To the Editor,

I have a concern regarding a recently published article in European Radiology entitled 'Diagnostic accuracy of digital breast tomosynthesis versus digital mammography for benign and malignant lesions in breasts: a meta-analysis' [1]. The authors claim to investigate one-view digital breast tomosynthesis (DBT) versus digital mammography (DM) in a meta-analysis based on seven studies [2–8], but in fact, only four of the studies [2,3,7,8] investigate these particular imaging techniques. Out of these four studies, only one study showed significant improvement using one-view DBT in comparison with regular mammography (DM). That study material was enriched with subtle cases, more difficult than typically encountered in the clinical practice, with a relatively low frequency of DCIS. These factors may have increased the effect sizes. However, when the studies are stratified and analysed according to projection views, the pooled sensitivity and specificity may not correlate with the conclusions drawn, e.g. ‘the present study revealed that one-view DBT had higher sensitivity and specificity for the diagnosis of benign and malignant lesions in the breasts. These results illustrated the superior diagnostic accuracy of DBT relative to DM using meta-analysis’. Readers of the article might get the impression that one-view DBT alone is superior to DM in a general population of women, which may not be true. The following studies [4–6] did not investigate DBT in one view, but used other DBT techniques:

Michell et al. [6]: DM + SFM + DBT in two views versus DM;

Teerstra [4]: DBT in two views versus DM;

Gur et al. [5] DBT in two views versus DM.

Sincerely, Tony M. Svahn

P.S. I thank the authors for their thorough response.