Cetuximab for the treatment of locally advanced squamous cell cancer of the head and neck

The manufacturer's submission approached the decision problem by comparing cetuximab plus radiotherapy with radiotherapy alone. The manufacturer specified that the population under consideration consisted of people with locally advanced squamous cell cancer of the head and neck for whom chemotherapy is considered inappropriate but for whom radiotherapy is suitable. The outcome measures specified in the decision problem were duration of locoregional control, overall survival, progression-free survival and safety.

The manufacturer's submission presented evidence on the clinical effectiveness of cetuximab plus radiotherapy based on a single randomised controlled trial (RCT; the Bonner trial) that compared cetuximab plus radiotherapy with radiotherapy alone in people with stage III or IV non-metastatic squamous cell cancer of the oropharynx, hypopharynx or larynx. Criteria for eligibility included medical suitability for definitive radiotherapy, a Karnofsky performance-status score of at least 60%, and normal haematopoietic, hepatic and renal function. Patients were not included in the trial if they had undergone surgery or had previously received radiotherapy for head and neck cancer. The primary outcome measure was the duration of control of locoregional disease. The secondary endpoints were overall survival, progression-free survival, response rate and safety.

Final analyses of the trial showed that the 211 people in the cetuximab plus radiotherapy arm had a longer median duration of locoregional control than the 213 people in the radiotherapy-alone arm (24.4 versus 14.9 months, p=0.005, hazard ratio [HR] 0.68, 95% confidence interval [CI] 0.52 to 0.89) and greater median overall survival (49.0 versus 29.3 months, p=0.03, HR 0.74, 95% CI 0.57 to 0.97).

The manufacturer's submission presented a de novo economic analysis that compared cetuximab plus radiotherapy with radiotherapy alone. The model used individual patient data from the RCT to estimate costs and health effects during the trial period for each patient. When trial observations were censored, the model extrapolated costs and health effects.

The base-case analysis compared cetuximab plus radiotherapy with radiotherapy alone and resulted in an incremental cost-effectiveness ratio (ICER) of £6,400 per quality-adjusted life year (QALY) gained. The manufacturer undertook a univariate sensitivity analysis, which demonstrated that the model was not sensitive to change when assessing the effect of uncertainty in a variety of inputs. Relatively large variability was observed when the timeframe of the analysis changed from a lifetime to the period of the trial follow-up, resulting in an ICER of £20,000 per QALY gained.

The ERG reviewed the evidence on clinical and cost effectiveness submitted by the manufacturer. The ERG judged that the 1 trial included in the manufacturer's submission was well conducted and that the results for the primary endpoints appeared robust. However, the ERG noted that the majority of patients in the trial population had a good performance status (Karnofsky performance-status score ranged from 60% to 100% but was most commonly 90%), and chemotherapy would be expected to be suitable for them. Therefore, the population of the trial did not match the population described in the decision problem, that is, patients for whom chemoradiotherapy is considered inappropriate. Furthermore, there are differences between the radiotherapy regimens used in the trial and those most commonly used in UK clinical practice.

The ERG reviewed the economic model and identified a number of concerns. The most important of these was that the only RCT informing the economic analysis (the Bonner trial) did not match the patient population specified in the manufacturer's decision problem. The manufacturer provided a set of possible criteria for defining patients for whom chemoradiotherapy is inappropriate, based on discussions with a small number of oncologists. In addition, the manufacturer was asked to provide information on the number of patients in the trial for whom chemoradiotherapy was considered inappropriate. However, the manufacturer stated that it was unable to provide analyses based on these criteria as the RCT was not designed or statistically powered to assess subgroups of patients for whom chemoradiotherapy may be considered inappropriate.

In addition, the ERG identified a series of issues and uncertainties about the methods for extrapolation of the trial data, assessment of health-related quality of life (HRQoL), and estimation of resource use and costs. The ERG concluded that the methods used were probably appropriate but was unable to determine, in the majority of cases, the likely influence of using alternative methods on the results of the economic model. However, the ERG concluded that altering the method of extrapolation would be unlikely to cause the ICER to increase to above £20,000.

The ERG undertook additional work to examine the robustness of the base-case results to the assumptions made in the manufacturer's cost-effectiveness model about HRQoL, resource use and cost. The ERG concluded that any inaccuracies would have to be very large to have a material effect on the conclusions of the manufacturer's cost-effectiveness analysis.

The ERG felt that although the economic analyses undertaken by the manufacturer demonstrated that cetuximab in combination with radiotherapy was cost effective compared with radiotherapy alone under a broad range of different assumptions (assuming a threshold of £20,000 per QALY), the cost-effectiveness estimates might not be directly applicable to the population specified in the manufacturer's decision problem (that is, patients for whom chemoradiotherapy is considered inappropriate). This was because the clinical study on which the economic analysis was based included a substantial proportion of patients for whom chemoradiotherapy would be considered suitable.

Following an appeal hearing, the Appeal Panel requested that the manufacturer provide subgroup survival data (derived from the Bonner trial) for each of the separate Karnofsky performance-status score subgroups (Karnofsky performance-status scores of 100%, 90%, 80%, 70% and less than 70%). The manufacturer stated that the number of patients in some of the subgroups was small (numbers ranged from 12 to 91), and this should be taken into consideration when interpreting these data. For patients with Karnofsky performance-status scores of 100% and 90%, the survival HRs were in favour of cetuximab plus radiotherapy over radiotherapy alone (HR 0.61, 95% CI 0.28 to 1.31, and HR 0.58, 95% CI 0.39 to 0.88 for Karnofsky performance-status scores of 100% and 90%, respectively). For patients with Karnofsky performance-status scores of 80%, 70% and less than 70%, the survival HRs were in favour of radiotherapy alone over cetuximab plus radiotherapy (HR 1.11, 95% CI 0.69 to 1.77; HR 1.22, 95% CI 0.53 to 2.78; and HR 3.41, 95% CI 0.65 to 17.7, respectively).

The manufacturer was further asked by the Appeal Panel to provide cost-effectiveness estimates for the subgroup analyses described in section 3.11. The analyses were conducted using the manufacturer's original cost-effectiveness model. The manufacturer's analysis gave ICERs for cetuximab in combination with radiotherapy versus radiotherapy alone of £13,151 and £4,467 per additional QALY gained for patients with Karnofsky performance-status scores of 100% and 90%, respectively. For patients with Karnofsky performance-status scores of 70%, radiotherapy alone dominated cetuximab in combination with radiotherapy (that is, radiotherapy alone was more effective in terms of QALYs gained and was less expensive). For patients with Karnofsky performance-status scores of 80% and less than 70%, the manufacturer reported ICERs for cetuximab in combination with radiotherapy versus radiotherapy alone of £58,200 and £37,000 per additional QALY gained, respectively.

Following the appeal hearing, NICE invited the manufacturer and consultees and commentators to provide or highlight further evidence on the efficacy of carboplatin monotherapy in combination with radiotherapy, and on the safety or toxicity of carboplatin with fluorouracil and radiotherapy in patients with locally advanced squamous cell cancer of the head and neck. The manufacturer undertook a literature review and identified 22 studies on the efficacy of carboplatin monotherapy and radiotherapy, none of which were phase III studies or meta-analyses. Six of the 22 studies reported median overall survival estimates, which ranged from 6.7 months to 30 months. The manufacturer considered the median overall survival estimate of 30 months reported by Jeremic and colleagues (n=53) to be the most robust. The manufacturer further identified 9 published studies on the efficacy and safety of carboplatin with fluorouracil and radiotherapy, of which 3 were phase III trials. The phase III studies reported median overall survival estimates of 23 months, 20 months and 19 months (n=113, 109 and 64, respectively), and haematological toxicities (grade 3 or 4 acute toxicities) of 23% and 29.5% (n=113 and 64, respectively). Consultees highlighted that there was little published evidence on the efficacy of carboplatin-based chemoradiotherapy compared with cisplatin-based chemoradiotherapy or with radiotherapy alone, but that carboplatin-based chemoradiotherapy can be used as a treatment for patients for whom cisplatin-based chemoradiotherapy is not an option.

Full details of all the evidence are in the manufacturer's submission and the ERG report.