Paclitaxel for Breast Cancer: The Memorial Sloan-Kettering Cancer Center Experience

Introduction

In the management of patients with metastatic breast cancer, paclitaxel(Drug information on paclitaxel) (Taxol) has demonstrated safety and significant antitumor activity.[1-5] To fully characterize the drug's potential in the treatment of advanced breast cancer, we have conducted a series of phase II clinical trials evaluating its efficacy and tolerability as single-agent therapy (Table 1).

Paclitaxel has been studied in various doses and schedules among patients with metastases who have not received previous chemotherapy for advanced disease, as well as among those who have had moderate and extensive prior therapy.[1,3-5] In heavily pretreated patients, we have conducted a formal parallel investigation of quality-of-life (QOL) parameters to better assess the palliative potential of paclitaxel among patients who receive this agent with the primary intent of symptom control.[6] We are presently incorporating these variables into a prospective study that addresses economic (charges/costs) and QOL outcomes, in addition to classic tumor response and toxicity assessments, in patients receiving single-agent paclitaxel.[7]

Our studies were influenced by preclinical data demonstrating schedule-dependent cytokinetic variability in resistance profiles to the taxanes,[8-10] as well as clinical documentation of the efficacy and safety of a 96-hour infusion schedule against anthracycline-refractory breast cancer.[11] These investigations led us to perform a phase II and pharmacologic trial assessing the activity and toxicity profile of this prolonged infusion schedule in patients with documented progression of disease during shorter taxane exposure.[12]

Data supporting the feasibility and efficacy of more frequent, shorter-duration taxane infusions[13] has motivated our ongoing trial, which is evaluating a weekly 1-hour infusion paclitaxel schedule in patients who have received one or two prior regimens for breast cancer (including adjuvant therapy).[14] Collaborative laboratory investigations aimed at characterizing clinically relevant mechanisms of resistance to paclitaxel are well underway in studies involving biopsied human breast cancer tissue.

Combinations of paclitaxel and other cytotoxic agents are being studied. A variety of two- and three-drug paclitaxel-containing regimens are in clinical trials to define possible synergism, as well as to illuminate safety profiles. At Memorial Sloan-Kettering Cancer Center (MSKCC), we have investigated the combination of paclitaxel and edatrexate, a dihydrofolate reductase inhibitor with preclinical advantages over methotrexate(Drug information on methotrexate)[15] and proven single-agent activity against metastatic breast cancer.[16-17] The trial design derives from schedule-dependent synergy observed in vitro in mammary carcinoma cells.[18,19]

A promising avenue of translational research pertains to the observed synergy between paclitaxel and monoclonal antibodies (MoAbs) directed at various growth factor receptors in human breast carcinoma xenografts.[20,21] While laboratory studies addressing possible cellular mechanisms for this effect are ongoing, we are conducting a clinical trial combining paclitaxel with a MoAb directed against the epidermal growth factor.

The preserved activity of paclitaxel after anthracycline threapy has stimulated its evaluation as a component of an adjuvant sequential chemotherapy regimen for node-positive stage II-III resectable breast cancer.[22] We recently evaluated the optimal integration of paclitaxel into doxorubicin(Drug information on doxorubicin)/cyclophosphamide (Cytoxan, Neosar)-based adjuvant therapy in a randomized clinical trial.[23]

This article will review 6 years of experience with paclitaxel in the treatment of breast cancer at MSKCC and describe ongoing and planned studies.

Single-Agent Paclitaxel

First-Line Therapy

Subsequent to the M. D. Anderson Cancer Center report of the promising antitumor activity of paclitaxel against metastatic breast cancer,[2] we performed a confirmatory phase II trial of the agent as first-line chemotherapy for stage IV disease.[3] In this study, 28 patients who had not received prior chemotherapy for metastatic disease were treated with paclitaxel, 250 mg/m² via a 24-hour infusion every 3 weeks. Because the first two patients we treated on this protocol experienced significant myelosuppression (which was dose-limiting in the previous trial[2]), the protocol was amended so that all subsequent patients received prophylactic granulocyte colony-stimulating factor (G-CSF [Neupogen]), 5 µg/kg/d subcutaneously, on days 3 to 10 of each cycle.

A 62% response rate (95% confidence interval [CI], 41% to 80%) was noted, including three complete responses.[3] Responses were as common among women who had received prior adjuvant therapy, including doxorubicin-containing regimens, as among those who had not. Treatment was well tolerate; adverse effects included generalized alopecia in all patients, and grade 3 or 4 nonhematologic toxicities were uncommon. Of 178 cycles administered, there were 8 admissions (4%) for febrile neutropenia, involving 6 (21%) of 28 patients.

Administration of recombinant human granulocyte colony stimulating factor (rh-G-CSF) resulted in a median duration with an absolute neutrophil count less than 500 cells/µL of 2 days, which was shorter than that previously reported (7 days)[1] without concomitant growth factor support. Fifty-eight percent of cycles were delivered at modestly reduced doses, primarily due to significant neutropenia or febrile neutropenia. Because drug supply was a consideration at this time in the agent's development, the clinical trial design specified the length of treatment to be two cycles beyond the best response, with a maximum of 10 cycles per patient. Thus, response duration was not a valid end point of this trial.

After Prior Chemotherapy

Following this confirmation of paclitaxel's marked antitumor activity in patients with minimal prior therapy, we evaluated patients who had received extensive prior chemotherapy for metastatic breast cancer.[4] Fifty-one patients who had previously received two or more prior chemotherapy regimens for metastatic disease (median number of regimens, three; range, two to six; all with prior anthracycline therapy) entered our second phase II trial.

The median Karnofsky performance score (KPS) for these patients was 70% (range, 60% to 90%). Fourteen percent had received prior high-dose chemotherapy regimens sufficiently myelotoxic as to require the reinfusion of autologous bone marrow and/or peripheral blood progenitor cells; two-thirds of these patients had received radiotherapy for metastatic disease. Paclitaxel was administered at 200 mg/m² (a lower starting dose was chosen in anticipation of significant toxicity in these more heavily pretreated patients) via a 24-hour infusion every 3 weeks with G-CSF, as previously described.

Partial responses were observed in 14 patients (27.5%; 95% CI, 16% to 42%), with a median response duration of 7 months. Hospitalization for febrile neutropenia occurred in 24 (8%) of the first 312 cycles and in 9 (18%) of 51 patients. No patient was removed from the trial due to toxicity.

Our next trial evaluated the higher 250-mg/m² dose, again via a 24 hour infusion every 3 weeks, in patients who had received only one prior chemotherapy regimen for metastatic disease (with or without prior adjuvant therapy).[4] Nine partial and two complete responses were noted among 25 evaluable patients (44%; 95% CI, 24% to 65%).

Significantly, in this and the previous trial, prior demonstrated sensitivity or resistance to an anthracycline did not predict the likelihood of subsequent response to paclitaxel. This lack of significant clinical cross-resistance between paclitaxel and doxorubicin, an observation corroborated by others, was particularly gratifying because preclinical data suggest significant in vitro cross-resistance between paclitaxel and other agents for which p-glycoprotein-mediated multidrug resistance is considered relevant.[24,25]

These observations have motivated our ongoing studies characterizing pre- and post-paclitaxel human tumor tissue biopsies for mdr expression, tubulin alterations, and genomic changes. It is hoped that such studies will expand the understanding of clinically relevant mechanisms of taxane resistance, which may then guide the development of efficient strategies to overcome resistance and, potentially, the development of superior analogs.

Shorter Infusion Schedule

Renewed interest in the shorter, more convenient 3-hour infusion schedule coincided with our next two phase II trials, which addressed the safety and efficacy of this schedule as salvage chemotherapy in heavily treated patients and as initial chemotherapy for stage IV disease. In the previously treated patients (two or more prior regimens for metastatic disease, including an anthracycline), paclitaxel was administered at a starting dose of 175 mg/m² via a 3-hour infusion every 3 weeks. Since 3-hour infusions are associated with less significant myelosuppression than 24-hour infusions, G-CSF was not administered prophylactically.

Sixty-four percent of the patients had predominantly visceral disease, and the median KPS was 70%. After the 111 cycles were delivered (median, 3; range, 1 to 8), five partial responses were observed in 24 evaluable patients (20.8%; 95% CI, 7% to 42%),[5] with a median response duration of 4 months (range, 2 to 11 months).

Paclitaxel treatment was well tolerated; the only grade 3 or 4 nonhematologic toxicities noted were myalgia (4%) and mucositis (4%). Grade 3 or 4 neutropenia was seen in one-third of patients, grade 3 or 4 thrombocytopenia in 8%, and grade 3 or 4 anemia in 13%. Dose reduction was required in 21% of patients, and dose escalation was possible in only 4%.

We then evaluated a 250-mg/m² dose administered via a 3-hour infusion, again without prophylactic G-CSF, as first-line chemotherapy for metastatic breast cancer. Among 25 evaluable patients,1 complete and 7 partial responses were noted (32%; 95% CI, 15% to 53%).[5]

Myalgias, arthralgias, and neuropathy appeared to be more significant than in our prior experience with this dose delivered over 24 hours to a similar group of patients. Several patients experienced photopsia[26] at paclitaxel doses of 250 mg/m² or more over 3 hours, a phenomenon that may represent an optic neuropathy.[27] These trials provided pilot data for the design of ongoing randomized trials by the Cancer and Leukemia Group B (CALGB) and National Surgical Adjuvant Breast and Bowel Project (NSABP).

Additional Effects of Prolonged Treatment

In the course of these trials, it became obvious to us that many patients experiencing symptomatic relief of bone pain and radiographic healing of osteolytic metastases on plain x-rays, CT, or MRI had transient worsening on nuclear scintigraphic evaluation of the skeleton.[28] This radiographic "flare" on bone scan was followed by further clinical improvement in skeletal pain and recalcification of destructive bone lesions in one-third of patients.

Another phenomenon became apparent with prolonged paclitaxel treatment (in some cases, beyond 20 courses). Among a series of 52 patients experiencing continued response of visceral, osseous, and soft-tissue metastases during paclitaxel treatment, 6 experienced disease progression in the central nervous system in the absence of other evidence of treatment failure. This manifested as both parenchymal brain metastases and leptomeningeal disease. Thus, the central nervous system does appear to be a sanctuary site in many women receiving paclitaxel for the management of metastatic breast cancer.[29]

Schedule-Dependent Activity

Our next phase II clinical and pharmacologic trial was motivated by in vitro data showing less resistance to paclitaxel in p-glycoprotein-overexpressing MCF-7 breast cancer cells with longer drug exposure time.[8] Further support of this phenomenon came from other preclinical studies[9,10,30,31] and encouraging clinical experience with prolonged paclitaxel infusion in patients with anthracycline-resistant breast cancer.[11]

In this study, we evaluated the possibility of schedule-dependent activity by administering paclitaxel via a 96-hour continuous infusion specifically to patients with disease that had demonstrated clinical resistance to short taxane exposure.[12] A total of 27 such patients with disease progression had recently received 3-hour paclitaxel (N = 24), 1-hour docetaxel(Drug information on docetaxel) (Taxotere; N = 2), or both (N = 1). Per protocol, all patients received paclitaxel via a 96-hour infusion at 140 mg/m² (35 mg/m²/d ×4), with a starting dose of 120 mg/m²/d for patients with impaired hepatic function. Because early data had suggested that the omission of steroid and H1- and H2-receptor antagonist premedication was not associated with significant hypersensitivity-like reactions with this dose and schedule,[11] these drugs were not given in our study.

With 195 cycles administered, seven partial responses were noted in 26 evaluable patients (26.9%; 95% CI, 11.6% to 47.8%), with acceptable hematologic and nonhematologic toxicity. Despite the omission of standard premedication, no cases of hypersensitivity-like reactions occurred. This suggests that the slower rate of exposure to paclitaxel--and possibly to the polyoxyethylated castor oil (Cremophor EL) formulation--may not precipitate mast-cell degranulation and other cellular phenomena associated with hypersensitivity reactions.

Serum paclitaxel concentrations were assayed in 23 patients by high-performance liquid chromatography at 24, 48, 72, and 96 hours of the infusion. The median steady-state paclitaxel concentration (Css) was .047 µM (range, .023 to .176 µM). For 11 patients experiencing grade 4 neutropenia, median Css was .068 µM (range, .032 to .176 µM), as compared with .039 µM (.023 to .098 µM) in 12 patients with less severe neutropenia (P less than .05). Median Css and absolute neutrophil count (ANC) nadirs were .094 µM (range, .074 to .176 µM) and 300 cells/mm³, respectively, in four patients with baseline elevation of hepatic transaminases vs .041 µM (.023 to .102 µM) and 800 cells/mm³ in 19 patients with normal transaminases (Css, P less than .01; ANC, not significant).[12]

No relationship between steady-state level and tumor response was noted with this limited sample size. To further define the significance of infusion duration for paclitaxel, a recently completed multi-institution trial led by the M. D. Anderson Cancer Center randomized patients with refractory metastatic breast cancer to receive paclitaxel via either a 3- or 96-hour infusion schedule. It is noteworthy that there are greater pharmacologic differences between the 3- and 96-hour infusion schedules than between the 3- and 24-hour schedules, and many preclinical data suggest the importance of paclitaxel exposure duration in breast carcinoma cells. Thus, this trial should provide the most definitive information about the impact of paclitaxel infusion duration on efficacy and toxicity.