![Cure model survival analysis (progression-free survival [PFS] and overall survival [OS]). ERT, cetuximab in combination with radiotherapy; RT, radiotherapy.](https://www.researchgate.net/profile/Robert-Launois/publication/227508227/figure/fig1/AS:267554266546210@1440801316645/Cure-model-survival-analysis-progression-free-survival-PFS-and-overall-survival-OS_Q320.jpg)
Content uploaded by Robert Launois
Author content
All content in this area was uploaded by Robert Launois on May 30, 2022
Content may be subject to copyright.
An Economic Evaluation of Cetuximab Combined with
Radiotherapy for Patients with Locally Advanced Head and
Neck Cancer in Belgium, France, Italy, Switzerland, and the
United Kingdom
Ben Brown, BA,1Alexander Diamantopoulos, MSc,1Jacques Bernier, PhD,2Patrick Schöffski, PhD,3
Klaus Hieke, BA,4Lorenzo Mantovani, PhD,5,6 Robert Launois, PhD,7Ingolf Griebsch, MPH,8Paul Robinson, MSc1
1IMS Health, London, UK; 2Clinique de Genolier, Genolier, Switzerland; 3Leuven Cancer Institute, University Hospital Gasthuisberg, Catholic
University Leuven, Leuven, Belgium; 4NEOS Health, Binningen, Switzerland; 5Center of Pharmacoeconomics Federico II University of Naples,
Naples, Italy; 6Fondazione CHARTA, Milan, Italy; 7REES France, Paris, France; 8Merck KGaA, Darmstadt, Germany
[Correction added after online publication 28-May-2008:Table 6 caption has been updated]
ABSTRACT
Objectives: A phase III randomized trial that compared the
combination of cetuximab and radiotherapy to radiotherapy
alone in patients with locally advanced squamous cell carci-
noma of the head and neck provided a platform for a compre-
hensive economic evaluation. The study was conducted to
estimate the cost-effectiveness of cetuximab in combination
with radiotherapy compared to radiotherapy alone, for the
treatment of locally advanced head and neck cancer in patients
for whom chemoradiotherapy is inappropriate or intolerable.
Methods: Separate economic analyses were conducted for
Belgium, France, Italy, Switzerland, and the United Kingdom.
The economic model was based on individual patient data
extracted from an international phase III trial. Country-
specific costs of care from official sources were applied in
each analysis. Clinical expert panels supplemented resource
use estimates from the phase III trial and validated assump-
tions used to extrapolate costs and health outcomes beyond
the follow-up of the phase III trial.
Results: In the base-case analysis, the incremental cost per
quality-adjusted life-year for patients receiving radiotherapy
in combination with cetuximab compared to radiotherapy
alone among all countries was in the range of €7538 to
€10,836. Sensitivity analysis showed the results to be robust.
Conclusion: This cost-effectiveness analysis indicated that
the addition of cetuximab to high-dose radiotherapy offers a
good value-for-money alternative to radiotherapy alone in
the treatment of locally advanced head and neck cancer in
five European countries.
Keywords: cetuximab, economic, head and neck cancer,
models, quality-adjusted life-years, radiotherapy.
Introduction
Squamous cell carcinoma of the head and neck
(SCCHN) is a group of malignant tumors primarily
arising in the oral cavity, nasopharynx, pharynx, and
larynx [1]. SCCHN is the sixth most common form of
cancer worldwide representing approximately 5% of
all cancers. A total of 500,000 new cases are diagnosed
worldwide each year [2]. Tobacco and alcohol con-
sumption are etiological factors involved in the onset
of SCCHN, which commonly affects middle-aged or
older men. Furthermore, patients with head and neck
cancer are typically from lower-income groups and are
often socially marginalized, and approximately half of
the patients suffer numerous secondary health prob-
lems such as pulmonary, cardiovascular, hepatic, or
neurological disorders [1,3].
Chemotherapy with concurrent radiotherapy
(CRT) presently is considered to be standard care for
patients with good performance status [4]; however,
survival gains with CRT are almost universally accom-
panied by severe, acute dose-limiting toxicities and, in
some patients, a higher proportion of late toxicities
[5]. Radiotherapy alone, especially through the deliv-
ery of altered fractionation, remains the principal
treatment used in patients for whom CRT is contrain-
dicated or cannot be tolerated [4–6].
The epidermal growth factor receptor (EGFR) is
abnormally activated in patients with head and neck
cancers. The cells of almost all such neoplasms express
high levels of EGFR, a feature associated with a poor
clinical outcome [7]. Radiation increases the expres-
sion of EGFR in cancer cells, and blockade of EGFR
signaling sensitizes cells to the effects of radiation.
Cetuximab (Erbitux), an IgG1 monoclonal antibody
against EGFR, enhances the cytotoxic effects of radia-
tion in SCCHN [8]. In a recent phase III randomized
trial, the addition of cetuximab to high-dose radio-
therapy (ERT) significantly increased duration of
Address correspondence to: Alexander Diamantopoulos, Health
Economics and Outcomes Research, IMS Health, Level 3 Lee
House, 109 Hammersmith Road, London W14 0QH, UK.
E-mail: adiamantopoulos@uk.imshealth.com
10.1111/j.1524-4733.2007.00302.x
Volume 11 • Number 5 • 2008
VALUE IN HEALTH
© 2008, International Society for Pharmacoeconomics and Outcomes Research (ISPOR) 1098-3015/08/791 791–799 791
locoregional disease control, progression-free and
overall survival compared to high-dose radiotherapy
alone among patients with locally advanced head and
neck cancer treated with curative intent [9]. Notably,
cetuximab did not exacerbate the common toxic
effects associated with radiotherapy of the head and
neck. In the absence of a randomized trial between
ERT and CRT, it remains unclear how these two regi-
mens compare; however, it is reasonable for clinicians
to consider the use of ERT in patients who are not
good candidates for CRT or surgery [4].
Evidence of a new drug’s safety and efficacy is no
longer sufficient to ensure reimbursement for use in
public markets. Increasingly, funding of new drugs is
based on evidence of cost-effectiveness [10]. Many
countries in Europe have introduced formal require-
ments and/or voluntary guidelines for economic evi-
dence to be submitted as part of the pricing or
reimbursement decision. These include, among others,
Belgium, France, Italy, Switzerland, and the United
Kingdom (including the Scottish Medicines Consor-
tium and the National Institute for Health and Clinical
Excellence).
The current literature evaluating the cost-
effectiveness of interventions for patients with head
and neck cancer is extremely scarce. A recent review
could only identify publications that reported original
cost data on diagnosis, treatment, and management.
No economic evaluations have been published for
cetuximab in locally advanced head and neck cancer
[11]. The aim of this study was therefore to determine
if ERT is cost-effective compared to radiotherapy in
the treatment of locally advanced head and neck
cancer. Separate analyses were conducted for five
European countries: Belgium, France, Italy, Switzer-
land, and the United Kingdom.
Materials and Methods
Clinical Trial Design, Patient Population and Treatment
The economic evaluation was based on a phase III
randomized trial that compared ERT to radiotherapy
in patients with locally advanced SCCHN. Full details
of this study have been published elsewhere [9]. Briefly,
patients with stage III or IV, nonmetastatic, measurable
SCC of the oropharynx, hypopharynx, or larynx were
stratified by Karnofsky performance status (90–100 vs.
60–80), regional node involvement (positive vs. nega-
tive), tumor stage (T1–3 vs. T4), and radiation frac-
tionation (concomitant boost vs. once daily vs. twice
daily), and then randomized (1:1) to treatment with
radiotherapy alone for 6 to 7 weeks (n =213) or in
combination with weekly administered cetuximab
(n =211).
In the ERT group, administration of intravenous
cetuximab was initiated 1 week before radiotherapy at
a loading dose of 400 mg/m2of body surface area over
a period of 120 minutes, followed by weekly 60-
minute infusions of 250 mg/m2for the duration of
radiotherapy.
The primary end point was the duration of locore-
gional disease control; secondary end points were
overall survival, progression-free survival, the response
rate, safety, and quality of life. Routine assessments
were performed weekly during radiotherapy, and
disease assessments were performed at week 4 (except
CT scans or magnetic resonance imaging) and week
8 after radiotherapy, every 4 months thereafter for
2 years; and then semiannually during years 3, 4, and
5. Acute toxic effects were assessed through the
8 weeks after treatment. Late radiation effects were
assessed thereafter using the Radiation Therapy
Oncology Group toxicity scales. Quality-of-life ques-
tionnaires were included in the study (the EORTC
QLQ-C30, version 3.0 and the head and neck module
QLQ-H & N35).
Economic Model Overview
The economic model provides cost-utility analyses
undertaken from the perspective of the national
health system of each country, or where not appli-
cable, from the perspective of the major national
payer in the relevant country. The time horizon of the
model was patient lifetime (in the base case), which is
appropriate when evaluating two interventions with
different overall survival times. Costs and health out-
comes were discounted to present value at 3.5%.
The model was constructed at an individual patient
level using the trial data set provided by Merck
KGaA, Darmstadt, Germany. All study patients
were included in the model for costs and outcomes
(intention-to-treat analysis). Nevertheless, a few
patients did not record any adverse events or treat-
ment costs because of early withdrawal from the
trial, and consequently some cost items were zero for
these patients.
Health Outcomes in the Economic Evaluation
It is essential for economic evaluation to estimate
mean—not median—survival times for each treatment
group [12,13]. This involved an extrapolation of the
survival curves beyond the follow-up time in the trial
because the right tail of the survival distributions
included censored observations.
Parametric survival models were used to impute
censored survival times for progression-free and
overall survival; otherwise, the known time-to-event
data were used without any adjustment. The treatment
goal for patients with newly diagnosed but unresect-
able disease is cure [6] and in this context it was
deemed appropriate to fit a cure model [14]. Statistical
analysis of the hazard functions for the progression-
free survival and overall survival curves in the trial
supported this decision. The STATA command “cur-
792 Brown et al.
eregr” was used to estimate the proportion of patients
who were “cured” and thus would never experience
the event of interest (progression or death). The overall
survival probability of cured patients was estimated
from life tables together with an estimate of the pro-
portional increase in mortality hazard for patients who
have experienced locally advanced SCCHN, based on
10-year overall survival curves reported in a large
meta-analysis [15]. For estimating the survival of non-
cured patients, the goodness-of-fit of various models
was assessed using the log-likelihood statistic and
the results were similar across most of the models
(Weibull, log-normal, logistic, gamma, and exponen-
tial). The log-normal model was chosen because it is
appropriate for characterizing patterns of survival
where the hazard is initially increasing, but then begins
to decrease. The specific details of the cure model is
not the central point of this article, so it is important
to emphasize that two alternatives to the cure model
were tested for the estimation of progression-free and
overall survival times: first, an analysis limited to the
period of follow-up in the trial; and second, a Weibull
survival distribution. The results of these models are
presented in the sensitivity analyses.
The model estimates quality-adjusted life-years
(QALYs), life-years (LYs), and progression-free sur-
vival for each treatment group. The quality-of-life
instruments in the trial (EORTC QLQ-C30, QLQ-H
& N35) could not be used in cost-utility analysis
because there is currently no preference-based scaling
method to translate these data into an utility score
where full health =1 and dead =0 [13,16]. Utility
values for key health states were therefore estimated
in a separate study where 50 oncology nurses in the
United Kingdom evaluated quality of life using the
EQ-5D for eight key health states (Table 1). Recruit-
ment screening criteria identified nurses with experi-
ence of at least 2 years in medical oncology, seeing 11
patients or more with SCCHN in the last 3 months,
with experience in treating patients with radiotherapy,
chemotherapy, or chemoradiotherapy. The nurses were
asked to complete the EQ-5D for each health state,
which was described to them, such that their ratings
reflected their judgement about a patient’s health-
related quality of life when experiencing each health
state. Preference weights for the EQ-5D scoring func-
tion have been elicited using the time-trade-off tech-
nique on a random sample of 3395 members of the
adult population in the United Kingdom [17], and
therefore the scoring is representative of preferences of
the general public.
Patient time was allocated among the first six health
states (A–F, Table 1) according to adverse events
recorded in the patient-level data set. Patients with
more than one simultaneous adverse event were allo-
cated to the health state associated with the worst
quality of life. The average duration of each adverse
event health state was calculated using the trial data
set. The time in the acute treatment phase and the
average duration of each health state event applied to
the model for both treatment groups is presented in
Table 1.
After the acute phase, the remainder of each
patient’s overall survival was allocated between the
post-treatment health states according to their
progression-free survival (health state J) and survival
with progressive disease (health state K).
Resource Use
The patient-level data set provides the primary source
of evidence regarding resource use in the model. Nev-
ertheless, hospitalization in the trial was protocol-
driven and therefore may not entirely be representative
of local clinical practice. Furthermore, health-care
resource use data collection was not an explicit part of
the protocol and as a result information was not con-
sistently collected in the trial. Where the data set did
not provide adequate information about resource use,
such as drug administration costs, local expert clinical
opinion and published national reference costs were
used. A panel of clinical experts was convened in each
country to provide opinion about resource utilization
representative of local practice. The clinicians pro-
vided individual written answers to a questionnaire,
Ta b l e 1 Model health states and utilities
Health state Definition Utility Time (day)*
Acute phase health states
A General in-treatment—range of ⱕgrade 1 adverse events 0.659 n/a
B As health state A plus grade 3 or 4 mucositis, stomatitis, and dysphagia 0.062 55.43
C As health state A plus grade 2 mucositis, stomatitis, and dysphagia 0.608 34.46
D As health state A plus grade 3 or 4 nausea and vomiting 0.108 13.14
E As health state A plus grade 2 nausea and vomiting 0.573 29.82
F As health state A plus grade 3 or 4 acne and rash 0.226 72.92
Postacute phase health states
J Post-treatment locoregional disease control 0.862 n/a
K Post-treatment progressive or worsening disease 0.129 n/a
*An average time in adverse event health states B–F was calculated using patient-level data where a start and end date was known.Average time was then assigned to patients
having the adverse event. Health state A took the remaining time in the acute phase after allocation of time to other health states.Time in health states J and K was available
for each patient.
An Economic Evaluation of Cetuximab Combination 793
and numerical estimates were then negotiated in open
discussion until a consensus was reached. Table 2
presents the principal source of evidence for model
variables.
Radiotherapy Treatment
The trial included three radiotherapy treatment
regimens:
1. Once daily: 35 fractions, 5 fractions per week for
7 weeks (conventional regimen);
2. Twice daily: 60 to 64 fractions, 10 fractions per
week for 6 to 6.5 weeks;
3. Concomitant boost: 42 fractions, 5 fractions per
week for 3.5 weeks then 10 fractions per week for
2.5 weeks.
These latter schedules (2 and 3) were therefore two
variants of altered fractionation regimens.
Costs consisted of radiotherapy set-up, daily radio-
therapy, and administration. Members of expert clini-
cal panels provided information about the local setting
for radiotherapy administration (inpatient, day case,
outpatient) while the patient-level data set provided
information on the type and number of days of therapy
received per patient.
Cetuximab Cost and Administration
The patient-level data set recorded the exact dose of
cetuximab administered to each patient in the ERT
group. Each dose was rounded up to the nearest vial
size to account for wastage in the model. The cost of
cetuximab in each country is shown in Table 3.
The cost of administration for patients in the ERT
group was estimated by applying local clinical practice
norms that dictated settings—inpatient, day admis-
sion, or outpatient—and using published national
reference costs (Table 3). The cost of radiotherapy
administration was added to cetuximab administra-
tion costs in the ERT group.
Treatment of Adverse Events
Consultation with experts sought to identify adverse
events associated with the greatest cost significance
and possible groupings of these events (Table 3). The
patient-level data set provided incidence rates for the
selected adverse events.
Published national reference costs for diagnosis-
related groups of hospital admissions were used to
establish costs for these adverse events on an episode
basis. Members of the clinical expert panels estimated
the probability of hospitalization for each adverse
event. Table 3 presents the unit costs used for adverse
event categories in each country analysis.
Patient Imaging and Routine Monitoring
The model includes imaging and routine-monitoring
costs. The clinical expert panels provided estimates on
the types of scans performed and their typical fre-
quency, as well as information on the frequency of
specialist visits, for patients in this population. Table 3
presents the unit costs for each type of scan and out-
patient visit. Routine monitoring costs were applied
for each patient according to survival duration. Costs
incurred after the first year were discounted.
Salvage/Palliative Care
The model assumed that the pattern of care changed
after patients moved into progressive disease. Informa-
tion on rates of salvage surgery, secondary radio-
therapy, and secondary systemic therapy was obtained
from the trial. Expert opinion informed rates and use
of palliative care nursing.
Unit costs were sourced for these items to estimate
costs of salvage/palliative care and were applied in the
model after disease progression of surviving patients
(Table 3). The costs were discounted if that point
occurred after the first year.
Percutaneous Endoscopic Gastrostomy
The economic model considers the cost of percutane-
ous endoscopic gastrostomy (PEG), which may be per-
formed prophylactically on patients in this population.
The clinical expert panels provided estimates of the
frequency of PEG procedures. The cost of PEG in each
country is presented in Table 3.
Addressing Uncertainty
In order to address uncertainty around the observed
cost and effect values of the model cohort, stochastic
and one-way sensitivity analyses were performed on
the lifetime cost-utility analysis [18]. Individual patient
cost and health outcomes estimates were sampled by
applying a bootstrapping approach. A set of 2000
samples was obtained from the observed estimates in
both the ERT and radiotherapy groups.
Sensitivity Analyses
One-way and scenario-based sensitivity analysis was
also performed, where appropriate, on model variables
(discount rate, radiotherapy administration cost, costs
Ta b l e 2 Resource use categories and data sources
Resource Primary data source
Radiotherapy treatment
(type & days)
Pivotal trial data set
Study drug (vial usage) Pivotal trial data set
Administration of therapy Pivotal trial data set (frequency of
administration);
Expert panel (administration setting
and schedule)
Treatment of adverse effects Pivotal trial data set (frequency and
duration of event);
Expert panel (resource use)
Imaging Expert panel
Routine monitoring Expert panel
Procedures Expert panel
Salvage/palliative care Pivotal trial study report
794 Brown et al.
of adverse event treatment, acute health state utilities,
duration of health states, and postacute health state
utilities) and on different survival extrapolation
methods [13,19] for all country analyses.
Results
Base-Case Analysis
Clinical. As reported previously [9], with a median
follow-up of 54.0 months in the trial, the addition of
cetuximab to radiotherapy prolonged the duration of
locoregional control by almost 10 months compared
to radiotherapy alone, with a median duration of
24.4 months versus 14.9 months (P=0.005, 95% con-
fidence interval [CI] 0.52–0.89, hazard ratio 0.68). The
median survival time was 49.0 months among patients
treated with combined therapy and 29.3 months
among those who received radiotherapy alone (P=
0.03, 95% CI 0.57–0.97, hazard ratio 0.74), while the
median progression-free survival was 17.1 months for
ERT and 12.4 months for radiotherapy (P=0.006,
95% CI 0.54–0.90, hazard ratio 0.70).
Economic. After extrapolation of the cure model’s
survival curves, ERT had an estimated overall survival
advantage over radiotherapy of 12.7 months, or
0.92 years after discounting (discount rate applied
3.5% per annum, see Fig. 1, Table 4). In addition to
this survival advantage, the model estimates that
patients treated with ERT gain on average of 3.96
QALYs compared to 2.88 QALYs for those treated
Ta b l e 3 Unit costs by country (€, 2005/2006 values)
Parameter Belgium France Italy Switzerland United Kingdom
Cetuximab 100 mg 199 209 199 200 203
Cetuximab administration: outpatient visit 118 424.17 21 468 (1st)
312 (latervisits)
266 (1st);
186 (latervisits)
PEG 1,733 126 n/a n/a 1,118
Adverse events†‡
Acne/rash grade 3 2,514 2,037 1,727 1,914 2,173
Acne/rash grade 4 2,514 3,995 1,727 1,914 2,173
Anaemia grade 3 3,507 378§3,418 957 1,386
Anaemia grade 4 3,507 2,260 3,418 957 1,386
Dehydration grade 3 or 4 2,452 671 2,581 3,349 1,400
Febrile neutropenia grade 3 or 4 4,221 4,311 4,824 5,741 1,993
Fever/infection grade 3 4,001 277 5,983 4,784 3,288
Fever/infection grade 4 4,001 3,743 5,983 4,784 3,288
Mucositis/stomatitis/dysphagia grade 2 4,408 1,756 3,605 n/a 2,710
Mucositis/stomatitis/dysphagia grade 3 4,408 1,756 3,605 3,349 4,524
Mucositis/stomatitis/dysphagia grade 4 4,408 3,236 3,605 26,790|| 4,524
Nausea/vomiting grade 2 2,076 1,635 1,757 n/a 1,047
Nausea/vomiting grade 3 2,076 1,635 2,507 2,870 1,638
Nausea/vomiting grade 4 2,076 3,612 2,507 5,262 1,638
Thrombocytopenia grade 3 3,098 363* 4,824 1,914 1,993
Thrombocytopenia grade 4 3,098 6,580 4,824 1,914 1,993
Weight loss grade 3 3,198 1,981 n/a 10,046|| 2,264
Weight loss grade 4 3,198 3,697 n/a 10,046|| 2,264
Imaging
CT scan 112.51 25.00 137.90 146 87
MRI scan 94.82 69.00 249.45 229 327
PET scan 324.84 0.00 n/a 403 1,304
X-rays 29.59 0.00 n/a 33 48
Palliative/salvage care
Nurse (palliative) 54.63 37.70 n/a 35¶639#
Nurse (oncology) 10.91 7.25 n/a 35¶431#
Surgery: resection without severe
complications
12,109 6,785 3,605 2,629 18,674
Cisplatin 50 mg 35 3 14 67 38
*Unit cost is €13 per 5 minutes. Assumed 180 minutes for initial loading dose, 120 minutes per infusion thereafter.
†With severity according to NCI CTC/RTOG toxicity grades.
‡For Switzerland, all adverse event costs were estimated using the average cost per diem of canton hospitals multiplied by average length of stay estimated by the local clinical
expert panel for that adverse event.
§For France, the cost of anaemia grade 3 and thrombocytopenia grade 3 were day-case casemix codes.
||For Switzerland, the high cost of mucositis/stomatitis/dysphagia grade 4 is due to an estimated inpatient length of stay of 28 days.Similarly,weight loss grades 3–4 had an estimated
length of stay of 10.5 days.
¶For Switzerland, the hour rate for nursing care was estimated as the average tariff of a convenience sample of 10 cantons,assuming that typically 50% of real costs are covered
by the taxpayer.
#For the United Kingdom, the costs of nursing care are per episode of community/outreach nursing care.
Unit cost sources: Belgium (Official tarifs Belgium, Cecodi, 2005, http://www.riziv.fgov.be/prices 2005, INAMI/RIZIF;Technische cel MKG-MFG data. Site: http//tct.fgov.be); France
(T2A: casemix-based hospital financing system 2005);Italy (diagnosis-related groups);Switzerland (Bundesamt für Gesundheit, Spezialitätenliste, http://www.galinfo.net/sl/batchhtm/
substance.htm; http://www.fmh.ch/tarif-browser/de/); United Kingdom (NHS reference costs [health-related groups]; MIMS March 2005).
Currency analysis: Unit cost values were provided in Euros except for Switzerland (CHF) and the United Kingdom (GBP).Exchange rate: CHF to Euros 0.628955; GBP to Euros
1.49022.
MRI, magnetic resonance imaging; PEG, percutaneous endoscopic gastrostomy; PET, positron emission tomography.
An Economic Evaluation of Cetuximab Combination 795
with radiotherapy. The health benefits of the two treat-
ment groups in terms of QALYs, overall survival, and
progression-free survival are summarized in Table 4.
Moreover, ERT is associated with an increased cost
per patient. The expected cost values per patient for
each country are presented in Table 5. The key driver
of the cost difference is cetuximab acquisition, and
subsequently, the administration cost. All other cost
categories were comparable between treatment
groups.
Most of the cost was incurred during the first
5 years of therapy. Imaging and routine-monitoring
costs were accrued during the first 3 and 4 years,
respectively. Palliative care was assumed to be admin-
istered from the point of disease progression.
Based on these estimated costs and outcomes, the
incremental cost-effectiveness ratios (ICERs) were cal-
culated for the five countries in focus. The additional
costs per QALY gained ranged from €7000 to €11,000
(Table 6).
Addressing Uncertainty
The cost-effectiveness acceptability curves in Figure 2
present the probability that ERT is cost-effective
compared to radiotherapy across different cost-
effectiveness threshold values for all countries in focus.
In all countries, 98% or more simulations produced a
cost per incremental QALY of less than €30,000.
Sensitivity Analysis
The results of the study were most sensitive to the
choice of time horizon for the analysis (Table 7). The
most conservative scenario assumes that the accrued
survival benefit is immediately lost at the moment the
trial ends and therefore considers only the period from
the start of the study to the end of follow-up. This
extreme analysis resulted in estimated mean overall
survival times of 39.5 months and 34.9 months for
ERT and radiotherapy, respectively, a difference of
4.6 months in favor of ERT, increasing the incremental
costs per QALY up to €33,535.
Using the Weibull model instead of the cure
model to extrapolate the survival curves beyond trial
duration resulted in a small improvement of the
cost-effectiveness ratio in favor of cetuximab and
radiotherapy. This was due to an increased incremental
QALY gain for cetuximab under the Weibull survival
curves.
Substantial changes to the utility values for the post-
acute treatment phase also had a significant impact on
the ICER. Nonetheless, in all scenarios ERT still rep-
resents a cost-effective alternative to radiotherapy.
Discussion
The ERT significantly increased duration of locore-
gional disease control, progression-free survival, and
overall survival compared to high-dose radiotherapy
alone among patients with locally advanced head and
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
0 5 10 15 20 25
Years
Survival probability
RT - PFS
RT - OS
ERT - PFS
ERT - OS
Figure 1 Cure model survival analysis
(progression-free survival [PFS] and overall sur-
vival [OS]). ERT,cetuximab in combination with
radiotherapy; RT, radiotherapy.
Ta b l e 4 Health outcomes results (discounting applied*)
RT ERT Difference (95% CI†)
QALY 2.88 3.96 1.08 (0.42–1.80)
Overall survival (year) 4.18 5.10 0.92 (0.11–1.80)
Progression-free survival (year) 3.31 4.62 1.31 (0.53–2.15)
*Discounted at 3.5% per annum.
†95% Bias-corrected confidence interval (percentile method) based on 2000 boot-
strap iterations.
ERT, cetuximab in combination with radiotherapy;QALY, quality-adjusted life-year; RT,
radiotherapy.
796 Brown et al.
neck cancer treated with curative intent [9]. The find-
ings of the economic study presented here suggest that
the use of cetuximab with radiotherapy is a cost-
effective health-care intervention in the five countries
studied for patients with locally advanced head and
neck cancer who are not amenable to chemoradio-
therapy. Despite the increase of total treatment costs
through the addition of cetuximab to radiotherapy, the
incremental costs per additional QALY range from
€7538 (Italy) to €10,836 (France), which is far below
commonly accepted cost-effectiveness thresholds in
Europe [10]. Published estimates of a cost-effectiveness
threshold for the United Kingdom [20,21] converted
into Euros provide values of €30,000 to €45,000; this
threshold range is higher than the estimated ICER for
any country-specific analysis. Sensitivity analysis indi-
cated the robustness of these results.
The objective of this economic evaluation was to
estimate the cost-effectiveness of ERT compared to
radiotherapy in the treatment of locally advanced
head and neck cancer for those patients who presently
are treated with radiotherapy alone. Chemoradio-
Ta b l e 5 Estimated costs by country (€)
Belgium France Italy Switzerland United Kingdom
RT ERT RT E RT RT E RT RT ERT RT E RT
Cetuximab 0 8,002 0 8,405 0 8,002 0 8,063 0 8,180
Radiotherapy 2,596 2,534 5,995 5,909 2,072 2,048 2,420 2,363 3,965 3,871
Administration 137 966 145 3,396 171 330 2,243 4,851 908 2,416
Adverse events 3,213 3,395 1,479 1,614 2,371 2,282 3,276 3,765 1,133 1,136
Imaging 1,022 1,187 454 475 787 823 444 458 437 437
Monitoring 253 275 197 215 238 244 553 574 2,417 2,766
Procedures 433 433 31 31 0 0 0 0 167 167
Palliative/salvage 1,741 1,846 1,072 1,019 519 561 615 620 1,685 1,603
Total 9,395 18,638 9,373 21,064 6,158 14,291 9,552 20,695 10,710 20,575
INCREMENT 9,244 11,691 8,133 11,143 9,865
ERT, cetuximab in combination with radiotherapy;RT, radiotherapy.
Ta b l e 6 Incremental cost-effectiveness ratios (€)
Analysis Belgium France Italy Switzerland United Kingdom
Cost per QALY gained 8,568 10,836 7,538 10,328 9,144
Cost per LY gained 10,086 13,367 8,874 12,159 10,765
Cost per PFS year gained 7,041 8,907 6,195 8,489 7,515
1€=0.6289 CHF; 1€=0.671 GBP (November 2006).
LY, life-year; PFS, progression-free survival; QALY, quality-adjusted life-year.
[Correction added after online publication 28-May-2008:Table6 caption has been updated].
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
0
Willingness to pay (€/QALY, 2006 values)
Probability of cost-effectiveness
Belgium
France
Italy
Switzerland
UK
30,00025,00020,00015,00010,0005,000
Figure 2 Cost-effectiveness acceptability cur-
ves. Curve shows the probability that a cetux-
imab in combination with radiotherapy is
cost-effective versus radiotherapy alone for a
range of decision-makers’ maximum willing-
ness to pay per quality-adjusted life-year
(QALY).
An Economic Evaluation of Cetuximab Combination 797
therapy with cisplatin is the recommended standard
of care for patients who can tolerate it [4–6]. This
recommendation is not affected by the conclusions of
our study: such a change to clinical practice normally
would need a clinical trial directly comparing ERT to
CRT. Nevertheless, our conclusion applies to the situ-
ation when CRT is inappropriate, such as when there
are coexisting medical conditions in the light of
chronic alcohol and tobacco use; contraindications to
chemotherapy (e.g., renal failure, cardiac dysfunction,
poor hearing, tinnitus, neuropathy); or in patients
with poor performance status; and when there are
other factors such as socioeconomic factors (e.g., low
social support) and patient choice [5,9,22]. The trial
that this economic evaluation is based on was not
designed or statistically powered to assess specifically
those subgroups of patients who may be inappropri-
ate for CRT treatment. Nevertheless, in our opinion
the use of the whole trial data set is justified because
the clinical factors that lead to CRT being considered
inappropriate are highly variable and these factors are
not necessarily likely to modify the treatment benefit
of ERT.
The economic model is conceptually simple. Costs
are similar for both treatment groups, except for the
additional cost of cetuximab, and the differences in
health outcomes (progression-free survival and overall
survival) were proven within the trial. The costs of
managing the adverse effects of treatment in SCCHN
are high [23,24]; however, cetuximab did not exacer-
bate the common toxic effects associated with radio-
therapy of the head and neck. Because the cost of
cetuximab is the key driver of the cost difference
between the treatment groups, it is useful that patient-
level data were used to accurately estimate drug usage
and costs.
Our analysis has several limitations. First, the impu-
tation of censored progression-free and overall sur-
vival data is always subject of some uncertainty.
Nevertheless, this should not be overstated, because
sensitivity analyses show that the cost-effectiveness of
cetuximab is robust to various methods of extrapola-
tion, and in fact, the ICER for cetuximab is below
common cost per QALY threshold even under the con-
servative situation of zero data imputation.
A second limitation is that utility data were not
available directly from the trial. Therefore, it was nec-
essary to estimate utilities based on a separate health
state valuation study and a mapping exercise using
patient-level data from the trial. Sensitivity analyses
revealed that the economic model results were not
sensitive to utility values based on adverse event status;
however, they were somewhat sensitive to the use of
extreme, albeit implausible, values for locoregional
disease control and progressive disease.
A final concern is that some of the unit costs avail-
able for health economic evaluation varied substan-
tially across the countries. Nevertheless, the absolute
level of most unit costs does not differentially affect the
treatment arms. The cost items that drive the difference
between the treatment arms are the price of cetuximab
and its administration costs.
In conclusion, the study shows that the addition of
cetuximab to radiotherapy is a cost-effective treatment
regimen in locally advanced head and neck cancer
patients in Belgium, France, Italy, Switzerland, and the
United Kingdom, who are not amenable to chemora-
diotherapy, and for whom radiotherapy is the best
option.
We wish to acknowledge Professor Chris Boshoff, Medical
Oncologist, University College Hospital London, UK, for
providing expert advice for the economic evaluation. We are
particularly grateful to the following participants of the
Advisory Board Meetings for their invaluable contributions:
Belgium: Vincent Gregoire, Jan Vermorken, Joseph Kerger,
Ahmad Awada, Danielle Van den Weyngaert; France: Lau-
rence Bozec, Martin Housset, Xavier Pivot, Florian Scotte,
Tierry Toubiana; Italy: Giovanni Almadori, Lucio Ascani,
Marco Benasso, Renzo Corvi, Lisa Licitra, Lorenzo Manto-
vani, Marco Merlano, Patrizia Olmi; the United Kingdom:
Liz Junor, David Morgan, A. Biswas, N. Shihari, Sarah Little,
Ta b l e 7 Sensitivity analysis on incremental cost per QALY gained (€)
Analysis Belgium France Italy Switzerland United Kingdom
Base case 8,568 10,836 7,538 10,328 9,144
Extrapolation of survival
Restricted means survival data* 26,731 33,535 23,740 32,440 28,729
Extrapolation using the Weibull distribution 7,970 10,394 6,959 9,524 8,407
Discounting
0% discount rate (costs & benefits) 6,820 8,500 5,934 8,119 7,233
5% discount rate (costs & benefits) 9,393 11,925 8,297 11,373 10,050
Utility values
Halve increment between locoregional
control and progressive disease utility
(HS J 0.679; HS K 0.312)
12,078 15,430 10,626 14,560 12,890
Administration cost for cetuximab
Increased by 100% 9,406 13,849 7,685 12,706 10,537
*Analysis restricted to the study follow-up period.
HS, health state; QALY, quality-adjusted life-year.
798 Brown et al.
Geoff Saunders; Switzerland: Pia Huegenin, Mahmut
Ozsahin, Razvan Popescu, Nicolas Mach. (ClinicalTrials.gov
number, NCT00004227)
Source of financial support: This study was commissioned by
Merck KGaA, Darmstadt, Germany.
References
1 Lefebvre JL. Current clinical outcomes demand new
treatment options for SCCHN. Ann Oncol 2005;
16(Suppl. 6):S7–i12.
2 Parkin DM, Pisani P, Ferlay J. Estimates of the world-
wide incidence of 25 major cancers in 1990. Int J
Cancer 1999;80:827–41.
3 Argiris A. Update on chemoradiotherapy for head and
neck cancer. Curr Opin Oncol 2002;14:323–9.
4 Seiwert TY, Cohen EE. State-of-the-art management
of locally advanced head and neck cancer. Br J Cancer
2005;92:1341–8.
5 Harari PM, Ritter MA, Petereit DG, Mehta MP.
Chemoradiation for upper aerodigestive tract cancer:
balancing evidence from clinical trials with individual
patient recommendations. Curr Probl Cancer 2004;
28:7–40.
6 National Comprehensive Cancer Network. NCCN
Clinical Practice Guidelines in Oncology: Head and
Neck Cancers, Report No.: V.1.2006. Jenkintown,
PA: National Comprehensive Cancer Network, 2006.
7 Ang KK, Berkey BA, Tu X, et al. Impact of epidermal
growth factor receptor expression on survival and
pattern of relapse in patients with advanced head and
neck carcinoma. Cancer Res 2002;62:7350–6.
8 Bernier J, Schneider D. Cetuximab combined with
radiotherapy: an alternative to chemoradiotherapy for
patients with locally advanced squamous cell carcino-
mas of the head and neck? Eur J Cancer 2007;43:35–
45.
9 Bonner JA, Harari PM, Giralt J, et al. Radiotherapy
plus cetuximab for squamous-cell carcinoma of the
head and neck. N Engl J Med 2006;354:567–78.
10 Drummond MF, Mason AR. European perspective on
the costs and cost-effectiveness of cancer therapies. J
Clin Oncol 2007;25:191–5.
11 Lee JM, Turini M, Botteman MF, et al. Economic
burden of head and neck cancer. A literature review.
Eur J Health Econ 2004;5:70–80.
12 Drummond MF, Sculpher MJ, Torrance GW, et al.
Methods for the Economic Evaluation of Health Care
Programmes (3rd ed.). Oxford: Oxford University
Press, 2005.
13 Tappenden P, Chilcott J, Ward S, et al. Methodologi-
cal issues in the economic analysis of cancer treat-
ments. Eur J Cancer 2006;42:2867–75.
14 Perperoglou A, Keramopoullos A, van Houwelingen
HC. Approaches in modelling long-term survival: an
application to breast cancer. Stat Med 2007;6:2666–
85.
15 Pignon JP, Bourhis J, Domenge C, Designe L. Chemo-
therapy added to locoregional treatment for head and
neck squamous-cell carcinoma: three meta-analyses of
updated individual data. MACH-NC Collaborative
Group. Meta-analysis of Chemotherapy on Head and
Neck Cancer. Lancet 2000;355:949–55.
16 Brazier J, Deverill M, Green C, et al. A review of the
use of health status measures in economic evaluation.
Health Technol Assess 1999;3:1–164.
17 Dolan P. Modeling valuations for EuroQol health
states. Med Care 1997;35:1095–108.
18 Briggs AH. Handling uncertainty in economic evalu-
ation. In: Drummond MF, McGuire A, eds. Economic
Evaluation in Health Care. Oxford, NY: Oxford Uni-
versity Press, 2001.
19 Neymark N, Adriaenssen I, Gorlia T, et al. Estimating
survival gain for economic evaluations with survival
time as principal endpoint: a cost-effectiveness analy-
sis of adding early hormonal therapy to radiotherapy
in patients with locally advanced prostate cancer.
Health Econ 2002;11:233–48.
20 Rawlins MD, Culyer AJ. National Institute for
Clinical Excellence and its value judgments. BMJ
2004;329:224–7.
21 Eichler HG, Kong SX, Gerth WC, et al. Use of cost-
effectiveness analysis in health-care resource alloca-
tion decision-making: how are cost-effectiveness
thresholds expected to emerge? Value Health 2004;
7:518–28.
22 Posner MR, Wirth LJ. Cetuximab and radiotherapy
for head and neck cancer. N Engl J Med 2006;
354:634–6.
23 Braaksma MAM, Nijdam W, et al. Costs of treatment
intensification for head and neck cancer: concomitant
chemoradiation randomised for radioprotection with
amifostine. Eur J Cancer 2005;41:2102–11.
24 Elting LS, Cooksley C, Chambers M, et al. The
burdens of cancer therapy. Clinical and economic out-
comes of chemotherapy-induced mucositis. Cancer
2003;98:1531–9.
An Economic Evaluation of Cetuximab Combination 799
