Format

Send to

Choose Destination
Eur Radiol. 2014 Mar;24(3):595-602.

Diagnostic accuracy of digital breast tomosynthesis versus digital mammography for benign and malignant lesions in breasts: a meta-analysis.

Abstract

OBJECTIVE:

To evaluate the diagnostic performance of digital breast tomosynthesis (DBT) and digital mammography (DM) for benign and malignant lesions in breasts.

METHODS:

Document retrieval was conducted on PubMed, EMBASE, the Cochrane Library, Web of Science and Chinese Biomedical Literature Database, etc., from 1950 to June 2013. Metadisc1.4 software was used to analyse the pooled sensitivity, specificity, diagnostic odds ratio (DOR), and positive and negative likelihood ratio. The heterogeneity was assessed using forest plots and the inconsistency index (I(2)). Before statistical comparison, the area under (AUC) the summary receiver-operating characteristic curve (SROC) of two different diagnostic methods was calculated respectively.

RESULTS:

A total of seven studies involving 2,014 patients and 2,666 breast lesions were included. Compared with the gold standard (histological results), the pooled sensitivity and specificity of DBT were 90.0% and 79.0%, and for DM they were 89.0% and 72.0%, respectively. The pooled positive likelihood ratio of DBT and DM was 3.50 and 2.83; the pooled negative likelihood ratio of DBT and DM was 15% and 18%; the pooled DOR for DBT and DM was 26.04 and 16.24, respectively.

CONCLUSIONS:

Digital breast tomosynthesis has a higher sensitivity and specificity in breast diagnosis than digital mammography.

KEY POINTS:

• Digital breast tomosynthesis has high sensitivity and specificity in breast diagnosis.• DBT appears to have superior diagnostic accuracy relative to digital mammography. • DBT images were captured at a lower dose than 2D images. • DBT displays abnormal features of lesions more clearly than DM. • Digital breast tomosynthesis could become the first choice for assessing breast lesions.

PMID:
24121712
DOI:
10.1007/s00330-013-3012-x
[Indexed for MEDLINE]

Supplemental Content

Full text links

Icon for Springer
Loading ...
Support Center