A prospectively applied treatment policy for breast-conserving therapy used margin assessment as the exclusive guide to the intensity of radiation therapy directed at the tumor bed.

From 1982 to 1994, 498 women with 509 Stage I/II breast carcinomas with a median follow-up of 121 months were treated. Final margin status (FMS) categories were defined as greater than 5 mm, greater than 2-5 mm, greater than 0-2 mm, and positive. For margins less than or equal to 2 mm or indeterminate, reexcisions were performed if feasible. All patients received whole breast irradiation to 50.0-50.4 Gy. Final tumor bed boosts as a function of FMS were as follows: no residual on reexcision, no boost performed; FMS greater than 5 mm, boost of 10 Gy; FMS greater than 2-5 mm, boost of 14 Gy; FMS greater than 0-2 mm or positive, boost of 20 Gy. Cases were analyzed for local failure with respect to histology, age, tumor size, excision volume, reexcision, and total dose.

FMS was positive, greater than 0-2 mm, greater than 2-5 mm, and greater than 5 mm, and no residual tumor on reexcision in 21%, 20%, 17%, 14%, and 28% of cases, respectively. At 12 years, Kaplan-Meier local failure rates were 17% for FMS positive, 9% for FMS greater than 0-2 mm, 5% for FMS greater than 2-5 mm, 0% for FMS greater than 5 mm, and 6% for specimens without evidence of residuum on reexcision (P = 0.009). Patients 45 years old and younger had a 12-year local failure rate of 15% whereas patients older than 45 years had a 12-year local failure rate of 6% (P = 0.01). On multivariate analysis, young age (P = 0.03) predicted increased local failure rate, whereas margins that were less than or equal to 2 mm or positive predicted late (> 5 years) but not early (< or = 5 years) recurrence (P = 0.003).

Graded tumor bed dose escalation in response to FMS results in very low rates of local failure over the first 5 years for all FMS categories. However, tumors with close/positive margins have significantly increased local failure rates after 5 years of follow-up even with increased radiation boost dose. In addition, graded tumor bed dose escalation does not fully overcome the adverse influence of young age.