Purpose: To examine the long-term pattern and frequency of recurrences after breast-conserving therapy and whether the outcome was influenced by the era of treatment.
Methods and materials: From 1968 to 1986, 1870 patients with unilateral Stage I or II breast cancer were treated at the Joint Center for Radiation Therapy. Of these, 1628 underwent gross tumor excision and received a dose of > 60 Gy to the tumor bed and constituted the study population. Patients were classified as without evidence of disease, dead from other causes (DOC), or by their first site of recurrence. First sites of recurrent disease were categorized as distant/regional (DF/RNF) or local (LR). Local recurrence was defined as the detection of any invasive or in situ carcinoma occurring in the ipsilateral breast and was further categorized as: true recurrence (TR), marginal miss (MM), skin recurrence (S), or elsewhere in the breast (E). Median follow-up in survivors was 116 months. Eighty patients (4.9%) were lost to follow-up at 3-175 months. The population was divided into two time cohorts: 1968-1982 (n = 810), with a median follow-up time of 143 months, and 1983-1986 (n = 792), with a median follow-up time of 95 months.
Results: The overall crude rates of ipsilateral breast recurrence were 7.4 and 13.3% at 5 and 10 years, respectively. Crude rates at 5 and 10 years were 5.7 and 9.3% for TR/MM and were 0.9 and 2.8% for E recurrences, respectively. The annual incidence rates for all LR ranged from 0.5-2.4% and was relatively constant after the first year. The annual incidence rates for TR/MM ranged from 0.4 to 1.9%, whereas for E recurrences the range was 0.1-0.7%. The crude rates of DF/RNF were 16.6 and 23.1% at 5 and 10 years, respectively. The annual incidence rates for DF/RNF ranged from 1-5% over all years. Although the magnitude of the incidence was different, DF/RNF recurrence predominated in years 1-3 for both node-positive and node-negative patients. For the 1968-1982 and 1983-1986 cohorts, the 5-year crude rates of ipsilateral breast recurrence were 8.8 and 5.9%, respectively.
Conclusions: Distant and regional nodal failures were the predominant form of recurrence. The annual incidence rate of LR was relatively constant over the first decade. True recurrence/marginal miss was the most frequent type of ipsilateral breast recurrence and was highest during years 2 through 7. The risk of a recurrence elsewhere in the breast increased with longer follow-up and was highest during years 8 through 10. The 5-year crude rate of ipsilateral breast recurrence appeared lower in the 1983-1986 patient cohort compared to the 1968-1982 patient cohort (8.8% vs. 5.9%), but the distributions of site of first failure did not differ significantly (p = 0.13). Any decrease in ipsilateral breast recurrence likely reflects improvements in mammographic and pathologic evaluation, patient selection, and the increased use of reexcision.