In community practice in the U.S., data have shown that around 90% of abnormal screening mammograms do not end in a cancer diagnosis, or are false positive (FP). The long term goal of this project is to reduce FP results in screening mammography, both to reduce the burden of extra unnecessary imaging work-ups and unnecessary biopsies.
Our specific aims are to: 1) identify the characteristics of women, mammographic examinations, radiologists and practices that discriminate a FP from a TP mammogram; and 2) identify the descriptive characteristics of abnormalities seen on mammograms that will improve the ability to discriminate a FP from TP mammography result when added to the findings of Aim 1. This study will use the infrastructure of the Carolina Mammography Registry (CMR), a population-based registry that collects prospective data on women and mammography examinations in mammography facilities in North Carolina. Data routinely collected by CMR that can be used for this study include demographic and health history on the women and the results and recommendations from imaging studies performed in the screening work-up. Mammography data are linked to cancer outcome data from the NC Central Cancer Registry and pathology laboratories, providing outcome data form the screening work-up. An active review of the films of positive mammograms identified from CMR will be conducted by outside reviewers, to classify the findings on the mammograms in a standardized way. The routinely collected CMR data will be combined with the descriptive data from the film review to test the ability to discriminate a FP from TP result, using GEE logistic analyses. In addition, other variables that could affect the prediction of a FP or TP that will be used in analyses include: mammography practice procedures; radiologist characteristics; and practice setting and type. The goal is to be improve the ability to identify false positive results, to ultimately reduce recall rates, increase positive predictive value, without a decrease in cancer detection rates. Work that has the potential to reduce the extra imaging studies and procedures that women experience in screening for breast cancer would go a long way toward making screening mammography more cost effective in terms of the monetary, physical and psychological costs to women, and the economic costs to the health care system. ? ? ?
| Yankaskas, Bonnie C; May, Ryan C; Matuszewski, Jeanine et al. (2011) Effect of observing change from comparison mammograms on performance of screening mammography in a large community-based population. Radiology 261:762-70 |
| Halladay, Jacqueline R; Yankaskas, Bonnie C; Bowling, J Michael et al. (2010) Positive predictive value of mammography: comparison of interpretations of screening and diagnostic images by the same radiologist and by different radiologists. AJR Am J Roentgenol 195:782-5 |
