| Outcome | Comparison or Biopsy Method | Overall Rating | Key Findings and Comments |
|---|---|---|---|
| Test performance of individual biopsy methods | Freehand | Low | – Sensitivity: 0.91 (0.80 to 0.96) – Specificity: 0.98 (0.95 to 1.00) |
| Ultrasound, automated | Moderate | – Sensitivity: 0.99 (0.98 to 0.99) – Specificity: 0.97 (0.95 to 0.98) | |
| Ultrasound, vacuum-assisted | Moderate | – Sensitivity: 0.97 (0.92 to 0.99) – Specificity: 0.98 (0.96 to 0.99) | |
| Stereotactically guided, automated | Moderate | – Sensitivity: 0.97 (0.95 to 0.98) – Specificity: 0.97 (0.96 to 0.98) | |
| Stereotactically guided, vacuum-assisted | Moderate | – Sensitivity: 0.99 (0.98 to 0.99) – Specificity: 0.92 (0.89 to 0.94) | |
| MRI-guided, automated | Insufficient | – Sensitivity: 0.90 (0.57 to 0.99) – Specificity: 0.99 (0.91 to 1.00) | |
| MRI-guided, vacuum-assisted | Insufficient | – Sensitivity: 1.00 (0.98 to 1.00) – Specificity: 0.91 (0.54 to 0.99) | |
| Comparison of test performance among alternative biopsy methods | Ultrasound-guided, automated vs. vacuum-assisted | Low | – Difference in sensitivity: 0.01 (−0.01 to 0.06) [no difference] – Difference in specificity: −0.01 (−0.03 to 0.01) [no difference] |
| Stereotactically guided, automated vs. vacuum-assisted | Low | – Difference in sensitivity: −0.02 (−0.04 to −0.01) [vacuum-assisted is better] – Difference in specificity: 0.05 (0.02 to 0.08) [automated is better] | |
| MRI-guided, automated vs. vacuum-assisted | Insufficient | – Difference in sensitivity: −0.10 (−0.43 to −0.01) [vacuum-assisted is better] – Difference in specificity: 0.07 (−0.03 to 0.43) [no difference] | |
| Modifiers of test performance for women at average and high risk of breast cancer | All biopsy methods | Insufficient | – Few studies provided within-sample information for each modifier of interest; meta-regression results rely on cross-study comparisons so consistency of effects cannot be assessed – Within-study (direct) evidence was sparse; between study evidence relied on indirect comparisons across studies – In meta-regression analyses CrIs were wide; extreme odds ratio values were often observed because sensitivity and specificity for all tests were very close to 1 (see Results) |
CrIs = credible interval; MRI = magnetic resonance imaging.
From: Discussion
NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.