A comparison of cyclophosphamide,adriamycin, and 5-fluorouracil (CAF) and cyclophosphamide,methotrexate, 5-fluorouracil,vincristine, and prednisone (CMFVP) in patients with advanced breast cancer

Summary The Southeastern Cancer Study Group, in a prospectively randomized study involving patients with advanced breast cancer, has compared a low dose intermittently administered five-drug regimen including cyclophosphamide, methotrexate, 5-fluorouracil, vincristine, and prednisone (CMFVP) with an aggressively administered three-drug regimen including cyclophosphamide, doxorubicin, and 5-fluorouracil (CAF). CAF induced more responses and more complete responses and a longer duration of disease control. However, only a marginal difference was demonstrated in overall survival between the two regimens (p = <0.10).Patients with a good risk pattern of metastases, i.e. those with nodular local chest wall recurrence, nodular pulmonary metastases, or bone-only metastases were more likely to achieve a response and survive longer than those with a poor risk pattern, i.e. lymphangitic pulmonary metastases, pleural effusion with chest wall ulceration, or widespread metastases, including hepatic. All of the differences between the two regimens were noted in the good risk pattern groups. CAF provided no additional benefit to patients presenting with poor risk patterns of metastases.Adriamycin-containing combinations may achieve a greater degree of tumor cell kill in certain subsets of patients with advanced breast cancer, but this provides only a marginal increase in survival. Address for reprints: Richard V. Smalley, M.D., National Cancer Institute, Frederick Cancer Research Facility, Building 426, Rm. 1, Frederick, MD 21701, USA.     

CAP (cyclophosphamide,adriamycin, platinum) vs CMFVP (cyclophosphamide,methotrexate, 5-fluorouracil,vincristine, prednisolone) combination chemotherapy in untreated metastatic breast cancer

Summary The prospective controlled phase III clinical trial compared the therapeutic value of the cis-platinum — adriamycin — cyclophosphamide combination (CAP) and that of the combination of cyclophosphamide, methotrexate, 5-fluorouracil, vincristine, and prednisolone (CMFVP) in untreated metastatic breast cancer. Seventy-two patients (>2 cycles) were evaluated: 36 had received CAP and 36, CMFVP. An objective response (CR+PR) to CAP combination chemotherapy was achieved in 75% of patients (27 of 36), with a high rate (42%) of complete remissions. In terms of metastatic site, the response rate appeared to be particularly high in soft tissue and visceral organ (lung, liver) metastases. In the CMFVP group, an objective response was noted in 16 of 36 patients (44%) with 19% complete remissions. Overall therapeutic response and the complete remission rate were better with CAP regimen (statistically significant; P<0.01). The duration of remissions was 4–16+months (M=12) for CAP and 2–12+months (M=8) for CMFVP. Toxic side-effects were more pronounced in the CAP group, particularly myelosuppression, and anemia was prevalent. Side-effects of CMFVP treatment were mild. In 11 CMFVP-resistant cases CAP was administered as second-line treatment, and an objective response was observed in 45% of cases (5 of 11). The preliminary results of this controlled trial show the advantage of the CAP combination in the treatment of metastatic breast cances.     

The Eastern Cooperative Oncology Group experience with cyclophosphamide, adriamycin, and 5-fluorouracil (CAF) in patients with metastatic breast cancer

Data on 162 women (90 premenopausal and 72 postmenopausal) with metastatic breast cancer randomized to receive cyclophosphamide, Adriamycin (doxorubicin) and 5-fluorouracil (CAF) on two Eastern Cooperative Oncology Group (ECOG) protocols were analyzed. Twenty-three percent had complete remission; 39% had partial remission; 28% had no change; and 3% had disease progression. Of those patients in whom receptors were known, response rates were 65% for estrogen (ER)-receptor positive and 70% for ER-negative patients. The median duration of response was 11.4 months. The median survival time from the start of CAF was 20.2 months. The response rate, time to treatment failure (TTF), and median survival time were superior in the premenopausal women. These differences ceased, however, to be statistically significant in logistic models. Factors significantly associated with longer TTF and longer survival were as follows: one or two organs with metastases (TTF, P less than 0.0001; survival, P less than 0.0001); dominant site other than soft tissue (TTF, P less than 0.0001; survival, P = 0.05); and an initial good performance status (TTF, P = 0.007; survival, P = 0.02). Patients with ER-positive disease had a significantly longer median survival time (P = 0.003).     

Low dose chemotherapy of metastatic breast cancer with cyclophosphamide, adriamycin, methotrexate, 5-fluorouracil (CAMF) versus sequential cyclophosphamide, methotrexate, 5-fluorouracil (CMF) and adriamycin.

Seventy-eight advanced breast cancer patients with hormone-resistant disease or visceral metastases were randomized to receive either of two low dose regimens consisting of cyclophosphamide (C), methotrexate (M), 5-fluorouracil (F), and Adriamycin (A) as their initial chemotherapy. One group was treated with CAMF, and the other with CMF until progression, followed by A (CMF leads to A). C was given at 50 mg/m2, po, days 1-14; M at 20 mg/m2, F at 300 mg/m2, and A at 20 mg/m2, iv, days 1 and 8 of each 28-day cycle. The response rates for CAMF vs. CMF did not differ significantly (complete and partial responses-62% vs. 49%; stabilizations-23% vs. 31%). Responses by site of metasis, median times to progression and median survivals were similar for both groups. Poor and good risk partial responders had similar survivals. Twelve percent of CMF patients treated with Adriamycin at the time of progression had partial responses with an associated improved survival. Since CMF is as effective as CAMF, but has less toxicity, low dose therapy with CMF is more acceptable than CAMF as an initial chemotherapy regimen for metastatic breast cancer. Adriamycin may be reserved for subsequent regression induction.     

A phase III randomized trial of cyclophosphamide,mitoxantrone, and 5-fluorouracil (CNF) versus cyclophosphamide,adriamycin, and 5-fluorouracil (CAF) in patients with metastatic breast cancer

Summary One hundred patients with metastatic breast cancer were randomly selected to receive combined chemotherapy treatment with adriamycin (50 mg/m²) or mitoxantrone (12 mg/m²) associated with 5-fluorouracil (600 mg/m²) and cyclophosphamide (600 mg/m²) administered intravenously every 21 days with a maximum of ten cycles. All patients included in this study were under 75 years of age and had ECOG performance status of less than 4. They had not been treated previously with chemotherapy for metastatic disease. Patients treated with adjuvant chemotherapy, which could not have included anthracyclines, had to have relapsed at least 12 months after the completion of therapy. There were no statistically significant differences in pretreatment characteristics or metastatic disease location between the two groups. Ninety-four patients were assessable for response. No differences were observed in response rate or in survival between the groups. The response rate (complete response (CR) and partial response (PR)) was 68% (13% CR and 55% PR for CAF; 0% CR and 68% PR for CNF). Median survival for all patients was 19 months (18 months with CAF and 19 months with CNF). All patients were assessable for toxicity. There were no differences in gastrointestinal and cardiac toxicity. More grade I-II hematologic toxicity episodes (p < 0.001) and treatment delays (p = 0.05) due to leucopenia were observed with the CNF group, and more grade III alopecia (p < 0.001) was observed with the CAF group. Patients received further therapeutic manoeuvres after finishing the study with a sequential treatment consisting of hormonal therapy and chemotherapy with mitomycin (M) -vinblastine (Vbl) (M 10 mg/m² day 1, Vbl 5 mg/m² days 1, 15 and 29; maximum 5 cycles). This chemotherapy treatment was received by 32 patients, with a response rate of 34% and grade III-IV hematologic toxicity of 37%. Treatment with CNF can be considered a good alternative to CAF for first-line treatment of metastatic breast cancer. M-Vbl treatment is useful as second-line treatment in patients with prior adriamycin exposure.     

5-Fluorouracil, adriamycin, cyclophosphamide (FAC) vs. 5-fluorouracil, epirubicin, cyclophosphamide (FEC) in metastatic breast cancer

94 evaluable patients with metastatic breast cancer were randomly assigned to 5-fluorouracil, adriamycin, and cyclophosphamide (FAC) or 5-fluorouracil, epirubicin, and cyclophosphamide (FEC), with cycles repeated every 3 weeks. The objective response rate to FAC was 46% versus 44% to FEC. There was no significant difference in the median duration of response and median survival for the two regimens. Toxicity was more frequent and more pronounced in patients receiving FAC. Results indicate therapeutic equivalence of the two regimens and reduced toxicity of the epirubicin arm.     

A randomized trial of cyclophosphamide, doxorubicin, and prednisone versus cyclophosphamide, 5-fluorouracil, and prednisone in patients with metastatic breast cancer

Ninety-four patients were entered in a clinical trial assessing the clinical activity of cyclophosphamide, doxorubicin, and prednisone (CAP) versus a combination of cyclophosphamide. 5-Fluorouracil, and prednisone (CFP) in patients with advanced breast cancer. Objective response rates were comparable, 49% for CFP and 46% for CAP. There was no statistical difference between the duration of response of the two regimens or in time to progression. Most importantly, survival differences were not apparent. Both regimens were clinically tolerable and toxicities, for the most part, were comparable. Thus, no therapeutic advantage existed for either of these polychemotherapy regimens in patients with advanced breast cancer.     

Dose-intense weekly cyclophosphamide, methotrexate, 5-fluorouracil, vincristine and prednisolone (CMFP) in advanced breast cancer

Weekly chemotherapy with cyclophosphamide 80 mg m-2 day-1 p.o. continuously, methotrexate 35 mg m-2 week-1 i.v., 5-fluorouracil 500 mg m-2 week-1 i.v., vincristine 1.4 mg m-2 i.v. every two weeks and prednisolone 20 mg m-2 day-1 p.o. continuously (CMFVP) was prospectively studied in 45 previously untreated outpatients with advanced breast cancer to determine the feasibility of delivering a dose-intense regimen. Of 40 evaluable patients, complete response (CR) occurred in one patient, partial response (PR) in 20 (CR + PR 53%), stable in eight, progression in 11 and five were unevaluable for response. The median relapse-free survival for responders was 25 weeks and median survival for all patients was 31 weeks. The mean dose intensity relative to the Cooper regimen fell from 1.02 to 0.6 within the first 4 weeks of treatment and the median dose intensity achieved for all patients on study was only 0.52. Eighty-seven per cent of patients had treatment delays with a mean of 3.9 delays per patient and 71% had dose reductions. Neutropenia was the major toxicity with WHO grade 3 or 4 neutropenia (less than 1.0 x 10(9) l-1) in 62% of patients and three septic deaths while neutropenic. Dose-intense weekly CMFVP in this schedule cannot be delivered to previously untreated outpatients with advanced breast cancer.     

An evaluation of the effect of vincristine added to cyclophosphamide, 5-fluorouracil,methotrexate, and prednisone in advanced breast cancer

Summary A multi-institutional randomized clinical trial was carried out to evaluate the effect of vincristine (V) added to cyclophosphamide, methotrexate, 5-fluorouracil, and prednisone (CMFP) for the treatment of metastatic breast cancer. There were 427 patients entered into the study and randomly assigned to one of the two treatments, i.e. the five drug therapy CMFPV or the four drug therapy CMFP. The differences in patient survival and tumor response between the two treatment groups were not statistically significant. The data were also analyzed using multivariate procedures to determine those factors ascertained at entry into the study which were predictors of survival or predictors of response to therapy. The one factor that predicted both response and survival was performance status. An additional important predictor of survival was sites of metastatic involvement. Other significant predictors of response were menopausal age, BUN, and hematocrit.Deceased, February, 1985     

Randomized trial comparing adriamycin vincristine (av) cyclophosphamide methotrexate 5-Fluorouracil prednisone (cmfp) hybrid versus av-cmfp monthly alternated in metastatic breast-cancer

194 metastatic breast cancer patients with no prior chemotherapy for advanced disease were randomized to one of two alternating schedules, fulfilling the requisites of Goldie and Coldman's hypothesis to evaluate if the earlier alternation of two non-cross resistant regimens is superior in terms of response (R), duration of R (DR), and survival (SV). Treatment: arm A: Adriamycin (A) 60 mg/m2 IV day (d) 1 and vincristine (V) 1.4 mg/m2 IV d 1 and 8 monthly alternated with cyclophosphamide (C) 100 mg/m2 p.o. d 1-14; methotrexate (M) 30 mg/m2 IV d 1 and 8; 5-fluorouracil (F) 600 mg/m2 IV d 1 and 8 and prednisone (Pr) 40 mg/m2 p.o. d 1-14. Arm B (hybrid): A 60 mg/m2 IV d 1; V 1.4 mg/m2 IV d 1; C 100 mg/m2 p.o. d 8-14; M 30 mg/m2 IV d 8; F 600 mg/m2 IV d 8 and Pr 40 mg/m2 p.o. d 8-14. Results: 87 and 89 patients are evaluable for R. Arm A: R= 59% (51/87); median DR= 13 months (m); median SV= 25 m. Arm B: R= 69% (61/89); median DR= 15 m.; median SV= 29 m. Myelosuppression was slightly more marked in arm B. Three patients had toxic-related deaths (arm A: 1; arm B: 2). Conclusions: a trend favoring an earlier alternation and higher dose intensity (DI) was found regarding to R, DR and SV. However, differences were not statistically significant.     

Combination chemoimmunotherapy of metastatic breast cancer with 5-fluorouracil, adriamycin, cyclophosphamide, and BCG.

One hundred five patients with metastatic breast cancer were treated with 5-fluorouracil, Adriamycin, cyclophosphamide and BCG (FAC-BCG). The results were compared to those observed in a group of 44 patients treated with FAC chemotherapy alone. Although the overall response rates were similar (76% for FAC-BCG and 73% for FAC), the duration of remission was of 9 months for FAC and 14 months for FAC-BCG (p = 0.008). Similarly, survival of responding patients treated with FAC-BCG was significantly longer (24 months) than that observed in the chemotherapy alone treated group (15 months). There was no difference in survival or duration on study for nonresponders. Analysis of response rates by known prognostic factors was unrewarding. The duration of remission and survival, however, were significantly longer for patients with bone soft tissue involvement than for patients with visceral metastasis. Similarly patients with 1 or 2 metastatic sites survived significantly longer than those with more than 3 organ sites involved (p = 0.02). This chemotherapeutic combination is highly effective in inducing remissions. In addition, nonspecific immunotherapy with BCG appears to prolong duration of remission and survival for responding patients.     

Combination chemoimmunotherapy of metastatic breast cancer with 5-fluorouracil, adriamycin, cyclophosphamide, and BCG.

One hundred and five patients with metastatic breast cancer were treated with 5-fluorouracil, Adriamycin, cyclophosphamide and BCG (FAC-BCG). The results were compared to those observed in a group of 44 patients treated with FAC chemotherapy alone. Although the overall response rates were similar (76% for FAC-BCG and 73% for FAC), the duration of remission was of 9 months for FAC and 14 months for FAC-BCG (p = 0.04). Similarly, survival or responding patients treated with FAC-BCG was significantly longer (24 months) than that observed in the chemotherapy alone treated group (15 months). There was no difference in survival or duration on study for non-responders. Response rates were not influenced by dominant site of disease, menopausal status or disease-free interval. The duration of remission and survival, however, were significantly longer for patients with bone and soft tissue involvement than for patients with visceral metastasis. Similarly patients with 1 or 2 metastatic sites survived significantly longer than those with more than 3 organ sites involved (p = 0.02). This chemotherapeutic combination is highly effective in inducing remissions. In addition, nonspecific immunotherapy with BCG appears to prolong duration of remission and survival for responding patients.     

Intravenous administration of cyclophosphamide, methotrexate and 5-fluorouracil in metastatic breast cancer. A pilot study

Sixty-two women with advanced breast cancer were admitted to a pilot study in which a modified CMF regimen was administered. Cyclophosphamide was administered i.v. at a dosage of 600 mg/m2 on the same day as fluorouracil (600 mg/m2/i.v.) and methotrexate (40 mg/m2/i.v.). The therapy was recycled on the 21st day and in the presence of myelosuppression, the administration of the drugs was delayed for 1-2 weeks recovery of the hematologic values. CR + PR were obtained in 42% of patients and no change in 32% (U.I.C.C. criteria). Metastases to soft tissues showed CR + PR in 55% of the cases, bone in 33% and viscera in 35%. The menopausal status, the disease-free interval and the number of involved sites did not influence statistically the percentage of responses; however, the response rate was influenced statistically by previous treatment. The median duration of response was 7.5 months; the median overall survival of the 60 evaluable patients was 18 months. Due to myelosuppression, CMF i.v. administration was delayed 90/620 times (14%). Toxicity was acceptable and had a lower incidence than that reported in the literature in different series of CMF administered p.o. Nausea and vomiting, in particular, were limited to 24-48 h after administration of the drugs, and alopecia was seldom observed.     

Results of clinical trials with a CMitF (cyclophosphamide, mitoxantrone, and 5-fluorouracil) regimen versus a CAF (cyclophosphamide, adriamycin, and 5-fluorouracil) regimen in advanced/relapsed breast cancer

Randomized clinical trials were conducted in patients with advanced/relapsed breast cancer, using CMitF (mitoxantrone, cyclophosphamide, and 5-fluorouracil) regimen in comparison with CAF (cyclophosphamide, adriamycin, and 5-fluorouracil) regimen. The response rate was 50% (13/26) for the CMitF group and 50% (10/20) for the CAF group. Classified by disease site, CAF group showed a tendency for a higher response rate in soft tissue, but in bone and viscera the rates were similar for both regimen groups. The median weeks to response (range) was 5 (3-21) weeks for the CMitF group, and 8 (3-22) weeks for CAF group, a tendency for slightly earlier response thus being shown in the CMitF group. The median duration of response (range) was 10 (4-47) weeks for the CMitF group, and 9 (4-50) weeks for CAF group, whereas by the method of Kaplan-Meier, at 40 weeks after the start of therapy, the proportion showing response still continued to be 60% for the CMitF group, which was a higher rate than the 27% for CAF group. The survival rate at 48 weeks was for both groups with no difference shown. Leukopenia, gastrointestinal symptoms and alopecia were major side effects in both groups, but the incidence and severity grade of gastrointestinal symptoms (nausea and vomiting) and alopecia were significantly less in the CMitF group than in CAF group. In both groups, the WBC nadir was 2,100/microliter and the time to nadir was 14 days, with no difference shown in hematological toxicity between the two groups. No severe cardiac, hepatic, or renal toxicity was noted.     

The effects of multiple combination chemotherapy with vincristine,cyclophosphamide (Endoxan), methotrexate, 5-fluorouracil,adriamycin and prednisolone (VEMFAH) for advanced breast cancer

Thirty-eight patients with advanced breast cancer were treated with the 'VEMFAH' multiple-drug combination chemotherapy, consisting of vincristine (V), cyclophosphamide (Endoxan; E), methotrexate (M), 5-fluorouracil (F), adriamycin (A), and prednisolone (H). Disease response was evaluated by the UICC criteria. Among the 35 evaluable cases, 4 complete responses (CR), 23 partial responses (PR), 2 cases of no change (NC), and 6 of progressive disease (PD) were observed. The response rate (CR + PR) was 77.1%. The median duration of response was 52 weeks (8-192 weeks) or 12 months. In 32 patients who received more than two courses of therapy the 50% survival time of responders was 27.0 months, which was significantly longer than the 10.3 months of nonresponders (P less than 0.05). Except for 2 patients who developed myocardial damage, the therapy was never terminated because of side effects. Cumulative cardiotoxicity was not apparent in this study. This multiple-drug combination chemotherapy with 'VEMFAH' is concluded to be an effective treatment for advanced and disseminated breast cancer.     

Biological monitoring screening of patients provided antineoplastic drugs including adriamycin, cyclophosphamide, 5-fluorouracil, methotrexate, and vincristine

The aim of the present study was to establish screening biomarkers of exposure to antineoplastic drugs administered to 11 patients undergoing cancer chemotherapy. Among the anticancer drugs administered were cyclophosphamide (all), Adriamycin (5 of 11), methotrexate (3 of 11), 5-fluorouracil (4 of 11), vincristine (3 of 11), megestrol acetate (1 of 11), and procarbazine (1 of 11). The noninvasive urinary parameters investigated were thioethers, D-glucaric acid, elements, and forward and reverse mutagenesis using bacterial bioassays. The data were analyzed in terms of the observed concentrations and those corrected for personal baseline. Personal baseline correction for parameters with significant nonexposure baseline levels was essential. While glucaric acid and thioethers were increased by the drug treatments, the correlations with baseline-uncorrected data showing an inverse relationship proved spurious, because saturation of the detoxification systems occurred at the high doses administered. Glucaric acid was also influenced by methotrexate and vincristine. Thioether content was affected by cyclophosphamide only. The forward mutagenesis assay was directly correlated to cyclophosphamide dose but the reverse assay was not, in the presence or absence of rat S9 fraction. The forward assay was not sensitive to the effects of smoking. Relative to controls, the elements changed by cyclophosphamide were K, S, and P. Those affected by Adriamycin were Ca, Mg, and Na; 5-fluorouracil affected Ca, Mg, Na, and C; methotrexate changed P and S. The forward mutagenesis assay and D-glucaric acid concentrations were the screening biomarkers best suited to monitoring for extent of exposure to these antineoplastic drugs.     

A phase III clinical trial comparing the combination cyclophosphamide, adriamycin, cisplatin with cyclophosphamide, 5-fluorouracil, prednisone in patients with advanced breast cancer

We assessed the efficacy of a regimen consisting of four cycles of cyclophosphamide, Adriamycin (Adria Laboratories, Inc, Columbus, Ohio), cisplatin (CAP) followed by maintenance with cyclophosphamide, 5-fluorouracil, prednisone (CFP) compared with CFP alone in a randomized trial of 86 patients with advanced breast cancer. The objective regression rates were 46% (CFP) and 49% (CAP with CFP) which included complete regression rates of 7% (CFP) and 4% (CAP with CFP). The median time to progression was nine months for CFP and six months for CAP with CFP. Median survival in the CFP group was 18 months v 11 months in the CAP recipients. Due to the therapeutic trend in favor of patients receiving CFP, we terminated the study before achieving our initially projected accrual. We observed over a twofold excess of substantial nausea and vomiting among patients receiving the platinum-based regimen. In our view, the CAP followed by the CFP regimen is a more toxic program that offers no clinically meaningful improvement over CFP to patients with advanced breast cancer.     

A randomized trial of two regimens of cyclophosphamide,methotrexate, 5-fluorouracil,and prednisone in advanced breast cancer

Summary One-hundred evaluable patients with progressive advanced breast carcinoma untreated by cytotoxic chemotherapy but resistant to hormone therapy and irradiation were randomly allocated to receive either a combination of cyclophosphamide (600 mg/m²), methotrexate (40 mg/m²), 5-fluorouracil (600 mg/m²) IV every 3 weeks and prednisone 20 mg/m² PO daily, with diminishing doses (intermittent group), or a combination of cyclophosphamide (100 mg/m² PO on days 1–15, alternating with a 15-day rest period), methotrexate 20 mg/m² IV, 5-fluorouracil 500 mg/m² IV weekly for 20 weeks and prednisone 20 mg/m² PO daily, with diminishing doses in the remission induction period, followed by a maintenance regimen of cyclophosphamide 100 mg/m² PO on days 1–15, methotrexate 20 mg/m² IV on days 1, 8, and 15, 5-fluorouracil 500 mg/m² IV on days 1, 8, and 15, and prednisone 20 mg/m² PO on days 1–15, with a 3-week rest period between the courses (intensive group). Entry was from 1 December 1982 to 30 November 1983. Objective responses were seen in 20/49 (41%) patients in the intermittent group and 34/51 (67%) in the intensive group (²=6.72; P<0.01). The estimated median duration of response was 11 months in the intermittent group and 14 months in the intensive group. The estimated median survival was greater in the intensive group, but the difference was not statistically significant, although this parameter can be influenced with alternative additional chemotherapy. Toxicity was similar in both groups. These data suggest there are no therapeutic and survival advantages to the 3-weekly IV protocol compared with our previous regimen CMFP.     

Cyclophosphamide, methotrexate and infusional 5-fluorouracil (infusional CMF) in metastatic breast cancer.

Bolus 5-fluorouracil (5-FU) is a phase-specific drug with a short plasma half-life that is used in combination with bolus cyclophosphamide and methotrexate in the treatment of breast cancer. The efficacy of 5-FU can be improved by continuous intravenous infusion using portable infusion pumps (infusional 5-FU). Infusional 5-FU, 200 mg m(-2) day(-1), in combination with standard doses of bolus cyclophosphamide and methotrexate, was evaluated in a phase I/II dose-finding study. The cyclophosphamide and methotrexate were administered in 28-day cycles as follows: cohort 1, cyclophosphamide 600 mg m(-2), days 1 and 8, and methotrexate 40 mg m(-2), day 1; cohort 2, cyclophosphamide 400 mg m(-2), days 1 and 8, and methotrexate 40 mg m(-2), day 1; cohort 3, cyclophosphamide 480 mg (m-2), days 1 and 8, and methotrexate 40 mg m(-2), day 1; cohort 4, cyclophosphamide 480 mg m(-2), days 1 and 8, and methotrexate 40 mg m(-2), days 1 and 8. Median overall survival was 10 months (range 3-21 months). Objective tumour responses were seen in 9 of 25 patients (36%, 95% CI 18-58%), including 3 of 13 patients (23%) previously treated for metastatic disease. Cohorts 1 and 4 proved to be too toxic, with five of six patients in cohort 1 and three of four in cohort 4 developing grade III/IV neutropenia. The dose intensity of cyclophosphamide achieved was as follows: cohort 1, 82%; cohort 2, 86%; cohort 3, 97%; cohort 4, 90%. Infusional 5-FU can be administered safely and is effective in combination with cyclophosphamide 480 mg m(-2), days 1 and 8, and methotrexate 40 mg m(-2), day 1, in the treatment of metastatic breast cancer.