Time-course of local recurrence following conservative surgery and radiotherapy for early stage breast cancer.

Abstract

The time-course of local failure following conservative surgery and radiotherapy (S+RT) for early breast cancer is not well established. We therefore examined the time-course and patterns of breast recurrence as a first site of treatment failure in a group of 607 AJCC clinical Stage I or II invasive breast carcinomas treated from 1968-81. Sixty-seven patients had a breast failure (11%), corresponding to 5- and 10-year actuarial rates of 10% and 16%. The hazard rate (i.e., the risk per unit time of a failure) for any breast failure increased over the first 2 years, was fairly constant at about 2.5%/year over the period from 2 to 6 years after treatment, and then decreased to about 1%/year at 8 years. The majority of failures were at or near the primary tumor site (33 true recurrences (TR) and 15 marginal misses (MM). In addition there were 12 failures at sites at least several cm from the boosted volume (E), 6 recurrences in the skin without a parenchymal mass (S), and 1 patient with an unclassifiable failure. Recurrences at or near the primary site (TR/MM) occurred earlier (median 38.5 mo, range 12-87 mo) than recurrences at distant sites in the breast (E) (median 64.5 mo, range 26-90). The hazard rate for TR/MM increased over the first 2-1/2 years to reach approximately 2%/year, remained at that level till about 5 years after treatment, and then decreased to about 0.5%/year at 8 years following RT. By contrast, the hazard rate for E increased slowly with time to approximately 1%/year at 5 years, with little change in the rate after that time. We conclude that the time-course of the development of local recurrence after S+RT is protracted. The majority of failures appear at or near the primary tumor site; these are seen mainly in the first 7 years following RT. Recurrences at distant sites in the breast have an even more protracted time-course. Such recurrences are rare in the first 4 years following RT. Our results emphasize the need to obtain long follow-up in these patients, both to detect these recurrences promptly and to properly evaluate the results of S+RT.