Surveillance Mammography Failed to Lower Advanced Second Cancer Risk in Older Survivors
Among older women who had already lived through a first breast cancer, continued surveillance mammography did not translate into fewer advanced second primary tumors. The cohort was large, the follow-up was long, and the question was clinically practical: does routine imaging years after treatment meaningfully shift outcomes, or does it mainly add tests in a population with limited absolute risk?
The answer from this retrospective longitudinal analysis was sobering. Advanced second primary breast cancer remained uncommon, and the cumulative incidence was similar whether women kept up surveillance or stopped after the five-year mark. In an era when breast imaging is often treated as a default habit, these data push the discussion toward selectivity rather than reflex.
What the Medicare-SEER Cohort Revealed About Late Surveillance
Investigators drew from Surveillance, Epidemiology, and End Results Program-Medicare data and identified women with nonmetastatic breast cancer diagnosed from 2004 to 2008 at age 67 years or older. All had a surveillance mammogram at approximately 5 years after the initial diagnosis, which created a clinically relevant cohort of survivors already beyond the early high-risk window for recurrence. Follow-up continued until second breast cancer diagnosis, death, or the end of the study period in December 2021.
The analytic approach relied on cause-specific hazard models with inverse probability of treatment weighting, a design meant to reduce confounding in this nonrandomized setting. Women who continued surveillance mammography more than 5 years after the first diagnosis were compared with women who did not. The main outcome was stage-specific cumulative incidence of a second primary breast cancer, separated into in situ, localized, and advanced disease, with the advanced category combining regional and distant stage.
The study population included 9888 women. Mean age was 81.1 years, and the cohort was predominantly White, with smaller proportions of Hispanic and non-Hispanic Black women. Most initial cancers were in situ or localized, and most tumors were hormone receptor positive. Overall, 80.3% of the cohort continued surveillance at study entry, so the comparison reflected a common clinical pattern rather than an unusual exposure group.
Advanced second primary breast cancer was rare in both groups. The cumulative incidence was 1.65% among women who continued surveillance and 2.15% among those who did not, with a hazard ratio of 0.76 and a 95% confidence interval of 0.46 to 1.26. The confidence interval crossed 1, leaving no statistical evidence that ongoing mammography reduced the risk of advanced disease. The same pattern held in subgroups defined by hormone receptor status of the original tumor.
Why the Biology Makes Sense When Risk Is This Low
The logic behind surveillance mammography in survivors is straightforward: detect a new cancer early enough to prevent progression to regional or distant disease. In principle, repeated imaging should improve stage at diagnosis if a second primary tumor develops and grows silently over time. In practice, the benefit depends on how often dangerous second cancers arise, how quickly they become clinically evident, and whether the survivor has enough life expectancy to realize a downstream advantage from earlier detection.
This cohort suggests a narrow window for benefit. When advanced second cancers are uncommon to begin with, even a sensitive screening strategy may have little room to improve absolute outcomes. In older survivors, competing mortality, prior treatment burden, and the natural history of new breast lesions all shape whether surveillance changes anything that matters. A mammogram can find a lesion earlier, but if the likelihood of that lesion progressing to advanced stage is already low, the clinical payoff shrinks quickly.
The biology of hormone receptor status did not materially change the picture. Similar results in HR-positive and HR-negative groups imply that the absence of benefit was not limited to one tumor subtype. This matters because long-term survivor care often assumes that certain biologic features justify more aggressive imaging. Here, the data do not support a broad claim that late surveillance mammography prevents advanced second cancers across these common subgroups.
Absolute Risk, Not Habit, Should Drive the Imaging Conversation
The most important number in this study may be the small absolute difference between the groups. A 1.65% versus 2.15% cumulative incidence sounds measurable, but the event rate was so low that the practical impact of continued screening is limited. For older women, each additional mammogram also carries the burden of false positives, extra biopsies, anxiety, transport, and the cascade of follow-up testing that can follow an equivocal image.
These tradeoffs are especially relevant in women in their 80s, who made up much of this cohort. In that age range, the question is rarely whether imaging can detect a lesion; it is whether detection changes meaningful outcomes enough to justify ongoing surveillance. The present findings argue for individualized decisions anchored in life expectancy, comorbidity, prior cancer features, patient preferences, and tolerance for downstream intervention.
The results also fit a broader clinical reality: surveillance imaging is often easier to continue than to stop. Once a survivor enters a routine of annual or biennial mammography, the inertia can be strong, even when the expected benefit becomes uncertain. This analysis gives clinicians concrete evidence to revisit that routine rather than assuming it remains useful indefinitely.
What These Data Do and Do Not Prove
This was an observational cohort study, so causality cannot be claimed. Residual confounding may still be present despite weighting methods, and women who continued surveillance may have differed in ways that were not fully captured in claims and registry data. The cohort was also limited to women who had already survived to roughly five years after diagnosis and had a mammogram at that point, which means the findings apply to a specific survivorship phase rather than to all breast cancer survivors.
Even with those limitations, the consistency of the findings is hard to ignore. Advanced second primary breast cancer was uncommon, surveillance did not lower its risk, and the same pattern appeared across hormone receptor subgroups. For older survivors, that combination of low event rates and no detectable stage benefit should temper enthusiasm for automatic long-term imaging.
The study does not eliminate the possibility that selected patients could still benefit from surveillance mammography. It does, however, show that broad continuation of imaging in older women with prior breast cancer is unlikely to prevent advanced second primary disease at a population level. The next clinical step is not more routine testing, but more selective use of surveillance in women most likely to live long enough and remain healthy enough to benefit from it.