Brought to you by

Reply to the Letter to the Editor: “Diagnostic accuracy of cone-beam breast computed tomography: a systematic review and diagnostic meta-analysis”

by Johannes Uhlig, Annemarie Uhlig, Lorenz Biggemann, Uwe Fischer, Joachim Lotz (johannes.uhlig@med.uni-goettingen.de)

Diagnostic accuracy of cone-beam breast computed tomography: a systematic review and diagnostic meta-analysis

Dear Editor,

With great interest we have read the “Letter to the Editor” by Yueqiang Zhu et al. regarding our earlier meta-analysis [1]. We appreciate the feedback given by the authors on minor inconsistencies in data-extraction, as provided in the Erratum to our study. However, none of these changed the results of the statistical analyses or the overall conclusion of the manuscript. ΔCT values were only provided by 2 included studies and were therefore not considered for statistical analyses in our manuscript. We fully agree with Yueqiang Zhu et al. that the ambiguous term “positive CBBCT” has no coherent definition in the literature, which could impact diagnostic measures. However, providing a comprehensive analysis of varying cutoff values for definition of “positive CBBCT” was beyond the scope of our study.

We appreciate that Yueqiang Zhu and colleagues highlight their recently published CBBCT studies [2, 3] and a study by Jung et al. from 2017 [4]. These studies would have met the criteria for inclusion in our meta-analysis: yet, as detailed in our manuscript, the systematic literature search has been conducted in April 2018, and therefore did not identify the aforementioned studies with publication in late 2018 [2, 3] and MEDLINE listing in May 2018 [4]. An update of our meta-analysis is planned once studies have been published providing a relevant change in the pooled estimates of diagnostic CBBCT accuracy.

Specific responses to comments/Erratum:
1. The total number of patients included in our study was n = 559 from 6 studies (39 + 235 + 120 + 41 + 59 + 65), given the conservative estimate of n = 120 for He et al.
2. agreed
3. agreed
4. While n = 212 is true to be the overall sample size, we simplified to n = 120 in the CE-CBBCT subgroup. Differences in sample size will, however, not affect the provided point-estimates and confidence intervals.
5. agreed
6. agreed
7. agreed
8. agreed
9. agreed (given changes mentioned above)