X-ray characterisation of normal and neoplastic breast tissues

Estimation of compressed breast thickness during mammography

To estimate radiation dose during mammography the breast thickness must be known. We present a new method for estimating the thickness of a compressed breast using only the breast image as projected onto a mammogram, calibration data such as the mAs value and image processing techniques. The method proves to be of high accuracy (+/- 0.2 cm for craniocaudal mammograms) and has the advantage over other methods of allowing retrospective estimation of thickness.

X-ray characterization of breast phantom materials

A pulse-height spectroscopic technique is used to measure the linear attenuation coefficients of commercially available composite phantom materials designed to simulate the attenuation characteristics of breast fat and breast glandular tissue. The manufacturers have specified the composition of these materials with the goal of matching the linear attenuation coefficients of breast tissues, calculated using the mixture rule. Over the energy range 18 to 100 keV, measurements from these materials are in close agreement with manufacturers' predictions and with previously measured linear attenuation coefficients of breast tissue samples.

Adaptive mammographic image enhancement using first derivative and local statistics

This paper proposes an adaptive image enhancement method for mammographic images, which is based on the first derivative and the local statistics. The adaptive enhancement method consists of three processing steps. The first step is to remove the film artifacts which may be misread as microcalcifications. The second step is to compute the gradient images by using the first derivative operators. The third step is to enhance the important features of the mammographic image by adding the adaptively weighted gradient images. Local statistics of the image are utilized for adaptive realization of the enhancement, so that image details can be enhanced and image noises can be suppressed. The objective performances of the proposed method were compared with those by the conventional image enhancement methods for a simulated image and the seven mammographic images containing real microcalcifications. The performance of the proposed method was also evaluated by means of the receiver operating-characteristics (ROC) analysis for 78 real mammographic images with and without microcalcifications.

Optimal technique factors for magnification mammography

RATIONALE AND OBJECTIVES

Use of small focal spots with low x-ray tube currents may result in very long exposure times and thus result in motion blur in magnification mammography. The authors investigated the reduction in exposure time with increasing x-ray tube kVp and the corresponding decrease in perceived visibility of low-contrast objects in phantom images.

METHODS

Exposure times required to radiograph an RMI 156 phantom in a magnification geometry were measured as a function of x-ray tube kVp when operated under automatic exposure control. Magnification images of the RMI 156 phantom were obtained at x-ray tube voltages ranging from 28 to 34 kVp. Five radiology residents ranked the visibility of two borderline fibers and six borderline microcalcification specks using a 5-point scale ranging from excellent to barely visible.

RESULTS

Between 28 and 34 kVp, the density of the RMI phantom images was nearly constant with a mean value of 1.32 +/- 0.04. Increasing the x-ray tube voltage from 28 kVp to 34 kVp reduced the exposure time from 1.27 seconds to 0.66 seconds. Image quality at 30 and 32 kVp was not significantly worse than that achieved at 28 kVp. Increasing the x-ray tube voltage to 34 kVp, however, resulted in a statistically significant (P < 0.001) deterioration in the relative visibility of fibers and microcalcification specks.

CONCLUSIONS

Magnification mammography performed at 32 kVp will decrease exposure times significantly and result in a microcalcification and fiber visibility that is similar to that achieved at 28 kVp.

Determination of average glandular dose with modern mammography units for two large groups of patients

Until recently, for mammography Mo anode-Mo filter x-ray tube assemblies were almost exclusively used. Modern mammography units provide the possibility to employ a variety of anode-filter combinations with the aim of adapting the x-ray spectrum to compressed breast thickness and composition. The present contribution provides information on the radiation exposure of two large groups of patients (one of 1678 and one of 945 women) who were mammographed with modern x-ray equipment, and on the dosimetry necessary for the evaluation. For dosimetric purposes spectral information is essential. X-ray spectra have been determined for various anode-filter combinations from measurements with a Ge detector. Based on these spectra, conversion factors from air kerma free in air to average glandular dose (g factors) have been calculated for different anode-filter combinations, compressed breast thickness ranging from 2 to 9 cm and breast compositions varying from 0 to 100% glandular tissue. Determinations of various quantities, including entrance surface air kerma (ESAK), tube output, tube loading (TL), fraction of glandular tissue (FGL) and compressed breast thickness, were made during actual mammography. Average glandular dose (AGD) was determined using g factors corrected for tissue composition as well as g values for standard breast composition, i.e. 50% adipose tissue and 50% glandular tissue by mass. It is shown that, on average, the influence of the actual breast composition causes variations of the order of about 15%. For group 1 and group 2, the mean values of average glandular dose (using g factors corrected for tissue composition) were 1.59 and 2.07 mGy respectively. The number of exposures per woman was on average 3.4 and 3.6 respectively. The mean value of compressed breast thickness was 55.9 and 50.8 mm respectively. The mean age of group 1 was 53.6 years (for group 2 the age was not recorded). The fraction by mass of glandular tissue FGL decrease with increasing compressed breast thickness and age of patient (from 75% at 25 mm to 20% at 80 mm, and from 65% at 20 years to 30% at 75 years). For a medium-sized breast, i.e. a compressed breast thickness of 55 mm, FGL is about 35%, indicating that the standard mix (FGL = 50%) might need some modification, particularly because of additional evidence from another investigation with similar results on FGL.

Current status of digital mammography

Some inherent limitations to further technical improvement in film-screen mammography exist. Many of these limitations can be overcome effectively with digital mammography, in which image acquisition, display, and storage are performed independently, thus allowing the optimization of each. Presented is a brief background of digital and analog imaging with emphasis on the features and drawbacks of digital mammography systems. Image storage, processing, and display, computer-aided detection and diagnosis, as well as telemammography are also discussed.

Photon linear attenuation coefficients and water content of normal and pathological breast tissues

Normal and pathological breast tissue samples were scanned using a Photon Transmission Tomography (PTT) technique in order to determine their averaged photon linear attenuation coefficients (mu). Subsequent to being freeze-dried the samples were examined, using a high purity germanium detector (HPGe) and the gamma-rays of energy 59.5 keV from an americium source, and the results were corrected for the water reduction by the use of the Mixture Rule. The ratio of our experimental findings to the published data for mu for various breast tissues were 88, 96 and 88% for adipose, glandular and tumour tissues, respectively. The mean accuracy in our study, investigated relative to standard chemical compounds, was about 3%. The water content of each tissue type was determined as the weight loss during the freeze drying process. This work was initiated in order to evaluate the suitability of new tissue substitute materials for mammography applications.

Recent developments in breast imaging

A review of breast imaging has already appeared in 1982 in this journal. Consequently, the present article concentrates on a discussion of only those developments of a more recent nature. Although the emphasis is placed on the physical aspects of the different imaging methods concerned, the essential factors relating to the clinical background and the associated radiation risk are also outlined. The completeness of detail depends on the present clinical importance of the method under discussion. X-ray mammography, which is still the most important breast imaging technique and has proved to be an effective method for breast cancer screening, is therefore treated in greater detail. Since the early 1980s, ultrasound B-mode scanning has evolved to an indispensable adjunct to x-ray mammography. For Doppler sonography, diaphanography, contrast-enhanced MRI, CT and DSA, the visualization of a tumour depends essentially on the enhanced vascularity of the lesion. Whether this will prove to be a reliable indicator for malignancy remains to be shown in controlled clinical studies. Common to all imaging systems is the increasing use of digital methods for signal processing, which also offers the possibility of computer-aided diagnosis by texture analysis and pattern recognition.

A representation for mammographic image processing

Mammographic image analysis is typically performed using standard, general-purpose algorithms. We note the dangers of this approach and show that an alternative physics-model-based approach can be developed to calibrate the mammographic imaging process. This enables us to obtain, at each pixel, a quantitative measure of the breast tissue. The measure we use is h(int) and this represents the thickness of 'interesting' (non-fat) tissue between the pixel and the X-ray source. The thicknesses over the image constitute what we term the h(int) representation, and it can most usefully be regarded as a surface that conveys information about the anatomy of the breast. The representation allows image enhancement through removing the effects of degrading factors, and also effective image normalization since all changes in the image due to variations in the imaging conditions have been removed. Furthermore, the h(int) representation gives us a basis upon which to build object models and to reason about breast anatomy. We use this ability to choose features that are robust to breast compression and variations in breast composition. In this paper we describe the h(int) representation, show how it can be computed, and then illustrate how it can be applied to a variety of mammographic image processing tasks. The breast thickness turns out to be a key parameter in the computation of h(int), but it is not normally recorded. We show how the breast thickness can be estimated from an image, and examine the sensitivity of h(int) to this estimate. We then show how we can simulate any projective X-ray examination and can simulate the appearance of anatomical structures within the breast. We follow this with a comparison between the h(int) representation and conventional representations with respect to invariance to imaging conditions and the surrounding tissue. Initial results indicate that image analysis is far more robust when specific consideration is taken of the imaging process and the h(int) representation is used.

A multiparameter optimization of digital mammography

Multiparameter optimizations have been carried out for a wide range of digital mammography system configurations and requirements, with the aim of optimizing the image quality for a given patient dose. These conditions include a range of slot widths for scanning mammography systems, exposure times from 1 to 10 s, focal spot sizes from 80 to 800 microns, a range of detector resolutions and noise levels, dose restrictions, patient thicknesses and targets, and x-ray tube targets. The influences of these on the optimum system configuration in terms of tube potential, filtration, source to patient distance and target magnification are discussed. It is demonstrated that x-ray tube power constraints can significantly restrict the optimum magnification for slot scanning systems, with the result that poor-resolution detectors are not suited for use in a scanning configuration, and that large-focal-spot-good-detector resolution combinations are more suitable. The use of a detector with increased width, raised tube potential and reduced amount of added filtration is shown to be helpful in reducing x-ray tube power limitations. It is shown that, in many cases, correct optimization can bring the detail SNR for an examination using a given detector-x-ray tube configuration to within 10-15% of the SNR achieved with the optimum combination. This gives the designer some scope to consider other factors such as cost and the implications of image size on storage space.

Role of equalisation mammography of dense breasts

Parenchymal patterns characteristic of dense breasts are known to degrade the mammographic detection of small breast cancers and microcalcifications. This arises from large variations in exposure of the film, resulting in reduced image contrast over areas of suboptimal exposure. Based on sensitometric measurements of mammograms from a typical patient population, it is shown that over 60% of a typical mammogram in Wolfe's DY classification was found to be exposed suboptimally, suggesting a significant margin for improving mammography for these patients. In order to address this problem, a prototype mammographic version of scanning equalisation radiography (MSER) has been developed, which delivers a patient-specific spatially non-uniform distribution of breast exposure, adjusted to maintain optimal film exposure and contrast over the entire mammogram. Anthropomorphic phantom MSER images show a marked improvement in subjective image quality relative to conventional mammograms, while exhibiting a similar radiation risk. The detection of small microcalcifications and fibrils over clinically significant breast densities is found to be improved by factors eight and four, respectively. Such a system may be clinically practical through the use of multiple-beam equalisation methods with available X-ray tube technology.

Computing the scatter component of mammographic images

The authors build upon a technical report (Tech. Report OUEL 2009/93, Engng. Sci., Oxford Uni., Oxford, UK, 1993) in which they proposed a model of the mammographic imaging process for which scattered radiation is a key degrading factor. Here, the authors propose a way of estimating the scatter component of the signal at any pixel within a mammographic image, and they use this estimate for model-based image enhancement. The first step is to extend the authors' previous model to divide breast tissue into "interesting" (fibrous/glandular/cancerous) tissue and fat. The scatter model is then based on the idea that the amount of scattered radiation reaching a point is related to the energy imparted to the surrounding neighbourhood. This complex relationship is approximated using published empirical data, and it varies with the size of the breast being imaged. The approximation is further complicated by needing to take account of extra-focal radiation and breast edge effects. The approximation takes the form of a weighting mask which is convolved with the total signal (primary and scatter) to give a value which is input to a "scatter function", approximated using three reference cases, and which returns a scatter estimate. Given a scatter estimate, the more important primary component can be calculated and used to create an image recognizable by a radiologist. The images resulting from this process are clearly enhanced, and model verification tests based on an estimate of the thickness of interesting tissue present proved to be very successful. A good scatter model opens the was for further processing to remove the effects of other degrading factors, such as beam hardening.

Mammographic feature enhancement by multiscale analysis

Introduces a novel approach for accomplishing mammographic feature analysis by overcomplete multiresolution representations. The authors show that efficient representations may be identified within a continuum of scale-space and used to enhance features of importance to mammography. Methods of contrast enhancement are described based on three overcomplete multiscale representations: 1) the dyadic wavelet transform (separable), 2) the phi-transform (nonseparable, nonorthogonal), and 3) the hexagonal wavelet transform (nonseparable). Multiscale edges identified within distinct levels of transform space provide local support for image enhancement. Mammograms are reconstructed from wavelet coefficients modified at one or more levels by local and global nonlinear operators. In each case, edges and gain parameters are identified adaptively by a measure of energy within each level of scale-space. The authors show quantitatively that transform coefficients, modified by adaptive nonlinear operators, can make more obvious unseen or barely seen features of mammography without requiring additional radiation. The authors' results are compared with traditional image enhancement techniques by measuring the local contrast of known mammographic features. They demonstrate that features extracted from multiresolution representations can provide an adaptive mechanism for accomplishing local contrast enhancement. By improving the visualization of breast pathology, one can improve chances of early detection while requiring less time to evaluate mammograms for most patients.

Measurement of small-angle photon scattering for some breast tissues and tissue substitute materials

For photon energies encountered in diagnostic radiology the shape of the scattering distributions for low-atomic-number media exhibits peaks in intensity close to the forward direction that are not predicted by conventional theoretical models. The positions and shapes of the peaks depend upon the interatomic and intermolecular configurations of the scatterers. The phenomenon is of particular interest because of its relevance to the understanding and modelling of x-ray imaging processes and the possibility that the peaking may be characteristic of tissue type. In the present study, peaks in the forward scattering distributions have been demonstrated for 19 samples of breast tissue and three tissue substitute materials using a position-sensitive photon detector and a 60 kVp x-ray source. Prominent features were observed for all samples investigated. Large differences were found in the shapes of the distributions between adipose and fibroglandular tissues and only small differences were found between carcinomas and fibroglandular tissues.

Generation of "soft x-rays" by using the free electron laser as a proposed means of diagnosing and treating breast cancer

The diagnosis and treatment of breast lesions may be markedly enhanced by the use of a unique new source of near-monochromatic x-rays. Concentric beams of near-monochromatic x-ray photons may be generated by collision of the free electron laser (FEL) electron beam with the optical beam in an interaction zone that delivers the x-rays to a shirtsleeve environment. The absence of Compton scatter and the photoelectric interaction within tissues improves conspicuity of lesions by two to six times. Increased attenuation of x-rays in malignant vs. normal tissues makes tumors more obvious. K-edge subtraction allows chemical analysis of tumors in vivo--all at radiation doses that are one-tenth to one-fiftieth that delivered by the lowest-dose mammographic x-ray technique available. This allows for an increased sensitivity and specificity and permits prediction of histology, negating necessity for biopsies. Selective bond-breaking at depth in tissues as well as x-ray-activated photodynamic therapy are also being explored.