Background: Few studies have used SEER-Medicare data to describe recurrence of breast cancer after primary treatment for U.S. women.
Methods: We used SEER-Medicare data to estimate the annual hazard rate (HR) of recurrence for women with breast cancer between 1991 and 1997 with 10 years of follow-up. The Kaplan-Meier method was used to derive the HR. Multivariate Cox proportional hazards model was used to estimate the relative hazard of the recurrence-associated prognostic factors.
Results: Of 20,027 women, 36.8% had recurrence within 10 years, with most of these recurrences (81.9%) occurring within 5 years after diagnosis. Women with stage III cancer showed the highest HR peak and largest magnitude than women with stage I or II disease (both P < 0.01) within the first 5 years. Women with negative tumor hormone receptor status had a higher peak hazard of developing recurrence within the first 5 years (P < 0.01), but the hazards were remarkably lower beyond 5 years of follow-up than in women with positive or unknown hormone receptor status (P > 0.05). Women with poorly differentiated histologic grade tumors showed higher HR in the first 5 years than women with other grades after primary treatment (both P < 0.01). The increased risk of recurrence of breast cancer was associated with advanced stage, moderate and poorly differently grades, and negative hormone receptor status (all P < 0.01).
Conclusion: The HRs of the recurrence are dynamic over 10 years and are markedly determined by prognostic factors at diagnosis.
Impact: Our study suggests that the optimal follow-up may differ among women.