Characterization of breast calcification types using dual energy x-ray method.

Calcifications are products of mineralization whose presence is usually associated with pathological conditions. The minerals mostly seen in several diseases are calcium oxalate (CaC2O4), calcium carbonate (CaCO3) and hydroxyapatite (HAp). Up to date, there is no in vivo method that could discriminate between minerals. To this aim, a dual energy x-ray method was developed in the present study. An analytical model was implemented for the determination of the Calcium/Phosphorus mass ratio ([Formula: see text]). The simulation was carried out using monoenergetic and polyenergetic x-rays and various calcification thicknesses (100-1000 [Formula: see text]) and types (CaC2O4, CaCO3, HAp). The experimental evaluation of the method was performed using the optimized irradiation conditions obtained from the simulation study. X-ray tubes, combined with energy dispersive and energy integrating (imaging) detectors, were used for the determination of the [Formula: see text] in phantoms of different mineral types and thicknesses. Based on the results of the experimental procedure, statistical significant difference was observed between the different types of minerals when calcification thicknesses were 300 [Formula: see text] or higher.

Three-dimensional cascaded system analysis of a 50 µm pixel pitch wafer-scale CMOS active pixel sensor x-ray detector for digital breast tomosynthesis.

High-resolution, low-noise x-ray detectors based on the complementary metal-oxide-semiconductor (CMOS) active pixel sensor (APS) technology have been developed and proposed for digital breast tomosynthesis (DBT). In this study, we evaluated the three-dimensional (3D) imaging performance of a 50 µm pixel pitch CMOS APS x-ray detector named DynAMITe (Dynamic Range Adjustable for Medical Imaging Technology). The two-dimensional (2D) angle-dependent modulation transfer function (MTF), normalized noise power spectrum (NNPS), and detective quantum efficiency (DQE) were experimentally characterized and modeled using the cascaded system analysis at oblique incident angles up to 30°. The cascaded system model was extended to the 3D spatial frequency space in combination with the filtered back-projection (FBP) reconstruction method to calculate the 3D and in-plane MTF, NNPS and DQE parameters. The results demonstrate that the beam obliquity blurs the 2D MTF and DQE in the high spatial frequency range. However, this effect can be eliminated after FBP image reconstruction. In addition, impacts of the image acquisition geometry and detector parameters were evaluated using the 3D cascaded system analysis for DBT. The result shows that a wider projection angle range (e.g. ±30°) improves the low spatial frequency (below 5 mm-1) performance of the CMOS APS detector. In addition, to maintain a high spatial resolution for DBT, a focal spot size of smaller than 0.3 mm should be used. Theoretical analysis suggests that a pixelated scintillator in combination with the 50 µm pixel pitch CMOS APS detector could further improve the 3D image resolution. Finally, the 3D imaging performance of the CMOS APS and an indirect amorphous silicon (a-Si:H) thin-film transistor (TFT) passive pixel sensor (PPS) detector was simulated and compared.

Evaluation of the 3D spatial distribution of the Calcium/Phosphorus ratio in bone using computed-tomography dual-energy analysis.

PURPOSE: The Calcium/Phosphorus (Ca/P) ratio was shown to vary between healthy bones and bones with osteoporotic symptoms. The relation of the Ca/P ratio to bone quality remains under investigation. To study this relation and determine if the ratio can be used to predict bone fractures, a non-invasive 3D imaging technique is required. The first aim of this study was to test the effectiveness of a computed-tomography dual-energy analysis (CT-DEA) technique developed to assess the Ca/P ratio in bone apatite (collagen-free bone) in identifying differences between healthy and inflammation-mediated osteoporotic (IMO) bones. The second aim was to extend the above technique for its application to a more complex structure, intact bone, that could potentially lead to clinical use. METHODS: For the first aim, healthy and IMO rabbit cortical bone apatite samples were assessed. For the second aim, some changes were made to the technique, which was applied to healthy and IMO intact bone samples. RESULTS: Statistically significant differences between healthy and IMO bone apatite were found for the bulk Ca/P ratio, low Ca/P ratio proportion and interconnected low Ca/P ratio proportion. For the intact bone samples, the bulk Ca/P ratio was found to be significantly different between healthy and IMO. CONCLUSIONS: Results show that the CT-DEA technique can be used to identify differences in the Ca/P ratio between healthy and osteoporotic, in both bone apatite and intact bone. With quantitative imaging becoming an increasingly important advancement in medical imaging, CT-DEA for bone decomposition could potentially have several applications.

50 µm pixel pitch wafer-scale CMOS active pixel sensor x-ray detector for digital breast tomosynthesis.

Wafer-scale CMOS active pixel sensors (APSs) have been developed recently for x-ray imaging applications. The small pixel pitch and low noise are very promising properties for medical imaging applications such as digital breast tomosynthesis (DBT). In this work, we evaluated experimentally and through modeling the imaging properties of a 50 µm pixel pitch CMOS APS x-ray detector named DynAMITe (Dynamic Range Adjustable for Medical Imaging Technology). A modified cascaded system model was developed for CMOS APS x-ray detectors by taking into account the device nonlinear signal and noise properties. The imaging properties such as modulation transfer function (MTF), noise power spectrum (NPS), and detective quantum efficiency (DQE) were extracted from both measurements and the nonlinear cascaded system analysis. The results show that the DynAMITe x-ray detector achieves a high spatial resolution of 10 mm(-1) and a DQE of around 0.5 at spatial frequencies <1 mm(-1). In addition, the modeling results were used to calculate the image signal-to-noise ratio (SNRi) of microcalcifications at various mean glandular dose (MGD). For an average breast (5 cm thickness, 50% glandular fraction), 165 µm microcalcifications can be distinguished at a MGD of 27% lower than the clinical value (~1.3 mGy). To detect 100 µm microcalcifications, further optimizations of the CMOS APS x-ray detector, image aquisition geometry and image reconstruction techniques should be considered.

Performance of a novel wafer scale CMOS active pixel sensor for bio-medical imaging.

Recently CMOS active pixels sensors (APSs) have become a valuable alternative to amorphous silicon and selenium flat panel imagers (FPIs) in bio-medical imaging applications. CMOS APSs can now be scaled up to the standard 20 cm diameter wafer size by means of a reticle stitching block process. However, despite wafer scale CMOS APS being monolithic, sources of non-uniformity of response and regional variations can persist representing a significant challenge for wafer scale sensor response. Non-uniformity of stitched sensors can arise from a number of factors related to the manufacturing process, including variation of amplification, variation between readout components, wafer defects and process variations across the wafer due to manufacturing processes. This paper reports on an investigation into the spatial non-uniformity and regional variations of a wafer scale stitched CMOS APS. For the first time a per-pixel analysis of the electro-optical performance of a wafer CMOS APS is presented, to address inhomogeneity issues arising from the stitching techniques used to manufacture wafer scale sensors. A complete model of the signal generation in the pixel array has been provided and proved capable of accounting for noise and gain variations across the pixel array. This novel analysis leads to readout noise and conversion gain being evaluated at pixel level, stitching block level and in regions of interest, resulting in a coefficient of variation ⩽1.9%. The uniformity of the image quality performance has been further investigated in a typical x-ray application, i.e. mammography, showing a uniformity in terms of CNR among the highest when compared with mammography detectors commonly used in clinical practice. Finally, in order to compare the detection capability of this novel APS with the technology currently used (i.e. FPIs), theoretical evaluation of the detection quantum efficiency (DQE) at zero-frequency has been performed, resulting in a higher DQE for this detector compared to FPIs. Optical characterization, x-ray contrast measurements and theoretical DQE evaluation suggest that a trade off can be found between the need of a large imaging area and the requirement of a uniform imaging performance, making the DynAMITe large area CMOS APS suitable for a range of bio-medical applications.

Medicine, material science and security: the versatility of the coded-aperture approach.

The principal limitation to the widespread deployment of X-ray phase imaging in a variety of applications is probably versatility. A versatile X-ray phase imaging system must be able to work with polychromatic and non-microfocus sources (for example, those currently used in medical and industrial applications), have physical dimensions sufficiently large to accommodate samples of interest, be insensitive to environmental disturbances (such as vibrations and temperature variations), require only simple system set-up and maintenance, and be able to perform quantitative imaging. The coded-aperture technique, based upon the edge illumination principle, satisfies each of these criteria. To date, we have applied the technique to mammography, materials science, small-animal imaging, non-destructive testing and security. In this paper, we outline the theory of coded-aperture phase imaging and show an example of how the technique may be applied to imaging samples with a practically important scale.

A new technique for the assessment of the 3D spatial distribution of the calcium/phosphorus ratio in bone apatite.

The value and distribution of calcium/phosphorus (Ca/P) ratio in bone vary between healthy and osteoporotic bone. The purpose of this study was the development of a technique for the assessment of the 3D spatial distribution of Ca/P ratio in bone apatite, which could eventually be implemented through a conventional computed tomography (CT) system. A three-material mass-fraction decomposition CT dual energy analysis was optimized. The technique was validated using ten bone phantoms of different, known Ca/P ratio. Their measured average Ca/P ratio showed a mean/maximum deviation from the expected Ca/P ratio of 0.24/0.35. Additionally, three healthy and three inflammation-mediated osteoporotic (IMO) collagen-free rabbit tibia bone samples were assessed, providing promising preliminary results on real bone tissue. The average Ca/P ratios in all IMO samples (1.64-1.65) were found to be lower than in healthy samples (1.67-1.68). Osteoporotic regions in IMO samples were located using Ca/P ratio colour maps and Ca/P ratio values as low as 1.40 ± 0.26 were found. The low Ca/P ratio volume proportion in IMO samples (12.8%-13.9%) was found to be higher than in healthy (5.8%-8.3%) samples. A region growing technique showed a higher homogeneity of Ca/P ratio in healthy than in IMO bone samples.

Method for automatization of the alignment of a laboratory based x-ray phase contrast edge illumination system.

Here we present a general alignment algorithm for an edge illumination x-ray phase contrast imaging system, which is used with the laboratory systems developed at UCL. It has the flexibility to be used with all current mask designs, and could also be applied to future synchrotron based systems. The algorithm has proved to be robust experimentally, and can be used for the automatization of future commercial systems through automatic alignment and alignment correction.

Low-dose phase contrast mammography with conventional x-ray sources.

PURPOSE: To provide an x-ray phase contrast imaging (XPCI) method working with conventional sources that could be readily translated into clinical practice. XPCI shows potential in synchrotron studies but attempts at translating it for use with conventional sources are subject to limitations in terms of field of view, stability, exposure time, and possibly most importantly, delivered dose. METHODS: Following the adaptation of our "edge-illumination" XPCI technique for use with conventional x-ray sources through the use of x-ray masks, the authors have further modified the design of such masks to allow further reducing the dose delivered to the sample without affecting the phase sensitivity of the method. RESULTS: The authors have built a prototype based on the new mask design and used it to image ex vivo breast tissue samples containing malignant lesions. The authors compared images acquired with this prototype to those obtained with a conventional system. The authors demonstrate and quantify image improvements, especially in terms of microcalcification detection. On calcifications detected also by the conventional system, the authors measure contrast increases from five to nine fold; calcifications and other features were also detected which are completely invisible in the conventional image. Dose measurements confirmed that the above enhancements were achieved while delivering doses compatible with clinical practice. CONCLUSIONS: The authors obtained phase-related image enhancements in mammography by means of a system built with components available off-the-shelf that operates under exposure time and dose conditions compatible with clinical practice. This opens the way to a straightforward translation of phase enhanced imaging methods into clinical practice.

Image quality evaluation of breast tomosynthesis with synchrotron radiation.

PURPOSE: This study investigates the image quality of tomosynthesis slices obtained from several acquisition sets with synchrotron radiation using a breast phantom incorporating details that mimic various breast lesions, in a heterogeneous background. METHODS: A complex Breast phantom (MAMMAX) with a heterogeneous background and thickness that corresponds to 4.5 cm compressed breast with an average composition of 50% adipose and 50% glandular tissue was assembled using two commercial phantoms. Projection images using acquisition arcs of 24°, 32°, 40°, 48°, and 56° at incident energy of 17 keV were obtained from the phantom with the synchrotron radiation for medical physics beamline at ELETTRA Synchrotron Light Laboratory. The total mean glandular dose was set equal to 2.5 mGy. Tomograms were reconstructed with simple multiple projection algorithm (MPA) and filtered MPA. In the latter case, a median filter, a sinc filter, and a combination of those two filters were applied on the experimental data prior to MPA reconstruction. Visual inspection, contrast to noise ratio, contrast, and artifact spread function were the figures of merit used in the evaluation of the visualisation and detection of low- and high-contrast breast features, as a function of the reconstruction algorithm and acquisition arc. To study the benefits of using monochromatic beams, single projection images at incident energies ranging from 14 to 27 keV were acquired with the same phantom and weighted to synthesize polychromatic images at a typical incident x-ray spectrum with W target. RESULTS: Filters were optimised to reconstruct features with different attenuation characteristics and dimensions. In the case of 6 mm low-contrast details, improved visual appearance as well as higher contrast to noise ratio and contrast values were observed for the two filtered MPA algorithms that exploit the sinc filter. These features are better visualized at extended arc length, as the acquisition arc of 56° with 15 projection images demonstrates the highest image reconstruction quality. For microcalcifications, filtered MPA implemented with a combination of median and sinc filters indicates better feature appearance due to efficient suppression of background tissue. The image quality of these features is less sensitive to the acquisition arc. Calcifications with size ranging from 170 to 500 µm, like the ones presently studied, are well identified and visualized for all arcs used. The comparison of single projection images obtained under different beam conditions showed that the use of monochromatic beam can produce an image with higher contrast and contrast to noise ratio compared to an image corresponding to a polychromatic beam even when the latter is acquired with double incident exposure. CONCLUSIONS: Filter optimization in respect to the type of feature characteristics is important before the reconstruction. The MPA combined with median and sinc filters results in improved reconstruction of microcalcifications and low-contrast features. The latter are better visualized at extended arc length, while microcalcifications are less sensitive to this acquisition parameter. Use of monochromatic beams may result in tomographic images with higher contrast acquired at lower incident exposures.

Effects of signal diffusion on x-ray phase contrast images.

We discuss the problem of signal diffusion among neighbouring pixels in x-ray phase contrast imaging (XPCi) specifically for coded-aperture (CA) XPCi, but many of the discussed observations are directly transferable to other XPCi modalities. CA XPCi exploits the principle of pixel edge illumination by means of two CA masks. The first mask, placed in contact with the detector, creates insensitive regions between adjacent pixels; the second one, placed immediately before the sample, creates individual beams impinging on the boundaries between sensitive and insensitive regions on the detector, as created by the detector mask. In this way, edge illumination is achieved for all pixels of an area detector illuminated by a divergent and polychromatic beam generated by a conventional source. As the detector mask redefines the resolution properties of the detector, sample dithering can be used to effectively increase the system spatial resolution, without having to apply any post-processing procedure (e.g., deconvolution). This however creates artifacts in the form of secondary fringes (which have nothing to do with phase-related secondary fringes) if there is signal diffusion between adjacent pixels. In non-dithered images, signal diffusion between adjacent pixels causes a reduction in image contrast. This effect is investigated both theoretically and experimentally, and its direct implications on image quality are discussed. The interplay with the sample positioning with respect to the detector pixel matrix, which also has an effect on the obtained image contrast, is also discussed.

Performance evaluation of a pixellated Ge Compton camera.

An ongoing project is being carried out to develop a high purity germanium (HPGe) Compton camera for medical applications. The Compton camera offers many potential advantages over the conventional gamma camera. The camera reported in this paper comprises two pixellated germanium detector planes housed 9.6 cm apart in the same vacuum housing. The camera has 177 pixels, 152 in the scatter detector and 25 in the absorption detector. The pixels are 4 × 4 mm(2) with a thickness of 4 mm in the scatter detector and 10 mm in the absorption detector. Images have been taken for a variety of test objects including point sources, a ring source and a Perspex phantom. The measured angular resolution is 9.4° ± 0.4° for a 662 keV gamma-ray source at 3 cm. Due to the limited number of readout modules a multiple-view technique was used to image the source distributions from different angles and simulate the pixel arrangement in the full camera.

Pixellated Cd(Zn)Te high-energy X-ray instrument.

We have developed a pixellated high energy X-ray detector instrument to be used in a variety of imaging applications. The instrument consists of either a Cadmium Zinc Telluride or Cadmium Telluride (Cd(Zn)Te) detector bump-bonded to a large area ASIC and packaged with a high performance data acquisition system. The 80 by 80 pixels each of 250 µm by 250 µm give better than 1 keV FWHM energy resolution at 59.5 keV and 1.5 keV FWHM at 141 keV, at the same time providing a high speed imaging performance. This system uses a relatively simple wire-bonded interconnection scheme but this is being upgraded to allow multiple modules to be used with very small dead space. The readout system and the novel interconnect technology is described and how the system is performing in several target applications.

Characterization of breast tissue using energy-dispersive X-ray diffraction computed tomography.

A method for sample characterization using energy-dispersive X-ray diffraction computed tomography (EDXRDCT) is presented. The procedures for extracting diffraction patterns from the data and the corrections applied are discussed. The procedures were applied to the characterization of breast tissue samples, 6mm in diameter. Comparison with histological sections of the samples confirmed the possibility of grouping the patterns into five families, corresponding to adipose tissue, fibrosis, poorly differentiated cancer, well differentiated cancer and benign tumour.

Evaluation of digital breast tomosynthesis reconstruction algorithms using synchrotron radiation in standard geometry.

PURPOSE: In this article, the image quality of reconstructed volumes by four algorithms for digital tomosynthesis, applied in the case of breast, is investigated using synchrotron radiation. METHODS: An angular data set of 21 images of a complex phantom with heterogeneous tissue-mimicking background was obtained using the SYRMEP beamline at ELETTRA Synchrotron Light Laboratory, Trieste, Italy. The irradiated part was reconstructed using the multiple projection algorithm (MPA) and the filtered backprojection with ramp followed by hamming windows (FBR-RH) and filtered backprojection with ramp (FBP-R). Additionally, an algorithm for reducing the noise in reconstructed planes based on noise mask subtraction from the planes of the originally reconstructed volume using MPA (MPA-NM) has been further developed. The reconstruction techniques were evaluated in terms of calculations and comparison of the contrast-to-noise ratio (CNR) and artifact spread function. RESULTS: It was found that the MPA-NM resulted in higher CNR, comparable with the CNR of FBP-RH for high contrast details. Low contrast objects are well visualized and characterized by high CNR using the simple MPA and the MPA-NM. In addition, the image quality of the reconstructed features in terms of CNR and visual appearance as a function of the initial number of projection images and the reconstruction arc was carried out. Slices reconstructed with more input projection images result in less reconstruction artifacts and higher detail CNR, while those reconstructed from projection images acquired in reduced angular range causes pronounced streak artifacts. CONCLUSIONS: Of the reconstruction algorithms implemented, the MPA-NM and MPA are a good choice for detecting low contrast objects, while the FBP-RH, FBP-R, and MPA-NM provide high CNR and well outlined edges in case of microcalcifications.

Experimental validation of a radiographic simulation code using breast phantom for X-ray imaging.

Computer models and simulations of X-ray imaging systems are becoming a very precious tool during the development and evaluation of new X-ray imaging techniques. To provide, however, a faithful simulation of a system, all components must be accurately modelled and tested, followed by verification through experimental measurements. This paper presents a validation study of the XRayImagingSimulator, an in-house developed X-ray imaging simulator, which is extensively used as a basic tool in carrying out complex breast imaging simulations. The approach followed compares results obtained via an experimental setup for breast phantom (CIRS 011A) imaging, using synchrotron radiation (SYRMEP beamline at ELETTRA), with those from its simulated setup under the same conditions. The study demonstrated a very good agreement between experimental and simulated images compared both in terms of subjective and objective criteria. The combination of the XRayImagingSimulator with our BreastSimulator provides a powerful tool for in silico testing of new X-ray breast imaging approaches.

Quantitative contrast-enhanced mammography for contrast medium kinetics studies.

Quantitative contrast-enhanced mammography, based on a dual-energy approach, aims to extract quantitative and temporal information of the tumour enhancement after administration of iodinated vascular contrast media. Simulations using analytical expressions and optimization of critical parameters essential for the development of quantitative contrast-enhanced mammography are presented. The procedure has been experimentally evaluated using a tissue-equivalent phantom and an amorphous silicon active matrix flat panel imager. The x-ray beams were produced by a tungsten target tube and spectrally shaped using readily available materials. Measurement of iodine projected thickness in mg cm(-2) has been performed. The effect of beam hardening does not introduce nonlinearities in the measurement of iodine projected thickness for values of thicknesses found in clinical investigations. However, scattered radiation introduces significant deviations from slope equal to unity when compared with the actual iodine projected thickness. Scatter correction before the analysis of the dual-energy images provides accurate iodine projected thickness measurements. At 10% of the exposure used in clinical mammography, signal-to-noise ratios in excess of 5 were achieved for iodine projected thicknesses less than 3 mg cm(-2) within a 4 cm thick phantom. For the extraction of temporal information, a limited number of low-dose images were used with the phantom incorporating a flow of iodinated contrast medium. The results suggest that spatial and temporal information of iodinated contrast media can be used to indirectly measure the tumour microvessel density and determine its uptake and washout from breast tumours. The proposed method can significantly improve tumour detection in dense breasts. Its application to perform in situ x-ray biopsy and assessment of the oncolytic effect of anticancer agents is foreseeable.

Deconvolution of x-ray phase contrast images as a way to retrieve phase information lost due to insufficient resolution.

When free-space propagation x-ray phase contrast imaging is implemented outside synchrotron radiation facilities, the combined effect of detector resolution and source size swamps the fine phase contrast fringes, often making them almost undetectable. In an attempt to mitigate this effect, a simple deconvolution procedure based on division in the Fourier space plus multiplication by an appropriate filter was applied to experimental x-ray phase contrast images of a simple geometric phantom. The filter parameter was varied in order to assess its impact on the level of retrieved phase signal. The deconvolved images were compared to simulated ones obtained under different resolution conditions, showing that this simple procedure provided signals equivalent to those that would be obtained with a detector with three times better resolution. By accepting an increase in the overall image noise, the method also appears to bring up secondary phase contrast fringes, which are not visible in the unprocessed signal.

Phase contrast imaging of breast tumours with synchrotron radiation.

Even though the potential of phase contrast (PC) imaging has been demonstrated in a number of biological tissue samples, the availability of free-space propagation phase contrast images of real breast tumours is still limited. The aim of this study was to obtain phase contrast images of two different pathological breast specimens containing tumours of differing morphological type at two synchrotron radiation (SR) facilities, and to assess any qualitative improvements in the evaluation and characterisation of the masses through the use of phase contrast imaging. A second aim was to assess the effects of parameters such as detector resolution, beam energy and sample-to-detector distance on image quality using the same breast specimens, as to date these effects have been modelled and discussed only for geometric phantoms. At each synchrotron radiation facility a range of images was acquired with different detectors and by varying the above parameters. Images of the same samples were also acquired with the absorption-based approach to allow a direct comparison and estimation of the advantages specifically ascribable to the PC technique.

Signal and noise transfer properties of CMOS based active pixel flat panel imager coupled to structured CsI:Tl.

Complementary metal-oxide-semiconductors (CMOS) active pixel sensors can be optically coupled to CsI:Tl phosphors forming a indirect active pixel flat panel imager (APFPI) for high performance medical imaging. The aim of this work is to determine the x-ray imaging capabilities of CMOS-based APFPI and study the signal and noise transfer properties of CsI:Tl phosphors. Three different CsI:Tl phosphors from two different vendors have been used to produce three system configurations. The performance of each system configuration has been studied in terms of the modulation transfer function (MTF), noise power spectra, and detective quantum efficiency (DQE) in the mammographic energy range. A simple method to determine quantum limited systems in this energy range is also presented. In addition, with aid of monochromatic synchrotron radiation, the effect of iodine characteristic x-rays of the CsI:Tl on the MTF has been determined. A Monte Carlo simulation of the signal transfer properties of the imager is also presented in order to study the stages that degrade the spatial resolution of our current system. The effect of using substrate patterning during the growth of CsI:Tl columnar structure was also studied, along with the effect of CsI:Tl fixed pattern noise due to local variations in the scintillation light. CsI:Tl fixed pattern noise appears to limit the performance of our current system configurations. All the system configurations are quantum limited at 0.23 microC/kg with two of them having DQE (0) equal to 0.57. Active pixel flat panel imagers are shown to be digital x-ray imagers with almost constant DQE throughout a significant part of their dynamic range and in particular at very low exposures.