Impact of false-positive mammography on subsequent screening attendance and risk of cancer

Abstract

Background: One area of concern within the largely successful UK National Health Service breast screening programme is the relatively high proportion of women showing mammographic abnormalities who undergo further diagnostic tests that prove negative. Previous studies suggest that, in addition to increasing anxiety, such false-positive mammography is associated with increased risk of subsequent interval cancer. In the present article, we quantify this increased risk, investigate whether it extends to cancers detected at rescreening, and determine whether cancers differ between women who have, and have not, experienced false-positive mammography.

Methods: This was a retrospective cohort study of 140,387 women aged 49-63 years routinely invited for first screening by the East Anglian National Health Service breast screening programme. Proportions reattending, and subsequent risk and pathological attributes of cancer were compared between women who underwent further (negative) assessment following false-positive mammography and women mammographically normal at first screen.

Results: At first screen, 108,617 (91.9%) of the screened women were mammographically normal, 4278 (3.6%) were assessed and then judged normal, and 514 (0.4%) underwent benign biopsy. Compared with nonassessed normal women, reattendance was lower among assessed women: 83.1% (95% confidence interval [CI], 82.0-84.1) versus 85.7% (95% CI, 85.5-85.9) (odds ratio [OR], 0.82; 95% CI, 0.76-0.89). Assessed women were at greater risk of interval cancer (rate per 1000 women screened, 9.6 [95% CI, 6.8-12.4] versus 3.0 [95% CI, 2.7-3.4]; OR, 3.19 [95% CI, 2.34-4.35]), and also of cancer detected at second screen (rate per 1000, 8.4 [95% CI, 5.8-10.9] versus 3.9 [95% CI, 3.5-4.3]; OR, 2.15 [95% CI, 1.55-2.98]). More cancers in assessed women measured >or = 20 mm (OR, 1.59; 95% CI, 0.99-2.55).

Conclusions: Women undergoing false-positive mammography at first screen were less likely to reattend for subsequent screens than were nonassessed women, yet they were more likely to develop interval cancers or cancers at second screen, and their cancers were larger. Factors predisposing for false-positive mammography require investigation. Women should be encouraged to continue with screening.