Pairing current reconstructed 2D images and digital breast tomosynthesis (DBT) is comparable to combining standard full-field digital mammography (FFDM) and DBT, according to research published in the June issue of Radiology.
Tomosynthesis has been implemented into many practices with FFDM due to its potential ability to simultaneously increase cancer detection rates, particularly for invasive ones, and reduce recall rates. However, concerns about increases in interpretation time and radiation dose have arisen. Research has been conducted to create alternate possibilities for DBT, one of which is pairing DBT images with 2D projection images synthetically reconstructed from DBT data. This option is desirable, as it eliminates the need for double radiation exposure. The quality of reconstruction has improved as well, leading to approved software for clinical use.
Lead author Per Skaane, MD, PhD, of the Olso University Hospital Ullevaal, and colleagues compared the performance of two versions of reconstructed 2D images in combination with DBT with the performance of standard FFDM in combination with DBT as part of a large prospective screening trial. Prospective examination of 24,901 women between the ages of 50 and 69 years was performed as part of the researchers’ screening trial. The trial was broken into two periods: period one, which ran from November 22, 2010, through December 21, 2011, and period two, which ran from January 30, 2012, to December 19, 2012. Two versions, one initial and one current, of reconstructed 2D images were used during the trial periods. They were compared for cancer detection and false-positive rates with corresponding FFDM plus DBT interpretations.
In period one, the cancer detection rate was 8.0 per 1,000 screened for FFDM plus DBT and 7.8 per 1,000 in period two. Cancer detection rates for initial reconstructed 2D images plus DBT were 7.4 per 1,000 screened in period one and 7.7 per 1,000 for current reconstructed 2D images plus DBT in period two. The false-positive scores were 5.3 percent, 4.6 percent, 4.6 percent and 4.5 percent, respectively. Corresponding reader-adjusted paired comparisons of false-positive scores indicated significant differences for period one but not for period two.
“The performance improvement noted in these studies suggest that one can use this improvement to find additional cancers, to eliminate the need for double readings, and/or for some combination of both (eg, double read only specific types of examinations, such as those in dense breasts),” wrote Skaane and colleagues. “The two-view combination of a synthesized 2D view plus DBT should be considered acceptable for routine use in mammographic screening,” they concluded.