From 1962 to 1992, 279 patients with recurrent breast cancer were treated. Of these, the cases with resected lesions, imperfectly evaluated cases and cases where information was lacking were excluded, and 185 recurrent breast cancers were evaluable at the first line or subsequent treatment. Sixty-seven (36.2%) out of 185 cases responded to first line treatment, while the remaining 118 cases did not respond. In the 67 responder cases at the first line treatments, 16 (23.9%) responded to second line or third line, however, in the 118 nonresponder cases, only 13 (11.0%) responded to next or subsequent treatment. The incidence of responder in the second line or third line treatment was significantly higher in the first line responder group than in the first line nonresponder group. Furthermore, overall survival of first line responders was significantly better than that of first line nonresponders. There was no significant difference in the survival after recurrence between responders and nonresponders in the first line chemotherapy, chemoendocrine treatment or radiotherapy, however, a significant difference was seen between the two groups in the first line endocrine treatment. These results suggest that first line treatment may select the treatment sensitive (especially, truly hormone-dependent) recurrent breast cancers which show a better prognosis.