Prospective evaluation of carcinoembryonic antigen levels and alternating chemotherapeutic regimens in metastatic breast cancer.

Ninety-seven eligible and evaluable women with metastatic breast cancer were placed on a prospective clinical protocol to evaluate the use of continuous cyclic therapy with dibromodulcitol, doxorubicin, vincristine, tamoxifen, and fluoxymesterone (DAVTH) v DAVTH alternating with cyclophosphamide, methotrexate, 5-fluorouracil, and prednisone (CMFP); and the use of pretreatment and serial carcinoembryonic antigen (CEA) levels in these patients. Continuous DAVTH and DAVTH/CMFP were equivalent therapies with respect to response rates, time to treatment failure (TTF), and survival. Pretreatment CEA levels were elevated (greater than 5 ng/mL) in 42/97 patients and less than 5 ng/mL in the remaining patients. Patients with elevated pretreatment CEA levels were more likely to be estrogen receptor (ER) positive (P = .006), to have prolonged disease-free intervals (P = .017), to have hepatic (P = .004) and/or osseous (P = .01) metastases, and to have multiple sites of metastatic disease (P = .004). Pretreatment CEA levels did not significantly predict for overall response rates, TTF, or survival; nonetheless, those patients with low pretreatment CEA levels had more complete responses (CRs) (16/55 v 4/42; P = .02). Serial CEA levels during therapy revealed a number of interesting patterns. During the first 4 months of treatment, serial CEA levels in responding patients either (1) progressively declined (15/29 women with elevated pretreatment CEA levels), or (2) initially rose significantly (mean, 243% of pretreatment value) and then declined (14/29 women with elevated pretreatment CEA levels). Peak CEA levels in the latter patients were seen 27 to 135 days following initiation of cytotoxic therapy. In some patients the initial increase in the CEA level was incorrectly interpreted as evidence of impending disease progression. CEA levels frequently increased around the time of clinical disease progression. However, rising CEA levels rarely provided a clinically meaningful lead time before the appearance of other clinical evidence of disease progression. These data suggest that routine pretreatment and monthly serial CEA levels in metastatic breast cancer patients have minimal use in clinical practice. Two further noteworthy findings were observed in this prospective study. First, patients with an unknown ER status had a prolonged median survival when compared with patients with ER positive or negative tumors; this appeared to be related to prolonged disease-free intervals in ER unknown patients. Second, two case of secondary acute leukemia were seen in patients treated with continuous DAVTH therapy.