Diagnostic yield of surveillance mammography among older women.

BACKGROUND Mammographic surveillance is common among older women with a history of breast cancer, but little is known about imaging outcomes in this population. The goal of this study was to describe short-term outcomes after surveillance mammography among older women, including use of subsequent imaging, cancer diagnosis, and positive predictive value. METHODS This was a longitudinal cohort study using SEER-Medicare data. We followed women diagnosed with breast cancer between 2003-2007 at age 67 or older until diagnosis of a second primary breast cancer, death, or end of follow up (2019). We used claims to identify surveillance mammograms. Outcomes included subsequent imaging, second primary breast cancer diagnosis, and positive predictive value after surveillance. Outcomes were identified based on proximity to surveillance mammography. We stratified analyses by age and time since diagnosis. RESULTS The cohort included 33, 607 women who received 160,637 mammograms over a median 12.1 years. Subsequent imaging was performed after 115 per 1000 mammograms (95% CI 114-117) and was similar across ages and time since diagnosis. Second primary breast cancer was diagnosed after 7.1 per 1000 mammograms (95% CI 6.7-7.5) and was similar across ages but generally higher beginning 5 years post diagnosis. Positive predictive value among women with subsequent imaging ranged from 3.7 to 9.2%. Most second primary breast cancers were either in-situ (23.3%) or stage I (57.0%). CONCLUSIONS Although surveillance mammography may detect breast cancer even years after diagnosis, surveillance also leads to substantial subsequent testing. Whether continued surveillance at older ages improves health outcomes is uncertain.