Breast-conservation treatment for early invasive breast cancer: prognostic factors for survival after salvage treatment of local recurrence

The aim of the study was to investigate prognosis of patients who develop an isolated local recurrence (ILR) after conservative surgery (CS) for early-stage invasive breast cancer. Between 1983 and 1987, 415 patients with stage I-II breast cancer were treated with CS. Of these patients, 68 developed an ILR. The mean follow-up time after ILR was 167 months. Cox models taking potential prognostic factors into account were used to estimate the risk of death. On univariate analysis, age (< or =40 vs. >40 years) at first treatment, time to ILR (< or =24 vs. >24 months), type of recurrence (true vs. new primary tumor, NP), and extent of recurrence (operable vs. inoperable) were, but initial tumor stage (pT1 vs. pT2), initial lymph node stage (pN-negative vs. -positive), adjuvant radiotherapy (yes vs. no), type of salvage surgery (wide excision vs. mastectomy), and recurrent tumor grade (1-2 vs. 3) were not significant predictors of post-recurrence survival. On multivariate analysis only time to ILR proved independent predictor of survival (relative risk: 3.2; 95% confidence interval: 1.4-7.3; p = 0.0051), and the age of the patients showed borderline significance (p = 0.0659). The 15-year actuarial rate of cause-specific survival after ILR was 58% for all patients, 60% and 0% for patients with operable or inoperable recurrence, 30% and 71% for patients with age < or =40 or >40 years, 25% and 72% for patients with time to ILR < or =24 or >24 months, 54% and 88% for patients with true recurrence or NP, and 92% for patients with age >40 years with NP, respectively. The rate of second local recurrence after salvage mastectomy or repeated wide excision was 16% and 28%, respectively (p = 0.2265). As a conclusion, many patients with ILR have good prognosis, particularly those with operable recurrence with prolonged time to ILR, or with NP. Salvage mastectomy is not mandatory for all CS patients.