Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.

RSNA: van Colen: Breast Density


Published on

  • Be the first to comment

RSNA: van Colen: Breast Density

  1. 1. Variation in Reported Breast Density Among Radiologists Stephanie van Colen DO, Carol Hulka MD, Janet Baum MD
  2. 2. Introduction The evaluation of breast density is an important part of evaluating a mammogram. It has long been recognized that the denser the tissues, the more difficult it is to obtain optimal mammograms and to visualize abnormalities within the breast. (1,2,3) The ACR, in its mammography lexicon, (4) has recommended that density be estimated by the radiologist and included in the mammography report to provide the clinician with an understanding of some of the limitations in evaluating the breasts on mammography.
  3. 3. <ul><li>Dense breast tissue limits evaluation </li></ul>
  4. 4. Introduction The ACR lexicon is very similar to an older method of describing breast density by Dr. John Wolfe in the era of xeromammography and early film mammography where he described the densities of the breasts as fatty (N1), P1 or P2 types of density or dysplastic (the most dense breast tissue). (5)
  5. 5. Introduction 40 years ago, Dr. Wolfe also proposed that density was related to the incidence of breast cancer with the incidence of breast cancer increasing with each level of density. This theory was not proven for many years. Recently several studies have shown a definite and significant relationship between overall breast density and the incidence of breast cancer across all ages. (6,7)
  6. 6. Introduction For the relationship between dense breast tissue and the increased risk of cancer to be proven, breast density definitions must be standardized between film and digital mammography including processed or enhanced digital mammography. One of the limitations in proving this theory is how breast density is determined.
  7. 7. Purpose Interpreting radiologists determine breast density during the review of mammograms. However, there is no standardized method for estimating breast density.
  8. 8. Purpose <ul><ul><li>The ACR recommends describing breast density using criteria of the 4 established basic density groups in the mammography report: </li></ul></ul><ul><ul><li>relatively fatty/fatty (less than 25% dense) </li></ul></ul><ul><ul><li>scattered fibroglandular tissues (26-50% dense) </li></ul></ul><ul><ul><li>heterogeneously dense (51-75% dense) </li></ul></ul><ul><ul><li>extremely dense (76-100% dense) </li></ul></ul>
  9. 9. Purpose <ul><ul><li>The ACR density groups: </li></ul></ul><ul><ul><li>relatively fatty (less than 25% dense) </li></ul></ul><ul><ul><li>scattered fibroglandular tissues (26-50% dense) </li></ul></ul>
  10. 10. Purpose <ul><ul><li>The ACR density groups: </li></ul></ul><ul><ul><li>heterogeneously dense (51-75% dense) </li></ul></ul><ul><ul><li>extremely dense (76-100% dense) </li></ul></ul>
  11. 11. Purpose In our study, we proposed that radiologists’ density estimation will range only 1 to 2 categories for most patients. We also proposed that there would be agreement among the majority of the radiologists on most exams.
  12. 12. Methods and Materials <ul><li>Eight radiologists reviewed 25 to 50 previously interpreted mammograms at a time, for a total of 250 exams. </li></ul><ul><li>Using a preprinted answer sheet, the radiologists selected one tissue density option for each exam from the ACR recommended categories. </li></ul>
  13. 13. Methods and Materials For the study we pulled digital screening mammograms from PACS at our institution ’ s three campuses from August 2009 to May 2010 for review. The studies were viewed by all radiologists on a GE mammography reading station with flat panel monitors in the same reading room.
  14. 14. Methods and Materials The initial 150 exams consisted of a total of 1 cranial caudal view and 1 mediolateral oblique view of each breast. We did this to simplify the review of each exam for the participating radiologists who were reviewing 25-50 exams at a time. We decided to increase the number of total images per exam to 6 for the final 100 studies, so that we would have a more representative sample, specifically including both some of the larger fatty breasts and some dense breasts which frequently have one additional view of each breast.
  15. 15. Methods and Materials <ul><li>We determined: </li></ul><ul><li>the density assigned to each exam by each radiologist noting how many readers assigned the same density to a given exam </li></ul><ul><li>how many exams varied by greater than 1 density category </li></ul><ul><li>how many exams varied by greater than 2 density categories </li></ul><ul><li>the range of densities assigned in the study </li></ul><ul><li>the discrepancy between assigned fatty or scattered fibroglandular tissues and heterogeneously dense or dense categories </li></ul>
  16. 16. Results We found great variability in breast density reporting among the radiologists in our study. In approximately 28% of exams there was total agreement (8/8) on density. In approximately 46% of exams a large majority (7/8, 24% and 6/8, 22 %) agreed on density.    In 16% of exams, 5 of 8 radiologists agreed. The remaining 10% of mammograms had 50% or less concordance of density reporting.
  17. 17. Results
  18. 18. Results In approximately 32% of cases, there was some radiologist disagreement between the categories > 50% density (dense and heterogeneously dense) and < 50% density (fatty and scattered fibroglandular tissue) . This disagreement may have included only one radiologist who described the breasts as either denser or less dense than the all of the other radiologists. In some of the cases more than one radiologist assigned a density greater than 50% on the same exam that others assigned a density less than 50%.
  19. 19. Results This is one case where there was discrepancy between scattered fibroglandular tissue and heterogeneously dense.
  20. 20. Results <ul><li>Assigned densities: </li></ul><ul><li>fatty - 83 (33%) </li></ul><ul><li>scattered fibroglandular tissues - 129 (52%) </li></ul><ul><li>heterogeneously dense - 35 (14%) </li></ul><ul><li>dense - 3 (1%) </li></ul><ul><li>If there was even assignment of 2 different categories, (4/4), we arbitrarily assigned the breast density as the lower density category. For example, in 6 cases there was even assignment of scattered fibroglandular and heterogeneously dense, and the category assigned was scattered fibroglandular. </li></ul>
  21. 21. Results
  22. 22. Results In 66 % of cases, there was variation by one adjacent density category by at least one radiologist. Three different adjacent density categories were assigned by various radiologists in 16 cases (6%). In no exam were all 4 density categories selected.
  23. 23. Discussion Many of the radiologists commented that they began to give more thought to estimating breast density. Some noted that the most difficult decisions included discerning between heterogeneously dense and scattered fibroglandular tissue. This distinction is important, because it affects a patient’s screening and diagnostic recommendations and risk assessment for developing breast cancer. Recent legislation in Connecticut states that patients must be made aware of their breast density on screening mammograms. If a patient has dense breast tissue, further imaging with ultrasound or MRI may be indicated, and insurance must cover the screening ultrasound in most cases.
  24. 24. Discussion Breast density is not yet incorporated into risk stratification models, such as the Gail model, (8,9) for assessing a woman’s risk for breast cancer. Larger studies are needed to validate the risk factors associated with breast density, and standardization of density estimation is necessary to establish the extent of risk.
  25. 25. Conclusion There is greater variation in breast density reporting among radiologists than we anticipated. It is difficult to detect early breast cancer in mammograms of patients with dense breast tissue. The estimated breast density of each mammogram may affect a woman’s risk cancer assessment and therefore screening and follow-up recommendations. Standardization of breast density reporting is necessary and may be achieved through additional radiologist education and/or use of computer breast density estimation software.
  26. 26. Conclusion There are several computer aided density estimation programs under development and some that are already commercially available but not yet routinely used which may help standardize breast density determinations. Software can be more easily used now that more and more practices have converted to digital mammography.
  27. 27. Thank you <ul><li>Janet Baer MD </li></ul><ul><li>Arthur Chang MD </li></ul><ul><li>Gregory Harrington MD </li></ul><ul><li>Joseph Sequeira MD </li></ul><ul><li>Franklin Zweiman MD </li></ul>
  28. 28. References <ul><li>1 Norman F. Boyd, M.D., et al., Mammographic Density and the Risk and Detection of Breast Cancer , New Engl J Med 2007;356:227-36. </li></ul><ul><li>2 Kerlikowske K, Grady D, Barclay J, Sickles EA, Ernster V. Effect of age, breast density, and family history on the sensitivity of first screening mammography. JAMA 1996;276:33-38.  </li></ul><ul><li>3 Mandelson MT, Oestreicher N, Porter PL, et al. Breast density as a predictor of mammographic detection: comparison of interval- and screen-detected cancers. J National Cancer Inst. 2000;92:1081–1087. </li></ul><ul><li>4 The ACR BIRAD Committee, The ACR Breast Imaging Reporting and Data System (BI-RADS). 4th ed. Reston, Va.: American College of Radiology, 2003. </li></ul><ul><li>5 Wolfe, J.N., Breast Patterns as an Index of Risk for Developing Breast Cancer , American Journal of Roentgenology, 126:1130-1139, 1976 </li></ul><ul><li>6 G. J. R. Porter, A. J. Evans, E. J. Cornford, H. C. Burrell, J. J. James, A. H. S. Lee, and J. Chakrabarti Influence of Mammographic Parenchymal Pattern in Screening-Detected and Interval Invasive Breast Cancers on Pathologic Features, Mammographic Features, and Patient Survival Am. J. Roentgenol., March 1, 2007; 188(3): 676 - 683. </li></ul>
  29. 29. References <ul><li>7 Jeffrey A. Tice, Steven R. Cummings, Rebecca Smith-Bindman, Laura Ichikawa, William E. Barlow, and Karla Kerlikowske, Using Clinical Factors and Mammographic Breast Density to Estimate Breast Cancer Risk: Development and Validation of a New Predictive Model, Ann Intern Med March 4, 2008 148:337-347 </li></ul><ul><li>8 William E. Barlow, Emily White, Rachel Ballard-Barbash, et al., Prospective Breast Cancer Risk Prediction Model for Women Undergoing Screening Mammography, Journal of the National Cancer Institute, Sept 2006 98(17):1204-1214; doi:10.1093/jnci/djj331 </li></ul><ul><li>9 Palomares MR, Machia JR, Lehman CD, Daling JR, McTiernan A. Mammographic density correlation with Gail model breast cancer risk estimates and component risk factors . Cancer Epidemiol Biomarkers Prev, July 2006;15(7). </li></ul>