAD's results were analyzed. The information loss to breast composed by 80 percent adipose tissue was 0,0561 per each centimeter increased in the breast's thickness. We calculated these same data considering a proportion variation of adipose tissue and considering the breast composition of 90 percent and 70 percent the loss it would be of 0.0504 and 0.07559 per increased cm, respectively. We can increase the wanted scattered radiation to any image with its own characteristics and analyze the disturbances that it can bring to the visual inspection or the automatic detection (CAD system) efficiently.

This paper presents the development of a prototype Tactile Mapping Device (TMD) system comprised mainly of a tactile sensor array probe (TSAP), a 3-D camera, and a force/torque sensor, which can provide the means to produce tactile maps of the breast lumps during a breast palpation. Focusing on the key tactile topology features for breast palpation such as spatial location, size/shape of the detected lesion, and the force levels used to demonstrate the palpable abnormalities, these maps can record the results of clinical breast examination with a set of pressure distribution profiles and force sensor measurements due to detected lesion. By combining the knowledge of vision based, neural networks and tactile sensing technology; the TMD is integrated for the investigation of soft tissue interaction with tactile/force sensor, where the hard inclusion (breast cancer) can be characterized through neural network learning capability, instead of using simplified complex biomechanics model with many heuristic assumptions. These maps will serve as an objective documentation of palpable lesions for future comparative examinations. Preliminary results of simulated experiments and limited pre-clinical evaluations of the TMD prototype have tested this hypothesis and provided solid promising data showing the feasibility of the TMD in real clinical applications.

4D-CT scanner is a high-speed cone-beam CT with good low- contrast detectability. In a single rotation, a voxel data set with higher spatial resolution over a wide z-axis can be obtained. Using continuous rotation, temporarily continuous voxel data sets and 3D dynamic images can be acquired. We have developed a 4D-CT prototype system using a large area 2D detector with high speed rotating gantry. The key technologies for 4D-CT scanner are a large area 2D detector, high-speed and continuous rotating gantry with high-speed data transfer. The full size 2D detector for prototype system has 912 channels, 256 detector rows for 0.5 mm slice thickness, and the sampling speed is 900 view/s. The rotation speed of the gantry is 1.0 s/rotation, and the speed of the data transfer system must be more than 5Gbps. The data transfer system consists of laser diode and photo diode. We have got real-time volume images of crayfish, moving jaw and breathing lung. We will start clinical research at next stage, and continue to improve the image quality.

Currently employed megavoltage photon detector systems in radiotherapy suffer from their low detection efficiency. This paper presents a new, more efficient approach. A dense converter material is placed along the direction of the photon beam together with the interspersed active medium. The high-energy electrons, generated in the converter material, create a detectable signal in the active medium. The characteristics of single detector element prototypes consisting of different brass tubes filled with either air or xenon were investigated. The significance of the wall thickness and material of a two-dimensional array of such elements was analyzed by Monte Carlo studies. Saturation curves and the signal as a function of the gas pressure illustrate the equivalence of the prototypes to conventional ionization chambers. A high-density gas like xenon at a high pressure is required to get a sufficiently large signal-to-noise ratio. Acquired MVCT slice images revealed contrasts of - 2.4 percent at a dose of 40 cGy. Monte Carlo calculations suggest an optimal wall thickness of a few hundred micrometers for an element width of 1.5 mm. The superiority of high density and high atomic number materials like tungsten in terms of the amount of primary signal recorded and the extent of crosstalk is demonstrated.

A dynamic flat detector with optimized overall system design has been developed by Trixell together with its partners. This detector enables the application of flat detector technology in cardiac procedures, including difficult applications like low-dose fluoroscopy, and later extending to digital subtraction angiography and high-speed bi-plane. The key parameters for the optimized design are the high DQE over a large dose range, the excellent linearity and the temporal behavior of the detector. High DQE performance could be realized best by using a thick CsI layer for high absorption. This guarantees a salient MTF of 65 percent at 1 lp/mm and 37 percent at 2 lp/mm even for thick layers because the detrimental effects of the dead layer are avoided. The well known temporal artifacts problems associated with a:Si-CsI technology have been solved using flashes of light in combination with a reset sequence at the beginning of every frame, which keeps the traps in the a-Si photodiodes constantly filled. Switching dose levels therefore leads only to a marginal change in the trap occupation, resulting in a low temporal cross-talk. Experimental data for the mentioned applications are demonstrated.

A physical characterization for different modern digital detectors (transparent imaging plate, CsI flat panel detector) as well as of an analogue screen-film system, a selenium based system and a 'normal CR system are presented. The detectable information in radiographs produced with digital systems is strongly dependent on the image processing used. The effect of an optimization of this processing corresponding to the physical parameters of different detectors was evaluated. To do so, VGA1, 2 and MS-ROC3 have been used. All optimizations led to 'better' images concerning the information content which can be used by the radiologist. For the flat-panel detector it led to a reduction of images which must be manually postprocessed and even of repeated radiographs. This was achieved by using an adaptive autowindow algorithm which is able to distinguish between anatomical structures and implants. For another detector the detection of unsharp lung nodules could be improved by far. Information can not be added by image processing, it is only possible to make it better visible. To do so, the optimizing of the processing for each kind of detector is much more important than an optimization concerning different organ programs.

Electromagnetic imaging has been used over the past few years for detecting and monitoring the treatment of leukemia. In previous works, TM polarized waves were used and the problem was studied by optimization method. Here we study the problem for TE polarized waves using the same type of method and propose an iterative method. Our numerical results for determining the values of the index of refraction inside inhomogeinities are obtained by using the optimization method.