- © 2010 by American Society of Clinical Oncology
Epidermal Growth Factor Receptor Inhibitor Gefitinib Added to Chemoradiotherapy in Locally Advanced Head and Neck Cancer
- Ezra E.W. Cohen⇓,
- Daniel J. Haraf,
- Rangesh Kunnavakkam,
- Kerstin M. Stenson,
- Elizabeth A. Blair,
- Bruce Brockstein,
- Eric P. Lester,
- Joseph K. Salama,
- Allison Dekker,
- Rosalyn Williams,
- Mary Ellyn Witt,
- Tatyana A. Grushko,
- James J. Dignam,
- Mark W. Lingen,
- Olufunmilayo I. Olopade and
- Everett E. Vokes
- From the University of Chicago and University of Chicago Cancer Research Center, Chicago; North Shore University Health System, Evanston, IL; and Oncology Care Associates, St Joseph, MI.
- Corresponding author: Ezra E.W. Cohen, MD, Department of Medicine, University of Chicago, 900 East 57th St, Chicago, IL; e-mail: ecohen{at}medicine.bsd.uchicago.edu.
-
Previously presented at the 41st Annual Meeting of the American Society of Clinical Oncology, May 13-17, 2005, Orlando, FL, and the 43rd Annual Meeting of the American Society of Clinical Oncology, June 1-5, 2007, Chicago, IL.
Abstract
Purpose Assess efficacy and toxicity of gefitinib, an epidermal growth factor receptor (EGFR) inhibitor, added to, and in maintenance after, concurrent chemoradiotherapy (CCRT) in locally advanced head and neck cancer (LA-HNC) and correlate outcomes with EGFR gene copy number alterations.
Patients and Methods Patients with stage III to IV LA-HNC received two cycles of carboplatin/paclitaxel induction chemotherapy (IC) followed by split-course CCRT with fluorouracil, hydroxyurea, twice daily radiotherapy (FHX), and gefitinib (250 mg daily) followed by continued gefitinib for 2 years total. The primary end point was complete response (CR) rate after CCRT. EGFR gene copy number was assessed by fluorescent in situ hybridization.
Results Sixty-nine patients (66 with stage IV disease, 37 with oropharynx primary tumors, and 67 with performance status 0 to 1) were enrolled with a median age of 55 years. Predominant grade 3 or 4 toxicities during IC and CCRT were neutropenia (n = 20) and in-field mucositis (n = 59) and dermatitis (n = 23), respectively. CR rate after CCRT was 90%. After median follow-up of 3.5 years, 4-year overall, progression-free, and disease-specific survival rates were 74%, 72%, and 89%, respectively. To date, one patient has developed a second primary tumor in the aerodigestive tract. In 31 patients with available tissue, high EGFR gene copy number was associated with worse overall survival (P = .02).
Conclusion Gefitinib can be administered with FHX and as maintenance therapy for at least 2 years, demonstrating CR and survival rates that compare favorably with prior experience. High EGFR gene copy number may be associated with poor outcome in patients with LA-HNC treated with this regimen.
INTRODUCTION
Locally advanced squamous cell carcinoma of the head and neck (LA-SCCHN) is a global public health issue, with more than 500, 000 cases diagnosed per year.1 Primary therapy for LA-SCCHN often uses a multimodality approach consisting of concurrent chemoradiotherapy (CCRT) with or without surgery.2 The administration of induction chemotherapy (IC) has recently gained prominence as supported by meta-analysis data.3 In addition, phase III trials have demonstrated that the addition of a taxane to cisplatin and fluorouracil yields improvements in complete response rates, survival, and larynx preservation.4–6
An association between epidermal growth factor receptor (EGFR) expression and survival has been noted in SCCHN.7 A randomized study demonstrated that the addition of cetuximab to radiotherapy improves overall survival (OS),8 whereas phase I studies have demonstrated the feasibility of administering EGFR inhibitors concurrently with chemoradiotherapy.9–11 Gefitinib is an EGFR small-molecule tyrosine kinase inhibitor (TKI) with radiosensitizing properties consistently demonstrating tumor growth delay and enhancement of apoptosis in preclinical models.12 Furthermore, single-agent activity has been demonstrated in recurrent and metastatic disease.13
Previous investigations have demonstrated that patients with LA-SCCHN have high survival and organ preservation rates after therapy with sequential carboplatin/paclitaxel IC and CCRT with paclitaxel/fluoruracil/hydroxyurea and hyperfractionated radiotherapy (TFHX).14 We hypothesized that gefitinib, as a targeted agent, would be well tolerated and a less toxic radiation enhancer than paclitaxel. Furthermore, we hypothesized that administration of gefitinib as a maintenance agent would be feasible with the potential to decrease failure rates. We therefore undertook a phase II study to incorporate gefitinib into a standard chemoradiotherapy regimen in patients with LA-SCCHN.
PATIENTS AND METHODS
Treatment-naive patients, 18 years or older with Eastern Cooperative Oncology Group performance status ≤ 2, were required to have histologically confirmed SCCHN, stage III (only tongue base or hypopharyngeal primaries) or IV, without evidence of metastatic disease, and intact organ function as previously described.14 Postoperative patients were eligible only if surgery was organ-preserving, consisting of simple excision of a primary lesion or selective neck dissection. Institutional review board approval was obtained, and all patients provided informed consent.
IC consisted of two cycles of paclitaxel (100 mg/m2 days 1, 8, and 15) and carboplatin (area under the curve of 6, day 1) as previously described.14 CCRT, scheduled to begin 1 to 2 weeks after the last dose of paclitaxel, consisted of 4 to 5 14-day cycles (5 days of 500 mg of hydroxyurea orally every 12 hours, 600 mg/m2/d of continuous-infusion fluorouracil, and 1.5 Gy of radiation twice per day followed by 9 days without therapy) based on extent of preprotocol surgery as previously described.14 Gefitinib (AstraZeneca, London, United Kingdom), 250 mg every day orally, was begun on the first day of radiotherapy and continued for a maximum of 2 years. Radiotherapy and surgical guidelines are described fully in the Appendix, online only.
Laboratory Analyses
Fluorescent in situ hybridization (FISH), detection of human papillomavirus (HPV) DNA, and EGFR immunohistochemistry (IHC) methods are fully described in Appendix (online only). Tumor biopsy was performed and tissue collected as previously described.15 FISH was performed using the LSI Locus Specific Identifier DNA Probes (Abbot Molecular, Des Planes, IL). Tumors with an EGFR:CEP7 signal ratio less than 2 were considered nonamplified, whereas those with a ratio of ≥ 2 were considered amplified. The alterations in EGFR signals due to alterations in chromosome 7 copy number were classified as previously described,16 and FISH positive was defined by amplification or high polysomy. For HPV analysis, in situ hybridization was performed according to manufacturer's guidelines using the Benchmark automated slide-staining system (Ventana Medical Systems, Tucson, AZ). For IHC, rabbit primary antibody (Santa Cruz Biotechnology, Santa Cruz, CA, catalog no. sc-03) was applied at 1:75 dilution and scored on a 0 to 3+ scale.
Statistical Considerations
The primary end point was complete response (CR) rate achieved 1 month after CCRT using Response Evaluation Criteria in Solid Tumors (RECIST). Patients who achieved a CR at the primary site, either clinically or radiologically (by computed tomography [CT] or magnetic resonance imaging), had that confirmed with biopsy whenever possible. Patients with less than CR at the primary site underwent biopsy. If this was negative, the patient was considered a complete responder. If this was positive or not performed, response was scored as the worst of either clinical or radiologic response assessment. Patients with less than CR in the neck underwent neck dissection, and the same algorithm was followed as for primary site response assessment. The overall response assessment consisted of both primary site and neck nodes using the worst response categorization in these areas. All patients who received at least one dose of therapy and had measurable disease, including those with incompletely resected disease before study entry, were assessed for response. All patients who received at least one dose of therapy were included in survival analyses. The study was conducted as a single arm, two-stage, noninferiority phase II trial with a total sample size of 59 evaluable patients, which was sufficient to detect a 15% lower CR rate than 80% achieved with TFHX historical controls14 and α and β of 0.05 and 0.20, respectively. During the first stage, 25 evaluable patients were recruited, with a plan to terminate if ≤ 16 CRs were observed; otherwise, an additional 34 patients would be recruited. The new treatment would be deemed not inferior to historical if ≥ 45 CRs were observed. The false-negative rate in the first stage was less than 0.05.
OS, progression-free survival (PFS), and disease-specific survival (DSS) were calculated using the Kaplan-Meier product-limit estimate and expressed as probabilities with 95% CIs. Univariate and multivariate analyses of patient and disease factors including age, race, sex, weight loss, performance status, stage, and tumor site in relation to failure risk were conducted using Cox proportional hazards regression models. EGFR status was evaluated for prognostic effect on OS and PFS by the log-rank test. Further detailed statistical analysis is provided in the Appendix (online only).
RESULTS
Patients Characteristics
From February 2003 to October 2004, 70 patients signed consent at three participating institutions, although one patient withdrew before initiating therapy and is not included in any analyses (Fig 1, Table 1, and Appendix Table A1, online only). As expected, the majority of HPV-positive tumors (15 of 17) were from oropharynx primary sites. Conversely, of the 22 oropharynx tumors for which tissue was available for HPV testing, 68% were HPV positive. Ten patients underwent surgery to remove all gross disease before initiating study and were thus unevaluable for response.
Flow diagram of all patients who signed informed consent. IC, induction chemotherapy; CRT, chemoradiotherapy.
Patient and Tumor Characteristics
Treatment Delivery and Toxicity
The frequency and grade of adverse events observed during IC and CCRT are presented in Table 2. All patients, except two, completed both cycles of IC, and 57 patients did so within the scheduled 8 weeks. One patient voluntarily withdrew from study during IC, and another was diagnosed with a duodenal ulcer. As expected, the most frequent toxicity encountered during IC was neutropenia. Twelve patients required dose modification during IC, in the majority of cases for neutropenia (10 of 12).
Number of Patients With Respective Toxicity Grade Observed During Induction Chemotherapy and Concurrent Chemoradiotherapy Listed by Frequency (n = 69)
The predominant toxicity observed during CCRT was radiation mucositis, which had a characteristic pattern, with a plateau in intensity midway through treatment and gradual abatement (Fig 2). Typical EGFR TKI toxicity was also observed, including rash and diarrhea, but neither was dose-limiting. Sixty-six of 69 patients were treated with intensity-modulated radiotherapy techniques, whereas 64 received all intended radiotherapy fractions as scheduled. Two patients declined to receive gefitinib during chemoradiotherapy, whereas 46 of 67 received 90% of all planned doses of systemic agents as scheduled. Twelve subjects received RBC transfusion at some point during CCRT. There were four treatment-related deaths during or immediately after CCRT: two from bacterial sepsis without neutropenia, one cardiac event, and one sudden death after planned neck dissection.
Mean toxicity grade of (A) acute in-volume mucositis and (B) dermatitis over time. The planned duration of chemoradiotherapy was 7 to 9 weeks. Toxicities were graded on a weekly basis starting from initiation of concurrent chemoradiotherapy to at least 4 weeks after completing chemoradiotherapy. Common Terminology Criteria of Adverse Events (CTCAE) version 3.0 was used to grade toxicity. Error bars represent standard deviation.
After completion of radiotherapy, patients were eligible to continue on gefitinib alone. The median duration of gefitinib maintenance therapy was 667 days. Nine subjects never received gefitinib after CCRT (four patients died during or after CCRT, one patient had progressive disease, three patients voluntarily refused, and one patient developed respiratory failure secondary to exacerbation of chronic obstructive pulmonary disease. This latter patient underwent CT scanning, which demonstrated no evidence of interstitial lung disease, but this patient did not restart gefitinib). Forty-eight of the 60 patients who began maintenance were able to complete 2 years of gefitinib; 12 patients discontinued gefitinib maintenance for elevated liver transaminases (n = 3), rash (n = 2), diarrhea (n = 1), xerostomia (n = 1), nausea (n = 1), comorbid illness (n = 2), voluntary refusal (n = 1), and disease recurrence (n = 1).
Institutional policies regarding gastrostomy tube placement included placement only when necessary. Nine patients underwent gastrostomy tube insertion before initiating therapy, whereas 25 patients required gastrostomy tube placement during or shortly after therapy, and 35 patients never required feeding tubes. In patients who survived at least 1 year from completion of CCRT, 11 patients still required a gastrostomy tube to maintain nutrition. Three of these patients were able to swallow, and another two required tube maintenance secondary to disease recurrence. In total, five patients underwent esophageal dilation to address dysphagia; three of these patients never required tube placement. Other long-term toxicities included two patients who required surgery for osteoradionecrosis of the mandible, whereas another patient underwent hyperbaric oxygen for bone exposure.
Efficacy
Ten patients who underwent resection of disease before initiating therapy were inevaluable for response (Fig 1). Fifty-nine patients were evaluable for response to IC (Fig 1), and consistent with prior studies, the overall response rate was 90%, with seven CRs (Appendix Table A2, online only). CR was observed in 52 patients after CCRT (Appendix Table A2), including one patient who underwent surgical resection of gross disease between IC and CCRT and is included in the response analysis to CCRT as stable disease. Neck dissection was performed on 36 patients after completion of CCRT; 32 patients were pathologically free of disease. Of the four patients with evidence of cancer in the neck dissection specimen, three have died of disease.
With a median follow-up of 3.54 years (range, 0.3 to 4.7 years), there have been 18 deaths and 10 patients with documented progressive disease: two local only, one regional only, two locoregional, one regional with distant metastasis, and four distant metastatic only, yielding 4-year OS and PFS rates of 74% (95% CI, 62% to 83%) and 72% (95% CI, 60% to 81%), respectively (Fig 3A and 3B). In addition, the 4-year DSS rate was 89% (95% CI, 77% to 95%; Fig 3C), with eight patients dying of SCCHN. One subject underwent laryngectomy for recurrent disease and is disease-free. The causes of non-SCCHN–related mortality were treatment related (four described above), respiratory disease (three unrelated to treatment with no evidence of interstitial lung disease on CT scan), cardiac disease (one patient), and second malignancy (one patient with hepatocellular carcinoma and one patient with esophageal carcinoma). One patient with local progression underwent salvage laryngectomy and is alive, whereas one patient is alive with pulmonary metastases. No patients developed a second primary SCCHN or lung cancer.
Kaplan-Meier survival curves. Survival estimates of (A) overall survival, (B) progression-free survival, and (C) disease-specific survival for all patients.
Univariate and multivariate analyses were performed to assess whether patient or disease characteristics were associated with OS or PFS (Appendix Table A3, online only). In univariate analysis, increasing age, black race, and nonoropharynx primary tumors were consistently associated with worse outcome. Appendix Figure A1 (online only) displays OS, PFS, and DSS in patients with oropharynx and nonoropharyx primary tumors. In multivariate analysis, only age and race were associated with both outcomes, whereas patients with nonoropharynx primary tumors maintained a statistically significantly worse PFS.
Laboratory Correlatives
EGFR protein expression and, more recently, gene copy number have been suggested as negative prognostic markers in patients with SCCHN.7,16–18 We hypothesized that administration of an EGFR inhibitor within a chemoradiotherapy platform would abrogate the negative consequences of increased EGFR expression. Tumor tissue collected before initiating treatment was adequate for EGFR FISH and IHC analysis in 31 and 21 patients, respectively (Appendix Table A4, online only). Quantification of EGFR gene copy number displayed a great deal of intra- and inter-tumor heterogeneity (Fig 4). EGFR amplification was present in three cases, whereas 10 cases had elevated EGFR copy number per cell due to chromosome 7 polysomy. There was a statistically significant association between EGFR gene copy number and IHC staining intensity (Appendix Table A5, online only; P = .02). We grouped patients with EGFR amplification and high polysomy together as previously described and defined these samples as EGFR FISH positive. EGFR status was not associated with patient sex, race, age, primary tumor site, prior alcohol or tobacco use, or weight loss before study entry. Moreover, when examining a relationship between EGFR status and outcome, there was no association between gene copy number and response to chemoradiotherapy; however, we found that EGFR FISH-positive status was associated with shorter OS time (P = .021), as well as a trend toward earlier time to recurrence (P = .067; Appendix Table A6, online only).
Photomicrographs (×1,200) for EGFR (red) and CEP7 (green). (A) Disomy. (B) EGFR trisomy in 15% (arrow) and tetrasomy in 2% of cells (arrowhead), classified EGFR fluorescent in situ hybridization (FISH) negative. (C) Cells are trisomic (arrow) or highly polysomic (arrowhead), classified EGFR FISH positive. (D) High polysomy in 83% (arrowhead), classified EGFR FISH positive. (E, F) Low and high EGFR amplification, classified EGFR FISH positive.
DISCUSSION
This multi-institutional phase II study in LA-SCCHN reports the efficacy, toxicity, and correlative results of EGFR TKI administration with chemoradiotherapy. The observed CR rate and other efficacy outcomes compare favorably with our prior experience,14 with dramatically reduced rates of neuropathy and myelosuppression (any grade neuropathy 36% v 17% and grade 3 to 4 neutropenia 30% v 16%, respectively, for paclitaxel- and gefitinib-containing regimens).
Since initiation of the present study, further data have been reported for administering EGFR inhibitors in LA-SCCHN8,9,11,19,20 supporting feasibility, whereas promising preliminary data have been presented combining cetuximab or panitumumab with platinum-based chemoradiotherapy platforms. Argiris et al21 reported 88% 2-year survival in 39 patients receiving cisplatin/docetaxel/cetuximab IC followed by cisplatin/cetuximab/radiotherapy and continuing cetuximab for a maximum of 6 months. The Eastern Cooperative Oncology Group22 observed a 76% 1-year survival in 60 patients with unresectable LA-SCCHN receiving cisplatin/cetuximab/radiotherapy. A phase I study combining panitumumab with concurrent paclitaxel, carboplatin, and radiotherapy demonstrated tolerability, and after a median follow-up of 21 months, 95% of patients remained disease-free.23 Cumulatively, these data suggest a genuine improvement in efficacy when EGFR inhibitors are added to chemoradiotherapy regimens, and we await ongoing phase III trials to establish the validity of this hypothesis.
The optimal dose of gefitinib in SCCHN has been controversial. Single-arm phase II studies in recurrent or metastatic SCCHN suggested that 500 mg was more efficacious than 250 mg,13,24 the standard dose in lung cancer, and a phase III study comparing gefitinib to methotrexate in recurrent or metastatic SCCHN confirmed a higher response rate of 500 mg compared with 250 mg.25 The current study was planned and completed before availability of the phase III data, and one could speculate that 500 mg would have been more effective. However, in the context of CCRT, it is unclear whether the higher dose is indicated, especially considering its greater associated toxicity. Moreover, four treatment-related deaths were observed on this study, and any measure that further increased toxicity should be undertaken with caution. One could argue that efforts now need to concentrate on reducing toxicity of CCRT regimens in selected patients.
Acute in-volume radiation toxicity observed during this study was manageable and comparable with that encountered during TFHX. It is notable that acute mucositis and dermatitis associated with FHX regimens have a unique pattern, with a peak midway through treatment, a plateau lasting to the end of radiotherapy, and then a gradual abatement (Fig 2). This contrasts with regimens administering uninterrupted single daily or accelerated fractionated radiation that peak late or after completion of treatment. Several investigators have suggested that measuring the highest grade toxicity during treatment can be misleading without reflecting duration of and recovery from adverse effects.26 In this study, we measured acute mucositis and dermatitis as a function of time during and after treatment, which allows more thorough evaluation, and we believe that these metrics will serve as better indicators of toxicity, allowing valid comparisons between regimens.
Gefitinib was feasibly administered in the majority of patients during the maintenance phase, with overall excellent adherence. The study design does not allow determination of this component's contribution to survival; however, it is intriguing to note that with 42 months median follow-up, only 10 patients have had disease recurrence. As a cautionary note, however, similar attempts in locally advanced non–small-cell lung cancer in which gefitinib was also administered after CCRT did not improve survival.27
EGFR gene copy number was a poor prognostic feature in this study, associated with shorter PFS and OS. We anticipated that EGFR FISH-positive patients would benefit most from EGFR TKI, with the hypothesis that their cancers depend on EGFR signaling for growth and survival. Recently, an analysis of associations between EGFR gene copy number and outcomes in a phase III study in recurrent or metastatic SCCHN adding cetuximab to platinum/FU chemotherapy found no predictive value of this parameter.28 Moreover, another randomized in similar patients found that patients with high EGFR expression by IHC fared worse when receiving cetuximab as compared with those with moderate or low expression.29 Finally, the randomized study comparing gefitinib to methotrexate in patients with recurrent or metastatic SCCHN reported that high EGFR gene copy number portended a poorer prognosis but was not predictive of EGFR TKI efficacy.25 Taken together with our study, EGFR gene copy number has prognostic value in SCCHN but may not help in selecting patients for treatment with EGFR inhibitors. The search for validated predictive biomarkers in this arena is still ongoing.
There are inherent limitations to comparing our results with historical controls, which is especially relevant with the realization that the majority of oropharynx primary tumors are HPV related.30 These cancers are also being detected with increasing frequency and carry a better prognosis.30 Within our study samples evaluated for HPV, 68% of oropharynx tumors were positive, an incidence consistent with large epidemiologic and retrospective series. Extrapolation to the entire study population approximates 25 patients or 36% (68% of 37 oropharynx tumors) being HPV positive. Along those lines, survival in patients with oropharynx tumors was significantly higher than in those with nonoropharynx. Without knowing the incidence of HPV positivity in prior studies, it is difficult to draw definitive conclusions comparing regimens. Moreover, although the sample size in the nonoropharynx subgroup does not allow adequately powered comparisons with older regimens, the efficacy observed in these patients is encouraging.
This study confirms the feasibility of administering an EGFR TKI with chemoradiotherapy and as maintenance to patients with LA-SCCHN. In addition, we found promising efficacy of this approach, highlighted by a 72% 4-year survival rate. Further pursuit of combination EGFR TKI and chemoradiotherapy is warranted, and randomized studies assessing the contribution of such agents (eg, lapatinib) are being pursued.
AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST
Although all authors completed the disclosure declaration, the following author(s) indicated a financial or other interest that is relevant to the subject matter under consideration in this article. Certain relationships marked with a “U” are those for which no compensation was received; those relationships marked with a “C” were compensated. For a detailed description of the disclosure categories, or for more information about ASCO's conflict of interest policy, please refer to the Author Disclosure Declaration and the Disclosures of Potential Conflicts of Interest section in Information for Contributors.
Employment or Leadership Position: None Consultant or Advisory Role: Ezra E.W. Cohen, AstraZeneca (C); Everett E. Vokes, AstraZeneca (C) Stock Ownership: None Honoraria: Ezra E.W. Cohen, AstraZeneca Research Funding: None Expert Testimony: None Other Remuneration: None
AUTHOR CONTRIBUTIONS
Conception and design: Ezra E.W. Cohen, Daniel J. Haraf, Kerstin M. Stenson, Everett E. Vokes
Administrative support: Allison Dekker
Provision of study materials or patients: Ezra E.W. Cohen, Daniel J. Haraf, Kerstin M. Stenson, Elizabeth A. Blair, Bruce Brockstein, Eric P. Lester, Joseph K. Salama, Mark W. Lingen, Everett E. Vokes
Collection and assembly of data: Ezra E.W. Cohen, Rangesh Kunnavakkam, Allison Dekker, Rosalyn Williams, Mary Ellyn Witt, Tatyana A. Grushko, Mark W. Lingen
Data analysis and interpretation: Ezra E.W. Cohen, Daniel J. Haraf, Rangesh Kunnavakkam, Tatyana A. Grushko, James J. Dignam, Mark W. Lingen, Olufunmilayo I. Olopade, Everett E. Vokes
Manuscript writing: Ezra E.W. Cohen, Daniel J. Haraf, Rangesh Kunnavakkam, Bruce Brockstein, Eric P. Lester, Joseph K. Salama, Tatyana A. Grushko, James J. Dignam, Mark W. Lingen, Everett E. Vokes
Final approval of manuscript: Ezra E.W. Cohen, Daniel J. Haraf, Rangesh Kunnavakkam, Kerstin M. Stenson, Elizabeth A. Blair, Bruce Brockstein, Eric P. Lester, Joseph K. Salama, Allison Dekker, Rosalyn Williams, Mary Ellyn Witt, Tatyana A. Grushko, James J. Dignam, Mark W. Lingen, Olufunmilayo I. Olopade, Everett E. Vokes
Acknowledgment
We thank Dezheng Huo, Shang Lin, Ann Mauer, and Syed Ahmed.
Appendix
Radiotherapy and Surgical Guidelines
All patients underwent computed tomography–based treatment planning both before and after induction chemotherapy. For each patient, four customized, risk-defined volumes were designed, as previously described14 (gross tumor volume, consisting of all known disease, clinically and radiographically; planned treatment volume [PTV] 1, consisting of the gross tumor volume plus a margin for setup uncertainty and organ motion; PTV2, including PTV1 as well as regions at high risk for microscopic involvement; and PTV3, including PTV2 as well as other regions with moderate risk for microscopic involvement). PTV1, PTV2, and PTV3 received 72 Gy, 51 Gy, and 36 Gy, respectively, by three-dimensional conformal or intensity-modulated radiotherapy techniques. Patients who underwent resection of all gross disease received 60 to 66 Gy. The dose limit to the spinal cord was less than 40 Gy with three-dimensional radiation and 46 Gy with intensity-modulated radiotherapy. RBC transfusion was recommended for hemoglobin ≤ 10 g/dL.
Simple excision of a primary lesion or selective neck dissection was allowed before starting study if preserving organ function was possible. When neck dissection was not performed initially, selective neck dissection was offered after concurrent chemoradiotherapy (CCRT) for any residual nodal disease and in the absence of macroscopic residual disease for patients initially staged N2b, N2c, or N3. In patients with pathologically proven residual disease, complete excision of disease was undertaken.
Fluorescent In Situ Hybridization
Fluorescent in situ hybridization was performed on frozen tissue sections adjacent to those stained with hematoxylin and eosin using the LSI Locus Specific Identifier DNA Probes (Abbot Molecular, Des Planes, IL), which is a hybridization mixture of a EGFR probe labeled with SpectrumOrange and chromosome 7 enumeration probe CEP7, labeled with SpectrumGreen. Slides were fixed in 4% formaldehyde in phosphate-buffered saline (PBS) at 0°C for 15 minutes, washed in PBS, and processed for fluorescent in situ hybridization according to manufacturer protocol. In each tumor sample, an average of 60 (range, 30 to 94) well-defined malignant nuclei were scored. The absolute number of EGFR and CEP7 signals, the mean copy number of each signal per cell the ratio of EGFR signals to CEP7 signals, and the percentage of cells with given copy number of each signal per cell were recorded.
Detection of Human Papillomavirus DNA
Four-micrometer sections were cut and mounted on positively charged glass slides. In situ hybridization was performed according to manufacturer's guidelines using the Benchmark automated slide-staining system (Ventana Medical Systems, Tucson, AZ). The Inform human papillomavirus (HPV) III probes set (Ventana Medical Systems), which detects 13 types of high-risk HPV (types 16, 18, 31, 33, 35, 45, 45, 51, 52, 56, 58, 59, 68, and 70), was used as the probe. The HPV 3 in 1 System control slides (Ventana Medical Systems) were used for positive and negative controls. Nuclear staining was considered a positive result for HPV DNA.
Epidermal Growth Factor Receptor Immunohistochemistry
Frozen tissue sections were air dried at room temperature and fixed in cold acetone for 10 minutes. After fixation, the rabbit primary antibody (Santa Cruz Biotechnology, Santa Cruz, CA; catalog no. sc-03) was applied at a dilution of 1:75 in PBS for 1 hour at room temperature. Antibody binding was visualized with antirabbit polymer–labeled horseradish peroxidase–bound secondary reagent (EnVision+, DAKO, Carpinteria, CA). All three immunohistochemistry (IHC) stains were developed with DAB chromogen and counterstained with hematoxylin. Corresponding negative control experiments were performed by omitting the incubation step with the primary antibody.
Statistical Considerations
Overall survival (OS) was measured from the date of study entry to the date of death or the date of the last follow-up visit. Patients who had not died or who were lost to follow-up were censored for OS at the time last known to be alive. Progression-free survival (PFS) was measured from the date of study entry to first evidence of disease recurrence, death from any cause, or last follow-up. Patients who were alive, had not experienced disease progression, or were lost to follow-up were censored for PFS at the time last known to be alive and progression-free. Disease-specific survival was measured from the date of study entry to the date of death or the date of last follow-up visit. Non–cancer-related and non–treatment-related deaths were censored.
Univariate and multivariate analyses of patient and disease factors, including age, race, sex, weight loss, performance status, stage, and tumor site in relation to failure risk, were conducted using Cox proportional hazards regression models. For multivariate Cox models, independent variables with significance levels of P > .1 were excluded sequentially from the models, except for tumor stage, which was considered essential to control for when interpreting other factors in the model. To study the relationship between IHC score and EGFR gene copy number, a proportional odds logistic regression model was used to take into account the ordinal nature of the dependant variable (IHC score).
EGFR status was evaluated for prognostic effect on OS and PFS by the log-rank test. Because the number of patients with EGFR information and number of failure events was small, evaluating EGFR adjusting for other covariates was not feasible. However, EGFR status was not found to be associated with other patient and disease characteristics. To deal with the small sample size, a permutation test approach was taken, replicating the log-rank tests for survival and PFS 5,000 times each under random group assignment to determine how likely the observed data test results were under the null hypothesis of no EGFR effect on outcomes.
Survival differences in patients with oropharynx and non-oropharynx primary tumors. Kaplan-Meier survival estimates of (A) overall survival, (B) progression-free survival, and (C) disease-specific survival for patients with oropharynx (gold line) and non-oropharynx (blue line) primary tumors. Corresponding P values using the log-rank test are shown.
TN Staging of All Patients: American Joint Committee on Cancer Staging System, 6th Edition
Best Response to Therapy
Univariate and Multivariate Analyses of Factors Associated With Overall and Progression-Free Survival
EGFR Gene Copy Number and IHC Staining Frequency in Patients With Available Tissue
EGFR IHC Staining Frequency and Its Correlation with EGFR Gene Copy Number by FISH in Patients With Available Tissue for Both Procedures
Relationship Between EGFR Gene Copy and Progression-Free and Overall Survival
Footnotes
-
Supported by Astra-Zeneca Pharmaceuticals; University of Chicago Cancer Research Center (UCCRC) Pharmacology Core Facility through the UCCRC clinical core support grant; Grant No. P30 CA14599 from the National Institutes of Health; University of Chicago Human Tissue Resource Core; and The Valda and Robert Svedsen Foundation.
-
Authors' disclosures of potential conflicts of interest and author contributions are found at the end of this article.
- Received November 12, 2009.
- Accepted March 30, 2010.
