False positive recall rate statistical model for digital breast tomosynthesis
The accuracy of digital breast tomosynthesis (DBT) has been shown in many clinical studies to be superior to digital mammography in helping radiologists identify early stage breast cancer. DBT reduces the effects of overlapping tissues that camouflage very small lesions. In part because of this, false positive recall rates tend to be higher. But by how much?
A Swedish research team conducting the 15,000-women Malmö Breast Tomosynthesis Screening Trial (MBTST) also created an easy-to-understand statistical model for clinical use. It identifies the factors that are statistically associated with unnecessary recalls. Their study, published in the October issue of The Breast, is based on data of the first 7,500 women recruited.
The primary objective of the MBTST population-based study is to compare the performance of one-view DBT as a single screening modality to two-view digital mammography. The trial’s preliminary findings of the results of the first 7,500 participants showed that the detection rate of early stage breast cancer was better with images of a one-view DBT than two-view digital mammography. Sixty-eight cancers were detected, 67 by DBT and 47 by digital mammography. Three-hundred-fifty-two women were recalled for additional imaging.1
They determined that the most important factor influencing a recall was breast density. Women with a BI-RADS breast density of 3 or 4 had a 4.8% chance of recall with DBT and 3.5% with digital mammography. By comparison, women with a breast density of 1 or 2 had a 2.1% and 1.4% chance of recall respectively.2
The authors of the study observed that because the diagnostic accuracy was better for DBT compared to digital mammography for experienced mammographers, there is the possibility that one-view DBT could replace conventional two-view digital mammography screening. They note that additional large studies with less experienced radiologists would be needed to confirm this.
They concluded that “the main conclusion is that both imaging modalities have limitations for women with dense breasts. The results [presented in the article] provide an important piece of information to be considered in the discussion about implementing DBT in breast cancer screening.”
REFERENCES
- Lång K, Andersson I, Rosso A, et al. Performance of one-view breast tomosynthesis as a stand-alone breast cancer screening modality: results from the Malmö Breast Tomosynthesis Screening Trial, a population-based study. Eur Radiol. Published online May 1, 2015 (Open Access).