Diagnostic Performance and Accuracy of the 3 Interpreting Methods of Breast Strain Elastography: A Systematic Review and Meta-analysis.

Diagnostic Performance and Accuracy of the 3 Interpreting Methods of Breast Strain Elastography: A Systematic Review and Meta-analysis.

Barr Richard G; De Silvestri Annalisa; Scotti Valeria; Manzoni Federica; Rebuffi Chiara; Capittini Cristina; Tinelli Carmine

Journal of ultrasound in medicine : official journal of the American Institute of Ultrasound in Medicine

2018

2018-11

There are 3 methods of interpreting breast strain elastography: the elastographic-to-B-mode length ratio (E/B), a 5-point color scale (5P), and the strain ratio (SR). This meta-analysis assessed which method is superior to the others. A systematic search of the medical literature was performed in July 2017. Studies were eligible for inclusion if they fulfilled the following criteria: (1) had biopsy-proven or long-term stability as the reference standard; (2) used either the E/B, 5P, or SR to interpret results; and (3) had at least 50 cases. A total of 220 records were retrieved; 60 full-text articles were examined, and 46 were included in the meta-analysis. Publication years ranged from 2007 and 2017. The quality of studies was generally high. The mean age of women was 48 years; 12,398 lesions (4242 malignant) were analyzed. For the 5P method, the sensitivity was 77%; specificity, 87%; positive likelihood ratio (LR), 5.3; and negative LR, 0.24. For the SR method, sensitivity was 87%; specificity, 81%; positive LR, 4.8; and negative LR, 0.16. For the E/B method, sensitivity was 96%; specificity, 88%; positive LR, 7.1; and negative LR, 0.03. Of the 3 methods, the E/B had the highest sensitivity, and the E/B and 5P had the highest specificity. With a negative LR of 0.03, the E/B method can downgrade lesions with a pretest probability of 50% to a 2% probability of malignancy.

breast; breast cancer; breast tumors; elastography; sonoelastography

Article information provided for research and reference use only. All rights are retained by the journal listed under publisher and/or the creator(s).