Digital mammography screening: how many breast cancers are additionally detected by bilateral ultrasound examination during assessment?

Abstract

To assess the positive predictive values of incremental breast cancer detection (PPV1) in relation to the mammographic breast density and of performed biopsies (PPV3) resulting from supplemental bilateral physician-performed whole-breast ultrasound (US) at recall of a population-based digital mammography screening programme. A total of 2,803 recalled screening participants (50-69 years), who had additional bilateral US with prospectively completed documentation [grading of breast density (ACR 1-4), biopsy recommendation related to US and mammography], were included. The PPV1 of supplemental cancer detection only by US was 0.21 % (6/2,803) compared to 13.8 % (386/2,803) by mammography. The PPV1 of US-only cancer detection was 0 %, 0.16 % (2/1,220), 0.22 % (3/1,374) and 1.06 % (1/94) for women with breast density of ACR 1, ACR 2, ACR 3 and ACR 4, respectively. The PPV3 of US-only lesion detection was 33.3 % (9/27) compared to 38.0 % (405/1,066) by mammography. The proportion of invasive cancers no larger than 10 mm was 37.5 % (3/8) for US-only detection compared to 38.4 % (113/294) for mammographic detection. Bilateral ultrasound at recall, in addition to the assessment of screen-detected mammographic abnormalities, resulted in a low PPV of incremental cancer detection only by US, without a disproportional increase of false-positive biopsies.