Factors determining outcome for breast-conserving irradiation with margin-directed dose escalation to the tumor bed.

PURPOSE: A prospectively applied treatment policy for breast-conserving therapy used margin assessment as the exclusive guide to the intensity of therapy directed at the tumor-bearing quadrant. METHODS AND MATERIALS: From 1982-1994, there were 509 treated Stage I and II breast carcinomas with a median follow-up of 72 months. For operational purposes, tumor excision margins were prospectively defined as: > 5 mm, 2.1-5 mm, > 0 < or = 2 mm, and positive. If a margin was assessed as < or = 2 mm or indeterminate, and it was deemed cosmetically feasible, a reexcision of the tumor bed would be performed. All patients received whole breast irradiation to 50-50.4 Gy. The following scheme for tumor bed boost irradiation as a function of final margin status (FMS) was observed: (a) Minimal risk = no tumor found on reexcision, no boost performed; (b) low risk = FMS > 5 mm, boost of 10 Gy; intermediate risk = FMS 2.1-5 mm, boost to 14 Gy; high risk = FMS < or = 2 mm or positive, boost to 20 Gy. Cases were analyzed for local failure (LF) with respect to histology (invasive ductal (IDC), IDC with associated DCIS (IDC/DCIS), invasive lobular (ILC)), age, tumor size, total excision volume, reexcision, total dose, tamoxifen therapy, and chemotherapy.
RESULTS: There were 19 breast recurrences for a Kaplan-Meier local failure rate for all cases at 5 and 10 years of 2.7% and 7.1%, respectively. Local failure in the first 4 years of follow-up was rare, with a mean annual incidence rate of 0.25% that rose to a mean of 1.1% in subsequent years. Univariate results of Cox proportional hazards regression survival models found positive FMS (p = 0.02), IDC/DCIS (p = 0.04) and age (0.0006) as significantly associated with local failure. In a multivariable model of FMS and IDC/DCIS, FMS retained significance (p = 0.01) but IDC/DCIS was borderline (p = 0.06). When FMS and age were included in a multivariable model, there was a significant interaction (p = 0.01) between the two variables. There was a significant increase in the relative risk of LF for age < or = 45 years (range 11.1-17.4), irrespective of FMS category. Although excellent overall control rates were achieved for patients > 45 years, for younger patients LF rates appeared to remain proportional to the relative closeness of the FMS, despite rigorous dose escalation.
CONCLUSIONS: Graded tumor-bed dose escalation in response to FMS results in an exceptionally low risk of "early" local recurrence within the first 5 years of follow-up. However, this strategy is unable to completely overcome the longer term adverse influence of young age and positive FMS.