• © 2005 by American Society of Clinical Oncology

Cetuximab and Cisplatin for Chemotherapy-Refractory Squamous Cell Cancer of the Head and Neck

Squamous cell carcinoma of the head and neck remains a challenging clinical problem. The majority of patients who present with stage III and IV disease will not be cured by chemoradiotherapy or surgery, and metastatic disease is increasingly common with the use of successful regimens for treatment of locoregional disease. Systemic therapies have not been curative, and median survival times have remained less than a year for patients with recurrent or metastatic disease.

Head and neck cancers are rich in expression of the epidermal growth factor receptor (EGFR), which has been proposed as an important therapeutic target in these tumors.¹ Overexpression has been associated with worse prognosis.² Activated EGFR signals via the ras, ERK1/2, PI3K/Akt, and stat pathways and results in treatment resistance, angiogenesis, enhanced proliferation, and expression of mediators of the metastatic phenotype.^3-5 The two principal classes of EGFR inhibitors under clinical development are monoclonal antibodies, which bind to the extracellular domain with high specificity and inhibit activation of the molecule by its natural ligands, and quinazoline inhibitors of the intracellular kinase domain. Cohen et al⁶ reported responses in patients with squamous cell carcinoma of the head and neck treated with the kinase inhibitor gefitinib.

In this issue, Herbst et al⁷ and Baselga et al⁸ separately report on the antitumor activity of the monoclonal antibody cetuximab in patients with previously treated squamous cell head and neck cancers. Both studies combined cetuximab with a platinum-based chemotherapy in patients with demonstrated platinum-refractory incurable disease. Each found a modest objective response rate, and survival results were also similar.

Herbst et al⁷ prospectively studied patients with cisplatin-unresponsive or minimally responsive disease who, immediately after progression, were assigned to treatment with continued cisplatin-based chemotherapy combined with cetuximab. During this phase of the study, 51 patients with stable disease after two cycles of a cisplatin-based regimen and 23 patients with progressive disease were assigned to cetuximab plus chemotherapy. An amendment in 2001 allowed an additional 56 patients who had progressed off study on a platinum-based regimen to be enrolled. In these patients, the median cumulative prior cisplatin dose was greater, the median Karnofsky performance score was lower, and they were more likely to have received cisplatin in combination with fluorouracil than the prospective study population. All patients were treated with cetuximab 400 mg/m² for one dose, followed by 250 mg/m² weekly in conjunction with cisplatin at the dose and schedule that each patient had previously received. Partial responses were observed in 13% of patients. The highest disease control rate was observed in patients included in the analysis who had stable disease on cisplatin-based therapy, and the lowest disease control rate was observed in the postamendment cohort of patients with progressive disease. The heterogeneity of the study population, resulting from the amendment to the study design during the trial, is potentially a significant shortcoming. However, the differences in response rate and survival among the three populations included may generate hypotheses to identify predictive factors for patient selection and stratification in future studies.

Baselga et al⁸ conducted a multicenter phase II trial of the combination of cetuximab and platinum therapy in patients with platinum-refractory squamous cell carcinoma of the head and neck. Ninety-six patients with documented progressive disease who were not candidates for local therapy were enrolled as the intent-to-treat population (ITT). Of this group, 64 patients were confirmed to have progressive disease by the independent review committee (IRC-PD). Cetuximab dosing was identical to that used in the Herbst et al⁷ study. The overall response rate was 10% in the ITT population and 11% in the IRC-PD population, and disease control rates in these groups were 53% and 52%, respectively. Thus, although the high proportion of nonassessable patients was potentially a significant shortcoming of this study, the results were identical in the ITT and IRC-PD patients. Among all 96 patients, the median time to tumor progression and overall survival time were 85 and 183 days, respectively.

Toxicity was similar in the two studies. hypersensitivity reactions, including anaphylaxis or anaphylactoid reactions, were seen in each study. The most common cetuximab toxicity in each study was dermatologic; acneiform rashes and paronychia were commonly observed. Both groups report an association between the development of rash and more favorable outcome, confirming an association between rash and survival reported for cetuximab therapy in colon cancer and gefitinib therapy in head and neck cancer.6,9

Each study required sufficient tumor tissue for immunohistochemistry for EGFR. This was a sensible requirement. EGFR is a known prognostic factor in squamous cell carcinoma of the head and neck. The study populations (patients with treatment-refractory and metastatic or locally advanced cancers) would be expected to have EGFR expression, with many tumors staining for EGFR at high intensity and on a high percentage of cancer cells. Baselga et al⁸ present the results of EGFR immunohistochemistry, which was successfully performed on 86 of the 96 patients in their study; indeed, 2 to 3+ staining was present in 70 patients, and high-density staining on ≥ 75% of cells was present in 60 patients. These EGFR expression data support the poor prognosis of patients in both the ITT and IRC-PD cohorts in the Baselga et al⁸ study; this is reassuring in view of the methodologic question raised by the large proportion of patients in whom the independent response committee could not confirm prestudy progression. Comparable data from Herbst et al⁷ would have been welcome, in light of the midstudy change in eligibility criteria. The ability to compare the populations of the studies with regard to EGFR expression or the expression of other biomarkers would also have been of interest because of the concordance in the results of the studies.

EGFR expression could also have been evaluated as a possible predictive factor for response to EGFR inhibitors, recognizing the limitations such exploratory analyses would have had in moderate-sized, single-arm studies. The experience with trastuzumab, a monoclonal antibody to the related tyrosine kinase growth factor receptor Her-2, initially led investigators to believe that EGFR inhibitors would be of greatest benefit when EGFR expression was greatest, and some studies required EGFR overexpression for eligibility. Two phase II trials in treatment-refractory colon cancer enrolled only patients with immunohistochemical evidence of EGFR expression and demonstrated activity for cetuximab in combination with irinotecan chemotherapy.^9,10 Neither study discovered any correlation between EGFR expression and response rate or survival. Objective responses to cetuximab have now been reported in chemotherapy-refractory colorectal cancer even when EGFR expression was not detected by immunohistochemistry.¹¹ Thus, in colorectal cancer, testing of single samples and archival material, sampling of the primary tumor in patients with metastatic disease, and the use of immunohistochemical assays that fail to detect low-level EGFR expression all limit our ability to rule out a role for cetuximab in patients with negative EGFR stains.

What about patients with high levels of EGFR expression? An exploratory analysis of E5397, a randomized first-line study comparing cisplatin with cisplatin plus cetuximab in squamous cell carcinoma of the head and neck, suggested that the treatment effect of cetuximab is greater among patients with moderate EGFR immunoreactivity compared with patients with 3+ immunohistochemical staining on ≥ 80% of cells.¹² This finding may indicate that patients with high expression are underdosed with current regimens or that ligand-independent signaling, which might be more effectively inhibited with a quinazoline, is of greater importance in this subset. The discrepancy in regard to the predictive value of EGFR expression may have its origin in methodology, because the colon studies did not use an integrated immunoreactivity score composed of staining intensity and density and none of the studies used methods with improved quantitation, such as automated image analysis. The discrepancy may also reflect differences in EGFR expression and function among tumor types. EGFR expression in head and neck cancer is 3+ in the majority of tumors, and far more patients with head and neck cancer than colon cancer will exhibit this intensity of staining on ≥ 80% of cells. Future studies will focus on confirming the significance of high EGFR expression and pinpointing mechanisms of resistance, so that strategies using higher cetuximab dosing, combinations of EGFR inhibitors, or combinations of these inhibitors with agents targeting antiapoptotic mechanisms can be explored.

The activity of cetuximab plus cisplatin in the unselected chemotherapy-refractory patients studied here is modest. In practice, until we have better means of selecting patients likely to benefit from this regimen, there will be only a minority of patients with advanced disease who are well enough at the time of disease progression on platinum-based treatment to receive second-line therapy containing cisplatin. This combination will be the reasonable choice for such patients. The two reports are most interesting because of the evidence they provide that cetuximab may overcome chemotherapy resistance. In this regard, they provide a strong rationale for continued study of cetuximab in combined-modality and systemic therapy earlier in the course of squamous cell head and neck cancer. It is likely that, in the future, EGFR inhibition will be a central strategy in the management of appropriately selected patients with all stages of head and neck cancer.

In conjunction with the publication of data on the efficacy of cetuximab in colorectal cancer, Schrag¹³ has editorialized on the price tag on progress, which is the steep increase in the cost of palliative therapy. As physicians, we will enjoy seeing meaningful responses in a proportion of the head and neck cancer patients we treat with EGFR inhibitors. Most of us will not enjoy the phone calls and appeal letters to insurance companies, practice managers, and others who will be charged with dealing with the economic impact of these high-cost therapies. Many patients with head and neck cancer are elderly or have limited financial resources. As with patients with colon cancer, for some head and neck cancer patients, the cost of these therapies will be problematic. In the aggregate, the costs will be burdensome to society. However, equitable access to the best treatment supported by current evidence is the unchanging ideal in our profession. To the extent that current pricing structures interfere with this, the voices of patients and medical professionals involved in all disease sites should join the national debate on how drugs are priced and reimbursed.