Background:Mammography has been found effective as the primary screening test for breast cancer, however, there is no clear evidence about the effectiveness of clinical examination. In Finland, collecting self-reported symptoms information andinspection of breasts by the radiographer are the part of mammography screening program. We estimated the cumulative probability of false positive screening test resultswith respect to symptom history reported at screen.
Methods: A historical prospective cohort study was done using individual screening data from 413,611 women aged 50-69 years with2,627,256 invitations for mammography screening between 1992 and 2012 in Finland. Symptoms (lump, retraction and secretion) were reported at56,805 visits and48,873 visits resulted in a false positive mammography result. Generalized linear models were used to estimate the probability of at least one false positive test and true positive at screening visits. The estimates were compared among women with and without symptoms history.
Results: The estimated cumulative probabilities were18% and 6% for false positiveand true positive results respectively. In women with a history of a lump, the cumulative probabilities of false positivetest and true positive were45% and 16% respectively, compared to 17% and 5% with no reported lump. In women with a history of any given symptom, the cumulative probabilities of false positive test and true positive were 38% and 13%. Likewise, women with a history of a ‘lump and retraction’ had the cumulative false positive probabilityof 56%.
Conclusion:The study showed higher cumulative risk of false positive tests and more cancers detected in women who reported symptoms compared to women who did not report symptoms at screen. The risk varies substantially, depending on symptom types and characteristics. Information on breast symptoms influences the balance of absolute benefits and harms of screening.