A search for optimal x-ray spectra in iodine contrast media mammography

The aim of this work was to search for the optimal x-ray tube voltage and anode-filter combination in digital iodine contrast media mammography. In the optimization, two entities were of interest: the average glandular dose, AGD, and the signal-to-noise ratio, SNR, for detection of diluted iodine contrast medium. The optimum is defined as the technique maximizing the figure of merit, SNR2/AGD. A Monte Carlo computer program was used which simulates the transport of photons from the x-ray tube through the compression plate, breast, breast support plate, anti-scatter grid and image detector. It computes the AGD and the SNR of an iodine detail inside the compressed breast. The breast thickness was varied between 2 and 8 cm with 10-90% glandularity. The tube voltage was varied between 20 and 55 kV for each anode material (Rh, Mo and W) in combination with either 25 microm Rh or 0.05-0.5 mm Cu added filtration. The x-ray spectra were calculated with MCNP4C (Monte Carlo N-Particle Transport Code System, version 4C). A CsI scintillator was used as the image detector. The results for Rh/0.3 mmCu, Mo/0.3 mmCu and W/0.3 mmCu were similar. For all breast thicknesses, a maximum in the figure of merit was found at approximately 45 kV for the Rh/Cu, Mo/Cu and W/Cu combinations. The corresponding results for the Rh/Rh combination gave a figure of merit that was typically lower and more slowly varying with tube voltage. For a 4 cm breast at 45 kV, the SNR2/AGD was 3.5 times higher for the Rh/0.3 mmCu combination compared with the Rh/Rh combination. The difference is even larger for thicker breasts. The SNR2/AGD increases slowly with increasing Cu-filter thickness. We conclude that tube voltages between 41 and 55 kV and added Cu-filtration will result in significant dose advantage in digital iodine contrast media mammography compared to using the Rh/Rh anode/filter combination at 25-32 kV.     

The effect of breast composition on absorbed dose and image contrast

We have studied the effect of breast composition on the average whole breast dose, average glandular dose, and image contrast in mammography, using both computational and experimental methods. Three glandular/adipose compositions were considered: 30/70, 50/50, and 70/30 by weight, for both 3- and 5-cm breast thickness. Absorbed dose was found to increase with greater glandular content and this increase is more pronounced for thick breasts and softer beams. For typical screen-film x-ray beams, the average dose to a highly glandular breast is nearly twice the dose to a highly adipose breast and the average glandular dose about 40% higher. Dose was reduced when higher energy beams were employed. The use of a grid increased the dose by a factor of 2.0 to 2.6. Finally, the measured image contrast decreases with increasing breast glandularity, to a greater extent in small breasts and when low energy beams were employed.     

Dual-energy approach to contrast-enhanced mammography using the balanced filter method: spectral optimization and preliminary phantom measurement

Dual-energy contrast agent-enhanced mammography is a technique of demonstrating breast cancers obscured by a cluttered background resulting from the contrast between soft tissues in the breast. The technique has usually been implemented by exploiting two exposures to different x-ray tube voltages. In this article, another dual-energy approach using the balanced filter method without switching the tube voltages is described. For the spectral optimization of dual-energy mammography using the balanced filters, we applied a theoretical framework reported by Lemacks et al. [Med. Phys. 29, 1739-1751 (2002)] to calculate the signal-to-noise ratio (SNR) in an iodinated contrast agent subtraction image. This permits the selection of beam parameters such as tube voltage and balanced filter material, and the optimization of the latter's thickness with respect to some critical quantity-in this case, mean glandular dose. For an imaging system with a 0.1 mm thick CsI:T1 scintillator, we predict that the optimal tube voltage would be 45 kVp for a tungsten anode using zirconium, iodine, and neodymium balanced filters. A mean glandular dose of 1.0 mGy is required to obtain an SNR of 5 in order to detect 1.0 mg/cm2 iodine in the resulting clutter-free image of a 5 cm thick breast composed of 50% adipose and 50% glandular tissue. In addition to spectral optimization, we carried out phantom measurements to demonstrate the present dual-energy approach for obtaining a clutter-free image, which preferentially shows iodine, of a breast phantom comprising three major components-acrylic spheres, olive oil, and an iodinated contrast agent. The detection of iodine details on the cluttered background originating from the contrast between acrylic spheres and olive oil is analogous to the task of distinguishing contrast agents in a mixture of glandular and adipose tissues.     

An edge spread technique for measurement of the scatter-to-primary ratio in mammography

An experimental measurement technique that directly measures the magnitude and spatial distribution of scatter in relation to primary radiation is presented in this work. The technique involves the acquisition of magnified edge spread function (ESF) images with and without scattering material present. The ESFs are normalized and subtracted to yield scatter-to-primary ratios (SPRs), along with the spatial distributions of scatter and primary radiation. Mammography is used as the modality to demonstrate the ESF method, which is applicable to all radiographic environments. Sets of three images were acquired with a modified clinical mammography system employing a flat panel detector for 2, 4, 6, and 8 cm thick breast tissue equivalent material phantoms composed of 0%, 43%, and 100% glandular tissue at four different kV settings. Beam stop measurements of scatter were used to validate the ESF methodology. There was good agreement of the mean SPRs between the beam stop and ESF methods. There was good precision in the ESF-determined SPRs with a coefficient of variation on the order of 5%. SPRs ranged from 0.2 to 2.0 and were effectively independent of energy for clinically realistic kVps. The measured SPRs for 2, 4, and 6 cm 0% glandular phantoms imaged at 28 kV were 0.21+/-0.01, 0.39+/-0.01, and 0.57+/-0.02, respectively. The measured SPRs for 2, 4, and 6 cm 43% glandular phantoms imaged at 28 kV were 0.20+/-0.01, 0.35+/-0.02, and 0.53+/-0.02, respectively. The measured SPRs for 2, 4, and 6 cm 100% glandular phantoms imaged at 28 kV were 0.22+/-0.02, 0.42+/-0.03, and 0.88+/-0.08, respectively.     

X-ray scattering from human breast tissues and breast-equivalent materials

The angular distributions of photons scattered by human breast tissues (adipose and glandular) and by eight breast-equivalent materials (water, polymethylmethacrylate, nylon, polyethylene and four commercial breast-equivalent materials simulating different glandular-adipose proportions) have been measured at a photon energy of 17.44 keV (Kalpha-radiation of Mo). Transmission target geometry has been used with an acceptance of +/- 0.6 degrees and an uncertainty of approximately 7%. Experimental molecular form factors were extracted from diffraction patterns normalizing the number of scattered photons with theoretical data in regions where no structure is expected. Linear attenuation coefficients have been measured for all samples at this energy. The results for water, polymethylmethacrylate, nylon and adipose tissue agree with former reported data. The results for human breast tissues at low and medium scattering angle (1-25 degrees, corresponding to the momentum transfer region between 0.2 and 3 nm(-1)) differ from the breast-equivalent materials. The results for adipose tissue are similar to the corresponding values from commercial breast-equivalent materials while the results for glandular tissue are similar to those for water.     

A new test phantom with different breast tissue compositions for image quality assessment in conventional and digital mammography

Our objective is to describe a new test phantom that permits the objective assessment of image quality in conventional and digital mammography for different types of breast tissue. A test phantom, designed to represent a compressed breast, was made from tissue equivalent materials. Three separate regions, with different breast tissue compositions, are used to evaluate low and high contrast resolution, spatial resolution and image noise. The phantom was imaged over a range of kV using a Contour 2000 (Bennett) mammography unit with a Kodak MinR 2190-MinR L screen-film combination and a Senograph 2000D (General Electric) digital mammography unit. Objective image quality assessments for different breast tissue compositions were performed using the phantom for conventional and digital mammography. For a similar mean glandular dose (MGD), the digital system gives a significantly higher contrast-to-noise ratio (CNR) than the screen-film system for 100% glandular tissue. In conclusion, in mammography, a range of exposure conditions is used for imaging because of the different breast tissue compositions encountered clinically. Ideally, the patient dose-image quality relationship should be optimized over the range of exposure conditions. The test phantom presented in this work permits image quality parameters to be evaluated objectively for three different types of breast tissue. Thus, it is a useful tool for optimizing the patient dose-image quality relationship.     

Scatter/primary in mammography: comprehensive results

Monte Carlo procedures using the SIERRA code (validated in a companion article) were used to investigate the scatter properties in mammography. The scatter to primary ratio (SPR) was used for quantifying scatter levels as a function of beam spectrum, position in the field, air gap, breast thickness, tissue composition, and the area of the field of view (FOV). The geometry of slot scan mammography was also simulated, and SPR values were calculated as a function of slot width. The influence of large air gaps (to 30 cm) was also studied in the context of magnification mammography. X-ray energy and tissue composition from 100% adipose to 100% glandular demonstrated little effect on the SPR. Air gaps over a range from 0 to 30 mm showed only slight effects. The SPR increased with increased breast thickness and with larger fields of view. Measurements from 82 mammograms provided estimates of the range of compressed breast thickness (median: 5.2 cm, 95% range: 2.4 cm to 7.9 cm) and projected breast area onto the film (left craniocaudal view, median: 146 cm2, 95% range: 58 cm2 to 298 cm2). SPR values for semicircular breast shapes, Mo/Mo spectra, and a 15 mm air gap were parametrized as a function of breast thickness and (semicircular) breast diameter. With the coefficients a = - 2.35452817439093, b = 22.3960980055927, and c = 8.85064260299289, the equation SPR= [a + b x (diameter in cm)--(-1.5) + c x (thickness in cm) --(-0.5)]-- -1 produces SPR data from 2 to 8 cm and from 3 to 30 cm breast diameters with an average error of about 1%.     

Evaluation of dual-energy subtraction of digital mammography images under conditions found in a commercial unit

Radiological contrast-to-noise ratio (CNR) is evaluated in subtracted images of microcalcifications in breast tissue. CNR is calculated for dual-kVp subtraction combining beams available in a Senographe 2000D, assuming single breast compression. Spectra were obtained from Boone et al (1997 Med. Phys. 24 1863-73), and the study was limited to lowest 25 kV Mo/Mo and highest 40 kV Rh/Rh beams, for 2.58 x 10(-4) C kg(-1) (1R) total exposure. For a standard case combining 25 kVp Mo/Mo and 40 kVp Rh/Rh beams, predicted maximum CNR for 300 microm calcification in 5 cm thick, 50% glandular, breast is about 1.2, below Rose's criterion for visualization. Total mean glandular doses are about 2.5 cGy for a standard case. The effect that input factors might have on predictions has been evaluated. Choice between alternative spectra can affect CNR by 50%. Assumed calcification composition leads to differences of 67% in calculated CNR, and assumed breast tissue composition can alter CNR by 45%; these results are weakly dependent on calcification or breast thickness, or on the assumed fraction of glandular tissue. CNR values are related to detected spectra effective energy. Calculations predict that above 37 kVp Mo/Mo beams are more energetic than Rh/Rh at the same kVp, due to beam hardening.     

The effect of breast compression on mass conspicuity in digital mammography

This study analyzed how the inherent quality of diagnostic information in digital mammography could be affected by breast compression. A digital mammography system was modeled using a Monte Carlo algorithm based on the Penelope program, which has been successfully used to model several medical imaging systems. First, the Monte Carlo program was validated against previous measurements and simulations. Once validated, the Monte Carlo software modeled a digital mammography system by tracking photons through a voxelized software breast phantom, containing anatomical structures and breast masses, and following photons until they were absorbed by a selenium-based flat-panel detector. Simulations were performed for two compression conditions (standard compression and 12.5% reduced compression) and three photon flux conditions (constant flux, constant detector signal, and constant glandular dose). The results showed that reduced compression led to higher scatter fractions, as expected. For the constant photon flux condition, decreased compression also reduced glandular dose. For constant glandular dose, the SdNR for a 4 cm breast was 0.60 +/- 0.11 and 0.62 +/- 0.11 under standard and reduced compressions, respectively. For the 6 cm case with constant glandular dose, the SdNR was 0.50 +/- 0.11 and 0.49 +/- 0.10 under standard and reduced compressions, respectively. The results suggest that if a particular imaging system can handle an approximately 10% increase in total tube output and 10% decrease in detector signal, breast compression can be reduced by about 12% in terms of breast thickness with little impact on image quality or dose.     

Contrast in diaphanography of the breast

Diaphanography is an imaging technique used in diagnosis of breast disease including cancer. The breast is illuminated with low intensity light and the transmission pattern of red and near-infrared radiation is detected, amplified, reconstructed and displayed in a monitor. The instrumentation for diaphanography has evolved empirically, mostly through clinical practice, without a very clear understanding of the scientific basis of the technique. This research is concerned with investigating theoretically the dependence of the contrast produced by a lesion in a diaphanography image on the size, depth at which a tumor is located, photon energy, and photon angular flux distribution. Contrast calculations using the DOT computer code in a two-dimensional geometry showed that decreasing the size of a tumor by 50% decreases the contrast by a factor of 3 and 4 for 695- and 853-nm photons, respectively. Decreasing the size of the normal tissue where a tumor is imbedded by 25% (from 4 to 3 cm) does not change the contrast very much (less than 20%) for both 695- and 853-nm photons. The contrast for 950- and 695-nm photons is comparable while the values for 853-nm photons are smaller by a factor of 5 for similar cases. The contrast was also found to be dependent on the angle at which the diffuse light is detected after it transverses the host tissue, maximum contrast was found for 695- and 853-nm photons at about 55 degrees. For a detection angle of 77 degrees the contrast observed is 3X and 12X smaller for 695- and 853-nm photons, respectively.(ABSTRACT TRUNCATED AT 250 WORDS)     

Signal-to-noise ratio evaluation in dual-energy radiography with synchrotron radiation

Dual-energy radiography is an effective technique that allows removal of contrast between pairs of materials in order to display details of interest on a uniform background. In mammographic images the detection of small nodules is often impeded by obscuring background 'clutter' resulting from the contrast between normal tissues (glandular and adipose) in their neighbourhood. We consider whether it could be possible to apply dual-energy radiography to the breast, which is hypothetically principally composed of three tissues, glandular, adipose and cancerous, in order to remove the contrast due to the distribution of normal tissues and, as a consequence, to enhance the intrinsic contrast of the pathology. The purpose of this work is to test the limitations of dual-energy radiography on a three-component phantom under optimum conditions of the source and detector. We use a synchrotron monochromatic beam, produced at the ELETTRA synchrotron facility (Trieste, Italy), and an imaging plate detector, in order to acquire two images at low and high energies of a phantom composed of polyethylene, plexiglas and water. For evaluation of the potential of this procedure we studied the signal-to-noise ratio (SNR) of polyethylene and water on a set of images obtained by applying the dual-energy procedure. We found that the SNR of polyethylene and water is around the detectability threshold (according to Rose's criteria) at the contrast cancellation angles. Finally we evaluated the air entrance dose required for this double exposure, resulting in 0.81 mGy for the low-energy image and 0.01 mGy for the high-energy image. To obtain the same image quality for a standard breast of 5.5 cm, mean glandular doses of 3.50 mGy and 0.03 mGy at 17 keV and at 34 keV, respectively, are required.     

Bone cement X-ray contrast media: a clinically relevant method of measuring their efficacy

It is important to compare different contrast media used in bone cement according to their ability to attenuate X-rays and thereby produce image contrast between bone cement and its surroundings in clinical applications. The radiopacity of bone cement is often evaluated by making radiographs of cement in air at an X-ray tube voltage of 40 kV. We have developed a method for ranking contrast media in bone cement simulating the clinical situation, by (1) choosing the same X-ray tube voltage as used in clinical work, and (2) using a water phantom to imitate the effects of the patients' soft tissue on the X-ray photons. In clinical work it is desirable to have low radiation dose, but high image contrast. The voltage chosen is a compromise, because both dose and image contrast decrease with higher voltage. Three contrast media (ZrO(2), BaSO(4), and Iodixanol) have been compared for degree of "image contrast." Comparing 10 wt % contrast media samples at an X-ray tube voltage of 40 kV, ZrO(2) produced higher image contrast than the other media. However, at 80 kV, using a water phantom, the results were reversed, ZrO(2) produced lower image contrast than both BaSO(4) and Iodixanol. We conclude that evaluations of contrast media should be made with voltages and phantoms imitating the clinical application.     

Investigation of the origin of scatter components transmitted through anti-scatter grids in X-ray Digital Imaging system using Monte Carlo Simulation

BackgroundProjection diagnostic X-ray images are inherently affected by the masking effects of transmitted scatter. Spatially distributed transmitted scatter degrades image quality engendering need for effective scatter correction protocol.ObjectivesTo investigate origin of scatter components transmitted through anti-scatter grids to the detector of digital radiography system using Monte Carlo simulation.MethodsOver 107 photons were exposed through the reconstructed MC simulation phantom. Transmitted photons (primary and scatter) were scored as fluence, dose and deposited energy. Scatter components were investigated analytically over varying phantom thickness, tube kV and grid characteristics. Test disks were exposed as ROI embedded in phantom to evaluate the potential contrast improvement in image quality with the proposed technique.ResultsSimulated and experimental results were comparable and in agreement with literature. SPR and SF mean values of 10.5, 0.314 and 7.96, 0.242 through grids of ratio 10:1 and 16:1 respectively was observed. Analysis of scatter components generation in object, grid''s assembly, and fluorescent yields gave mean values of 0.815, 0.167 and 0.017, respectively. Image contrast was observed to increase with tube voltage and grid ratio.ConclusionAchieving better image contrast, reduced patient dose and low scatter transmission while maintaining superior image quality, using grids with high grid ratio and selectivity is recommended.     

Grid line artifact formation: A comprehensive theory

Linear focused grids are commonly used in general radiography and mammography to control scatter. In these applications, if lines would be visible when the grid was stationary, then the grid is moved during the x-ray exposure to blur out grid lines. Presented is a theoretical framework for estimating grid line artifact magnitude and evaluating artifact suppression techniques. The framework takes as parameters the grid pitch, septum thickness, and exposure time, and allows for a variation in grid velocity and in x-ray tube output during the exposure. Grid line artifacts are evaluated for a variety of conditions. These include a stationary grid, a grid moving at a constant velocity with no kV ripple, a grid moving at a constant velocity with large kV ripple, and a grid moving with decreasing velocity and no kV ripple. Also evaluated are grid line artifacts for a novel suppression technique in which the grid moves at a constant velocity and the x-ray exposure waveform is "feathered," i.e., when the x-ray exposure waveform has a soft start and stop. Of practical interest is that it is possible to effectively eliminate grid line artifacts when the grid moves only a short distance with an appropriately "feathered" exposure waveform. This capability permits one to design efficient and compact coarse strip density grid systems.     

Normalized glandular dose (DgN) coefficients for flat-panel CT breast imaging

The development of new digital mammography techniques such as dual-energy imaging, tomosynthesis and CT breast imaging will require investigation of optimal camera design parameters and optimal imaging acquisition parameters. In optimizing these acquisition protocols and imaging systems it is important to have knowledge of the radiation dose to the breast. This study presents a methodology for estimating the normalized glandular dose to the uncompressed breast using the geometry proposed for flat-panel CT breast imaging. The simulation uses the GEANT 3 Monte Carlo code to model x-ray transport and absorption within the breast phantom. The Monte Carlo software was validated for breast dosimetry by comparing results of the normalized glandular dose (DgN) values of the compressed breast to those reported in the literature. The normalized glandular dose was then estimated for a range of breast diameters from 10 cm to 18 cm using an uncompressed breast model with a homogeneous composition of adipose and glandular tissue, and for monoenergetic x-rays from 10 keV to 120 keV. These data were fit providing expressions for the normalized glandular dose. Using these expressions for the DgN coefficients and input variables such as the diameter, height and composition of the breast phantom, the mean glandular dose for any spectra can be estimated. A computer program to provide normalized glandular dose values has been made available online. In addition, figures displaying energy deposition maps are presented to better understand the spatial distribution of dose in CT breast imaging.     

A large-scale study of the ultrawideband microwave dielectric properties of normal, benign and malignant breast tissues obtained from cancer surgeries

The development of microwave breast cancer detection and treatment techniques has been driven by reports of substantial contrast in the dielectric properties of malignant and normal breast tissues. However, definitive knowledge of the dielectric properties of normal and diseased breast tissues at microwave frequencies has been limited by gaps and discrepancies across previously published studies. To address these issues, we conducted a large-scale study to experimentally determine the ultrawideband microwave dielectric properties of a variety of normal, malignant and benign breast tissues, measured from 0.5 to 20 GHz using a precision open-ended coaxial probe. Previously, we reported the dielectric properties of normal breast tissue samples obtained from reduction surgeries. Here, we report the dielectric properties of normal (adipose, glandular and fibroconnective), malignant (invasive and non-invasive ductal and lobular carcinomas) and benign (fibroadenomas and cysts) breast tissue samples obtained from cancer surgeries. We fit a one-pole Cole-Cole model to the complex permittivity data set of each characterized sample. Our analyses show that the contrast in the microwave-frequency dielectric properties between malignant and normal adipose-dominated tissues in the breast is considerable, as large as 10:1, while the contrast in the microwave-frequency dielectric properties between malignant and normal glandular/fibroconnective tissues in the breast is no more than about 10%.     

Development of a patient-specific two-compartment anthropomorphic breast phantom

The purpose of this paper is to develop a technique for the construction of a two-compartment anthropomorphic breast phantom specific to an individual patient's pendant breast anatomy. Three-dimensional breast images were acquired on a prototype dedicated breast computed tomography (bCT) scanner as part of an ongoing IRB-approved clinical trial of bCT. The images from the breast of a patient were segmented into adipose and glandular tissue regions and divided into 1.59?mm thick breast sections to correspond to the thickness of polyethylene stock. A computer-controlled water-jet cutting machine was used to cut the outer breast edge and the internal regions corresponding to glandular tissue from the polyethylene. The stack of polyethylene breast segments was encased in a thermoplastic 'skin' and filled with water. Water-filled spaces modeled glandular tissue structures and the surrounding polyethylene modeled the adipose tissue compartment. Utility of the phantom was demonstrated by inserting 200?μm microcalcifications as well as by measuring point dose deposition during bCT scanning. Affine registration of the original patient images with bCT images of the phantom showed similar tissue distribution. Linear profiles through the registered images demonstrated a mean coefficient of determination (r(2)) between grayscale profiles of 0.881. The exponent of the power law describing the anatomical noise power spectrum was identical in the coronal images of the patient's breast and the phantom. Microcalcifications were visualized in the phantom at bCT scanning. The real-time air kerma rate was measured during bCT scanning and fluctuated with breast anatomy. On average, point dose deposition was 7.1% greater than the mean glandular dose. A technique to generate a two-compartment anthropomorphic breast phantom from bCT images has been demonstrated. The phantom is the first, to our knowledge, to accurately model the uncompressed pendant breast and the glandular tissue distribution for a specific patient. The modular design of the phantom allows for studies of a single breast segment and the entire breast volume. Insertion of other devices, materials and tissues of interest into the phantom provide a robust platform for future breast imaging and dosimetry studies.     

Normalized glandular dose (DgN) coefficients for arbitrary X-ray spectra in mammography: computer-fit values of Monte Carlo derived data

Normalized glandular dose (DgN) values have been reported by several investigators for specific spectra, however for unconventional or unanticipated x-ray spectra considered for use in mammography, practical methods are not available for DgN computation. In this study, the previously validated SIERRA Monte Carlo code was used to compute the normalized glandular dose coefficients for monoenergetic energies from 8 keV to 50 keV. The overall mammography geometry used was a 65 cm source to image distance, a 1.2 cm air gap between the breast and the detector, and breast thicknesses ranging from 2 to 9 cm. A 4 mm layer of skin was also modeled, and semicircular breast radii of 8.5 cm and 10.0 cm were studied. Breast compositions of 0% glandular, 50% glandular, and 100% glandular were evaluated. The Monte Carlo derived DgN results demonstrated coefficients of variation less than 0.3%. The monoenergetic DgN values, DgN(E), were computer fit using commercial software and the best fit equations are reported. All fits resulted in r2 values of 0.9999 or better. The computer fit equations, along with easy to use spectral modeling routines, are available electronically on the web.     

Classification of breast computed tomography data.

Differences in breast tissue composition are important determinants in assessing risk, identifying disease in images and following changes over time. This paper presents an algorithm for tissue classification that separates breast tissue into its three primary constituents of skin, fat and glandular tissue. We have designed and built a dedicated breast CT scanner. Fifty-five normal volunteers and patients with mammographically identified breast lesions were scanned. Breast CT voxel data were filtered using a 5 pt median filter and the image histogram was computed. A two compartment Gaussian fit of histogram data was used to provide an initial estimate of tissue compartments. After histogram analysis, data were input to region-growing algorithms and classified as to belonging to skin, fat or gland based on their value and architectural features. Once tissues were classified, a more detailed analysis of glandular tissue patterns and a more quantitative analysis of breast composition was made. Algorithm performance assessment demonstrated very good or excellent agreement between algorithm and radiologist observers in 97.7% of the segmented data. We observed that even in dense breasts the fraction of glandular tissue seldom exceeded 50%. For most individuals the composition is better characterized as being a 70% (fat)-30% (gland) composition than a 50% (fat)-50% (gland) composition.