A new generation of digital breast tomosynthesis system has been designed and is commercially available outside the US.
The system has both a 2D mode and a 3D mode to do either conventional mammography or tomosynthesis. Uniquely, it
also has a fusion mode that allows both 3D and 2D images to be acquired under the same breast compression, which results in co-registered images from the two modalities. The aim of this paper is to present a technical description on the design and performance of the new system, including system details such as filter options, doses, AEC operation, 2D and 3D images co-registration and display, and the selenium detector performance. We have carried out both physical and clinical studies to evaluate the system. In this paper the focus will be mainly on technical performance results.
Clinical studies have correlated a high breast density to a women's risk of breast cancer. A breast density measurement
that can quantitatively depict the volume distribution and percentage of dense tissues in breasts would be very useful for
risk factor assessment of breast cancer, and might be more predictive of risks than the common but subjective and
coarse 4-point BIRADS scale. This paper proposes to use a neural-network mapping to compute the breast density
information based upon system calibration data, x-ray techniques, and Full Field Digital Mammography (FFDM)
images. The mapping consists of four modules, namely, system calibration, generator of beam quality, generator of
normalized absorption, and a multi-layer feed-forward neural network. As the core of breast density mapping, the
network accepts x-ray target/filter combination, normalized x-ray absorption, pixel-wise breast thickness map, and x-ray
beam quality during image acquisition as input elements, and exports a pixel-wise breast density distribution and a
single breast density percentage for the imaged breast. Training and testing data sets for the design and verification of
the network were formulated from calibrated x-ray beam quality, imaging data with a step wedge phantom under a
variety x-ray imaging techniques, and nominal breast densities of tissue equivalent materials. The network was trained
using a Levenberg-Marquardt algorithm based back-propagation learning method. Various thickness and glandular
density phantom studies were performed with clinical x-ray techniques. Preliminary results showed that the neural
network mapping is promising in accurately computing glandular density distribution and breast density percentage.
A second-generation digital breast tomosynthesis system is used for a screening study comparing tomosynthesis with conventional two-view mammography with matched x-ray dose. The system acquires 15 projections of a breast at different angles using a digital detector. This work explores acquisition techniques that optimize the quality of projection images at low x-ray exposure. The system provides three target-filter combinations (Mo-Mo, Mo-Rh and Rh-Rh) and the recommended tube voltage range is from 25 to 40kVp. A thin disk was put on top of slabs of breast tissue equivalent materials (20 to 85mm). Contrast-to-noise ratio of the disk was measured from projections acquired with different kVp and target-filter combinations. The squared CNR normalized by average glandular dose was used to compare the quality/dose efficiency of different techniques. The optimal quality/dose efficiency was achieved as the detector entrance exposure was in the range of 5-30mR. Within this range, Mo-Mo gives the highest quality for 20mm; results are very close for 30mm; Rh-Rh is slightly better for 45mm and apparently better than others for 65 and 85mm. However, sufficient detector entrance exposure cannot be guaranteed for all cases due to the total dose limit and the system limit. For some cases, the detector is operated slightly off its optimal performance range. The kVp does not show an impact except for 85 mm, in which the quality/dose efficiency slightly increases at higher kVp. Rh-Rh is selected for > 40mm thickness; Mo-Mo is selected for 20mm thickness; and Mo-Rh is selected for 30 and 40mm.
KEYWORDS: Breast, Monte Carlo methods, Signal attenuation, 3D image processing, Tissues, Photons, X-rays, 3D image reconstruction, 3D metrology, Sensors
We are developing a breast specific scatter correction method for digital beast tomosynthesis (DBT). The 3D breast volume was initially reconstructed from 15 projection images acquired from a GE prototype tomosynthesis system without correction of scatter. The voxel values were mapped to the tissue compositions using various segmentation schemes. This voxelized digital breast model was entered into a Monte Carlo package simulating the prototype tomosynthesis system. One billion photons were generated from the x-ray source for each projection in the simulation and images of scattered photons were obtained. A primary only projection image was then produced by subtracting the scatter image from the corresponding original projection image which contains contributions from the both primary photons and scatter photons. The scatter free projection images were then used to reconstruct the 3D breast using the same algorithm. Compared with the uncorrected 3D image, the x-ray attenuation coefficients represented by the scatter-corrected 3D image are closer to those derived from the measurement data.
The Maximum Likelihood Expectation Maximization (MLEM) algorithm has been shown to produce the highest quality Digital Breast Tomosynthesis (DBT) images. MLEM, however, is computationally intensive. Single-processor image reconstruction times for each breast were on the order of several hours. In order for DBT to be clinically useful, faster reconstruction times using cost-effective software/hardware solutions are needed. We have implemented the MLEM reconstruction algorithm for use with DBT on a graphics processing unit (GPU). Compared to a single optimized 2.8GHz Pentium system this enabled a 113-fold speedup in processing time, while maintaining high image quality. Subsequently, we added various additional processing steps to the reconstruction algorithm in order to improve image quality and diagnostic properties. Since the performance of commercial GPUs increases rapidly, with little change in cost, the increased sophistication in processing does not entail an increase in system cost. The use of GPUs for reconstruction represents a technical breakthrough in the cost-effective application of MLEM to Digital Breast Tomosynthesis.
We describe what is, to the best of our knowledge, the first pilot study of coregistered tomographic x-ray and optical breast imaging. The purpose of this pilot study is to develop both hardware and data processing algorithms for a multimodality imaging method that provides information that neither x-ray nor diffuse optical tomography (DOT) can provide alone. We present in detail the instrumentation and algorithms developed for this multimodality imaging. We also present results from our initial pilot clinical tests. These results demonstrate that strictly coregistered x-ray and optical images enable a detailed comparison of the two images. This comparison will ultimately lead to a better understanding of the relationship between the functional contrast afforded by optical imaging and the structural contrast provided by x-ray imaging.
We developed a two-stage computerized mass detection algorithm
for digital tomosynthesis images of the breast. Rather than
analyze the reconstructed 3D breast volume, our algorithm operates
on each of the 2D projection images directly. We chose this approach
because reconstruction algorithms for breast tomosynthesis are still
being optimized, which can alter the appearance of the 3D reconstructed
breast volume. Furthermore this approach allows us to take advantage
of mass detection methods already developed for conventional two-view
projection mammography, which are similar to projection images for digital
tomosynthesis. We applied our algorithm to two tomosynthesis image sets,
one of which was a computer simulated 3D breast phantom, and
one was a clinical image set. In both cases, the lesion was detected in
the first stage of the algorithm, while the second stage of the algorithm
efficiently reduced false positive detections.
A parallel reconstruction method, based on an iterative maximum likelihood (ML) algorithm, is developed to provide fast reconstruction for digital tomosynthesis mammography. Tomosynthesis mammography acquires 11 low-dose projections of a breast by moving an x-ray tube over a 50° angular range. In parallel reconstruction, each projection is divided into multiple segments along the chest-to-nipple direction. Using the 11 projections, segments located at the
same distance from the chest wall are combined to compute a partial reconstruction of the total breast volume. The shape of the partial reconstruction forms a thin slab, angled toward the x-ray source at a projection angle 0°. The reconstruction of the total breast volume is obtained by merging the partial reconstructions. The overlap region between neighboring partial reconstructions and neighboring projection segments is utilized to compensate for the incomplete data at the boundary locations present in the partial reconstructions. A serial execution of the reconstruction is compared
to a parallel implementation, using clinical data. The serial code was run on a PC with a single PentiumIV 2.2GHz CPU. The parallel implementation was developed using MPI and run on a 64-node Linux cluster using 800MHz Itanium CPUs. The serial reconstruction for a medium-sized breast (5cm thickness, 11cm chest-to-nipple distance) takes 115 minutes, while a parallel implementation takes only 3.5 minutes. The reconstruction time for a larger breast using a
serial implementation takes 187 minutes, while a parallel implementation takes 6.5 minutes. No significant differences
were observed between the reconstructions produced by the serial and parallel implementations.
Mark Williams, Guimin Zhang, Mitali More, Allen Goode, Stan Majewski, Randy Wojcik, Brian Kross, Vladimir Popov, Andrew Weisenberger, Martin Stanton, Walter Phillips, Alexander Stewart, Thomas McCauley, Tao Wu, Edward DiBella
We are developing a scanner for simultaneous acquisition of x-ray computed tomography (CT) and single photon emission tomography (SPECT) images of small animals such as mice and rats. The scanner uses a cone beam geometry for both the x- ray transmission and gamma emission projections by using an area x-ray detector and pinhole collimator, respectively. The CT and SPECT data set are overlaid to form a coregistered structural-functional 3D image. The CT system includes a single CCD-based x-ray detector and a microfocus x-ray source. The SPECT scanner utilizes tungsten pinhole collimators and arrays of CsI(Tl) scintillation detectors. We describe considerations and the early performance of a prototype scanner.
KEYWORDS: Sensors, Breast, Image resolution, Signal to noise ratio, 3D image processing, Mammography, Image restoration, Data modeling, Reconstruction algorithms, X-rays
Typically in three-dimensional (3D) computed tomography (CT) imaging, hundreds or thousands of x-ray projection images are recorded. The image-collection time and patient dose required rule out conventional CT as a tool for screening mammography. We have developed a CT method that overcomes these limitations by using (1) a novel image collection geometry, (2) new digital electronic x-ray detector technology, and (3) modern image reconstruction procedures. The method, which we call Computed Planar Mammography (CPM), is made possible by the full-field, low-noise, high-resolution CCD-based detector design that we have previously developed. With this method, we need to record only a limited number (10 - 50) of low-dose x- ray images of the breast. The resulting 3D full breast image has a resolution in two orientations equal to the full detector resolution (47 microns), and a lower, variable resolution (0.5 - 10 mm) in the third orientation. This 3D reconstructed image can then be viewed as a series of cross- sectional layers, or planes, each at the full detector resolution. Features due to overlapping tissue, which could not be differentiated in a conventional mammogram, are separated into layers at different depths. We demonstrate the features and capabilities of this method by presenting reconstructed images of phantoms and mastectomy specimens. Finally, we discuss outstanding issues related to the further development of this procedure, as well as considerations for its clinical implementation.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.