Radiology
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

DOI: 10.1148/radiol.2232010482
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Shepherd, J. A.
Right arrow Articles by Cummings, S. R.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Shepherd, J. A.
Right arrow Articles by Cummings, S. R.
(Radiology 2002;223:554-557.)
© RSNA, 2002

Technical Developments

Measurement of Breast Density with Dual X-ray Absorptiometry: Feasibility1

John A. Shepherd, PhD, Karla M. Kerlikowske, MD, Rebecca Smith-Bindman, MD, Harry K. Genant, MD and Steve R. Cummings, MD

1 From the Department of Radiology, Osteoporosis and Arthritis Research Group (J.A.S., H.K.G.) and Department of Medicine and Epidemiology and Biostatistics (K.M.K., S.R.C.), University of California at San Francisco, 350 Parnassus Ave, Suite 607, San Francisco, CA 94143; Veterans Affairs Medical Center, San Francisco, Calif (K.M.K.); and Department of Radiology, UCSF/Mt Zion Medical Center, San Francisco, Calif (R.S.B.). From the 2000 RSNA scientific assembly. Received February 19, 2001; revision requested March 28; final revision received November 28; accepted December 10. Supported by Breast Cancer Research Program Concept Award DAMD17-00-1-0612 from the Department of Defense and by UCSF Academic Senate Committee on Research. Address correspondence to J.A.S. (e-mail: john.shepherd@oarg.ucsf.edu).


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 Materials and Methods
Results
Discussion
 REFERENCES
 
Dual x-ray absorptiometry (DXA) was used to quantify breast density with a phantom and with cadaveric breasts. With DXA, percentage of fat correlated with percentage of glandular density of the phantom (r > 0.998) and with density at mammography (radjusted = 0.83). DXA precision (SD) was 0.5% without and 1.1% with breast repositioning. DXA devices can be used to accurately and precisely estimate breast tissue density.

© RSNA, 2002

Index terms: Breast • Breast neoplasms, 00.30 • Cancer screening • Phantoms


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
Materials and Methods
Results
Discussion
 REFERENCES
 
Perhaps the least recognized factor for breast cancer risk is breast density. Other than age, breast density has been shown to be one of the strongest indicators of breast cancer risk. Women who have greater than 50% of total breast area that is mammographically dense have a three- to fivefold greater risk for breast cancer than women with less than 25% mammographically dense breasts (1–4). Recently, breast density has been linked to specific biological processes in the breast that give rise to histologic features (5), such as atypical hyperplasia and carcinoma in situ, that are known to be related to increased breast cancer risk. For these reasons, breast density may be an important measure to monitor in clinical drug trials and epidemiologic cancer risk studies.

Breast density was initially described with a semi-quantitative classification system that categorized breast density into one of four categories by taking into account the quantitative (amount) and qualitative (diffuse or pronounced ductal structures and dense parenchymal patterns) nature of the density (3,4,6,7).

A more quantitative approach is to measure the area of mammographically dense breast relative to the total projected breast area. In this article, we will refer to this as mammographic density (3,8,9). Mammographic density (10) is a quantitative continuous grading from 0% to 100% measured by means of delineating the radiographically dense areas on the mammogram from the entire breast area and providing a percentage of breast density. The percentage of breast density is calculated as follows: (high radiographic density area)/(total breast area) x 100. Although quantitative, this method is limited by the fact that films are calibrated for optical density, not mass density, and a unique threshold of breast density is selected by a reader for each image. In addition, the total and dense projected areas will change on the basis of the amount of breast compression. The reproducibility (coefficient of variation) of delineating dense regions combined with patient repositioning errors is generally approximately 5% or more (11). Therapy with tamoxifen, which reduces cancer risk, decreases breast density by 4.3% per year in patients with cancer (12). Thus, the sensitivity of the percentage of breast density in the prediction of risk of breast cancer or in the detection of response to therapy may be similar to that of categorical methods and insufficient to monitor therapeutic changes in breast density for individuals.

There are compelling reasons to use dual x-ray absorptiometry (DXA) techniques to measure breast composition: (a) DXA is the reference standard for measuring whole-body composition because of its low radiation dose and high accuracy and precision (13). (b) The precision and accuracy of DXA have been characterized in small animals that are less than 600 g (14), which is similar to the size of a human breast. (c) The technique does not require a subjective interpretation of results. (d) Breast compression is not required. (e) The technique is readily available throughout the world for measuring bone density and diagnosing osteoporosis. Our specific goals for this study were to calibrate a commercially available DXA device to measure breast glandular density, to quantify in vitro precision by using cadaveric breasts, and to compare our measurements with conventional mammographic density measurements.


   Materials and Methods
TOP
 ABSTRACT
 INTRODUCTION
 Materials and Methods
Results
Discussion
 REFERENCES
 
Devices
For this study, a DXA densitometer (Hologic QDR-4500A; Hologic, Bedford, Mass) was used. The software version was V9.10, and scans were acquired with the small animal/rat whole body scanning protocol. The scanning area was 18 x 36 cm2 with a pixel size of 1.0 x 1.5 mm. Low- and high-energy images were acquired at 100 and 140 kVp, respectively, with the scanner. The entrance dose was 0.3 mGy (30 mR). Values from the densitometer were recorded as percentage of fat, total fat, total lean, and total mass. The objects were placed on the table and scanned with the x-ray gantry in the standard vertical position.

Phantom Measurements
For this study, a phantom (M17; CIRS, Norfolk, Va) was used as a breast density calibration tool. This model is a density-step phantom of constant thickness that simulates different ratios of breast glandular tissue and adipose fat (Fig 1). This phantom is an approximate atomic equivalent to adipose fat and breast glandular tissue, as reported by Hammerstein et al (15). The phantom’s attenuation coefficients are within 1% of their respective fat-gland ratios from 10 to 200 keV. The density ranged from 0% to 100% glandular density in six steps. The inner clear acrylic section was not included in our comparison, since acrylic is not a stable representation of tissue across a wide x-ray energy range. The phantom was scanned 10 times with the DXA scanner without repositioning, and the average percentage of fat value from each density step was determined.



View larger version (123K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1. Density-step phantom used to calibrate the DXA device to units of percentage of glandular density. Phantom has a six-step range in percentage of glandular density from 0% (pure adipose fat) to 100% (pure breast glandular tissue).

 
Cadaveric Breasts
Four whole cadaveric breast pairs (eight breasts) were examined by using both DXA and mammography. The breasts were excised whole, and all residual muscle tissue was removed. Each breast was scanned by using DXA twice without repositioning and once after flipping the breast 180° on the table to simulate repositioning owing to a second visit. The breasts were positioned in as close to a craniocaudal view as could be constructed. The precision of the DXA scanner was defined by the variance (PROC GLM; SAS Institute, Cary, NC) of the breast tissue scans without repositioning. The precision associated with repositioning, an estimate of the expected in vivo precision for follow-up examinations, was defined with PROC GLM in the same way by using the first scan and the last scan with repositioning.

For mammographic density, the films of the cadaveric breasts were acquired with a mammography machine (Sensorgraphe DMR; GE Medical Systems, Waukesha, Wis). The films were read by a trained radiologist, and the mammographically dense regions were delineated on the film with a wax pencil. The films were digitized at 100-µm spatial resolution with a digitizer (Lumisys 200; Lumisys, Sunnyvale, Calif). The mammographic density was then quantified by a research assistant with a workstation designed by Swarnakar et al (16) by tracing the pencil lines with a cursor.


   Results
TOP
 ABSTRACT
 INTRODUCTION
 Materials and Methods
Results
Discussion
 REFERENCES
 
Calibration of Percentage of Glandular Density
The percentage of glandular density as defined by the phantom was compared with the percentage of fat measured with the DXA scanner. Each step was analyzed by using an 8.7-cm2 region of interest. Figure 2 shows a plot of the reported percentage of glandular density of the phantom and the percentage of fat measured with the phantom for each density step. The relationship can be expressed with the following equation: percentage of glandular density = -0.627 x percentage of fat at DXA + 72.5. The slope of -0.627 shows a compression of the range of the percentage of glandular density to that of the percentage of fat such that 100% fat is 9.8% glandular density and 0% fat is 72.5% glandular density.



View larger version (41K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 2. Plot shows results of calculation of 1 minus the percentage of fat measured with the DXA device and the percentage of glandular density reported with the phantom. Error bars are shown but do not extend beyond the dimension of the plot symbol.

 
Characteristics of Cadaveric Breasts
The characteristics of cadaveric breasts are shown in Table 1. There was a wide range of values of total mass, fat mass, lean mass, and percentage of fat. The precision of the DXA scanner was characterized by using the cadaveric breast without repositioning. Table 2 presents a summary of the precision results with the DXA scanner. The precision (determined with the SD) is approximately 1 g for a mean mass of 1 kg, or 0.1% coefficient of variation. By using the previously noted equation to convert percentage of fat to percentage of glandular density, the precision of the percentage of glandular density is 0.5% SD, with a mean value of 32%. The precision results with repositioning in cadaveric breasts are shown in Table 3. The SD for all values increased slightly. Thus, the overall precision was limited by repositioning more so than by x-ray noise. The cadaveric breast mammographic density and the percentage of breast glandular tissue density were correlated with an r value adjusted for the sample size, or radjusted = 0.83. However, because the number of samples was small, the regression analysis relationship was highly influenced by each data point.


View this table:
[in this window]
[in a new window]

  		TABLE 1. Ranges of Mass and Percentage of Fat in Cadaveric Whole Breasts

 

View this table:
[in this window]
[in a new window]

  		TABLE 2. Precision Results with DXA Device without Repositioning in Cadaveric Breasts

 

View this table:
[in this window]
[in a new window]

  		TABLE 3. Precision Results with DXA Device with Repositioning in Cadaveric Breasts

 

   Discussion
TOP
 ABSTRACT
 INTRODUCTION
 Materials and Methods
Results
Discussion
REFERENCES
 
We found that the percentage of fat of breast tissue measured with DXA was highly correlated and linearly related to the percentage of breast glandular tissue density measured with a standard mammographic phantom. In addition, the percentage of breast density was moderately to highly correlated with the percentage of glandular density measured by using DXA.

It is of interest that the percentage of breast glandular tissue density does not equal one minus the percentage of fat. This is most likely, since the percentage of fat is measured relative to a two-compartment model of fat and muscle, not fat and glandular tissue. Even so, this relationship should result in a slope difference between the percentage of fat and the percentage of breast glandular tissue density. The offset is most likely caused by the DXA densitometer being calibrated to the in vivo four-compartment model of body composition. With this model, underwater weighing is used to derive body density, which has a known offset to absolute standards of fat and lean tissue (17). We expect the in vivo precision to be similar to the cadaveric precision with repositioning, since flipping the breast 180° would be a worst-case repositioning error.

There are other methods available to estimate body fat, but only DXA (18), computed tomography (CT) (19), and magnetic resonance (MR) imaging (20) can be used to measure the tissue composition of isolated body regions. Lee et al (20) reported a 2% accuracy of segmenting breast fat from glandular and ductal tissue in phantoms by using whole-breast MR imaging. The sections were individually segmented into two compartments, and measurements in all sections were summed to calculate a total percentage of breast fat. In 40 women, the SD of the mean value for the group was 18% compared with 30% for mammographic density in the same women, although the techniques were correlated (r = 0.6). This suggests that mammographic density is related to segmented compositional density but with a variance that is influenced by nondensity features.

CT can provide a precise measure of tissue composition calibrated to electron density or absolute references. Kalef-Ezra et al (19) described the normal breast electron density from CT volume scans in pre- and postmenopausal women. However, the whole-organ radiation dose with CT limits its usefulness as a screening tool. Neither technique, to the authors’ knowledge, has been used to quantify cancer risk on the basis of breast density.

Dual-energy mammographic imaging is not a new concept and has been used for selecting calcifications and for improving imaging contrast (21–25). Breitenstein and Shaw (26) and Shaw and Plewes (27) reported on theoretical calculations of signal-to-noise ratios for single- versus dual-energy mammography to quantify tissue density with idealized phantoms. However, there is a substantial effort necessary to generate precise and accurate quantitative DXA images with standard mammography equipment (ie, filtering, poor dynamic range of film, availability of digital mammography units, x-ray tube stability). The widespread use of digital detectors and the replacement of x-ray film should make DXA imaging more feasible with standard digital mammography machines.

Our study had several limitations. First, there were only a small number of cadaveric breasts available for our estimates of precision and the regression analysis statistics. Second, we used a conventional DXA device optimized for bone density and whole-body composition measurements. In contrast to CT and the more specialized scanning modes possible with digital mammography machines, the DXA images acquired in this study have no diagnostic value beyond their use in determining tissue density and mass. Last, choosing alternative DXA energies may improve the technique’s tissue selectivity even further.

In conclusion, findings of this study show that conventional DXA devices can be calibrated to measure the percentage of glandular density; with DXA, breast density can be quantified to approximately 1% precision limited principally by repositioning. The agreement between mammographic density and the percentage of breast glandular tissue density was moderately to highly correlated. Thus, compositional densitometry may be more accurate and precise than mammographic density for quantifying breast cancer risk. However, it has not been demonstrated whether compositional breast density measured by using any technique is more predictive or discriminating than mammographic density in determining breast cancer risk. In vivo studies to quantify the percentage of breast glandular tissue density and cancer risk are warranted.


   ACKNOWLEDGMENTS
 
The authors acknowledge the UCSF Ambulatory Care Center and Department of Anatomy for their participation in the data collection.


   FOOTNOTES
 
Abbreviation: DXA = dual x-ray absorptiometry

Author contributions: Guarantor of integrity of entire study, J.A.S.; study concepts, J.A.S., K.K., S.R.C.; study design, J.A.S.; literature research, J.A.S., K.M.K.; experimental studies, J.A.S.; data acquisition, J.A.S., R.S.B.; data analysis/interpretation, R.S.B., K.M.K., J.A.S.; statistical analysis, J.A.S.; manuscript preparation, J.A.S., K.M.K.; manuscript definition of intellectual content, S.R.C., K.M.K., J.A.S., H.K.G.; manuscript editing, K.M.K.; manuscript revision/review, H.K.G., S.R.C., R.S.B.; manuscript final version approval, all authors.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 Materials and Methods
Results
Discussion
 REFERENCES
 

  1. Boyd NF, Byng JW, Jong RA, et al. Quantitative classification of mammographic densities and breast cancer risk: results from the Canadian National Breast Screening Study. J Natl Cancer Inst 1995; 87:670-675.[Abstract/Free Full Text]
  2. Byrne C, Schairer C, Wolfe J, et al. Mammographic features and breast cancer risk: effects with time, age, and menopause status. J Natl Cancer Inst 1995; 87:1622-1629.[Abstract/Free Full Text]
  3. Wolfe JN, Saftlas AF, Salane M. Mammographic parenchymal patterns and quantitative evaluation of mammographic densities: a case-control study. AJR Am J Roentgenol 1987; 148:1087-1092.[Abstract/Free Full Text]
  4. Saftlas AF, Szklo M. Mammographic parenchymal patterns and breast cancer risk. Epidemiol Rev 1987; 9:146-174.[Free Full Text]
  5. Boyd NF, Jensen HM, Cooke G, Han HL, Lockwood GA, Miller AB. Mammographic densities and the prevalence and incidence of histological types of benign breast disease: Reference Pathologists of the Canadian National Breast Screening Study. Eur J Cancer Prev 2000; 9:15-24.[CrossRef][Medline]
  6. Wolfe JN. Breast patterns as an index of risk for developing breast cancer. AJR Am J Roentgenol 1976; 126:1130-1137.[Abstract]
  7. American College of Radiology. Illustrated breast imaging reporting and data system (BI-RADS) 3rd ed. Reston, Va: American College of Radiology, 1998.
  8. Boyd NF, O’Sullivan B, Campbell JE, et al. Mammographic signs as risk factors for breast cancer. Br J Cancer 1982; 45:185-193.[Medline]
  9. Brisson J, Merletti F, Sadowsky NL, Twaddle JA, Morrison AS, Cole P. Mammographic features of the breast and breast cancer risk. Am J Epidemiol 1982; 115:428-437.[Abstract/Free Full Text]
  10. Byng J, Boyd N, Fishell E, John R, Yaffe M. The quantitative analysis of mammographic densities. Phys Med Biol 1994; 39:1629-1638.[CrossRef][Medline]
  11. Sivaramakrishna R, Obuchowski NA, Chilcote WA, Powell KA. Automatic segmentation of mammographic density. Acad Radiol 2001; 8:250-256.[CrossRef][Medline]
  12. Chow CK, Venzon D, Jones EC, Premkumar A, O’Shaughnessy J, Zujewski J. Effect of tamoxifen on mammographic density. Cancer Epidemiol Biomarkers Prev 2000; 9:917-921.[Abstract/Free Full Text]
  13. Laskey MA. Dual-energy x-ray absorptiometry and body composition. Nutrition 1996; 12:45-51.[CrossRef][Medline]
  14. Makan S, Bayley HS, Webber CE. Precision and accuracy of total body bone mass and body composition measurements in the rat using x-ray-based dual photon absorptiometry. Can J Physiol Pharmacol 1997; 75:1257-1261.[CrossRef][Medline]
  15. Hammerstein GR, Miller DW, White DR, Masterson ME, Woodard HQ, Laughlin JS. Absorbed radiation dose in mammography. Radiology 1979; 130:485-491.[Abstract]
  16. Swarnakar V, Prevrhal P, Kerlikowske K, Cummings S, Genant HK, Shepherd JA. A mammographic density reading service for clinical drug trials (abstr). Radiology 2000; 217(P):701.[Abstract/Free Full Text]
  17. Kelly T. Whole body enhancements: free software upgrades available for QDR-4500A and QDR-4500W users with and without body composition option In: QDR Insights Newsletter: New Developments in Bone Mineral Measurement. Vol 7. Bedford, Mass: Hologic, 1996; 15.
  18. Sutcliffe JF. A review of in vivo experimental methods to determine the composition of the human body. Phys Med Biol 1996; 41:791-833.[CrossRef][Medline]
  19. Kalef-Ezra JA, Karantanas AH, Koligliatis T, Boziari A, Tsekeris P. Electron density of tissues and breast cancer radiotherapy: a quantitative CT study. Int J Radiat Oncol Biol Phys 1998; 41:1209-1214.[CrossRef][Medline]
  20. Lee NA, Rusinek H, Weinreb J, et al. Fatty and fibroglandular tissue volumes in the breasts of women 20–83 years old: comparison of x-ray mammography and computer-assisted MR imaging. AJR Am J Roentgenol 1997; 168:501-506.[Abstract/Free Full Text]
  21. Johns PC, Drost DJ, Yaffe MJ, Fenster A. Dual-energy mammography: initial experimental results. Med Phys 1985; 12:297-304.[CrossRef][Medline]
  22. Asaga T, Masuzawa C, Yoshida A, Matsuura H. Dual-energy subtraction mammography. J Digit Imaging 1995; 8(1 suppl 1):70-73.[Medline]
  23. Johns PC, Yaffe MJ. Theoretical optimization of dual-energy x-ray imaging with application to mammography. Med Phys 1985; 12:289-296.[CrossRef][Medline]
  24. Chakraborty DP, Barnes GT. An energy sensitive cassette for dual-energy mammography. Med Phys 1989; 16:7-13.[CrossRef][Medline]
  25. Asaga T, Chiyasu S, Mastuda S, et al. Breast imaging: dual-energy projection radiography with digital radiography. Radiology 1987; 164:869-870.[Abstract/Free Full Text]
  26. Breitenstein DS, Shaw CC. Comparison of three tissue composition measurement techniques using digital mammograms: a signal-to-noise study. J Digit Imaging 1998; 11:137-150.
  27. Shaw CG, Plewes DB. Effects of scattered radiation and veiling glare in dual-energy tissue-bone imaging: a theoretical analysis. Med Phys 1987; 14:956-967.[CrossRef][Medline]



This article has been cited by other articles:


Home page
 RadiologyHome page
P. R. Bakic, A.-K. Carton, D. Kontos, C. Zhang, A. B. Troxel, and A. D. A. Maidment
Breast Percent Density: Estimation on Digital Mammograms and Central Tomosynthesis Projections
Radiology, July 1, 2009; 252(1): 40 - 49.
[Abstract] [Full Text] [PDF]

Home page
 Cancer Epidemiol. Biomarkers Prev.Home page
J. A. Shepherd, S. Malkov, B. Fan, A. Laidevant, R. Novotny, and G. Maskarinec
Breast Density Assessment in Adolescent Girls Using Dual-Energy X-ray Absorptiometry: A Feasibility Study
Cancer Epidemiol. Biomarkers Prev., July 1, 2008; 17(7): 1709 - 1713.
[Abstract] [Full Text] [PDF]

Home page
 RadiologyHome page
J. A. Harvey and V. E. Bovbjerg
Quantitative Assessment of Mammographic Breast Density: Relationship with Breast Cancer Risk
Radiology, January 1, 2004; 230(1): 29 - 41.
[Abstract] [Full Text] [PDF]

Home page
 JNCI J Natl Cancer InstHome page
E. Ziv, J. Shepherd, R. Smith-Bindman, and K. Kerlikowske
RESPONSE: Re: Mammographic Breast Density and Family History of Breast Cancer
J Natl Cancer Inst, November 19, 2003; 95(22): 1726 - 1727.
[Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Shepherd, J. A.
Right arrow Articles by Cummings, S. R.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Shepherd, J. A.
Right arrow Articles by Cummings, S. R.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
RADIOLOGY RADIOGRAPHICS RSNA JOURNALS ONLINE