Paclitaxel plus high-dose cyclosphosphamide with G-CSF support in patients with relapsed and refractory aggressive non-Hodgkin's lymphoma


Dr Anas Younes Department of Hematology, Section of Lymphoma, Box 68, University of Texas M. D. Anderson Cancer Center, 1515 Holcombe Blvd, Houston, TX 77030, U.S.A.


Based on the single-agent activity of both paclitaxel and cyclophosphamide in the treatment of non-Hodgkin's lymphoma (NHL), we conducted a phase II study to evaluate the efficacy of the combination of the two drugs in patients with refractory and relapsed aggressive NHL. All patients received 900 mg/m2 bolus of cyclophosphamide intravenously daily for 3 consecutive days with a concurrent infusion of 150 mg/m2 of paclitaxel over 72 h (50 mg/m2/d). 24 h after the completion of chemotherapy, patients received subcutaneous injections of 5 μg/kg of granulocyte-colony stimulating factor (G-CSF) daily until white cell count recovery. Treatment was repeated every 3 weeks. Patients who had at least a partial response (PR) after two courses continued to receive a maximum of four courses. Patients with responding disease were allowed to undergo high-dose chemotherapy followed by stem-cell/bone marrow transplantation if they were eligible. Of the 77 patients who were eligible for the study, 74 (96%) were evaluable for toxicity and treatment response. The overall response rate was 45% (95% CI 33–57%). Patients who received treatment after their disease relapsed from a complete response (CR) had an 81% response rate (38% CRs), whereas those with primary refractory disease had a 22% response rate. Toxicities of >grade 2 included alopecia (100%) and stomatitis (25%). Neutropenic fever of grade >2 occurred after 18% of the courses, and platelet count of leqslant R: less-than-or-eq, slant20 × 109/l developed after 20% of the courses. Thus, the combination of paclitaxel plus high-dose cyclophosphamide is an effective new regimen in the treatment of refractory and relapsed NHL.

Although aggressive non-Hodgkin's lymphoma (NHL) was among the first human cancers to be cured with combination chemotherapy, the cure rate of this disease remains about 50%, which has not changed since the introduction of CHOP (cyclophosphamide, hydroxydoxorubicin, vincristine, prednisolone) combination chemotherapy in the 1970s ( Armitage, 1993). Later combination programmes failed to improve survival and were more toxic than CHOP ( Armitage, 1992; Fisher et al, 1993 ; Hryniuk, 1993). For patients whose disease fails to respond to primary therapy or relapses from a complete remission (CR), the prognosis remains poor. Various salvage programmes can achieve high response rates, but the responses achieved are rarely durable ( Rodriguez, 1995; Velasquez et al, 1994 ; Wilson et al, 1993 ), leaving high-dose chemotherapy and stem-cell transplantation as the only potentially curative treatment ( Armitage, 1993).

The success of stem cell and bone marrow transplantation in the treatment of relapsed lymphoma increased the demand for more effective salvage therapy that could reduce the tumour load and mobilize stem cells into peripheral blood prior to their collection. Although several salvage programmes have been reported to induce high response rates in patients with relapsed and refractory aggressive NHL, they predominantly targeted patients who had previously failed CHOP or CHOP-like therapy. Thus, new effective treatment programmes are needed for those who do not respond to or relapse after salvage therapy.

Paclitaxel (Taxol; Bristol-Myers Squibb, Princeton, N.J.) has been recently tested for the treatment of NHL ( Younes et al, 1997b ). In our experience with 44 previously treated patients with aggressive NHL, paclitaxel alone induced an overall response rate of 30% (16% CRs). The most favourable activity of paclitaxel was observed in patients with relapsed aggressive NHL who did not have a primary refractory disease as it induced a 50% response rate ( Younes et al, 1997a ), Based on this favourable single-agent activity, we became interested in evaluating paclitaxel-based new treatment programmes that can be tested in patients with relapsed and refractory aggressive NHL. Active regimens can then be tested in untreated patients with poor prognostic features ( Shipp et al, 1993 ). In this study we report our experience with paclitaxel combined with high-dose cyclophosphamide. The dose and schedule of this combination were adopted from a recently published phase I trial in patients with breast carcinoma ( Tolcher et al, 1996 ). The addition of high-dose cyclophosphamide with growth factor support to paclitaxel was appealing because it enhanced peripheral blood stem cell collection in responding patients prior to stem-cell transplantation.



Patients were eligible for this phase II clinical trial if they had histologically documented refractory or relapsed aggressive NHL, which included the following histologies according to the REAL classification: diffuse large B-cell, follicular centre-cell (follicular, grade III), peripheral T-cell, and anaplastic large-cell lymphoma ( Harris et al, 1994 ). All patients were required to have bidimensionally measurable disease, a bilirubin level <25 μmol/l, SGPT level <twice the upper normal value, a platelet count geqslant R: gt-or-equal, slanted100 × 109/l unless caused by marrow infiltration of NHL, and an absolute neutrophil count (ANC) geqslant R: gt-or-equal, slanted1 × 109/l. Patients were excluded if they had HIV infection, central nervous system lymphoma, prior treatment with paclitaxel or high doses of cyclophosphamide or ifosfamide, a life expectancy <12 weeks, or serious cardiac disease. Patients were registered after signing a consent form according to institutional guidelines.


Before starting therapy all patients had a complete physical examination, a chest X-ray, a CT scan of the chest, abdomen and pelvis, a gallium scan, and bilateral bone marrow biopsies. These tests were repeated after two courses and after the completion of therapy.


The dose and schedule of drugs used in this study were adopted from a recently published phase I trial in breast carcinoma patients ( Tolcher et al, 1996 ). Patients received 20 mg of dexamethasone orally 4 h before the chemotherapy infusion. 30 min before the infusion of paclitaxel, patients received premedication with intravenous injections of dexamethasone (20 mg), cimetidine (300 mg) and diphenhydramine (50 mg). Cyclophosphamide was given intravenously at a dose of 900 mg/m2/d on days 1, 2 and 3 (total dose 2700 mg/m2) with mesna given intravenously at a dose of 900 mg/m2/d by continuous infusion for 72 h. Paclitaxel was given intravenously at a dose of 50 mg/m2/d by continuous infusion for 72 h (total dose 150 mg/m2). 24 h after completion of the paclitaxel infusion, G-CSF (filgrastim, Amgen, Thousand Oaks, Calif.) was administered subcutaneously at a dose of 5 μg/kg/d for 14 d or until an ANC geqslant R: gt-or-equal, slanted10 × 109/l was achieved. Courses were repeated every 3 weeks.

Responding patients were allowed to receive consolidation with high-dose chemotherapy and bone marrow or stem cell transplantation after at least two courses of paclitaxel and cyclophosphamide. Patients who were not eligible for or refused bone marrow transplantation continued to receive a maximum of four courses. Patients whose disease progressed after any course were removed from the study.

Dose modification

Dose reduction was allowed if the ANC count on day 21 was <0.75 × 109/l or the platelet count was <100 × 109/l, and for patients who had grade III non-haematological toxicities other than nausea. The reduced doses consisted of cyclophosphamide at a dose of 700 mg/m2/d and a total paclitaxel dose of 125 mg/m2 over 72 h. Treatment was stopped if grade IV non-haematological toxicity developed.

Response criteria

A CR was defined as normalization of radiological, physical and bone marrow findings for at least 4 weeks, or stable residual radiographic abnormalities with no gallium uptake for at least 8 weeks. A partial response (PR) was defined as at least a 50% reduction in the sum of the products of the greatest cross-sectional diameters of measurable lesions without evidence of new lesions.


Patient characteristics

Between April 1995 and December 1996, 77 patients were eligible for the study. Two patients withdrew consent and one patient received one course and was lost to follow-up. Therefore 74 patients (96%) were evaluable for treatment response and toxicity, of whom 72% had diffuse large B-cell NHL and 16% had transformed NHL ( Table I). All patients had received prior combination chemotherapy, and 22 patients (23%) received radiation therapy. 41 patients (55%) had primary refractory disease (never achieved a CR with histologic evidence of active disease or persistent gallium activity) and 12 patients (16%) had a resistant relapse ( Table I).

Table 1. Table I. Characteristics of 74 evaluable patients. ASHAP: adriamycin, solumedrol, high-dose ara-C, platinum; MBACOS: methotrexate, bleomycin, adriamycin, cyclophosphamide, vincristine solumedrol; MINE: mesna, ifosphamide, novatrone, etoposide.* ASHAP/MBACOS/MINE is considered one regimen.Thumbnail image of

The median age of the patients was 59 years (range 24–75 years). 21 patients (28%) were >65 years and 36 patients (49%) were >60 years. 56 patients (76%) had elevated lactate dehydrogenase (LDH) values. The pretreatment B2 microglobulin level was determined in 44 patients, 15 of whom (36%) had elevated levels (geqslant R: gt-or-equal, slanted3 mg/l).

Treatment response

Thirty-three patients achieved a PR or CR for an overall response rate of 45% (95% CI 33–57%). The response rates varied according to how the disease responded to prior treatments. Patients who were treated after relapsing from a CR had a response rate of 81% (38% CRs and 43% PRs). Patients who had primary refractory disease (never achieved a CR) or resistant relapses (relapsed patients who failed to respond to their last therapy) had response rates of 22% and 58%, respectively ( Table II). Patients with non-primary refractory disease had a median time to treatment failure of 4.8 months and a median duration of response of 6.5 months (range 2–18 months) (Fig 1 ). Of the 33 responding patients, 12 were consolidated with high-dose chemotherapy and autologous stem-cell transplantation (eight patients were transplanted while in PR after two courses). The remaining 21 patients were either not eligible for or refused stem cell transplantation, and were observed until disease progression.

Table 2. Table II. Treatment responses.Thumbnail image of
Figure 1.

2 responding patients who were consolidated with bone marrow transplantation were censored.

Responses were similar in patients who had received prior ara-C/platinum-based therapy and in those who had not (47% and 44%, respectively). Patients whose disease responded to prior ara-C/platinum therapy had a response rate of 71% compared with a response rate of 29% for those whose disease was refractory to ara-C/platinum ( Table II).


A total of 178 courses were administered to 74 patients. 14 patients (19%) received only once course (they did not continue because their disease rapidly progressed), 33 (45%) received only two courses (because of either lack of response or they were crossed over to bone marrow transplantation), 10 patients (14%) received three courses, and 17 patients (23%) received four courses. Therefore 60 patients (81%) received at least two courses. In 27 courses (15%) the doses were reduced because of toxicities. All patients had alopecia. The most common toxicities were nausea, stomatitis and vomiting ( Table III). Neutropenic fever of grade >2 was observed after 18% of the courses, and stomatitis of grade >2 was seen in 25% after 10% of the courses ( Table III).

Table 3. Table III. Toxicity of cyclophosphamide plus paclitaxel in NHL patients.Thumbnail image of

Thrombocytopenia and neutropenia were not cumulative. 20% of the patients had a platelet count leqslant R: less-than-or-eq, slant20 × 109/l and 35% of the patients had an ANC leqslant R: less-than-or-eq, slant0.1 × 109/l. No treatment-related deaths were observed, but treatment was discontinued in 8% of the patients because of grade IV non-haematological toxicity ( Table III).


The activity of paclitaxel in patients with relapsed and refractory NHL has been studied by three independent groups using different doses and infusion schedules ( Press et al, 1998 ; Wilson et al, 1995 ; Younes et al, 1997a ). In patients with non-primary refractory aggressive NHL, paclitaxel induced a response rate of 25–50%, with higher and more complete responses were observed when higher doses of paclitaxel were infused over a shorter period of time ( Press et al, 1998 ; Younes et al, 1997a ). In this report we describe our experience with the first paclitaxel-based combination for the treatment of relapsed and refractory aggressive NHL.

The overall response rate in the current study was 45%, but patients who did not have a primary refractory disease had a response rate of 73% (33% CRs). Because responding patients were allowed to be consolidated with high-dose therapy and bone marrow transplantation after only two courses, the CR rate and response duration in this study may have been underestimated. In fact, eight patients were consolidated with stem cell/bone marrow transplantation after achieving PRs after only two courses of therapy. Although only 12 patients were treated in a setting of resistant relapse (did not respond to their last salvage therapy), it is interesting to note that 58% of these patients responded (25% CRs). However, consistent with previous observations, patients with primary refractory disease continue to be poor responders. It is for this reason that response rates should be reported separately for patients with primary refractory and non-primary refractory disease. Thus, trials that include high number of patients with primary refractory NHL are expected to have a lower overall response rate. Therefore the activity of a new drug or regimen should not be judged based on the overall response rate because such analysis may lead to excluding potentially active agents prematurely. On the other hand, a regimen that can induce a high response rate in NHL patients whose disease was not refractory to induction therapy should be tested in newly diagnosed patients with poor-risk features because these patients are infrequently cured with CHOP chemotherapy ( Shipp et al, 1993 ).

One important aspect of this programme was its ability to induce responses in patients who had previously received ara-C/platinum-based therapy, which is frequently used as a first-line salvage therapy. Although several other salvage programmes have been used to treat relapsed NHL, they were mainly aimed at CHOP or CHOP-equivalent failures and included patients with variable histologies ( Rodriguez et al, 1995 ; Velasquez et al, 1994 ; Wilson et al, 1993 ). These salvage programmes have also varied in the number of patients included with refractory disease, the number of prior treatment regimens, the median age, and LDH levels. Therefore it is difficult to compare the activities of these different programmes ( Table IV). In this study the patients were more homogenous because only those with aggressive histologies were included.

Table 4. Table IV. Salvage programmes for aggressive NHL. DHAP: dexamethasone, high-dose ara-C, platinum; ESHAP: DHAP + etoposide; MINE: mesna, ifosphamide, novantrone, etoposide; EPOCH: etoposide, prednisone, vincristine, cyclophosphamide, hydroxydoxorubicin.Thumbnail image of

In this study, treatment was reasonably well tolerated despite the fact that the median patient age was 59 years and that the patients were heavily pretreated. Dose reduction was done after only 15% of the courses. Although cardiac arrhythmia was noted in four patients, no treatment-related deaths occurred.

Based on this study, the combination of paclitaxel and cyclophosphamide seems to be active in the treatment of patients with relapsed aggressive NHL. Because 81% of the patients who had a relapse following a CR responded to this combination, it is possible that even more favourable activity and less toxic effects could be observed in less heavily pretreated patients. Because the combination of paclitaxel and cyclophosphamide plus G-CSF was reported to adequately mobilize stem cells into the peripheral blood ( Demirer et al, 1995 ; Fennelly et al, 1995 ), this combination can be used prior to high-dose chemotherapy and autologous stem-cell infusion. The real test of the activity of this regimen can only be determined in newly diagnosed patients with poor prognostic features.


Anas Younes is the recipient of an American Cancer Society Clinical Oncology Career Development Award. This work was supported in part by grants from Bristol-Myers Squibb Pharmaceutical and Amgen.