Breast cancer screening centers may want to consider implementing digital breast tomosynthesis (DBT) and digital mammography (DM) into their practice, according to results of a recent Radiology study. Pairing both modalities significantly increased the sensitivity and specificity of detecting breast cancers.
In the Oslo Tomosynthesis Screening Trial of more than 24,000 women with 281 cancers, pairing DBT with DM into population-based breast cancer screening resulted in a 70.5 percent sensitivity and 95 percent specificity, compared to 54 percent and 94 percent, respectively in DM alone.
Additionally, the predictive value of positive scores jumped when adding DBT, from 9.9 percent to 14 percent, and the predictive value of a negative score improved from 99.4 percent to 99.6 percent, reported first author Per Skaane, of the University of Oslo Breast Imaging Center in Norway, and colleagues.
The group also tested the diagnostic accuracy of adding computer-aided detection (CAD) to DM and how synthetic mammography paired with DBT impacted screening. Overall, the addition of CAD resulted in little difference to sensitivity and specificity, and synthetic mammography images combined with DBT also had little impact on the accuracy of screening.
“If SM could be used rather than DM with DBT, a screening examination could be performed at a radiation dose similar to that of DM,” Skaane et al. wrote. “There have been positive reports on the use of SM in screening but, to our knowledge, this study is the first large prospective trial comparing SM and DM in combination with DBT for the same set of cases.”
Importantly, a majority of cancers detected by adding DBT were small, low-grade and lymph-node-negative invasive cancers, which is aligned with the purpose of breast cancer screening, wrote Kristina Lång, of the diagnostic radiology and department of translational medicine at Lund University in Zurich Switzerland, in a related editorial.
Overall, Lång argued, the current study “strengthens” the assumption that DBT can improve breast cancer screening, but also questioned whether clinicians should rush to implement tomosynthesis and synthetic mammograms into breast cancer screening.
“We need to have evidence whether tomosynthesis improves the net benefit of screening before such a transition can be considered,” Lång wrote. “Analyzing interval cancer rates and tumor biology by pooling data from completed trials together with investigations on repeated tomosynthesis screening rounds could provide grounds for an implementation.”
Several large randomized breast tomosynthesis trials are already underway, including those in Germany, Italy the U.K. and United States. Once these are complete, Lång noted, we may have a final conclusion.
“Altogether, approximately half a million women will be enrolled in these trials. With these efforts, we will soon have a definitive answer.”