Fax: (412) 648-6579
Docetaxel and irinotecan in recurrent or metastatic head and neck cancer
A phase 2 trial of the Eastern Cooperative Oncology Group
Article first published online: 24 JUL 2009
Copyright © 2009 American Cancer Society
Volume 115, Issue 19, pages 4504–4513, 1 October 2009
How to Cite
Argiris, A., Buchanan, A., Brockstein, B., Kolesar, J., Ghebremichael, M., Pins, M., Hahn, K., Axelrod, R. and Forastiere, A. (2009), Docetaxel and irinotecan in recurrent or metastatic head and neck cancer. Cancer, 115: 4504–4513. doi: 10.1002/cncr.24528
- Issue published online: 17 SEP 2009
- Article first published online: 24 JUL 2009
- Manuscript Accepted: 11 MAR 2009
- Manuscript Revised: 28 FEB 2009
- Manuscript Received: 23 DEC 2008
- Sanofi-Aventis. Grant Numbers: CA-23,318, CA-21,115
- head and neck cancer;
- vascular endothelial growth factor
Docetaxel and irinotecan have single-agent antitumor activity in squamous cell carcinoma of the head and neck (SCCHN). The authors sought to evaluate their combination in the treatment of patients with recurrent or metastatic SCCHN.
Eligibility criteria included recurrent or metastatic SCCHN with measurable disease, good performance status, and adequate laboratory parameters. Patients received docetaxel 35 mg/m2 and irinotecan 60 mg/m2, intravenously, on Days 1 and 8, every 21 days, until disease progression. The authors assessed UGT1A1 genotype, vascular endothelial growth factor (VEGF) in serum, and cyclooxygenase-2 and VEGF in baseline tumor tissue.
Fifty-two patients were analyzable: 20 chemotherapy naive (Group A) and 32 previously treated with 1 chemotherapy regimen (Group B); 73% of patients had distant metastasis, and 60% were paclitaxel-exposed. In Group A, 3 (15%) patients achieved a partial response; in Group B, 1 (3%) patient achieved a partial response. Median progression-free survival (PFS) and overall survival were 3.3 and 8.2 months in Group A and 1.9 and 5.0 months in Group B, respectively. Common serious toxicities were diarrhea, fatigue, and anorexia. Patients with high serum VEGF had a median PFS of 2.8 months versus 1.7 months for patients with low VEGF (P = .085).
Docetaxel and irinotecan had acceptable toxicities, but efficacy results in unselected patients with recurrent or metastatic SCCHN did not suggest an advantage over docetaxel alone or platinum-based regimens. Cancer 2009. © 2009 American Cancer Society.
More than 45,000 new cases of head and neck cancer are diagnosed annually in the United States.1 Patients with recurrent or metastatic squamous cell carcinomas of the head and neck (SCCHN) have a poor prognosis.2 Cisplatin-based combination regimens given as first-line treatment of recurrent or metastatic SCCHN result in objective response rates of about 30% and median overall survival (OS) of 8 to 9 months.3 More recently, the epidermal growth factor receptor inhibitors were introduced in the systemic therapy of SCCHN.4 However, at present time, there is no standard regimen for the second-line treatment of recurrent or metastatic SCCHN.
Docetaxel is a semisynthetic taxane that acts as a mitotic spindle poison by promoting microtubule assembly but inhibiting tubulin depolymerization, which disrupts cell division. The major toxicity with the 3-week scheduling of docetaxel is neutropenia, which is ameliorated by weekly administration.5 Clinical studies have documented the efficacy of docetaxel in many solid tumors, even after previous treatment with paclitaxel.6 Weekly docetaxel at a dose of 30 mg/m2 was highly active in a phase 2 trial in chemonaive recurrent or metastatic SCCHN with a reported response rate of 42% and median OS of 11.3 months.7 A phase 2 randomized study of weekly docetaxel versus methotrexate showed higher response rates for docetaxel but comparable survival rates.8
Irinotecan is a water-soluble analogue of camptothecin that inhibits topoisomerase I, a critical enzyme for DNA replication and transcription. Irinotecan is metabolized in the liver to SN-38, an active metabolite that undergoes glucuronidation in the liver through uridine diphosphate glucuronosyltransferase isoform 1A1 (UGT1A1) to the relatively inactive SN-38G (SN-38 glucuronide). The major toxicities of irinotecan are neutropenia and diarrhea. Polymorphisms in UGT1A1 have been reported to result in increased incidence of irinotecan-related toxicities.9, 10 A phase 2 study of irinotecan given weekly at 125 mg/m2 for 4 weeks followed by 2 weeks of rest in recurrent or metastatic SCCHN reported a response rate of 26% in 19 evaluable patients.11 Because of toxicity, 14 subsequent evaluable patients were treated at a lower dose of 75 mg/m2, and the schedule was altered to 2 weeks on, 1 week off. The response rate in this cohort was 14%. However, irinotecan had no significant activity when given in the second-line therapy setting.11
The combination of irinotecan and docetaxel is supported by preclinical observations and showed promising activity in early clinical investigations in solid tumors.12 A phase 1 clinical trial established the recommended phase 2 dose as irinotecan 60 mg/m2 and docetaxel 35 mg/m2 on Days 1 and 8, repeated every 21 days.13 On the basis of these observations, we designed a phase 2 study to investigate the antitumor activity and toxicities with irinotecan and docetaxel in patients with recurrent or metastatic SCCHN. We also examined the potential correlation between UGT1A1 genotype and toxicity of the regimen. Cyclooxygenase-2 (COX-2) and vascular endothelial growth factor (VEGF) are overexpressed in SCCHN.14-16 Therefore, we also sought to evaluate the expression of COX-2 and VEGF in tumor tissue as well as serum VEGF as predictors of antitumor efficacy.
MATERIALS AND METHODS
Eligible patients were ≥18 years of age with recurrent or metastatic SCCHN considered incurable by means of locoregional therapies, Eastern Cooperative Oncology Group (ECOG) performance status of 0 or 1, and measurable disease according to Response Evaluation Criteria in Solid Tumors.17 Patients were enrolled simultaneously in 2 groups. Group A patients could not have received prior chemotherapy for locally recurrent or metastatic disease but may have received chemotherapy as part of primary curative therapy, if completed >6 months before registration. Group B patients had to have received 1 prior chemotherapy regimen for locally recurrent or metastatic disease, or chemotherapy as part of primary curative therapy <6 months before registration. Prior paclitaxel was permitted, but prior docetaxel or irinotecan at any time was not allowed. Other inclusion criteria included adequate hematologic and liver function test parameters and no peripheral neuropathy of grade 2 or worse. All patients signed informed consent, and the protocol was approved by each institution's human investigations committee.
Docetaxel (Sanofi-Aventis, Bridgewater, NJ) was administered as a 60-minute intravenous infusion at a dose of 35 mg/m2 followed by the administration of irinotecan (Pharmacia Corporation/Pfizer Inc, New York, NY) intravenously over 30 minutes at a dose of 60 mg/m2. Chemotherapy was given on Days 1 and 8 of a 21-day schedule and continued until progression of disease, unacceptable toxicity, or patient withdrawal. Patients received antiemetics and dexamethasone for a total of 3 doses, 12 hours before, 30 minutes before, and 12 hours after docetaxel. All toxicities were graded according to the National Cancer Institute's Common Toxicity Criteria (version 2.0). A dose reduction of docetaxel to 30 mg/m2 and irinotecan to 50 mg/m2 was allowed. High-dose loperamide, tincture of opium, and octreotide were used for treatment of delayed diarrhea.
Patient Assessments and Monitoring
Patients were evaluated by computed tomography (CT) of the chest and abdomen and CT or magnetic resonance imaging of the neck at baseline, within 4 weeks of registration, and then after every 3 cycles (9 weeks). Bone scan was performed at baseline and then as clinically indicated. When a patient was deemed to have an objective response, tumor measurements were to be repeated 4 to 6 weeks later to confirm the response. Complete blood counts were obtained on Days 1 and 8, and serum chemistry tests were on Day 1 of each cycle.
The primary endpoint was the overall objective response rate, which was defined as the proportion of patients with complete or partial response defined by Response Evaluation Criteria in Solid Tumors17 among eligible, treated patients, including patients not evaluable for response. In the first stage of a 2-stage design, 14 eligible patients (16 total to allow for a 10% ineligibility rate) were to be accrued to each of the 2 cohorts. Response rates of 40% and 20% were considered promising in Groups A and B, respectively. If at least 4 responses were seen in Group A and at least 1 response in Group B, accrual would continue to the second stage to accrue 18 additional patients in each group (ie, the total accrual goal was 32 eligible patients for each group). OS was defined as the time from registration to the date of death or last follow-up. Progression-free survival (PFS) was defined as the time from registration to disease progression or death from any cause or last follow-up. Two-stage confidence intervals, 2-stage power calculations, and the 2-stage stopping rules were used to analyze the data in regard to the study's 2-stage design.18 Fisher exact test19 was used to analyze the contingency tables of response and to compare the distribution of categorical data between groups. Cochran-Mantel-Haenszel test was used to assess the association between response and categories adjusting for the differences in prior treatment status. The Wilcoxon rank sum test20 was used to compare the distribution of continuous data between the 2 groups. Kaplan-Meier curves were plotted for OS and PFS.21 The log-rank test statistic was used to compare survival curves between categories, and a stratified log-rank test was used to adjust for differences in prior treatment status. Moreover, logistic and Cox proportional hazards regression models,22 respectively, were used to model objective response and survival data on covariates of interest while adjusting for prior treatment status. Mehta's exact test for ordered categorical data was used to test for associations between UGT1A1 genotype and toxicity severity.23 Two-sided P values are reported for all the statistical tests used in the analysis.
Immunohistochemistry for COX-2 and VEGF
Immunohistochemistry for the determination of COX-2 and VEGF in archival formalin-fixed paraffin-embedded tumor tissue was performed using commercially available antibodies: a monoclonal mouse antihuman COX-2 (Cayman Chemical, Ann Arbor, Mich), at 1:50 dilution, and a polyclonal rabbit antihuman VEGF(A-20) (Santa Cruz Biotechnology, Santa Cruz, Calif), at 1:500 dilution. For VEGF cytoplasmic staining, tumors were assigned a score of either 0 (negative), 1+ (weak, <1% of cells), 2+ (medium, 1%-10% of cells), or 3+ (strong, >10% of the tumor cells); for COX-2 membrane staining, tumors were assigned a score of either 0 (negative, or faint in <10%), 1+ (faint, >10% of cells), 2+ (moderate, >10% of cells), or 3+ (strong, complete in >10% of the tumor cells). The median intensity of VEGF was used to classify the cases into low (<1.5) or high (>1.5) VEGF categories. A cutoff of 2+ COX-2 intensity was used to classify the cases into low (≤2) or high (>2) categories.
Quantitative determination of human VEGF concentrations in serum was performed by enzyme-linked immunosorbent assay using either kit DY293B or DVE00 from R&D Systems (Minneapolis, Minn). Samples were measured in duplicate, and VEGF standard was included with every group of sera tested. The coefficient of variation for VEGF standard was ±10% (62.5-2000 pg/mL). The median serum VEGF score was used to classify the cases into low (≤394) or high (>394) serum VEGF categories.
Genomic DNA for UGT1A1 determination was prepared from whole blood (100-200 mL) using the QIAmp blood kit (Qiagen, Valencia, Calif). A standard polymerase chain reaction was performed using UGT1A1-specific primers that flanked the TATA region with the forward primer biotinylated. TA repeat number was determined by pyrosequencing with a PSQ 96MA pyrosequencer and software (Biotage, Uppsala, Sweden) by standard methods.24, 25 Genotypes for the cell lines DU145 and MCF-7 were previously determined to be 6 of 7 and 7 of 7, respectively. DU145 (n = 7) and MCF-7 (n = 3) were assayed with 100% accuracy for each gene. One or both of these cell lines were included as a positive control along with patient samples in each pyrosequencing reaction.
From October 2002 until August 2004, a total of 54 patients were enrolled, 18 in Group A and 36 in Group B. Two patients in Group B never started treatment; 1 patient withdrew consent and another signed consent but died before starting treatment. Four patients were reclassified with respect to their prior chemotherapy, 1 from Group A to B and 3 from Group B to A. As a result, there were 52 analyzable patients; 20 in Group A, and 32 in Group B. Although 4 objective responses were observed in Group A, 1 was not confirmed by repeat imaging, and accrual did not proceed to the second stage for that group. Patient characteristics and prior treatments are shown in Tables 1 and 2. Patients received a median of 3 cycles of irinotecan and docetaxel (range, 1-10 cycles).
|Group A, n=20||Group B, n=32|
|ECOG performance status|
|Primary tumor site|
|Salivary gland, parotid||0||0.0||1||3.1|
|Tumor cell differentiation|
|Weight loss in previous 6 months|
|<5% of body weight||12||70.6||14||46.7|
|5-<10% of body weight||3||17.6||5||16.7|
|10-<20% of body weight||1||5.9||8||26.7|
|≥20% of body weight||1||5.9||3||10.0|
|Pipe or cigar smoker only||0||0.0||1||3.4|
|Cigarette smoker <20 pack-years||2||10.0||2||6.9|
|Cigarette smoker 20-40 pack-years||4||20.0||11||37.9|
|Cigarette smoker >40 pack-years||12||60.0||12||41.4|
|<10 ounces of whiskey/wk*||14||73.7||20||69.0|
|10-32 ounces of whiskey/wk*||3||15.8||4||13.8|
|>32 ounces of whiskey/wk*||2||10.5||5||17.2|
|Metastatic site involvement|
|Group A, n=20||Group B, n=32|
|Received curative therapy only||6||100.0||20||62.5|
|Received palliative therapy only||—||—||7||21.9|
|Both palliative and curative||—||—||5||15.6|
|Prior biologic treatment||0||0.0||1||3.1|
In Group A, there were 3 partial responses for a response rate of 15% (90% confidence interval, 4.2%-34.4%) and in Group B, there was 1 partial response for a response rate of 3.1% (90% confidence interval 0.004%-19.6%). There was an additional unconfirmed partial response in each group (Table 3). Of the 52 patients, 13 were unevaluable because no post-treatment measurements were taken; 4 patients had symptomatic deterioration, 1 patient died before first follow-up assessment, 4 patients withdrew from study after only 1 cycle of treatment because of toxicity, 3 patients had inadequate data or were lost to follow-up, and 1 patient did not have a consistent method of evaluation.
|Group A, n=20||Group B, n=32|
|Best confirmed response*||No. (%)||No. (%)|
|Partial response||3 (15.0)||1 (3.1)|
|Stable disease||4 (20.0)||9 (28.1)|
|Progression||8 (40.0)||14 (43.8)|
|Unevaluable||5 (25.0)||8 (25.0)|
|Median progression-free survival, mo||3.3||1.9|
|1-Year progression-free survival||(10)||(6)|
|Overall survival, median, mo||8.2||5.0|
|1-Year overall survival||(35)||(16)|
OS and PFS
At the time of this analysis, all patients but 1 have progressed or died. For patients in Group A, the median OS was 8.2 months, and for patients in Group B, the median OS was 5.0 months (Table 3 and Fig. 1). In Group A, the median PFS was 3.3 months, and in Group B, the median PFS was 1.9 months. No baseline characteristic was found to be statistically significant in predicting survival, but the study was not powered to identify such factors.
Three deaths, all in Group B, were deemed possibly related to study treatment. Two were because of sepsis, in 1 case associated with neutropenia, diarrhea, and dehydration. The third patient presented after 2 weeks of treatment on cycle 1 with fever, chills, and dyspnea, refused treatment, and died at home 2 days later from presumed pneumonia. Two other patients, 1 in each group, died from grade 5 carotid hemorrhage without thrombocytopenia, which was attributed to disease progression. The most common grade 3 or 4 events in the 2 groups combined were diarrhea (grade 3, 21%; grade 4, 4%), fatigue (grade 3, 17%), anorexia (grade 3, 8%; grade 4, 4%), and neutropenia (grade 3, 8%; grade 4, 13%). Only 1 patient (2%) had febrile neutropenia (Table 4).
|Group A, n=20||Group B, n=32|
|Grade 3||Grade 4||Grade 5||Grade 3||Grade 4||Grade 5|
|Transfusion of packed red blood cells||1||5%||—||—||—||—||3||9%||—||—||—||—|
|Infection with grade 3-4 neutropenia||1||5%||1||5%||—||—||—||—||1||3%||1||3%|
|Infection with unknown neutrophil count||—||—||—||—||—||—||1||3%||—||—||—||—|
|Infection without neutropenia||—||—||—||—||—||—||—||—||2||6%||1||3%|
|Depressed level of consciousness||—||—||—||—||—||—||2||6%||—||—||—||—|
COX-2 and VEGF
Forty-one patients consented for the use of their samples for correlative studies. COX-2 and VEGF tumor expression data were available for 31 and 29 cases, respectively. We did not detect any significant differences in OS or PFS between groups on the basis of COX-2 or VEGF expression, presumably because of small sample sizes (Table 5). Baseline serum VEGF data were available for 18 patients. The association between serum VEGF at baseline and median PFS is shown in Table 5. Patients with high baseline VEGF levels had a median PFS of 2.84 months versus 1.73 months for patients with low VEGF (P = .085, using stratified log-rank test).
|No. of Patients||Median Months||P*|
|All Patients||Chemonaive Group A||Chemoexposed Group B|
We explored the association between polymorphisms in the UGT1A1 gene and race, neutropenia, diarrhea, and any toxicity among 35 patients with available data. There were no statistically significant differences in the pattern of worst degree toxicity, or in the grade intensity of neutropenia or diarrhea by TA repeat category (data not shown).
We evaluated the combination of docetaxel and irinotecan, a novel non–platinum-containing regimen, in the first- or second-line treatment of patients with recurrent or metastatic SCCHN. Both drugs were given on a weekly schedule of administration based on prior phase 1 experience.13 Phase 2 trials of docetaxel and irinotecan, using weekly or every 3 weeks schedules of administration, have been conducted in many other solid tumors.26-33 To the best of our knowledge, this is the only phase 2 study of docetaxel and irinotecan in recurrent or metastatic SCCHN. Although objective responses were observed, the prespecified criteria for efficacy were not met. In the first-line setting (Group A), 4 objective responses were observed as required per study design, but 1 was unconfirmed, which did not allow the study to accrue beyond the first stage of a 2-stage Simon design. The statistical design assumed a target response rate of 40% in Group A, which in retrospect was rather high for the cooperative group setting and with the application of Response Evaluation Criteria in Solid Tumors. Other cooperative group studies in comparable patient populations, such as E5397 and E1595, showed that single-agent chemotherapy with cisplatin achieves an objective response rate of 10% and median survival of 8 months (E5397),34 and that cisplatin doublets (cisplatin/5-FU or cisplatin/paclitaxel) result in objective response rates of 26% to 27% and median survival of 8.1 to 8.7 months (E1395).3 In the current study, docetaxel and irinotecan produced a response rate of 17% (22% counting an unconfirmed response), median PFS of 3.3 months, and median OS of 8.2 months. Therefore, survival results with docetaxel and irinotecan may be comparable to platinum-based combinations.
The group of patients treated in the second-line setting (Group B) is 1 of the largest that have been studied so far, and it was characterized by a high representation of patients with distant metastasis (78%) and previous treatment with paclitaxel (81%). In these patients, the efficacy of docetaxel and irinotecan, with a response rate of 3% (6% counting an unconfirmed response), median PFS of 1.9 months, and median OS of 5.0 months, cannot be considered satisfactory. As a single agent, irinotecan may be inactive in previously treated recurrent or metastatic SCCHN,11 whereas data with single-agent docetaxel is limited in a similar setting, but activity has been reported.35
The docetaxel and irinotecan regimen we used in our study was associated with expected toxicities, which were predominantly nonhematologic, including diarrhea, anorexia, and fatigue. Although most of the patients had received paclitaxel in the past, grade 3 or 4 neuropathy was not seen. Grade 4 toxicity was observed in 30% of chemotherapy-naive patients in this study versus 42% and 50% with cisplatin/paclitaxel and cisplatin/5-FU, respectively, in E1395, whereas there was no treatment-related death versus 5% and 7% with cisplatin/paclitaxel and cisplatin/5-FU, respectively.3 However, there were some differences in the toxicity criteria used between these ECOG studies, so the rates of grade 3 and 4 toxicities may not be directly comparable. Polymorphisms in the UGT1A1 gene have demonstrated racial variability and have been shown to be associated with differences in observed toxicities among patients treated with irinotecan. We could not demonstrate any correlation of toxicities with UGT1A1 genotypes, possibly because of the small sample size, or the low dose of irinotecan used in this study.
COX-2 and VEGF are overexpressed in SCCHN and have been suggested as potential predictors of outcome.14-16, 36 Moreover, COX-2 expression has been reported to correlate with the expression of VEGF in SCCHN. On the basis of preclinical observations, it has been proposed that docetaxel may have an antiangiogenic effect.37 We elected to assess VEGF as well as COX-2 on baseline tumor tissue and attempted to associate its expression with outcome. However, in the clinical setting examined, we could not demonstrate that expression of either COX-2 or VEGF correlated with worse survival, possibly because of insufficient sample size.38 Patients with high serum VEGF levels had a trend toward improved PFS with docetaxel and irinotecan versus patients with low levels at baseline, an observation that may require further evaluation in subsequent studies.
The docetaxel and irinotecan regimen we used was feasible and was associated with a toxicity profile potentially favorable to cisplatin-based combinations. However, its antitumor activity is unlikely to be superior to platinum-based combinations in the first-line treatment of recurrent or metastatic SCCHN, whereas its antitumor activity in the second-line setting was rather disappointing. It has been reported that selected patients, such as those with tumors with high levels of excision repair cross complementation Group 1, may benefit less from platinum-based chemotherapy.39 Whether docetaxel and irinotecan, a nonplatinum doublet, will be beneficial in selected patients, such as those with tumors with high excision repair cross complementation Group 1, is a worthwhile hypothesis to be evaluated in future clinical trials in patients with SCCHN.
Conflict of Interest Disclosures
This study was conducted by the Eastern Oncology Cooperative Group (Robert L. Comis, MD, Chair) and supported by Grants CA-23,318 and CA-21,115, and Sanofi-Aventis.
Drs. Argiris and Forastiere have served on Advisory Boards for Sanofi-Aventis.
- 4Open-label, uncontrolled, multicenter phase II study to evaluate the efficacy and toxicity of cetuximab as a single agent in patients with recurrent and/or metastatic squamous cell carcinoma of the head and neck who failed to respond to platinum-based therapy. J Clin Oncol. 2007; 25: 2171-2177., , , et al.
- 6Randomized phase III trial of docetaxel versus vinorelbine or ifosfamide in patients with advanced non-small-cell lung cancer previously treated with platinum-containing chemotherapy regimens. The TAX 320 Non-Small Cell Lung Cancer Study Group. J Clin Oncol. 2000; 18: 2354-2362., , , et al.
- 17New guidelines to evaluate the response to treatment in solid tumors. European Organization for Research and Treatment of Cancer, National Cancer Institute of the United States, National Cancer Institute of Canada. J Natl Cancer Inst. 2000; 92: 205-216., , , et al.
- 19Categorical data analysis. New York, NY: John Wiley & Sons; 1990..
- 22Regression models and life tables. J R Stat Soc Series B Stat Methodol. 1972; 23: 187-202..
- 39Excision repair cross complementation group 1 immunohistochemical expression predicts objective response and cancer-specific survival in patients treated by Cisplatin-based induction chemotherapy for locally advanced head and neck squamous cell carcinoma. Clin Cancer Res. 2007; 13: 3855-3859., , , et al.