Table 11:
Positive Predictive Value of Screening by Mammography and Ultrasound Compared With Mammography Alone in Women at High Risk for Breast Cancer
| Author, Year | Round | PPV, % (95% CI) | ||
|---|---|---|---|---|
| Mammography | Mammography + Ultrasound | Absolute Difference Between Groups | ||
| Positive Test: BI-RADS 4, 5 | ||||
| Riedl et al, 201541 | All | 28.3 (18, 41.6) | 26 (17.5, 36.7) | −2.3 (NR) |
| Kuhl et al, 201038 | All | 39.1 (20.4, 61.2) | 32.5 (19.1, 49.2) | −6.6 (NR) |
| Sardanelli et al, 201142 | All | 71.4 (53.7, 85.4) | 55.6 (41.4, 69.1) | −15.8 (NR) NSa |
| Kuhl et al, 200539 | All | 23.7 (13.6, 36.6)b | 11.9 (7.5,17.6)b | −11.8 (NR) |
| Positive Test: BI-RADS 3, 4, 5 | ||||
| Berg et al, 201236, | 1 | 6.5c (4.0, 9.9) | 4.8c (3.4 to 6.7) | −1.7c (−3.7, 0.1); P = .07 |
| 2,3 | 8.6c (6.2, 11.6) | 7.0c (5.4, 9.0) | −1.6c (−3.1, −0.2); P = .04 | |
| Berg et al, 201236, | 1 | 29.2d (18.6, 41.8) | 11.4d (7.9 to 15.8) | −17.8d (−26.7, −9.3); P < .001 |
| 2,3 | 38.1d (28.5, 48.6) | 16.2d (12.5, 20.6) | −21.9d (−28.7, −14.7); P < .001 | |
Abbreviations: BI-RADS, Breast Imaging Reporting and Data System, CI, confidence interval; NR, not reported; NS, not significant; PPV, positive predictive value.
The authors stated that the PPV for all included screening tests ranged from 53% to 71%, without statistically significant differences. It is unclear if all test combinations were considered and compared.
Calculated based on data provided in article
Calculated as the malignancy rate among cases that test positive (recommended for further testing, short-interval follow-up or biopsy) on screening.
Defined by authors as the malignancy rate among women with a positive screening test who underwent biopsy of the same lesion. These values could include biopsy resulting from a BI-RADS 3 diagnosis.